Home My Page Projects Code Snippets Project Openings SML/NJ
 Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

# SCM Repository

[smlnj] Annotation of /sml/trunk/HISTORY
 [smlnj] / sml / trunk / HISTORY

# Annotation of /sml/trunk/HISTORY

 1 : dbm 570 This is the HISTORY file for the Yale SML/NJ CVS repository. 2 : 3 : An entry should be made for _every_ commit to the repository. 4 : The entries in this file will be used when creating the README 5 : for new versions, so keep that in mind when writing the 6 : description. 7 : 8 : The form of an entry should be: 9 : 10 : mblume 1615 Name: 11 : mblume 1480 Date: yyyy/mm/dd 12 : Tag: 13 : Description: 14 : 15 : ---------------------------------------------------------------------- 16 : mblume 1702 Name: Matthias Blume (blume (at) tti - c (dot) org) 17 : mblume 1724 Date: 2004/12/13 14:45:00 CST 18 : Tag: blume-20041213-ml-makedepend 19 : Description: 20 : 21 : Fixed bug in code for ml-makedepend. 22 : 23 : ---------------------------------------------------------------------- 24 : Name: Matthias Blume (blume (at) tti - c (dot) org) 25 : mblume 1720 Date: 2004/12/09 16:30:00 CST 26 : Tag: blume-20041209-statistics 27 : Description: 28 : 29 : Added two simple but potentially useful statistics modules to SML/NJ Library. 30 : (See CHANGES file there.) 31 : 32 : ---------------------------------------------------------------------- 33 : Name: Matthias Blume (blume (at) tti - c (dot) org) 34 : mblume 1718 Date: 2004/12/01 16:50:00 CST 35 : Tag: blume-20041201-atom 36 : Description: 37 : 38 : smlnj-lib: 39 : 40 : Added function HashString.hashString' for substrings. 41 : Hand-inlined CharVector.foldl into HashString (for speed). 42 : Modified implementation of structure Atom to avoid extracting 43 : strings from substrings unless necessary. 44 : (Also see CHANGES file for smlnj-lib.) 45 : 46 : ---------------------------------------------------------------------- 47 : Name: Matthias Blume (blume (at) tti - c (dot) org) 48 : mblume 1717 Date: 2004/11/24 22:15:00 CST 49 : Tag: blume-20041124-cml 50 : Description: 51 : 52 : Made sure CML compiles when Position = Int64. 53 : 54 : ---------------------------------------------------------------------- 55 : Name: Matthias Blume (blume (at) tti - c (dot) org) 56 : mblume 1715 Date: 2004/11/24 14:30:00 CST 57 : Tag: blume-20041124-position 58 : Description: 59 : 60 : The compiler can now be compiled in a mode that makes structure Position 61 : equal to Int64. The default, however, is unchanged (Position = Int31) 62 : for the time being. 63 : 64 : To enable 64-bit positions, use the following procedure: 65 : 66 : 1. Start sml 67 : 2. Autoload $smlnj/cmb.cm (if not already autoloaded) 68 : 3. Type 69 : #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1); 70 : 4. Run CMB.make() as usual. 71 : 72 : This is barely tested. The only test so far was a little SML program 73 : counting the number of characters in an 8-gigabyte file by 74 : reading it character-by-character. That test was successful. 75 : 76 : In support of 64-bit positions, a number of new functions have been 77 : added to the runtime system. 78 : 79 : ---------------------------------------------------------------------- 80 : Name: Matthias Blume (blume (at) tti - c (dot) org) 81 : mblume 1716 Date: 2004/11/23 14:45:00 CST 82 : Tag: blume-20041123-useFile 83 : mblume 1712 Description: 84 : 85 : Fixed a problem with unhelpful error messages related to problems with 86 : .cm- or .sml files that appear as part of the sml command line. 87 : 88 : ---------------------------------------------------------------------- 89 : Name: Matthias Blume (blume (at) tti - c (dot) org) 90 : mblume 1703 Date: 2004/11/18 15:40:00 CST 91 : Tag: Release_110_51 92 : Description: 93 : 94 : New working version (110.51). NEW BOOTFILES! 95 : 96 : ---------------------------------------------------------------------- 97 : Name: Matthias Blume (blume (at) tti - c (dot) org) 98 : mblume 1702 Date: 2004/11/18 15:35:00 CST 99 : Tag: 100 : Description: 101 : 102 : Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5 103 : Joemann). 104 : 105 : ---------------------------------------------------------------------- 106 : allenleung 1701 Name: Allen Leung (leunga (at) reservoir (dot) com) 107 : Date: 2004/11/17 16:05:21 EST 2004 108 : Tag: leunga-20041117-mlrisc-live-kill 109 : Description: 110 : 111 : Added support for MLTree constructs LIVE and KILL 112 : to all the architectures. 113 : 114 : ---------------------------------------------------------------------- 115 : mblume 1687 Name: Matthias Blume (blume (at) tti - c (dot) org) 116 : mblume 1689 Date: 2004/11/13 00:20:00 CST 117 : Tag: blume-20041113-versiontool 118 : Description: 119 : 120 : - Stripped down the versiontool: It now only handles the version number. 121 : The date string is generated at bootstrap time (during makeml). 122 : 123 : - In a previous commit, fixed a minor issue with how polyequal is being 124 : translated. In particular, the code now "looks through" abstractions. 125 : This results in slightly fewer polyEqual warnings and hopefully slightly 126 : more efficient code. Important examples for where this matters are 127 : the new int64 and word64 types. 128 : 129 : ---------------------------------------------------------------------- 130 : Name: Matthias Blume (blume (at) tti - c (dot) org) 131 : mblume 1687 Date: 2004/11/12 00:30:00 CST 132 : Tag: blume-20041112-int64 133 : Description: 134 : 135 : Structure Int64 fully hooked in. (The implementation is not very 136 : efficient, though.) 137 : 138 : ---------------------------------------------------------------------- 139 : Name: Matthias Blume (blume (at) tti - c (dot) org) 140 : mblume 1686 Date: 2004/11/11 17:30:00 CST 141 : Tag: blume-20041111-more64 142 : Description: 143 : 144 : All the pieces of Word64 are now there, with the exception of the 145 : conversions from and to LargeWord. (Eventually these need to be identities, 146 : but for the time being they don't even make sense because LargeWord is 147 : 32-bit wide.) 148 : 149 : Also started to add similar support for Int64, but major pieces of that 150 : are still missing. 151 : 152 : ---------------------------------------------------------------------- 153 : mblume 1668 Name: Matthias Blume (blume (at) tti - c (dot) org) 154 : mblume 1684 Date: 2004/11/11 00:15:00 CST 155 : Tag: blume-20041111-word64 156 : Description: 157 : 158 : Structure Word64 is now (almost) complete, word literals and patterns 159 : seem to work. There are a few odd pieces missing. In particular, 160 : I didn't do the {from,to}LargeWord parts because LargeWord is still Word32 161 : at the moment. 162 : 163 : Making Word64 official would mean that LargeWord becomes Word64. But 164 : this requires extreme care because most word-word conversions have to 165 : go through LargeWord, so making a mistake means loss of efficiency or 166 : worse. Eventually there will be a solution similar to (but actually 167 : simpler than) what I did with IntInf. 168 : 169 : ---------------------------------------------------------------------- 170 : Name: Matthias Blume (blume (at) tti - c (dot) org) 171 : mblume 1683 Date: 2004/11/10 18:12:00 CST 172 : Tag: blume-20041110-64bit 173 : Description: 174 : 175 : More 64-bit hacking (but still not even half-way there yet). 176 : Also, some assorted improvements to the handling of 8-bit words. 177 : 178 : ---------------------------------------------------------------------- 179 : Name: Matthias Blume (blume (at) tti - c (dot) org) 180 : mblume 1682 Date: 2004/11/09 17:50:00 CST 181 : mblume 1683 Tag: 182 : mblume 1682 Description: 183 : 184 : Started adding some infrastructure for supporting 64-bit int- and 185 : word-types. (Still in its very early stages.) 186 : 187 : ---------------------------------------------------------------------- 188 : Name: Matthias Blume (blume (at) tti - c (dot) org) 189 : mblume 1672 Date: 2004/10/28 10:45:00 CDT 190 : Tag: Release_110_50 191 : Description: 192 : 193 : New working version (110.50). NEW BOOTFILES! 194 : 195 : ===================== 196 : 197 : Also: 198 : 199 : - Changend config/srcarchiveurl from a file just 200 : containing the URL string into a file containing 201 : shell script code. The code has access to the$VERSION variable. 202 : - Made corresponding changes to config/install.sh and config/unpack. 203 : - Default contents of config/srcarchiveurl uses $VERSION and 204 : normally does not have to be edited to reflect a version change. 205 : 206 : (As a result, a version change can be done by just editing 207 : config/version, the rest is now automatic.) 208 : 209 : ---------------------------------------------------------------------- 210 : Name: Matthias Blume (blume (at) tti - c (dot) org) 211 : mblume 1669 Date: 2004/10/27 17:50:00 CDT 212 : Tag: blume-20041027-btrace-msg 213 : Description: 214 : 215 : BackTrace.monitor now also reports the source of the exception that 216 : triggered the trace. 217 : 218 : ---------------------------------------------------------------------- 219 : Name: Matthias Blume (blume (at) tti - c (dot) org) 220 : mblume 1668 Date: 2004/10/27 17:20:00 CDT 221 : Tag: blume-20041027-x86-c-calls 222 : Description: 223 : 224 : This is the HISTORY entry for two earlier commits, both concerning 225 : the x86 c-calls code in MLRISC: 226 : 227 : - added a missing LOAD in the code that deals with struct arguments 228 : - made sure the caller does not add the wrong number of bytes to the 229 : stack pointer after a call of a function returning a struct 230 : (the callee already pops the implicit argument which points to 231 : the space reserved for the result) 232 : 233 : ---------------------------------------------------------------------- 234 : allenleung 1662 Name: Allen Leung (leunga (at) reservoir (dot) com) 235 : Date: 2004/10/24 14:00:00 EST 236 : Tag: leunga-20041024-x86-gas-fucomip 237 : Description: 238 : 239 : John discovered a bug in the syntax of fucomip. 240 : The opcodes FU?COMIP? have been changed to 241 : 242 : fu?comip? %st(i), %st 243 : 244 : ---------------------------------------------------------------------- 245 : mblume 1622 Name: Matthias Blume (blume (at) tti - c (dot) org) 246 : mblume 1656 Date: 2004/10/20 15:06:00 CDT 247 : Tag: blume-20041020-standalone-backtrace 248 : Description: 249 : 250 : Added a mechanism for getting back-trace information from standalone 251 : programs. Here is how it works: 252 : 253 : 1. The part of the program from which you want to get backtrace 254 : information (usually the whole program) should be wrapped with 255 : BackTrace.monitor. This is a (unit->'a)->'a function, and your 256 : main program could be modified from something like 257 : 258 : fun main (pgm, args) = ... 259 : 260 : to 261 : 262 : fun main (pgm, args) = BackTrace.monitor (fn () => ...) 263 : 264 : 2. To be able to access BackTrace.monitor, you have to add 265 : library$smlnj-tdp/plugins.cm to the .cm file that contains your 266 : main function. 267 : 268 : 3. Remove all compiled code (i.e., all the .cm/ subdirectories that 269 : CM might have created in the past for your project). 270 : 271 : 4. Build the system using this command line: 272 : 273 : ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \ 274 : myprog.cm MyProg.main myprog 275 : 276 : instead of the usual 277 : 278 : ml-build myprog.cm MyProg.main myprog 279 : 280 : I changed a library name: 281 : 282 :$/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm 283 : 284 : New libraries: 285 : 286 :$smlnj-tdp/back-trace.cm 287 : -- when loaded causes the back-trace plugin to be installed 288 : $smlnj-tdp/coverage.cm 289 : -- when loaded causes the coverage plugin to be installed 290 : 291 : ---------------------------------------------------------------------- 292 : Name: Matthias Blume (blume (at) tti - c (dot) org) 293 : mblume 1655 Date: 2004/10/18 16:45:00 CDT 294 : Tag: blume-20041018-groupowner 295 : Description: 296 : 297 : Added an "obsolete" warning for the "group owner" syntax to CM's parser. 298 : 299 : Eliminated group owner specs from .cm files throughout the source tree. 300 : 301 : ---------------------------------------------------------------------- 302 : Name: Matthias Blume (blume (at) tti - c (dot) org) 303 : mblume 1653 Date: 2004/10/15 15:45:00 CDT 304 : Tag: blume-20041015-coverage 305 : Description: 306 : 307 : * Test coverage tool added! 308 : * Further reorganization of tracing-, debugging-, and profiling support: 309 : 310 : - moved original BTImp -- now called BackTrace -- into a separate 311 : library called$/trace-debug-profile.cm 312 : - eliminated all mentions of BTrace from SMLofNJ.Internals 313 : - only the instrumentation mechanism is now left in the compiler proper 314 : - BackTrace module is a plugin which is NOT plugged in by default 315 : - Coverage module is another such plugin 316 : 317 : To get the benefits of any of these plugin modules, the code in 318 : question must be compiled with tdp instrumentation turned on. This 319 : can be done by setting SMLofNJ.Internals.TDP.mode to true. 320 : (The ref cell is also controlled via the -Ctdp.instrument=... switch.) 321 : 322 : Plugins are selected at link time. (Pre-compiled instrumented code 323 : can be re-loaded with different plugins in effect.) When an 324 : instrumented module is linked, whatever plugins are at that time 325 : enabled will come into effect for that module. 326 : 327 : To enable the back-trace plugin, load library $/trace-debug-profile.cm 328 : and invoke BackTrace.install() (e.g., from the interactive prompt). 329 : To enable the coverage plugin, load the same library and invoke 330 : Coverage.install(). 331 : 332 : Back-traces are generated automatically on uncaught exceptions and 333 : when the code in question explicitly invokes BackTrace.trigger(). 334 : 335 : Coverage (and execution frequency-) information must be queried 336 : explicitly by calling Coverage.not_covered and Coverage.hot_spots. 337 : 338 : ---------------------------------------------------------------------- 339 : Name: Matthias Blume (blume (at) tti - c (dot) org) 340 : mblume 1652 Date: 2004/10/14 17:40:00 CDT 341 : Tag: blume-20041014-tdp-core 342 : Description: 343 : 344 : Snapshot of a significant overhaul of how the trace/debug/profile support 345 : is hooked into the system (specifically: Core and SMLofNJ.Internals). 346 : 347 : ---------------------------------------------------------------------- 348 : Name: Matthias Blume (blume (at) tti - c (dot) org) 349 : mblume 1650 Date: 2004/10/13 16:34:00 CDT 350 : Tag: blume-20041013-tdp 351 : Description: 352 : 353 : Some rationalization of names: 354 : 355 : structure BTrace -> structure TDPInstrument 356 : etc. 357 : 358 : This is is preparation of using the original back-trace 359 : instrumentation for other purposes. "TDP" stands for 360 : Trace/Debug/Profile. 361 : 362 : The control flag controlling whether instrumentation is on or off is now 363 : registered under a different name, so instead of running sml as 364 : 365 : sml -Cinstrument.btrace-mode=true 366 : 367 : one has to say 368 : 369 : sml -Ctdp.instrument=true 370 : 371 : ---------------------------------------------------------------------- 372 : Name: Matthias Blume (blume (at) tti - c (dot) org) 373 : mblume 1645 Date: 2004/10/11 16:37:00 CDT 374 : Tag: blume-20041011-regions 375 : Description: 376 : 377 : Made some minor modifications to elabcore.sml to have source regions 378 : be propagated more tightly -- resulting in better (i.e., smaller) 379 : regions being reported in error- and debug messages. 380 : 381 : ---------------------------------------------------------------------- 382 : Name: Matthias Blume (blume (at) tti - c (dot) org) 383 : mblume 1644 Date: 2004/10/08 22:50:00 CDT 384 : Tag: blume-20041008-cmkw 385 : Description: 386 : 387 : Fixed handling of keywords in .cm files: After seeing "is" the lexer 388 : treats subsequent occurrences of "group", "library", "source", "is", 389 : "*", and "-" as ordinary identifiers rather than keywords. 390 : 391 : Most seriously, this fixes a problem with CM's "shell" tool. The tool 392 : is supposed to accept a tool argument called "source", but this did 393 : not work because of the clash with the keyword. 394 : 395 : ---------------------------------------------------------------------- 396 : Name: Matthias Blume (blume (at) tti - c (dot) org) 397 : mblume 1642 Date: 2004/10/07 16:00:00 CDT 398 : Tag: blume-20041007-cleanup 399 : Description: 400 : 401 : Assorted cleanup work: 402 : 403 : - got rid of intstrmap in favor of using the library's 404 : hash table implementation 405 : - threw out most of the pathnames stuff, as it was not used anyway 406 : - simplified tokentable implementation 407 : - fixed some minor spelling errors 408 : 409 : ---------------------------------------------------------------------- 410 : Name: Matthias Blume (blume (at) tti - c (dot) org) 411 : mblume 1641 Date: 2004/10/06 15:15:15 CDT 412 : Tag: blume-20041006-handler 413 : Description: 414 : 415 : Cleaned up the absyn to reflect the invariant that HANDLE always 416 : carries a FNexp as part of the type definition. This eliminates some 417 : superfluous sanity checks at runtime down the road. 418 : 419 : Some minor cleanup of the btrace code. 420 : 421 : ---------------------------------------------------------------------- 422 : Name: Matthias Blume (blume (at) tti - c (dot) org) 423 : mblume 1640 Date: 2004/10/01 10:20:30 CDT 424 : Tag: blume-20041001-slave 425 : Description: 426 : 427 : Added hack to make slave mode work in the presence of the version 428 : tool. (Still, since the master does two passes over the code for 429 : CMB.make, the release number gets bumped twice when slaves are 430 : attached. I don't know if this is worth fixing...) 431 : 432 : ---------------------------------------------------------------------- 433 : Name: Matthias Blume (blume (at) tti - c (dot) org) 434 : mblume 1639 Date: 2004/09/30 10:55:00 CDT 435 : Tag: blume-20040930-version 436 : Description: 437 : 438 : * Moved the "version" magic into its own little library under 439 : mblume 1672 src/system/smlnj/internal. This avoids expensive reconstruction of 440 : a stable src/compiler/core.cm. 441 : mblume 1639 442 : mblume 1672 * At the same time, structure CompilerVersion is now known as 443 : structure SMLNJVersion. 444 : mblume 1639 445 : * Arranged for the version tool to NOT kick in when rebuilding the system 446 : (makeml -rebuild, fixpt). Otherwise one would never reach a fixpoint. 447 : Also, loading the versiontool does not work when rebuilding the system 448 : because CM is not properly initialized at that time. 449 : 450 : ---------------------------------------------------------------------- 451 : Name: Matthias Blume (blume (at) tti - c (dot) org) 452 : mblume 1638 Date: 2004/09/29 14:00:00 CDT 453 : Tag: blume-20040929-autoversion 454 : Description: 455 : 456 : Implemented some CM magic to have 457 : file src/compiler/TopLevel/main/version.sml 458 : generated automagically. 459 : The version is taken from two files: config/version and config/release. 460 : The first is expected to contain a two-part version number such as 110.49. 461 : The second should contain a single number, but it may be missing. 462 : 463 : If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the 464 : time the version tool is loaded (which is the first time you say CMB.make), 465 : then the tool will increment the value stored in config/release every 466 : time CMB.make is invoked. 467 : 468 : The binfile format is now insensitive to anything beyond the first 469 : two components of a version number, so bumping the release does not render 470 : binfiles incompatible. Auto-bumping can be used to keep track of versions 471 : during development without invalidating existing binfiles. 472 : 473 : In any case, every CMB.make updates the date information in version.sml. 474 : (This is the date that is printed in the banner.) 475 : 476 : ---------------------------------------------------------------------- 477 : Name: Matthias Blume (blume (at) tti - c (dot) org) 478 : mblume 1634 Date: 2004/09/28 10:53:00 CDT 479 : Tag: blume-20040928-controls 480 : Description: 481 : 482 : Some cleanup of the controls code. 483 : 484 : ---------------------------------------------------------------------- 485 : Name: Matthias Blume (blume (at) tti - c (dot) org) 486 : mblume 1633 Date: 2004/09/27 22:08:00 CDT 487 : Tag: blume-20040927-controls 488 : Description: 489 : 490 : Added two pieces of functionality to the Controls interface: 491 : 492 : 1. val save'restore: 'a control -> unit -> unit 493 : 494 : grabs the current value of the control in stage 1 and restores it 495 : in stage 2. 496 : 497 : 2. val set' : 'a control * 'a -> unit -> unit 498 : 499 : stores the given value into the control in stage 2 (i.e., delayed) 500 : but does all error checking in stage 1. 501 : (This is for string controls that need to do parse their argument -- 502 : something that might fail. In some cases, notably in CM, one 503 : already knows the intended argument but wants to delay the actual 504 : assignment until a time when error recovery would be more difficult.) 505 : 506 : Changed the handling of controls in tool arguments to classes "sml" and 507 : "lazysml": 508 : - use Controls.save'restore as a more robust way of restoring the 509 : old value (in particular: without having to re-parse the string) 510 : - use controls to handle the "overload" keyword in the init group 511 : (I believe this change actually fixes a long-standing obscure bug.) 512 : 513 : ---------------------------------------------------------------------- 514 : Name: Matthias Blume (blume (at) tti - c (dot) org) 515 : mblume 1632 Date: 2004/09/27 17:00:00 CDT 516 : Tag: blume-20040927-lazysml 517 : Description: 518 : 519 : Added a new tool class called "lazysml" to CM's tool chest. The only 520 : difference to "sml" is that compilation is done with Control.lazysml 521 : set to true. A source of class "lazysml" is automatically recognized 522 : by a file name suffix of ".lml". 523 : 524 : In addition to the above feature, the original class "sml" now also 525 : supports a tool argument "lazy" which has the same effect. As a 526 : result, the following three lines are equivalent: 527 : 528 : foo.sml : lazysml 529 : foo.sml : sml (lazy) 530 : foo.sml (lazy) 531 : 532 : The setting goes into effect both during parsing and during 533 : compilation. The original setting is restored right after parsing and 534 : after compilation, respectively. 535 : 536 : In addition to all the above, there is also a general mechanism to set 537 : ANY of the "controls" that are available at the command line via 538 : "-C..." on a per-sml-file basis. The same rules that apply for "lazy" 539 : apply as well. (In fact, "lazy" is implemented as a special case of 540 : the general mechanism.) 541 : 542 : The .cm file syntax uses a new keyword tool argument called "with". 543 : There are several ways of indicating the desired settings: 544 : 545 : foo.sml (with:parser.quotations=true) 546 : foo.sml (with:(name:parser.quotations value:true)) 547 : foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...)) 548 : foo.sml (with:(name1=value1 name2=value2 ...)) 549 : foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...)) 550 : 551 : etc. 552 : 553 : Another possible abbreviation is to leave out the =v or value:v part 554 : if the name refers to a boolean control (in which case the value is 555 : taken to be true). Thus, one could get lazy sml also by saying: 556 : 557 : foo.sml (with:parser.lazy-keyword=true) 558 : foo.sml (with:parser.lazy-keyword) 559 : foo.sml (with:(name:parser.lazy-keyword value:true)) 560 : foo.sml (with:(name:parser.lazy-keyword)) 561 : 562 : ---------------------------------------------------------------------- 563 : Name: Matthias Blume (blume (at) tti - c (dot) org) 564 : mblume 1631 Date: 2004/09/24 16:22:00 CDT 565 : Tag: blume-20040924-ppc-long-branch 566 : Description: 567 : 568 : Turned message about "emiting long form of branch" off by default. 569 : Added a control flag to turn it back on when desired. 570 : 571 : ---------------------------------------------------------------------- 572 : Name: Matthias Blume (blume (at) tti - c (dot) org) 573 : mblume 1630 Date: 2004/09/24 16:05:00 CDT 574 : Tag: blume-20040924-rounding 575 : Description: 576 : 577 : Applied patch for setting rounding modes under Mac OS X. Thanks to 578 : Melissa O'Neill for providing the code! 579 : 580 : ---------------------------------------------------------------------- 581 : Name: Matthias Blume (blume (at) tti - c (dot) org) 582 : mblume 1629 Date: 2004/09/23 17:30:00 CDT 583 : Tag: blume-20040923-envvars 584 : Description: 585 : 586 : 1. Changed definition of type ControlRegistry.registry_tree to 587 : include control_info (i.e., the name of the controlling environment 588 : variable). 589 : 590 : 2. Added command-line flags -e and -E to print the names of environment 591 : variables that can be used to control internal settings. 592 : (This uses the new API mentioned in 1.) 593 : 594 : ---------------------------------------------------------------------- 595 : Name: Matthias Blume (blume (at) tti - c (dot) org) 596 : mblume 1622 Date: 2004/09/13 16:50:00 CDT 597 : Tag: Release_110_49 598 : Description: 599 : 600 : New working version (110.49). NEW BOOTFILES! 601 : 602 : ---------------------------------------------------------------------- 603 : mblume 1621 Name: Matthias Blume 604 : Date: 2004/09/13 16:20:00 CDT 605 : Tag: blume-20040913-config-mlrisc 606 : Description: 607 : 608 : Put target "mlrisc" back into the default list. 609 : (There is no harm in having it, and some users have expressed their 610 : wish to have "mlrisc" included by default.) 611 : 612 : ---------------------------------------------------------------------- 613 : jhr 1620 Name: John Reppy 614 : Date: 2004/09/13 615 : Tag: jhr-20040913-signals 616 : Description: 617 : 618 : Fixed the signal masking code to properly nest mask/unmask operations 619 : on a per-signal basis. 620 : 621 : ---------------------------------------------------------------------- 622 : mblume 1619 Name: Matthias Blume (blume (at) tti - c (dot) org) 623 : Date: 2004/09/08 13:20:00 CDT 624 : Tag: blume-20040908-heap-magic 625 : Description: 626 : 627 : Bumped the heap macig number to 0x09082004 to account for the changed 628 : layout of the ML frame under MacOS X. 629 : 630 : ---------------------------------------------------------------------- 631 : allenleung 1618 Name: Allen Leung (leunga (at) reservoir (dot) com) 632 : Date: 2004/09/03 11:26:00 EST 633 : Tag: leunga-20040903-cygwin-install 634 : Description: 635 : 636 : Added a patch to _arch-n-opsys to enable the Cygwin runtime. 637 : The Cygwin runtime is turned on by setting the environment 638 : variable SMLNJ_CYGWIN_RUNTIME to 1. 639 : 640 : ---------------------------------------------------------------------- 641 : mblume 1615 Name: Matthias Blume (blume (at) tti - c (dot) org) 642 : mblume 1617 Date: 2004/08/31 17:15:00 CDT 643 : Tag: blume-20040831-core 644 : Description: 645 : 646 : Added some exports to src/compiler/core.cm upon request by J. Joemann. 647 : 648 : ---------------------------------------------------------------------- 649 : Name: Matthias Blume (blume (at) tti - c (dot) org) 650 : mblume 1616 Date: 2004/08/30 17:55:00 CDT 651 : Tag: blume-20040830-installer 652 : Description: 653 : 654 : Upon request by Johannes Joemann: 655 : 656 : - improved ML code of installer to fall back to coping when renaming 657 : fails (i.e., when source and target are on different file systems); 658 : the code compiles but has yet to be tested in anger 659 : - removed mlrisc from list of default targets (config/targets) 660 : 661 : ---------------------------------------------------------------------- 662 : Name: Matthias Blume (blume (at) tti - c (dot) org) 663 : mblume 1615 Date: 2004/08/27 17:20:00 CDT 664 : Tag: blume-20040827-ptreql 665 : Description: 666 : 667 : Added ptreql primop to structure InlineT (upon request from Larry 668 : Paulson). 669 : 670 : ---------------------------------------------------------------------- 671 : allenleung 1613 Name: Allen Leung (leunga (at) reservoir (dot) com) 672 : allenleung 1614 Date: 2004/08/15 21:21:00 EST 673 : Tag: leunga-110_48-udgraph 674 : Description: 675 : 676 : Another bug fix from Carl Hauser: 677 : 678 : diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml 679 : > 48c48 680 : > < | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else 681 : > e::L) 682 : > --- 683 : > > | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L) 684 : > Without this, any deletion of an edge in an undirected graph does severe 685 : > violence to the graph. 686 : 687 : ---------------------------------------------------------------------- 688 : Name: Allen Leung (leunga (at) reservoir (dot) com) 689 : allenleung 1613 Date: 2004/08/10 23:35:00 EST 690 : Tag: leunga-110_48-ppc 691 : Description: 692 : 693 : The IBM/MacOS syntax switch on PPC was incorrectly swapped. 694 : 695 : ---------------------------------------------------------------------- 696 : mblume 1602 Name: Matthias Blume (blume (at) tti - c (dot) org) 697 : mblume 1615 Date: 2004/08/10 12:00:00 CDT 698 : mblume 1602 Tag: Release_110_48 699 : Description: 700 : 701 : New working version (110.48). NEW BOOTFILES! 702 : 703 : ---------------------------------------------------------------------- 704 : allenleung 1599 Name: Allen Leung (leunga (at) reservoir (dot) com) 705 : Date: 2004/08/09 12:21:00 EST 706 : Tag: leunga-110_47-dijsktra 707 : Description: 708 : 709 : Bug fix from Carl Hauser: 710 : 711 : single_source_shortest_paths in dijkstra.sml was observed to get wrong 712 : answers (by comparing to single_source_shortest_paths in bellman-ford.sml). 713 : 714 : The problem is that following the expression A.update(dist,s,Num.zero) 715 : it is necessary to update the priority queue using Q.decreaseWeight(Q,s). 716 : 717 : ---------------------------------------------------------------------- 718 : mblume 1595 Name: Matthias Blume (blume (at) tti - c (dot) org) 719 : Date: 2004/08/06 18:10:00 CDT 720 : Tag: blume-20040806-cmdline 721 : Description: 722 : 723 : Fiddled with handling of command-line options: 724 : 725 : * sml now quits after processing the command line 726 : if -H, -S, -h, or -s appears as the last 727 : command-line argument 728 : * a new option -q terminates the session when encountered on 729 : the command line; subsequent arguments will be ignored 730 : * bug fixes: short (erroneous) arguments are no longer ignored 731 : completely 732 : 733 : ---------------------------------------------------------------------- 734 : allenleung 1593 Name: Allen Leung (leunga (at) reservoir (dot) com) 735 : Date: 2004/08/04 18:17:00 EST 736 : Tag: leunga-110_47-ppc-ibm-asm 737 : Description: 738 : 739 : - Added minimal IBM assembly syntax support for PowerPC. 740 : 741 : - Cygwin: manually changed the file cygwin.def. Some exported symbols have 742 : been altered in the runtime. We need an automatic way to keep the file 743 : in sync. 744 : 745 : ---------------------------------------------------------------------- 746 : mblume 1548 Name: Matthias Blume (blume (at) tti - c (dot) org) 747 : mblume 1581 Date: 2004/08/04 14:00:00 CDT 748 : Tag: Release_110_47 749 : Description: 750 : 751 : New working version (110.47). NEW BOOTFILES! 752 : 753 : ---------------------------------------------------------------------- 754 : Name: Matthias Blume (blume (at) tti - c (dot) org) 755 : mblume 1580 Date: 2004/08/03 14:25:00 CDT 756 : Tag: blume-20040803-callingconv 757 : Description: 758 : 759 : Added low-level support for choosing C calling conventions by 760 : twiddling the type of rawccall. (See 761 : src/compiler/Semant/types/cproto.sml for details.) 762 : 763 : ---------------------------------------------------------------------- 764 : Name: Matthias Blume (blume (at) tti - c (dot) org) 765 : mblume 1579 Date: 2004/08/02 15:55:00 CDT 766 : Tag: blume-20040802-backout 767 : Description: 768 : 769 : Backed out of change to win32-filesys.c. The earlier patch to 770 : get_file_time caused CM to produce files with the wrong time stamp. 771 : 772 : ---------------------------------------------------------------------- 773 : Name: Matthias Blume (blume (at) tti - c (dot) org) 774 : mblume 1573 Date: 2004/08/02 14:45:00 CDT 775 : Tag: blume-20040802-nlffi-win32 776 : Description: 777 : 778 : Added NLFFI support for Win32, adapted from a patch provided by David 779 : Hansel. This is currently completely untested. Also, the issue 780 : concerning stdcall vs. ccall is still unresolved. 781 : 782 : ---------------------------------------------------------------------- 783 : Name: Matthias Blume (blume (at) tti - c (dot) org) 784 : mblume 1572 Date: 2004/07/30 17:55:00 CDT 785 : Tag: blume-20040730-various 786 : Description: 787 : 788 : Gearing up towards 110.47... 789 : 790 : - various minor bugfixes to ml-nlffigen 791 : - a beginning of a manual for nlffi 792 : 793 : - eliminated 'export name=value' in config/install.sh as this does 794 : not work with certain versions of /bin/sh 795 : (Thanks to David King at Motorola for catching this.) 796 : 797 : - several bugfixes provided or suggested by David Hansel at Reactive Systems: 798 : - added a test for tm==NULL to gmtime.c and localtime.c 799 : - applied patch for incorrect GetFileTime under win32 800 : - toSeconds -> toMilliseconds in Win32/win32-process.sml 801 : 802 : ---------------------------------------------------------------------- 803 : Name: Matthias Blume (blume (at) tti - c (dot) org) 804 : mblume 1557 Date: 2004/07/21 18:20:00 CDT 805 : Tag: blume-20040721-nlffigen 806 : Description: 807 : 808 : - Fixed minor issue in ml-nlffigen: Now generate structure T_foo 809 : for a typedef to an incomplete type, but leave out the "typ" member. 810 : (This is just for consistency.) 811 : - Started to produce what is supposed to become better (i.e., comprehensive) 812 : documentation of what ml-nlffigen does and produces. 813 : 814 : ---------------------------------------------------------------------- 815 : Name: Matthias Blume (blume (at) tti - c (dot) org) 816 : mblume 1548 Date: 2004/07/14 16:25:00 CDT 817 : Tag: blume-20040714-union 818 : Description: 819 : 820 : Added C_UNION to c-calls/c-types.sml and updated the machinery 821 : (ml-nlffigen, cproto.sml) that conveys C function interface 822 : information to the code generator. 823 : 824 : However, the actual architecture-specific implementation of function 825 : arguments and results that are C unions is still not implemented. 826 : 827 : ---------------------------------------------------------------------- 828 : allenleung 1539 Name: Allen Leung (leunga (at) reservoir (dot) com) 829 : allenleung 1546 Date: 2004/07/14 14:38:00 EST 830 : Tag: leunga-110_46_1-ppc-lwzu 831 : Description: 832 : 833 : Added these instructions to the PowerPC architecture: 834 : LBZU(X), LHZU(X), LWZU(X), 835 : STWU(X), STFDU, STFSU 836 : 837 : etc... 838 : 839 : Note: I haven't added their instruction encoding into the description. 840 : 841 : ---------------------------------------------------------------------- 842 : Name: Allen Leung (leunga (at) reservoir (dot) com) 843 : allenleung 1539 Date: 2004/07/13 15:04:00 EST 844 : Tag: leunga-110_46_1-ppc-lwarx 845 : Description: 846 : 847 : Added the two instructions LWARX and STWCX to the PowerPC 848 : instruction set. 849 : 850 : A (untested) rewrite of loop-structure.sml. The old version 851 : is completely broken. 852 : 853 : ---------------------------------------------------------------------- 854 : mblume 1479 Name: Matthias Blume (blume (at) tti - c (dot) org) 855 : mblume 1537 Date: 2004/07/13 13:50:00 CDT 856 : Tag: blume-20040713-nlffi 857 : Description: 858 : 859 : - use paramAlloc to report c-calls with too many arguments 860 : (for PPC version where parameter area is pre-allocated) 861 : - added ccall_maxargspace to machspec (to implement the above) 862 : - made "make" commend in CM's "make" tool configurable 863 : - added option (default: on) for passing the name of the SML/NJ's "bin" 864 : directory to "make"; the call looks like this: 865 : 866 : make SMLNJ_BINDIR= 867 : 868 : This can be used by the Makefile to, e.g., pick the "right" version 869 : of ml-nlffigen. 870 : - minor code tweaks 871 : 872 : ---------------------------------------------------------------------- 873 : Name: Matthias Blume (blume (at) tti - c (dot) org) 874 : mblume 1532 Date: 2004/07/12 22:50:00 CDT 875 : Tag: blume-110_46_1-macosx-nlffi 876 : Description: 877 : 878 : NLFFI under Mac OS X now working (sort of). This is largely untested, 879 : though. 880 : 881 : Note: 1. You have to make a new, clean build of the runtime system. 882 : 2. There are new BOOTFILES, you have to use them! 883 : (Doing the bootstrap process yourself would be *very* painful! 884 : If you absolutely have to do it, build the system under 885 : a different architecture and then cross-compile.) 886 : 887 : Version bumped to 110.46.1 to account for runtime data format changes. 888 : 889 : ---------------------------------------------------------------------- 890 : Name: Matthias Blume (blume (at) tti - c (dot) org) 891 : mblume 1515 Date: 2004/06/18 14:30:00 CDT 892 : Tag: blume-20040618-unix 893 : Description: 894 : 895 : Changed the implementation of structure Unix so that the same stream 896 : is returned every time one of the {text,bin}{In,Out}streamOf functions 897 : is invoked on the same proc. This is not what the spec currently 898 : says -- although IMO it arguably should. (See discussion below.) 899 : 900 : ---------------------------------------------------------------------- 901 : Name: Matthias Blume (blume (at) tti - c (dot) org) 902 : mblume 1507 Date: 2004/06/17 18:15:00 CDT 903 : Tag: Release_110_46 904 : Description: 905 : 906 : New working version (110.46). NEW BOOTFILES! 907 : 908 : ---------------------------------------------------------------------- 909 : Name: Matthias Blume (blume (at) tti - c (dot) org) 910 : mblume 1506 Date: 2004/06/17 17:20:00 CDT 911 : Tag: blume-20040617-timer-unix 912 : Description: 913 : 914 : Changed the interface of structures Timer and Unix to match the most 915 : recent Basis spec. 916 : 917 : In the case of Unix there still seems to be an open/weird issue: 918 : 919 : The {text,bin}{In,Out}streamOf functions are supposed to create 920 : fresh streams whenever they are called -- as opposed to have them 921 : return the same stream every time. This design is supposed to 922 : prevent space leaks caused by proc values hanging on to streams. 923 : 924 : The reap function, on the other hand, is supposed to close the 925 : streams. This cannot be done without having a handle on the 926 : stream in proc after all... 927 : 928 : I took the liberty to implement the following stopgap solution: 929 : 930 : The proc value hangs on to the most recently created stream(s). 931 : Reap closes those. If either or both of the two streams hadn't 932 : been created at all yet, then reap will close the corresponding 933 : file descriptors directly. 934 : 935 : PS: I don't understand the original space leak argument anymore. If 936 : a proc hangs on to the imperative stream, then I/O operations on those 937 : will advance the state of the cached stream and avoid the space leak. 938 : 939 : ---------------------------------------------------------------------- 940 : Name: Matthias Blume (blume (at) tti - c (dot) org) 941 : mblume 1500 Date: 2004/05/28 16:45:00 CDT 942 : Tag: blume-20040528-basis 943 : Description: 944 : 945 : Added signature PACK_REAL and exported functor PrimIO. 946 : 947 : ---------------------------------------------------------------------- 948 : Name: Matthias Blume (blume (at) tti - c (dot) org) 949 : mblume 1498 Date: 2004/05/25 16:00:00 CDT 950 : Tag: blume-20040525-group-owner 951 : Description: 952 : 953 : CM now ignores (but still accepts) the "owner" information in group 954 : descriptions. The owner of a group is its next enclosing 955 : library. Each group must have a unique owner. (There is a virtual 956 : mblume 1505 "toplevel" library that owns groups which are not nested within a real 957 : mblume 1498 library.) Previously, each group had to explicitly declare its owner, 958 : and CM would check that such a declaration is correct. The new scheme 959 : is to have CM check that for each group there is precisely one owning 960 : library. 961 : 962 : The advantage of the new scheme is that the programmer no longer needs 963 : to maintain the somewhat annoying owner information. The downside is 964 : that CM cannot enforce the ownership rule across multiple runs of 965 : CM.make. Fortunately, enclosing the same group in two different 966 : libraries A and B which are not part of the same program does not 967 : cause real problems. 968 : 969 : ---------------------------------------------------------------------- 970 : Name: Matthias Blume (blume (at) tti - c (dot) org) 971 : mblume 1496 Date: 2004/05/20 16:00:00 CDT 972 : Tag: blume-20040520-win32 973 : Description: 974 : 975 : Made the win32 version work again. (Strangely, a misplaced comma had 976 : slipped into win32-process.c which prevented the runtime from being 977 : compiled correctly.) 978 : 979 : Also, included a minor addition to ml-build.bat analogous to what was 980 : done in blume-20040519-ml-build. 981 : 982 : ---------------------------------------------------------------------- 983 : Name: Matthias Blume (blume (at) tti - c (dot) org) 984 : mblume 1494 Date: 2004/05/19 22:10:00 CDT 985 : Tag: blume-20040519-ml-build 986 : Description: 987 : 988 : Arranged for ml-build to clean up after itself a little bit better. 989 : The script generates a temporary SML source file and compiles it using 990 : CM, so CM generates metadata (GUID, SKEL, objectfile) for it. It now 991 : gets rid of those at the end, so they don't accumulate under .cm. 992 : 993 : This required a minor change to install.sh because the name of the 994 : metadata directory (default: .cm) is actually configurable at 995 : installation time. 996 : 997 : ---------------------------------------------------------------------- 998 : Name: Matthias Blume (blume (at) tti - c (dot) org) 999 : mblume 1491 Date: 2004/05/18 15:50:00 CDT 1000 : Tag: blume-20040518-mkreader 1001 : Description: 1002 : 1003 : Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective 1004 : implementations from internal modules PosixBinPrimIO and PosixTextPrimIO. 1005 : 1006 : ---------------------------------------------------------------------- 1007 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1008 : mblume 1489 Date: 2004/05/11 14:35:00 CDT 1009 : Tag: blume-20040511-win32sock 1010 : Description: 1011 : 1012 : Added previously missing support for many socket-related functions 1013 : under win32. Thanks to David Hansel 1014 : for the voluminous patch! 1015 : 1016 : (I have not tested this patch under win32 yet.) 1017 : 1018 : mblume 1490 Here is David's e-mail: 1019 : 1020 : Hi, 1021 : 1022 : Attached to this email you find a diff against sml/nj 110.45 1023 : that will enable socket support under Windows. 1024 : 1025 : To apply the patch (using unix or cygwin) 1026 : 1) gunzip runtime.diff.gz 1027 : 2) "cd" into "src/runtime" in the source tree of a fresh 1028 : 110.45 installation. 1029 : 3) patch -p 1 < [your/path/to]runtime.diff 1030 : 1031 : The code compiles fine but has NOT yet been extensively tested. 1032 : I only ran a few tests for basic socket client functionality 1033 : (which worked fine). Especially the functions that use ioctl 1034 : are not tested at all and might not work (see below). 1035 : 1036 : I implemented this since we want to move to a newer version of sml/nj 1037 : but need socket support in order to use it. This is the first time I 1038 : even had a look at the sml/nj source, so please review my changes 1039 : before making this part of the distribution! Here are a few issues 1040 : that I think might be better for someone to solve who is more 1041 : familiar with the sml/nj source (and socket programming): 1042 : 1043 : - getnetbyaddr.c and getnetbyname.c will raise a "not implemented" 1044 : exception since I could not figure out what the windows equivalent 1045 : of these functions is 1046 : 1047 : - In sockets-osdep.h there are a some #include statements that are 1048 : only used in a few files that include sockets-osdep.h 1049 : 1050 : - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but 1051 : does not process its return value since I don't know how to 1052 : report an error upwards. 1053 : 1054 : - It would probably be good to have a call to WSACleanup() when 1055 : the library is unloaded (if there is such a possibility). 1056 : Otherwise I think Windows will take care of this automatically 1057 : when the process finishes. 1058 : 1059 : - I used ioctlsocket() as a replacement for ioctl() but I have 1060 : no idea if that is actually the proper replacement on Windows. 1061 : 1062 : - All these issues are marked in the code by "FIXME" comments. 1063 : 1064 : We use sml/nj extensively in our products and are quite happy 1065 : with it. I hope this contribution will help you. 1066 : 1067 : Keep up the good work! 1068 : 1069 : David 1070 : 1071 : mblume 1489 ---------------------------------------------------------------------- 1072 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1073 : mblume 1488 Date: 2004/05/11 14:20:00 CDT 1074 : Tag: blume-20040511-installml 1075 : Description: 1076 : 1077 : Fixed two bugs in installml script. (Thanks to Vesa A. Norrman for 1078 : the patch.) 1079 : 1080 : ---------------------------------------------------------------------- 1081 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1082 : mblume 1487 Date: 2004/05/11 14:05:00 CDT 1083 : Tag: blume-20040511-nlffi-netbsd 1084 : Description: 1085 : 1086 : Added support for nlffi under netbsd. (Thanks to Vesa A. Norrman for 1087 : the patch.) 1088 : 1089 : ---------------------------------------------------------------------- 1090 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1091 : mblume 1486 Date: 2004/05/11 12:05:00 CDT 1092 : Tag: blume-20040511-exports 1093 : Description: 1094 : 1095 : As per request by Adam Chlipala , extended various 1096 : export lists in compiler-related .cm-files. 1097 : 1098 : ---------------------------------------------------------------------- 1099 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1100 : mblume 1485 Date: 2004/05/11 11:35:00 CDT 1101 : Tag: blume-20040511-allsource 1102 : Description: 1103 : 1104 : The installer now honors the "src-smlnj" target again, although its meaning 1105 : has changed from "all sources required for the compiler" to "all sources 1106 : the installer knows about". In other words, if you enable "src-smlnj" 1107 : in the "targets" file, then the installer will pull in sources for 1108 : everything. (Notice that this refers to source code only. Compiled 1109 : code is still only installed for modules that were requested explicitly 1110 : or which are required for other modules that were requested explicitly.) 1111 : 1112 : ---------------------------------------------------------------------- 1113 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1114 : mblume 1482 Date: 2004/04/23 17:40:00 CDT 1115 : Tag: blume-20040423-ieee-scan 1116 : Description: 1117 : 1118 : Fixed IEEEReal.scan (and .fromString) so that if there is an overflow 1119 : in the exponent calculation we get INF or ZERO (depending on the mantissa 1120 : and the sign of the exponent). 1121 : 1122 : ---------------------------------------------------------------------- 1123 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1124 : mblume 1480 Date: 2004/04/23 10:40:00 CDT 1125 : Tag: blume-20040423-ml-build 1126 : Description: 1127 : 1128 : The ml-build script now terminates with a non-0 status when something 1129 : goes wrong. 1130 : 1131 : ---------------------------------------------------------------------- 1132 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1133 : mblume 1479 Date: 2004/04/22 16:35:00 CDT 1134 : Tag: blume-20040422-Option 1135 : Description: 1136 : 1137 : Made exception Option to be the same as exception Option.Option 1138 : (as it should be). 1139 : 1140 : ---------------------------------------------------------------------- 1141 : allenleung 1468 Name: Allen Leung (leunga (at) reservoir (dot) com) 1142 : Date: 2004/03/19 14:40:00 EST 1143 : Tag: leunga-20040319-cygwin-nlffi 1144 : Description: 1145 : 1146 : Fixed the runtime so that ml-nlffi-lib runs on the cygwin version 1147 : of SML/NJ. The problem is that 1148 : 1149 : lib = dlopen(NULL, ...) 1150 : f = dlsym(lib, "malloc"); 1151 : 1152 : does not work on Windows unless we explicitly export symbols 1153 : such as 'malloc' during linking. We fixed this by explicitly 1154 : exporting the required symbols with the magic gcc incantation: 1155 : 1156 : -Wl,--export-all cygwin.def 1157 : 1158 : where cygwin.def is a file containing all the symbols that 1159 : we wish to export. 1160 : 1161 : I suspect this is a Windows problem and we'll have to 1162 : do the same (somehow with windows compilers) when 1163 : we build the native win32 version with the system 1164 : calls LoadLibrary/GetProcAddress. 1165 : 1166 : ---------------------------------------------------------------------- 1167 : mblume 1384 Name: Matthias Blume (blume (at) tti - c (dot) org) 1168 : mblume 1465 Date: 2004/03/04 16:35:00 CST 1169 : Tag: blume-20040304-intinf-fmt 1170 : Description: 1171 : 1172 : Fixed problem with IntInf.fmt (sign would show up on the right instead 1173 : of on the left for BIN, OCT, and HEX). 1174 : 1175 : ---------------------------------------------------------------------- 1176 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1177 : mblume 1464 Date: 2004/03/04 11:25:00 CST 1178 : Tag: blume-20040304-symlinks 1179 : Description: 1180 : 1181 : Fixed problem with installer script (unix only) where bin/ml-yacc and 1182 : friends pointed (via symlinks) to absolute locations instead of just 1183 : .run-sml. This was reported by Vesa A Norrman. 1184 : 1185 : ---------------------------------------------------------------------- 1186 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1187 : mblume 1455 Date: 2004/02/13 14:50:00 CST 1188 : Tag: Release_110_45 1189 : Description: 1190 : 1191 : New working version (110.45). New bootfiles. 1192 : 1193 : ---------------------------------------------------------------------- 1194 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1195 : mblume 1448 Date: 2004/01/26 15:15:15 CST 1196 : Tag: blume-20040126-toplevel 1197 : Description: 1198 : 1199 : Improved handling of exceptions at the interactive toplevel. 1200 : 1201 : ---------------------------------------------------------------------- 1202 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1203 : Date: 2004/01/26 11:25:00 1204 : mblume 1446 Tag: blume-20040126-app 1205 : Description: 1206 : 1207 : Type of top-level "app" corrected. 1208 : mblume 1447 Added code for setting vp_limitPtrMask to Win32-specific runtime. 1209 : mblume 1446 1210 : ---------------------------------------------------------------------- 1211 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1212 : mblume 1427 Date: 2003/11/18 17:10 CST 1213 : Tag: blume-20031118-basis-fiddle 1214 : Description: 1215 : 1216 : - changed Timer interface to what might become the spec 1217 : - POSIX_FLAGS -> BIT_FLAGS according to spec 1218 : - some other minor discrepancies wrt. spec eliminated 1219 : 1220 : ---------------------------------------------------------------------- 1221 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1222 : mblume 1414 Date: 2003/11/06 12:00:00 CST 1223 : Tag: Release_110_44 1224 : Description: 1225 : 1226 : New working version (110.44). New bootfiles. 1227 : 1228 : ---------------------------------------------------------------------- 1229 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1230 : mblume 1411 Date: 2003/11/04 11:50:00 CST 1231 : Tag: blume-20031104-move-libraries 1232 : Description: 1233 : 1234 : Eliminated the "dont_move_libraries" directive in config/targets. 1235 : (The mechanism was broken and could not be fixed easily. Moreover, 1236 : there does not seem to be any reason not to move all libraries into 1237 : lib during installation. I originally implemented this directive as a 1238 : backward-compatibility feature when I first introduced the new CM. 1239 : Now that things have been stable for a long time and going back to the 1240 : old CM is not an option, there is no reason to keep it around.) 1241 : 1242 : ---------------------------------------------------------------------- 1243 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1244 : mblume 1409 Date: 2003/11/03 16:00:00 CST 1245 : Tag: blume-20031103-installdir 1246 : Description: 1247 : 1248 : Made installer honor INSTALLDIR variable again. (Thanks to Chris 1249 : Richards for pointing out the problem and providing the solution.) 1250 : 1251 : ---------------------------------------------------------------------- 1252 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1253 : mblume 1400 Date: 2003/10/01 17:05:00 CDT 1254 : Tag: blume-20031001-lal-mlrisc 1255 : Description: 1256 : 1257 : MLRISC bug fix from Lal. 1258 : 1259 : ---------------------------------------------------------------------- 1260 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1261 : mblume 1399 Date: 2003/09/30 16:10:00 CDT 1262 : Tag: blume-20030930-primio-bat 1263 : mblume 1398 Description: 1264 : 1265 : 1. Added openVector, nullRd, and nullWr to PRIM_IO. 1266 : 2. Improved .bat files (for Win32 port) to make things work under Win95. 1267 : (thanks to Aaron S. Hawley for this one) 1268 : 1269 : ---------------------------------------------------------------------- 1270 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1271 : mblume 1395 Date: 2003/09/26 16:05:00 CDT 1272 : Tag: blume-20030926-wrappriv 1273 : Description: 1274 : 1275 : Added missing wrapper for privilege "primitive" in$smlnj/viscomp/core.cm. 1276 : 1277 : ---------------------------------------------------------------------- 1278 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1279 : mblume 1394 Date: 2003/09/26 15:00:00 CDT 1280 : Tag: blume-20030926-110_43_3 1281 : Description: 1282 : 1283 : - additional cleanup 1284 : - version number bump, NEW BOOTFILES 1285 : 1286 : ---------------------------------------------------------------------- 1287 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1288 : mblume 1393 Date: 2003/09/26 12:00:00 CDT 1289 : Tag: blume-20030926-ppautoload 1290 : Description: 1291 : 1292 : I modified the read-eval-print loop so that the autoloader gets 1293 : invoked whenever the prettyprinter tries to look up a symbol that 1294 : is not currently defined in the toplevel environment but which 1295 : appears in CM's autoload registry. As a result, we see far fewer of 1296 : those ?.Foo.Bar.xxx names in the prettyprinter's output. 1297 : 1298 : In addition to this I tried to clean up some pieces of the Basis 1299 : implementation (e.g., Socket, Word8Array) in order to prevent other 1300 : instances of these ?.Foo.Bar.xxx names from being printed. 1301 : 1302 : The mechanism that picks names for types still needs some work, though. 1303 : (Right now it seems that if there is a type A.t which is defined to 1304 : mblume 1394 be B.u, but B is unavailable at toplevel, then A.t gets printed as 1305 : "?.B.u" although the perhaps more sensible solution would be to use 1306 : mblume 1393 "A.t" in this case. In other words, the prettyprinter should follow 1307 : a chain of DEFtycs not farther than there are corresponding toplevel 1308 : names in the current environment.) 1309 : 1310 : ---------------------------------------------------------------------- 1311 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1312 : mblume 1391 Date: 2003/09/24 16:31:00 CDT 1313 : Tag: blume-20030924-installer 1314 : Description: 1315 : 1316 : Another installer tweak: All the ML code for the installer is now 1317 : compiled during CMB.make and put into a little library called 1318 : $smlnj/installer.cm. The installation then simply invokes 1319 : 1320 : sml -m$smlnj/installer.cm 1321 : 1322 : and everything happens automagically. 1323 : 1324 : Win32: ML code senses value of environment variable SMLNJ_HOME. 1325 : Unix: ML code senses values of environment variables ROOT, CONFIGDIR, 1326 : and BINDIR. 1327 : 1328 : The new scheme guarantees that the ML code responsible for the installation 1329 : is in sync with the APIs of the main system. Also, the installer is 1330 : somewhat faster because the installer script is precompiled. 1331 : 1332 : ---------------------------------------------------------------------- 1333 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1334 : mblume 1390 Date: 2003/09/24 15:35:00 CDT 1335 : Tag: blume-20030924-synsock 1336 : Description: 1337 : 1338 : Added a signature SYNCHRONOUS_SOCKET to basis.cm. This is like SOCKET 1339 : but excludes all non-blocking operations. Defined SOCKET (in Basis) 1340 : and CML_SOCKET in terms of SYNCHRONOUS_SOCKET. Removed superfluous 1341 : implementations of non-blocking operations from CML's Socket 1342 : structure. 1343 : 1344 : ---------------------------------------------------------------------- 1345 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1346 : mblume 1389 Date: 2003/09/24 15:10:05 CDT 1347 : Tag: blume-20030924-sockets 1348 : Description: 1349 : 1350 : 1. Fixed SOCKET API and implementation to match Basis spec. 1351 : This required changing the internal representation of sockets to one 1352 : that remembers (for each socket file descriptor) whether it is currently 1353 : blocking or non-blocking. This state is maintained lazily (i.e., a system 1354 : call is made only if the state actually needs to change). 1355 : 1356 : 2. OS-specific details of sockets were moved into separate files, thus 1357 : making it possible to unify the bulk of the socket implementations 1358 : between Unix and Win32. 1359 : 1360 : 3. CML's socket API changed accordingly. 1361 : (Note that we need to remove non-blocking functions from this API 1362 : since they are redundant in the case of CML!) 1363 : 1364 : 4. CML's socket implementation now makes use of non-blocking functions 1365 : provided by Basis, thus removing all OS-dependent code from this part 1366 : of CML. 1367 : 1368 : 5. Changed Real64.precision from 52 to 53. Minor cleanup in Real64 code. 1369 : 1370 : ---------------------------------------------------------------------- 1371 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1372 : mblume 1387 Date: 2003/09/22 12:10:00 CDT 1373 : mblume 1388 Tag: blume-20030922-110_43_2 1374 : mblume 1387 Description: 1375 : 1376 : Made a new interim version and bootfiles for developer's bootstrapping 1377 : convenience. 1378 : 1379 : 110.43.2 -- NEW BOOTFILES 1380 : 1381 : ---------------------------------------------------------------------- 1382 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1383 : mblume 1385 Date: 2003/09/19 15:55:00 CDT 1384 : Tag: blume-20030919-cmdir 1385 : Description: 1386 : 1387 : 1. new-install.sh -> install.sh 1388 : 2. changed default CM "metadata" directory name to ".cm" (instead of "CM") 1389 : 3. tweaked installer so that another name instead of .cm can be chosen 1390 : at install time (by setting the CM_DIR_ARC environment variable 1391 : during installation); once installation is complete, the name is 1392 : fixed 1393 : 1394 : ---------------------------------------------------------------------- 1395 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1396 : mblume 1384 Date: 2003/09/18 16:00:00 CDT 1397 : Tag: blume-20030918-110_43_1 1398 : Description: 1399 : 1400 : Made a new interim version and bootfiles for developer's bootstrapping 1401 : convenience. 1402 : 1403 : 110.43.1 -- NEW BOOTFILES 1404 : 1405 : ---------------------------------------------------------------------- 1406 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1407 : mblume 1383 Date: 2003/09/18 15:20:00 CDT 1408 : Tag: blume-20030918-misc 1409 : Description: 1410 : 1411 : 1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as 1412 : this seems to be controversial at the moment) and used those in 1413 : Posix.ProcEnv.times. 1414 : 1415 : 2. Added Time.{from,to}Nanoseconds to Time. 1416 : 1417 : 3. Improved Real.{from,to}LargeInt by avoiding needless calculations. 1418 : For example, fromLargeInt never needs to look at more than 3 "big 1419 : digits" to get its 53 bits of precision. 1420 : 1421 : ---------------------------------------------------------------------- 1422 : mblume 1384 Name: Matthias Blume (blume (at) tti - c (dot) org) 1423 : mblume 1381 Date: 2003/09/17 16:30:00 CDT 1424 : Tag: blume-20030917-real32-slices 1425 : Description: 1426 : 1427 : Added an entry to the primitive environment 1428 : (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and 1429 : added code to compiler/CodeGen/main/mlriscGen.sml to implement it. 1430 : 1431 : Removed some of the "magic" constants in real64.sml and replaced them 1432 : with code that generates these values from their corresponding 1433 : integer counterparts. 1434 : 1435 : Made all(?) the slice-related changes to the Basis and made everything 1436 : compile again... 1437 : 1438 : ---------------------------------------------------------------------- 1439 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org) 1440 : mblume 1375 Date: 2003/09/15 17:45:00 CDT 1441 : Tag: blume-20030915-rbase 1442 : Description: 1443 : 1444 : Fixed bug in Real.fromLargeInt. 1445 : 1446 : ---------------------------------------------------------------------- 1447 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1448 : mblume 1369 Date: 2003/09/13 18:11:00 CDT 1449 : Tag: blume-20030913-libinstall 1450 : Description: 1451 : 1452 : Minor bugfix in config/libinstall (set anchor with path to 1453 : standalone tool after installing it, otherwise libraries that 1454 : need ml-lex or ml-yacc won't compile the first time the installer 1455 : runs). 1456 : 1457 : ---------------------------------------------------------------------- 1458 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1459 : mblume 1368 Date: 2003/09/12 11:45:00 CDT 1460 : Tag: blume-20030912-various 1461 : Description: 1462 : 1463 : - fixed bug in Real.toLargeInt 1464 : - fixed bug in Posix.ProcEnv.times 1465 : - changed inputLine functions to return an option 1466 : - minor installer improvements / bugfixes 1467 : - changed default @SMLalloc parameter for x86/celeron to 64k 1468 : 1469 : ---------------------------------------------------------------------- 1470 : Name: Matthias Blume (blume (at) tti - c (dot) org) 1471 : mblume 1356 Date: 2003/09/09 22:00:00 CDT 1472 : Tag: Release_110_43 1473 : Description: 1474 : 1475 : New working release 110.43. New bootfiles. 1476 : 1477 : ---------------------------------------------------------------------- 1478 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org) 1479 : mblume 1354 Date: 2003/09/09 19:20:00 CDT 1480 : Tag: blume-20030909-installer 1481 : Description: 1482 : 1483 : Rewrote large parts of config/install.sh in SML (config/libinstall.sml). 1484 : Modified config/install.bat to take advantage of it. Also modified 1485 : config/install.sh (and called it config/new-install.sh) to take advantage 1486 : of it on Unix systems. (The SML code is (supposed to be) platform- 1487 : independent.) 1488 : 1489 : The installer can now install everything under Win32 1490 : as well as under *nix as long as it compiles. 1491 : 1492 : Other changes: 1493 : 1494 : - made CML compile again under Win32 1495 : - made eXene compile under Win32 (by providing a fake structure UnixSock 1496 : and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv) 1497 : - fixed a bug in nowhere: it assumed that type OS.Process.status is the 1498 : same as type int; under Win32 it isn't 1499 : - fixed some slice-related problems in the win32-specific parts of CML 1500 : - added a functor argument "sameVol" to os-path-fn.sml in the Basis 1501 : (under Win32, the volume name is case-insensitive, and the 1502 : OS.Path code compares volume names for equality) 1503 : 1504 : ---------------------------------------------------------------------- 1505 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org) 1506 : mblume 1353 Date: 2003/09/08 11:55:00 CDT 1507 : Tag: blume-20030908-fullpath 1508 : Description: 1509 : 1510 : Made Win32 version of OS.FileSys.fullPath return current directory 1511 : when given an empty string. This is what the spec says, and incidentally, 1512 : CM depends on it. (CM otherwise goes into an infinite loop in certain 1513 : cases when presented with the name of a non-existing .cm file.) 1514 : 1515 : ---------------------------------------------------------------------- 1516 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org) 1517 : mblume 1350 Date: 2003/09/04 16:30:00 CDT 1518 : mblume 1352 Tag: blume-20030905-slices-etc 1519 : mblume 1350 Description: 1520 : 1521 : 1. Changed interface to vectors and arrays in Basis to match 1522 : (draft) Basis spec. 1523 : 2. Added signatures and implementations of slices according to 1524 : Basis spec. 1525 : 3. Edited source code throughout the system to make it compile again 1526 : under 1. and 2. (In some cases code had to be added to have it 1527 : match the new signatures.) 1528 : 4. MLRISC should be backward-compatible: the copies of the originals 1529 : of files that needed to change under 3. were retained, the .cm files 1530 : check the compiler version number and use old versions when 1531 : appropriate. 1532 : 5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to 1533 : dirstream -> string option (in accordance with Basis spec). 1534 : 6. When generating code that counts lines, ml-lex used function 1535 : CharVector.foldli, taking advantage of its old interface. 1536 : This has been replaced with the corresponding code from 1537 : CharVectorSlice. (html-lex must be re-lexed!) 1538 : 7. BitArray in smlnj-lib/Util has been extended/modified to match the 1539 : new MONO_ARRAY signature. (Do we need BitArraySlice?) 1540 : 8. Removed temporary additions (fromInternal, toInternal) from the 1541 : (now obsolete) IntInf in smlnj-lib/Util. 1542 : 9. Cleaned up structure Byte. 1543 : 10. Added localOffset, scan, and fromString to Date (according to spec). 1544 : Cleaned/corrected implementation of Date. 1545 : (Still need to check for correctness; implement better canonicalizeDate.) 1546 : 11. Added "scan" to signature IEEE_REAL. 1547 : 12. Some improvements to IntInf [in particular: efficiency-hack for 1548 : mod and rem when second operand is 2 (for parity checks).] 1549 : 13. Changed representation of type Time.time, using a single IntInf.int 1550 : value counting microseconds. This considerably simplified the 1551 : implementation of structure Time. We now support negative time 1552 : values; scan and fromString handle signs. 1553 : 14. Functor PrimIO now takes two additional arguments (VectorSlice and 1554 : ArraySlice). 1555 : 1556 : ---------------------------------------------------------------------- 1557 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org) 1558 : mblume 1347 Date: 2003/08/28 17:00:00 CDT 1559 : Tag: blume-20030828-intinf 1560 : Description: 1561 : 1562 : This is a major update which comes with a version number bump 1563 : (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES, 1564 : and an implementation of IntInf in the Basis. 1565 : 1566 : There are a fairly large number of related changes and updates throughout 1567 : the system: 1568 : 1569 : Basis: 1570 : - Implemented IntInf. 1571 : - Made LargeInt a projection of IntInf (by filtering through INTEGER). 1572 : mblume 1348 - Added some missing Real64 operations, most notably Real.toLargeInt. 1573 : mblume 1347 - Added FixedInt as a synonym for Int32. 1574 : 1575 : compiler: 1576 : * Added support for a built-in intinf type. 1577 : - literals 1578 : - pattern matching 1579 : - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.) 1580 : - overloading on literals and operations 1581 : 1582 : This required adding a primitive type intinf, some additional 1583 : primops, and implementations for several non-trivial intinf 1584 : operations in Core. (The intinf type is completely abstract 1585 : to the compiler; all operations get delegated back to the Core.) 1586 : 1587 : * Intinf equality is handled by polyequal. However, the compiler 1588 : does not print its usual warning in this case (since polyequal 1589 : is the right thing to do there). 1590 : 1591 : * Improved the organization of structure InlineT. 1592 : 1593 : * A word about conversion primops: 1594 : If conversions involving intinf do not cancel out during 1595 : mblume 1348 CPS contract, then the compiler must insert calls to Core functions. 1596 : mblume 1347 Since all core access must be resolved already during the FLINT 1597 : translate phase, it would be too late a the time of CPS contract 1598 : mblume 1348 to add new Core calls. For this reason, conversion primops 1599 : mblume 1347 for intinf carry two arguments: 1. the numeric argument that 1600 : they are supposed to convert, and 2. the Core function that 1601 : can help with this conversion if necessary. If CPS contract 1602 : mblume 1348 eliminates a primop, then the associated Core function becomes 1603 : dead and goes away. Intinf conversion primops that do not get 1604 : eliminated by CPS contract get rewritten into calls of their 1605 : core functions by a separate, new phase. 1606 : mblume 1347 1607 : interactive system: 1608 : - Control.Print.intinfDepth controls max length of intinf constants 1609 : being printed. (Analogous to Control.Print.stringDepth.) 1610 : - Cleanup in printutil and pputil: got rid of unused stuff and 1611 : duplicates; replaced some of the code with code that makes better 1612 : use of library functionality. 1613 : 1614 : CM: 1615 : Bugfix: parse-errors in init group (system/smlnj/init/init.cmi) 1616 : are no longer silent. 1617 : 1618 : CKIT: 1619 : Fixed mismatched uses of Int32 and LargeInt. I always decided 1620 : in favor of LargeInt -- which is now the same as IntInf. 1621 : CKIT-knowledgable people should check whether this is what's 1622 : intended and otherwise change things back to using Int32 or 1623 : FixedInt. 1624 : 1625 : Throughout the code: 1626 : mblume 1348 Started using IntInf.int literals and built-in operations 1627 : mblume 1347 (e.g., comparison with 0) where this seems appropriate. 1628 : 1629 : 1630 : ---------------------------------------------------------------------- 1631 : macqueen 1343 Name: Dave MacQueen (dbm@cs.uchicago.edu) 1632 : Date: 2003/08/13 11:36:00 CDT 1633 : Tag: dbm-20030813-mcz-merge1 1634 : Description: 1635 : 1636 : Merging changes from the mcz-branch development branch into trunk. 1637 : These changes involve replacement of the emulated old prettyprinter 1638 : interface with direct use of the SML/NJ Lib PP library, and fixing 1639 : of a couple of bugs (895, 1186) relating to error messages. A new 1640 : prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml}) 1641 : has been added. 1642 : 1643 : ---------------------------------------------------------------------- 1644 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org) 1645 : mblume 1342 Date: 2003/08/11 15:45:00 CDT 1646 : Tag: blume-20030811-windows 1647 : Description: 1648 : 1649 : Version number bumped to 110.42.9. NEW BOOTFILES!!! 1650 : 1651 : http://smlnj.cs.uchicago.edu/dist/working/110.42.9/ 1652 : 1653 : This patch restores SML/NJ's ability to run under win32. There are a 1654 : number of changes, including fixes for several bugs that had gone 1655 : unnoticed until now: 1656 : 1657 : - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when 1658 : trying to run the win32 version from within cygwin.) 1659 : 1660 : - There are a number of simple .bat scripts that substitute for their 1661 : corresponding Unix shell-scripts. (See below.) 1662 : 1663 : - The internals of ml-build have been modified slightly. The main 1664 : difference is that instead of calling ".link-sml" (or link-sml.bat) 1665 : using OS.Process.system, the ML process delegates this task back 1666 : to the script. Otherwise problems arise in mixed environments such 1667 : as Cygwin where scripts look and work like Unix scripts, but 1668 : where OS.Process.system cannot run them. 1669 : 1670 : - In CM, the srcpath pickler used native pathname syntax -- which 1671 : is incorrect in the case of cross-compilation. The new pickle format 1672 : is independent of platform-specific naming conventions. 1673 : 1674 : - Path configuration files (such as lib/pathconfig) can now choose 1675 : between native and standard syntax. Placing a line of the form 1676 : 1677 : standard! 1678 : 1679 : into the file causes all subsequent paths to be interpreted using 1680 : CM standard pathname syntax (= Unix conventions); a line 1681 : 1682 : native! 1683 : 1684 : switches back to native style. This was needed so that 1685 : path config files can be written portably, see src/system/pathconfig. 1686 : 1687 : - Runtime system: 1688 : 1689 : - win32-filesys.c: get_file_time and set_file_time now 1690 : access modification time, not creation time. 1691 : 1692 : - I/O code made aware of new array representation. 1693 : 1694 : - Bug fixes in X86.prim.masm. 1695 : 1696 : - src/system/makeml made aware of win32. (For use under cygwin 1697 : and other Unix-environments for windows.) 1698 : 1699 : - In Basis, fixed off-by-one error in win32-io.sml (function vecF) 1700 : which caused BinIO.inputAll to fail consistently. 1701 : 1702 : .bat scripts: 1703 : 1704 : Windows .bat scripts assume that SMLNJ_HOME is defined. 1705 : 1706 : - sml.bat, ml-yacc.bat, ml-lex.bat: Driver scripts for standalone 1707 : applications (sml, ml-yacc, ml-lex). 1708 : - ml-build.bat: analogous to ml-build. 1709 : - config\install.bat: Analogous to config/install.sh. This requires 1710 : that SMLNJ_HOME is set and that Microsoft Visual C is ready to use. 1711 : (nmake etc. must be on the path, and vcvars32 must have been run.) 1712 : Moreover, sources for ml-lex and ml-yacc need to exist under src, 1713 : and the bootfile hierarchy must have been unpacked under 1714 : sml.boot.x86-win32. 1715 : The script is very primitive and does a poor job at error checking. 1716 : It only installs the base system, ml-lex, and ml-yacc. No other 1717 : libraries are being installed (i.e., you get only those that 1718 : are part of the compiler.) 1719 : - link-sml.bat: analogous to .link-sml, but not currently used 1720 : 1721 : Unrelated bug fixes: 1722 : 1723 : - ml-nlffigen now exports structures ST_* corresponding to incomplete 1724 : types. 1725 : - Added getDevice to PP/src/pp-debug-fn.sml. (Would not compile 1726 : otherwise.) 1727 : 1728 : ---------------------------------------------------------------------- 1729 : macqueen 1337 Name: Dave MacQueen (macqueen@cs.uchicago.edu) 1730 : Date: 2003/06/17 1731 : Tag: macqueen-20030617-bug895 1732 : Description: 1733 : 1734 : Modified compiler/Elaborator/print/pptype.sml to fix bug 895. 1735 : Tag will be used for new development branch (mcz-branch) for 1736 : use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago. 1737 : 1738 : ---------------------------------------------------------------------- 1739 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org) 1740 : mblume 1335 Date: 2003/05/27 16:55:00 CDT 1741 : Tag: blume-20030527-polyeq 1742 : Description: 1743 : 1744 : Tried to eliminated most cases of polymorphic equality. 1745 : 1746 : ---------------------------------------------------------------------- 1747 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org) 1748 : mblume 1334 Date: 2003/05/21 17:45:00 CDT 1749 : Tag: blume-20030517-complete 1750 : Description: 1751 : 1752 : Two changes: 1753 : 1754 : 1. Added a flag for controlling whether non-exhaustive bindings will 1755 : be treated as errors (default is false). 1756 : 2. Cleaned up the *entire* source tree so that CMB.make goes through 1757 : without a single non-exhaustive match- or bind warning. 1758 : 1759 : ---------------------------------------------------------------------- 1760 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org) 1761 : mblume 1332 Date: 2003/05/17 10:20:00 CDT 1762 : Tag: blume-20030517-absyn 1763 : Description: 1764 : 1765 : 1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn. 1766 : 1767 : This mainly affects the quality of error messages. However, some 1768 : of the code is now more straightforward than before. (Treatment of 1769 : the above four constructs in translate.sml is much simpler than 1770 : the "macro-expansion" that was going on before. Plus, the mach- 1771 : compiler no longer gets invoked just to be able to compile an 1772 : if-expression.) 1773 : 1774 : 2. The ErrorMsg.Error exception is now caught and absorbed by the 1775 : interactive loop. 1776 : 1777 : ---------------------------------------------------------------------- 1778 : allenleung 1327 Name: Allen Leung 1779 : Date: 2003/05/16 13:05:00 CDT 1780 : Tag: leunga-20030516-cygwin-runtime 1781 : Description: 1782 : 1783 : Ported the runtime system to cygwin, which uses the unix 1784 : x86-unix bin files. Missing/buggy features: 1785 : 1786 : o getnetbyname, getnetbyaddr: these functions seem to be missing in 1787 : the Cygwin library. 1788 : o Ctrl-C handling may be flaky. 1789 : o Windows system calls and Windows I/O are not supported. 1790 : 1791 : A new set of binfiles is located at: 1792 : 1793 : http://www.dorsai.org/~leunga/boot.x86-unix.tgz 1794 : 1795 : This is only needed for bootstrapping the cygwin version of smlnj. 1796 : Other x86 versions can use the existing binfiles. 1797 : 1798 : ---------------------------------------------------------------------- 1799 : mblume 1317 Name: Matthias Blume 1800 : Date: 2003/04/08 15:42:00 CDT 1801 : Tag: blume-20030408-listpair 1802 : Description: 1803 : 1804 : 1. Added a target 'mlrisc' to installer. 1805 : 1806 : 2. Added missing elements to structure ListPair. 1807 : 1808 : ---------------------------------------------------------------------- 1809 : leunga 1296 Name: Allen Leung 1810 : Date: 2003/01/07 10:40:00 EST 1811 : Tag: leunga-20030107-int-rem 1812 : Description: 1813 : 1814 : Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86. 1815 : The arguments to the SUBL instruction were swapped. 1816 : 1817 : ---------------------------------------------------------------------- 1818 : blume 1229 Name: Matthias Blume 1819 : blume 1294 Date: 2002/12/12 16:25:00 EST 1820 : Tag: blume-20021212-risc-ra 1821 : Description: 1822 : 1823 : Fixed a serious bug in the rewrite code for FP spilling/reloading that 1824 : sent the RA into an infinite loop when floating point registers get 1825 : spilled. (Because of this bug, e.g., nucleic stopped compiling between 1826 : 110.37 and 110.38.) 1827 : There was another set of potential problems related to the handling of 1828 : MLRISC annotations (but those did not yet cause real problems, apparently). 1829 : 1830 : ---------------------------------------------------------------------- 1831 : Name: Matthias Blume 1832 : blume 1291 Date: 2002/12/06 22:40:00 EST 1833 : Tag: blume-20021206-cm-fileid 1834 : Description: 1835 : 1836 : Added a call of SrcPath.sync at the beginning of Parse.parse (in CM). 1837 : This fixes the problem of CM getting confused by files that suddenly 1838 : change their identity (e.g., by getting unlinked and recreated by some 1839 : text editor such as vi). There might be a better/cheaper/cleaner way 1840 : of doing this, but for now this will have to do. 1841 : 1842 : ---------------------------------------------------------------------- 1843 : Name: Matthias Blume 1844 : blume 1288 Date: 2002/10/28 09:50:00 EST 1845 : Tag: blume-20021028-typecheck 1846 : Description: 1847 : 1848 : Exported structure Typecheck from $smlnj/viscomp/core.cm. 1849 : 1850 : ---------------------------------------------------------------------- 1851 : Name: Matthias Blume 1852 : blume 1282 Date: 2002/10/17 09:10:00 EDT 1853 : blume 1279 Tag: Release_110_42 1854 : Description: 1855 : 1856 : blume 1282 In good old tradition, there has been a slight hiccup so that we have 1857 : to patch 110.42 after the fact. The old release tag has been replaced 1858 : (see below). 1859 : 1860 : The change solves a problem with two competing approaches the 1861 : configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in 1862 : each other's way. 1863 : 1864 : This change only affects the runtime system code and the installer script. 1865 : (No new bootfiles.) 1866 : 1867 : ---------------------------------------------------------------------- 1868 : Name: Matthias Blume 1869 : Date: 2002/10/16 12:00:00 EDT 1870 : Tag: Release_110_42_removed 1871 : Description: 1872 : 1873 : blume 1279 New working release. New bootfiles. 1874 : 1875 : ---------------------------------------------------------------------- 1876 : Name: Matthias Blume 1877 : blume 1273 Date: 2002/10/10 13:10:00 EDT 1878 : Tag: blume-20021010-ppc-divs 1879 : Description: 1880 : 1881 : The mltree operator DIVS must be implemented with an overflow check on 1882 : the PPC because the hardware indicates divide-by-zero using "overflow" as 1883 : well. 1884 : 1885 : ---------------------------------------------------------------------- 1886 : Name: Matthias Blume 1887 : blume 1261 Date: 2002/07/23 11:45:00 1888 : Tag: blume-20020723-smlnj-home 1889 : Description: 1890 : 1891 : Sml now senses the SMLNJ_HOME environment variable. If this is set, 1892 : then the bin dir is assumed to be in$SMLNJ_HOME/bin and (unless 1893 : CM_PATHCONFIG is also set), the path configuration file is assumed 1894 : to be in $SMLNJ_HOME/lib/pathconfig. This way one can easily move 1895 : the entire tree to some other place and everything will "just work". 1896 : 1897 : (Companion commands such as ml-build and ml-makedepend also sense this 1898 : variable.) 1899 : 1900 : ---------------------------------------------------------------------- 1901 : Name: Matthias Blume 1902 : blume 1259 Date: 2002/07/12 21:19:00 EDT 1903 : Tag: blume-20020712-liveness 1904 : Description: 1905 : 1906 : Exported two useful "step" functions from liveness module (MLRISC). 1907 : 1908 : ---------------------------------------------------------------------- 1909 : Name: Matthias Blume 1910 : blume 1253 Date: 2002/07/05 16:00 EDT 1911 : Tag: Release_110_41 1912 : Description: 1913 : 1914 : New working release. New bootfiles. 1915 : 1916 : ---------------------------------------------------------------------- 1917 : Name: Matthias Blume 1918 : blume 1250 Date: 2002/07/05 10:25:00 EDT 1919 : Tag: blume-20020705-btimp 1920 : Description: 1921 : 1922 : Exported structure BTImp from$smlnj/viscomp/debugprof.cm so that 1923 : other clients can set up backtracing support. 1924 : 1925 : ---------------------------------------------------------------------- 1926 : Name: Matthias Blume 1927 : blume 1249 Date: 2002/06/25 17:23:00 EDT 1928 : Tag: blume-20020625-fpmax 1929 : Description: 1930 : 1931 : Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point 1932 : case. (The sense of the isNaN test was reversed -- which made min and 1933 : max always return their first argument.) 1934 : 1935 : ---------------------------------------------------------------------- 1936 : Name: Matthias Blume 1937 : blume 1240 Date: 2002/06/11 1938 : Tag: blume-20020611-unixpath 1939 : Description: 1940 : 1941 : Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch. 1942 : 1943 : ---------------------------------------------------------------------- 1944 : Name: Matthias Blume 1945 : blume 1238 Date: 2002/06/10 16:35:00 EDT 1946 : Tag: blume-20020610-ieeereal 1947 : Description: 1948 : 1949 : I back-ported my implementation of IEEEReal.fromString from the 1950 : idlbasis-devel branch so that we can test it. 1951 : 1952 : Another small change is that ppDec tries to give more information 1953 : than just "" in the case of functors. However, this code is 1954 : broken in some mysterious way if the functor's body's signature 1955 : has not been declared by ascription but gets inferred from the 1956 : implementation. This needs fixing... 1957 : 1958 : ---------------------------------------------------------------------- 1959 : Name: Matthias Blume 1960 : blume 1229 Date: 2002/05/31 1961 : Tag: blume-20020531-btrace-mode 1962 : Description: 1963 : 1964 : Resurrected SMLofNJ.Internals.BTrace.mode. (It accidentally fell by 1965 : the wayside when I switched over to using Controls everywhere.) 1966 : 1967 : ---------------------------------------------------------------------- 1968 : george 1222 Name: Lal George 1969 : Date: 2002/05/23 12:21:40 EDT 1970 : Tag: george-20020523-visual-labels 1971 : Description: 1972 : 1973 : Labels are now displayed in the graphical output to make 1974 : the fall-through and target blocks obvious. 1975 : 1976 : ---------------------------------------------------------------------- 1977 : blume 1204 Name: Matthias Blume 1978 : blume 1220 Date: 2002/05/22 11:03:00 EDT 1979 : Tag: blume-20020522-shrink 1980 : Description: 1981 : 1982 : John tweaked yesterday's fix for 1131 to handle an out-of-memory 1983 : situation that comes up when allocating huge arrays. 1984 : 1985 : ---------------------------------------------------------------------- 1986 : Name: Matthias Blume 1987 : blume 1215 Date: 2002/05/21 16:00:00 EDT 1988 : Tag: Release_110_40 1989 : Description: 1990 : 1991 : New working release (110.40). New bootfiles. 1992 : 1993 : [Also: John Reppy fixed GC bug 1131.] 1994 : 1995 : ---------------------------------------------------------------------- 1996 : Name: Matthias Blume 1997 : blume 1212 Date: 2002/05/21 12:35:00 EDT 1998 : Tag: blume-20020521-cmdoc 1999 : Description: 2000 : 2001 : CM documentation update. 2002 : 2003 : ---------------------------------------------------------------------- 2004 : Name: Matthias Blume 2005 : blume 1211 Date: 2002/05/21 10:55:00 EDT 2006 : Tag: blume-20020521-misc 2007 : Description: 2008 : 2009 : - John tweaked runtime to be silent on heap export (except when 2010 : GC messages are on). 2011 : - I added a few more things (cross-compiling versions of CMB) to 2012 : config/preloads (as suggestions). 2013 : 2014 : ---------------------------------------------------------------------- 2015 : Name: Matthias Blume 2016 : blume 1208 Date: 2002/05/20 22:25:00 EDT 2017 : Tag: blume-20020520-controls 2018 : Description: 2019 : 2020 : - Added ControlUtil structure to control-lib.cm. 2021 : - Use it throughout. 2022 : - Used Controls facility to define MLRISC controls (as opposed to 2023 : registering MLRISC control ref cells with Controls after the fact) 2024 : - Fixed messed-up controls priorities. 2025 : 2026 : * Removed again all the stuff from config/preloads that one wouldn't 2027 : be able to preload at the time the initial heap image is built. 2028 : (Many libraries, e.g., CML, do not exist yet at this time. The 2029 : only libraries that can be preloaded via config/preloads are those 2030 : that come bundled with the bootfiles.) 2031 : 2032 : ---------------------------------------------------------------------- 2033 : Name: Matthias Blume 2034 : blume 1204 Date: 2002/05/20 10:59:00 EDT 2035 : Tag: blume-20020520-preloads 2036 : Description: 2037 : 2038 : Added a lot of commented-out suggestions for things to be included 2039 : in config/preloads. 2040 : 2041 : ---------------------------------------------------------------------- 2042 : leunga 1203 Name: Allen Leung 2043 : Date: 2002/05/18 14:20:00 EDT 2044 : Tag: leunga-20020518-mdl 2045 : Description: 2046 : 2047 : o Made the mdl tool stuff compile and run again. 2048 : o I've disabled all the stuff that depends on RTL specifications; they 2049 : are all badly broken anyway. 2050 : 2051 : ---------------------------------------------------------------------- 2052 : blume 1201 Name: Matthias Blume 2053 : Date: 2002/05/17 16:49:00 EDT 2054 : Tag: blume-20020517-controls 2055 : Description: 2056 : 2057 : 0. John Reppy made several modifications to the SML/NJ library. 2058 : In particular, there is a shiny new controls-lib.cm. 2059 : 2060 : 1. Pushed new controls interface through compiler so that everything 2061 : compiles again. 2062 : 2063 : 2. Added FormatComb and FORMAT_COMB to the CML version of the 2064 : SML/NJ library (so that CML compiles again). 2065 : 2066 : 3. Modified init scripts because XXX_DEFAULT environment variables 2067 : are no longer with us. (Boot-time initialization is now done 2068 : using the same environment variables that are also used for 2069 : startup-time initialization of controls.) 2070 : 2071 : ---------------------------------------------------------------------- 2072 : george 1192 Name: Lal George 2073 : Date: 2002/05/15 09:20:10 EDT 2074 : Tag: george-20020515-pseudo-op-decls 2075 : Description: 2076 : 2077 : All pseudo-ops emitted before the first segment declaration 2078 : such as TEXT, DATA, and BSS directives are assumed to be global 2079 : declarations and are emitted first in the assembly file. This is 2080 : useful in a number of situations where one has pseudo-ops that are not 2081 : specific to any segment, and also works around the constraint that one 2082 : cannot have client pseudo-ops in the TEXT segment. 2083 : 2084 : Because no segment is associated with these declarations it is 2085 : an error to allocate any space or objects before the first segment 2086 : directive and an exception will be raised. However, we cannot make 2087 : this check for client pseudo-ops. 2088 : 2089 : These top level declarations are a field in the CFG graph_info. 2090 : In theory you can continue to add to this field after the CFG has been 2091 : built -- provided you know what you are doing;-) 2092 : 2093 : ---------------------------------------------------------------------- 2094 : blume 1175 Name: Matthias Blume 2095 : blume 1190 Date: 2002/05/13 16:40:00 EDT 2096 : Tag: blume-20020513-pp-etc 2097 : Description: 2098 : 2099 : A few minor bugfixes: 2100 : 2101 : - Stopgap measure for bug recently reported by Elsa Gunter (ppDec). 2102 : (Bogus printouts for redefined bindings still occur. Compiler 2103 : bug should no longer occur now. We need to redo the prettyprinter 2104 : from scratch.) 2105 : 2106 : - CM pathname printer now also adds escape sequences for ( and ) 2107 : 2108 : - commend and docu fixes for ml-nlffi 2109 : 2110 : ---------------------------------------------------------------------- 2111 : Name: Matthias Blume 2112 : blume 1189 Date: 2002/05/10 16:40:00 EDT 2113 : Tag: blume-20020510-erg-textio 2114 : Description: 2115 : 2116 : Applied the following bugfix provided by Emden Gansner: 2117 : 2118 : Output is corrupted when outputSubstr is used rather than output. 2119 : 2120 : The problem occurs when a substring 2121 : 2122 : ss = (s, dataStart, dataLen) 2123 : 2124 : where dataStart > 0, fills a stream buffer with avail bytes left. 2125 : avail bytes of s, starting at index dataStart, are copied into the 2126 : buffer, the buffer is flushed, and then the remaining dataLen-avail 2127 : bytes of ss are copied into the beginning of the buffer. Instead of 2128 : starting this copy at index dataStart+avail in s, the current code 2129 : starts the copy at index avail. 2130 : 2131 : Fix: 2132 : In text-io-fn.sml, change line 695 from 2133 : val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0) 2134 : to 2135 : val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0) 2136 : 2137 : ---------------------------------------------------------------------- 2138 : Name: Matthias Blume 2139 : blume 1186 Date: 2002/04/12 13:55:00 EDT 2140 : Tag: blume-20020412-assyntax 2141 : Description: 2142 : 2143 : 1. Grabbed newer assyntax.h from the XFree86 project. 2144 : 2. Fiddled with how to compile X86.prim.asm without warnings. 2145 : 3. (Very) Minor cleanup in CM. 2146 : 2147 : ---------------------------------------------------------------------- 2148 : Name: Matthias Blume 2149 : blume 1185 Date: 2002/04/01 (no joke!) 17:07:00 EST 2150 : Tag: blume-20020401-x86div 2151 : Description: 2152 : 2153 : Added full support for div/mod/rem/quot on the x86, using the machine 2154 : instruction's two results (without clumsily recomputing the remainder) 2155 : directly where appropriate. 2156 : 2157 : Some more extensive power-of-two support was added to the x86 instruction 2158 : selector (avoiding expensive divs, mods, and muls where they can be 2159 : replaced with cheaper shifts and masks). However, this sort of thing 2160 : ought to be done earlier, e.g., within the CPS optimizer so that 2161 : all architectures benefit from it. 2162 : 2163 : The compiler compiles to a fixed point, but changes might be somewhat 2164 : fragile nevertheless. Please, report any strange things that you might 2165 : see wrt. div/mod/quot/rem... 2166 : 2167 : ---------------------------------------------------------------------- 2168 : Name: Matthias Blume 2169 : blume 1184 Date: 2002/03/29 17:22:00 2170 : Tag: blume-20020329-div 2171 : Description: 2172 : 2173 : Fixed my broken div/mod logic. Unfortunately, this means that the 2174 : inline code for div/mod now has one more comparison than before. 2175 : Fast paths (quotient > 0 or remainder = 0) are not affected, though. 2176 : The problem was with quotient = 0, because that alone does not tell 2177 : us which way the rounding went. One then has to look at whether 2178 : remainder and divisor have the same sign... :( 2179 : 2180 : Anyway, I replaced the bootfiles with fresh ones... 2181 : 2182 : ---------------------------------------------------------------------- 2183 : Name: Matthias Blume 2184 : blume 1183 Date: 2002/03/29 14:10:00 EST 2185 : Tag: blume-20020329-inlprims 2186 : Description: 2187 : 2188 : NEW BOOTFILES!!! Version number bumped to 110.39.3. 2189 : 2190 : Primops have changed. This means that the bin/boot-file formats have 2191 : changed as well. 2192 : 2193 : To make sure that there is no confusion, I made a new version. 2194 : 2195 : 2196 : CHANGES: 2197 : 2198 : * removed REMT from mltree (remainder should never overflow). 2199 : 2200 : * added primops to deal with divisions of all flavors to the frontend 2201 : 2202 : * handled these primops all the way through so they map to their respective 2203 : MLRISC support 2204 : 2205 : * used these primops in the implementation of Int, Int32, Word, Word32 2206 : 2207 : * removed INLDIV, INLMOD, and INLREM as they are no longer necessary 2208 : 2209 : * parameterized INLMIN, INLMAX, and INLABS by a numkind 2210 : 2211 : * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including 2212 : floating point 2213 : 2214 : * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32, 2215 : and Real (but Real.abs maps to a separate floating-point-only primop) 2216 : 2217 : 2218 : TODO items: 2219 : 2220 : * Hacked Alpha32 instruction selection, disabling the selection of REMx 2221 : instructions because the machine instruction encoder cannot handle 2222 : them. (Hppa, PPC, and Sparc instruction selection did not handle 2223 : REM in the first place, and REM is supported by the x86 machine coder.) 2224 : 2225 : * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction 2226 : selection phase. (The two can be streamlined because the hardware 2227 : delivers both quotient and remainder at the same time anyway.) 2228 : 2229 : * Think about what to do with "valOf(Int32.minInt) div ~1" and friends. 2230 : (Currently the behavior is inconsistent both across architectures and 2231 : wrt. the draft Basis spec.) 2232 : 2233 : * Word8 should eventually be handled natively, too. 2234 : 2235 : * There seems to be one serious bug in mltree-gen.sml. It appears, though, 2236 : as if there currently is no execution path that could trigger it in 2237 : SML/NJ. (The assumptions underlying functions arith and promotable do not 2238 : hold for things like multiplication and division.) 2239 : 2240 : ---------------------------------------------------------------------- 2241 : Name: Matthias Blume 2242 : blume 1181 Date: 2002/03/27 16:27:00 EST 2243 : Tag: blume-20020327-mlrisc-divisions 2244 : Description: 2245 : 2246 : Added support for all four division operations (ML's div, mod, quot, 2247 : and rem) to MLRISC. In the course of doing so, I also rationalized 2248 : the naming (no more annoying switch-around of DIV and QUOT), by 2249 : parameterizing the operation by div_rounding_mode (which can be either 2250 : DIV_TO_ZERO or DIV_TO_NEGINF). 2251 : 2252 : The generic MLTreeGen functor takes care of compiling all four 2253 : operations down to only round-to-zero div. 2254 : 2255 : Missing pieces: 2256 : 2257 : * Doing something smarter than relying on MLTreeGen on architectures 2258 : like, e.g., the x86 where hardware division delivers both quotient and 2259 : remainder at the same time. With this, the implementation of the 2260 : round-to-neginf operations could be further streamlined. 2261 : 2262 : * Remove inlining support for div/mod/rem from the frontend and replace it 2263 : with primops that get carried through to the backend. Do this for all 2264 : int and word types. 2265 : 2266 : ---------------------------------------------------------------------- 2267 : Name: Matthias Blume 2268 : blume 1180 Date: 2002/03/25 17:25:00 EST 2269 : Tag: blume-20020325-divmod 2270 : Description: 2271 : 2272 : I improved (hopefully without breaking them) the implementation of Int.div, 2273 : Int.mod, and Int.rem. For this, the code in translate.sml now takes 2274 : advantage of the following observations: 2275 : 2276 : Let q = x quot y r = x rem y 2277 : d = x div y m = x mod y 2278 : 2279 : where "quot" is the round-to-zero version of integer division that 2280 : hardware usually provides. Then we have: 2281 : 2282 : r = x - q * y where neither the * nor the - will overflow 2283 : d = if q >= 0 orelse x = q * y then q else q - 1 2284 : where neither the * nor the - will overflow 2285 : m = if q >= 0 orelse r = 0 then r else r + y 2286 : where the + will not overflow 2287 : 2288 : This results in substantial simplification of the generated code. 2289 : The following table shows the number of CFG nodes and edges generated 2290 : for 2291 : fun f (x, y) = x OPER y 2292 : (* with OPER \in div, mod, quot, rem *) 2293 : 2294 : 2295 : OPER | nodes(old) | edges(old) | nodes(new) | edges(new) 2296 : -------------------------------------------------------- 2297 : div | 24 | 39 | 12 | 16 2298 : mod | 41 | 71 | 12 | 16 2299 : quot | 8 | 10 | 8 | 10 2300 : rem | 10 | 14 | 8 | 10 2301 : 2302 : 2303 : ---------------------------------------------------------------------- 2304 : Name: Matthias Blume 2305 : blume 1179 Date: 2002/03/25 22:06:00 EST 2306 : Tag: blume-20020325-cprotobug 2307 : Description: 2308 : 2309 : Fixed a bug in cproto (c prototype decoder). 2310 : 2311 : ---------------------------------------------------------------------- 2312 : Name: Matthias Blume 2313 : blume 1178 Date: 2002/03/25 16:00:00 EST 2314 : Tag: blume-20020325-raw-primops 2315 : Description: 2316 : 2317 : I did some cleanup to Allen's new primop code and 2318 : replaced yesterday's bootfiles with new ones. 2319 : (But they are stored in the same place.) 2320 : 2321 : ---------------------------------------------------------------------- 2322 : Name: Matthias Blume 2323 : blume 1175 Date: 2002/03/24 22:40:00 EST 2324 : Tag: blume-20020324-bootfiles 2325 : Description: 2326 : 2327 : Made the bootfiles that Allen asked for. 2328 : 2329 : ---------------------------------------------------------------------- 2330 : leunga 1174 Name: Allen Leung 2331 : Date: 2002/03/23 15:50:00 EST 2332 : Tag: leunga-20020323-flint-cps-rcc-primops 2333 : Description: 2334 : 2335 : 1. Changes to FLINT primops: 2336 : 2337 : (* make a call to a C-function; 2338 : * The primop carries C function prototype information and specifies 2339 : * which of its (ML-) arguments are floating point. C prototype 2340 : * information is for use by the backend, ML information is for 2341 : * use by the CPS converter. *) 2342 : | RAW_CCALL of { c_proto: CTypes.c_proto, 2343 : ml_args: ccall_type list, 2344 : ml_res_opt: ccall_type option, 2345 : reentrant : bool 2346 : } option 2347 : (* Allocate uninitialized storage on the heap. 2348 : * The record is meant to hold short-lived C objects, i.e., they 2349 : * are not ML pointers. With the tag, the representation is 2350 : * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8) 2351 : *) 2352 : | RAW_RECORD of {tag:bool,sz:int} 2353 : and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR 2354 : 2355 : 2. These CPS primops are now overloaded: 2356 : 2357 : rawload of {kind:numkind} 2358 : rawstore of {kind:numkind} 2359 : 2360 : The one argument form is: 2361 : 2362 : rawload {kind} address 2363 : 2364 : The two argument form is: 2365 : 2366 : rawload {kind} [ml object, byte-offset] 2367 : 2368 : 3. RAW_CCALL/RCC now takes two extra arguments: 2369 : 2370 : a. The first is whether the C call is reentrant, i.e., whether 2371 : ML state should be saved and restored. 2372 : b. The second argument is a string argument specifying the name of 2373 : library and the C function. 2374 : 2375 : These things are currently not handled in the code generator, yet. 2376 : 2377 : 4. In CProto, 2378 : 2379 : An encoding type of "bool" means "ml object" and is mapped into 2380 : C prototype of PTR. Note that "bool" is different than "string", 2381 : even though "string" is also mapped into PTR, because "bool" 2382 : is assigned an CPS type of BOGt, while "string" is assigned INT32t. 2383 : 2384 : 5. Pickler/unpicker 2385 : 2386 : Changed to handle RAW_RECORD and newest RAW_CCALL 2387 : 2388 : 6. MLRiscGen, 2389 : 2390 : 1. Changed to handle the new rawload/rawstore/rawrecord operators. 2391 : 2. Code for handling C Calls has been moved to a new module CPSCCalls, 2392 : in the file CodeGen/cpscompile/cps-c-calls.sml 2393 : 2394 : 7. Added the conditional move operator 2395 : 2396 : condmove of branch 2397 : 2398 : to cps. Generation of this is still buggy so it is currently 2399 : disabled. 2400 : 2401 : ---------------------------------------------------------------------- 2402 : george 1168 Name: Lal George 2403 : Date: 2002/03/22 14:18:25 EST 2404 : george 1171 Tag: george-20020322-cps-branch-prob 2405 : george 1168 Description: 2406 : 2407 : george 1171 Implemented the Ball-Larus branch prediction-heuristics, and 2408 : george 1168 incorporated graphical viewers for control flow graphs. 2409 : 2410 : george 1171 Ball-Larus Heuristics: 2411 : george 1168 --------------------- 2412 : See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml. 2413 : 2414 : By design it uses the Dempster-Shafer theory for combining 2415 : probabilities. For example, in the function: 2416 : 2417 : fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc) 2418 : 2419 : the ball-larus heuristics predicts that the n=0 is unlikely 2420 : (OH-heuristic), and the 'then' branch is unlikely because of the 2421 : RH-heuristic -- giving the 'then' branch an even lower combined 2422 : george 1171 probability using the Dempster-Shafer theory. 2423 : george 1168 2424 : Finally, John Reppy's loop analysis in MLRISC, further lowers the 2425 : probability of the 'then' branch because of the loop in the else 2426 : branch. 2427 : 2428 : 2429 : Graphical Viewing: 2430 : ------------------ 2431 : I merely plugged in Allen's graphical viewers into the compiler. The 2432 : additional code is not much. At the top level, saying: 2433 : 2434 : Control.MLRISC.getFlag "cfg-graphical-view" := true; 2435 : 2436 : will display the graphical view of the control flow graph just before 2437 : back-patching. daVinci must be in your path for this to work. If 2438 : daVinci is not available, then the default viewer can be changed 2439 : using: 2440 : 2441 : Control.MLRISC.getString "viewer" 2442 : 2443 : which can be set to "dot" or "vcg" for the corresponding viewers. Of 2444 : course, these viewers must be in your path. 2445 : 2446 : The above will display the compilation unit at the level of clusters, 2447 : many of which are small, boring, and un-interesting. Also setting: 2448 : 2449 : Control.MLRISC.getInt "cfg-graphical-view_size" 2450 : 2451 : will display clusters that are larger than the value set by the above. 2452 : 2453 : 2454 : ---------------------------------------------------------------------- 2455 : blume 1159 Name: Matthias Blume 2456 : Date: 2002/03/21 22:20:00 EST 2457 : Tag: blume-20020321-kmp-bugfix 2458 : Description: 2459 : 2460 : Changed the interface to the KMP routine in PreString and fixed 2461 : a minor bug in one place where it was used. 2462 : 2463 : ---------------------------------------------------------------------- 2464 : leunga 1156 Name: Allen Leung 2465 : leunga 1158 Date: 2002/03/21 20:30:00 EST 2466 : Tag: leunga-20020321-cfg 2467 : Description: 2468 : 2469 : Fixed a potential problem in cfg edge splitting. 2470 : 2471 : ---------------------------------------------------------------------- 2472 : Name: Allen Leung 2473 : leunga 1156 Date: 2002/03/21 17:15:00 EST 2474 : Tag: leunga-20020321-x86-fp-cfg 2475 : Description: 2476 : 2477 : 1. Recoded the buggy parts of x86-fp. 2478 : 2479 : a. All the block reordering code has been removed. 2480 : We now depend on the block placement phases to do this work. 2481 : 2482 : b. Critical edge splitting code has been simplified and moved into the 2483 : CFG modules, as where they belong. 2484 : 2485 : Both of these were quite buggy and complex. The code is now much, much 2486 : simpler. 2487 : 2488 : 2. X86 backend. 2489 : 2490 : a. Added instructions for 64-bit support. Instruction selection for 2491 : 64-bit has not been committed, however, since that 2492 : requires changes to MLTREE which haven't been approved by 2493 : Lal and John. 2494 : 2495 : b. Added support for FUCOMI and FUCOMIP when generating code for 2496 : PentiumPro and above. We only generate these instructions in 2497 : the fast-fp mode. 2498 : 2499 : c. Added cases for JP and JNP in X86FreqProps. 2500 : 2501 : 3. CFG 2502 : 2503 : CFG now has a bunch of methods for edge splitting and merging. 2504 : 2505 : 4. Machine description. 2506 : 2507 : John's simplification of MLTREE_BASIS.fcond broke a few machine 2508 : description things: 2509 : 2510 : rtl-build.{sig,sml} and hppa.mdl fixed. 2511 : 2512 : NOTE: the machine description stuff in the repository is still broken. 2513 : Again, I can't put my fixes in because that involves 2514 : changes to MLTREE. 2515 : 2516 : ---------------------------------------------------------------------- 2517 : blume 1145 Name: Matthias Blume 2518 : blume 1155 Date: 2002/03/20 15:55:00 EST 2519 : Tag: blume-20020320-kmp 2520 : Description: 2521 : 2522 : Implemented Knuth-Morris-Pratt string matching in PreString and used 2523 : it for String.isSubstring, Substring.isSubstring, and 2524 : Substring.position. 2525 : 2526 : (Might need some stress-testing. Simple examples worked fine.) 2527 : 2528 : ---------------------------------------------------------------------- 2529 : Name: Matthias Blume 2530 : blume 1152 Date: 2002/03/19 16:37:00 EST 2531 : Tag: blume-20020319-witnesses 2532 : Description: 2533 : 2534 : Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib. 2535 : 2536 : This implements a generic mechanism for changing constness qualifiers 2537 : anywhere within big C types without resorting to outright "casts". 2538 : (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you 2539 : modify the constness at the outermost level.) 2540 : The implementation of "convert" is based on the idea of "witness" 2541 : values -- values that are not used by the operation but whose types 2542 : "testify" to their applicability. On the implementation side, "convert" 2543 : is simply a projection (returning its second curried argument). With 2544 : cross-module inlining, it should not result in any machine code being 2545 : generated. 2546 : 2547 : ---------------------------------------------------------------------- 2548 : Name: Matthias Blume 2549 : blume 1148 Date: 2002/03/15 16:40:00 EST 2550 : Tag: blume-20020315-basis 2551 : Description: 2552 : 2553 : Provided (preliminary?) implementations for 2554 : 2555 : {String,Substring}.{concatWith,isSuffix,isSubstring} 2556 : 2557 : and 2558 : 2559 : Substring.full 2560 : 2561 : Those are in the Basis spec but they were missing in SML/NJ. 2562 : 2563 : ---------------------------------------------------------------------- 2564 : Name: Matthias Blume 2565 : blume 1145 Date: 2002/03/14 21:30:00 EST 2566 : Tag: blume-20020314-controls 2567 : Description: 2568 : 2569 : Controls: 2570 : --------- 2571 : 2572 : 1. Factored out the recently-added Controls : CONTROLS stuff and put 2573 : it into its own library /controls-lib.cm. The source tree for 2574 : this is under src/smlnj-lib/Controls. 2575 : 2576 : 2. Changed the names of types and functions in this interface, so they 2577 : make a bit more "sense": 2578 : 2579 : module -> registry 2580 : 'a registry -> 'a group 2581 : 2582 : 3. The interface now deals in ref cells only. The getter/setter interface 2583 : is (mostly) gone. 2584 : 2585 : 4. Added a function that lets one register an already-existing ref cell. 2586 : 2587 : 5. Made the corresponding modifications to the rest of the code so that 2588 : everything compiles again. 2589 : 2590 : 6. Changed the implementation of Controls.MLRISC back to something closer 2591 : to the original. In particular, this module (and therefore MLRISC) 2592 : does not depend on Controls. There now is some link-time code in 2593 : int-sys.sml that registers the MLRISC controls with the Controls 2594 : module. 2595 : 2596 : CM: 2597 : --- 2598 : 2599 : * One can now specify the lambda-split aggressiveness in init.cmi. 2600 : 2601 : ---------------------------------------------------------------------- 2602 : leunga 1142 Name: Allen Leung 2603 : Date: 2002/03/13 17:30:00 EST 2604 : Tag: leunga-20020313-x86-fp-unary 2605 : Description: 2606 : 2607 : Bug fix for: 2608 : 2609 : > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml 2610 : > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002 2611 : > - fun f(x,(y,z)) = Real.~ y; 2612 : > [autoloading] 2613 : > [autoloading done] 2614 : > fchsl (%eax), 184(%esp) 2615 : > Error: MLRisc bug: X86MCEmitter.emitInstr 2616 : > 2617 : > uncaught exception Error 2618 : > raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19 2619 : 2620 : The problem was that the code generator did not generate any fp registers 2621 : in this case, and the ra didn't know that it needed to run the X86FP phase to 2622 : translate the pseudo fp instruction. This only happened with unary fp 2623 : operators in certain situations. 2624 : 2625 : ---------------------------------------------------------------------- 2626 : blume 1137 Name: Matthias Blume 2627 : blume 1140 Date: 2002/03/13 14:00:00 EST 2628 : Tag: blume-20020313-overload-etc 2629 : Description: 2630 : 2631 : 1. Added _overload as a synonym for overload for backward compatibility. 2632 : (Control.overloadKW must be true for either version to be accepted.) 2633 : 2634 : 2. Fixed bug in install script that caused more things to be installed 2635 : than what was requested in config/targets. 2636 : 2637 : 3. Made CM aware of the (_)overload construct so that autoloading 2638 : works. 2639 : 2640 : ---------------------------------------------------------------------- 2641 : Name: Matthias Blume 2642 : blume 1138 Date: 2002/03/12 22:03:00 EST 2643 : Tag: blume-20020312-url 2644 : Description: 2645 : 2646 : Forgot to update BOOT and srcarchiveurl. 2647 : 2648 : ---------------------------------------------------------------------- 2649 : Name: Matthias Blume 2650 : blume 1137 Date: 2002/03/12 17:30:00 EST 2651 : Tag: blume-20020312-version110392 2652 : Description: 2653 : 2654 : Yet another version number bump (because of small changes to the 2655 : binfile format). Version number is now 110.39.2. NEW BOOTFILES! 2656 : 2657 : Changes: 2658 : 2659 : The new pid generation scheme described a few weeks ago was overly 2660 : complicated. I implemented a new mechanism that is simpler and 2661 : provides a bit more "stability": Once CM has seen a compilation 2662 : unit, it keeps its identity constant (as long as you do not delete 2663 : those crucial CM/GUID/* files). This means that when you change 2664 : blume 1145 an interface, compile, then go back to the old interface, and 2665 : blume 1137 compile again, you arrive at the original pid. 2666 : 2667 : There now also is a mechanism that instructs CM to use the plain 2668 : environment hash as a module's pid (effectively making its GUID 2669 : the empty string). For this, "noguid" must be specified as an 2670 : option to the .sml file in question within its .cm file. 2671 : This is most useful for code that is being generated by tools such 2672 : as ml-nlffigen (because during development programmers tend to 2673 : erase the tool's entire output directory tree including CM's cached 2674 : GUIDs). "noguid" is somewhat dangerous (since it can be used to locally 2675 : revert to the old, broken behavior of SML/NJ, but in specific cases 2676 : where there is no danger of interface confusion, its use is ok 2677 : (I think). 2678 : 2679 : ml-nlffigen by default generates "noguid" annotations. They can be 2680 : turned off by specifying -guid in its command line. 2681 : 2682 : ---------------------------------------------------------------------- 2683 : george 1136 Name: Lal George 2684 : Date: 2002/03/12 12 14:42:36 EST 2685 : Tag: george-20020312-frequency-computation 2686 : Description: 2687 : 2688 : Integrated jump chaining and static block frequency into the 2689 : compiler. More details and numbers later. 2690 : leunga 1025 2691 : george 1003 ---------------------------------------------------------------------- 2692 : george 1133 Name: Lal George 2693 : Date: 2002/03/11 11 22:38:53 EST 2694 : Tag: george-20020311-jump-chain-elim 2695 : Description: 2696 : 2697 : Tested the jump chain elimination on all architectures (except the 2698 : hppa). This is on by default right now and is profitable for the 2699 : alpha and x86, however, it may not be profitable for the sparc and ppc 2700 : when compiling the compiler. 2701 : 2702 : The gc test will typically jump to a label at the end of the cluster, 2703 : where there is another jump to an external cluster containing the actual 2704 : code to invoke gc. This is to allow factoring of common gc invocation 2705 : sequences. That is to say, we generate: 2706 : 2707 : f: 2708 : testgc 2709 : ja L1 % jump if above to L1 2710 : 2711 : L1: 2712 : jmp L2 2713 : 2714 : 2715 : After jump chain elimination the 'ja L1' instructions is converted to 2716 : 'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end 2717 : up being implemented in their long form (if L2 is far away) using: 2718 : 2719 : jbe L3 % jump if below or equal to L3 2720 : jmp L2 2721 : L3: 2722 : ... 2723 : 2724 : 2725 : For large compilation units L2 may be far away. 2726 : 2727 : 2728 : ---------------------------------------------------------------------- 2729 : blume 1132 Name: Matthias Blume 2730 : Date: 2002/03/11 13:30:00 EST 2731 : Tag: blume-20020311-mltreeeval 2732 : Description: 2733 : 2734 : A functor parameter was missing. 2735 : 2736 : ---------------------------------------------------------------------- 2737 : leunga 1130 Name: Allen Leung 2738 : leunga 1131 Date: 2002/03/11 10:30:00 EST 2739 : leunga 1142 Tag: leunga-20020311-runtime-string0 2740 : leunga 1131 Description: 2741 : 2742 : The representation of the empty string now points to a 2743 : legal null terminated C string instead of unit. It is now possible 2744 : to convert an ML string into C string with InlineT.CharVector.getData. 2745 : This compiles into one single machine instruction. 2746 : 2747 : ---------------------------------------------------------------------- 2748 : Name: Allen Leung 2749 : leunga 1130 Date: 2002/03/10 23:55:00 EST 2750 : Tag: leunga-20020310-x86-call 2751 : Description: 2752 : 2753 : Added machine generation for CALL instruction (relative displacement mode) 2754 : 2755 : ---------------------------------------------------------------------- 2756 : blume 1128 Name: Matthias Blume 2757 : Date: 2002/03/08 16:05:00 2758 : Tag: blume-20020308-entrypoints 2759 : Description: 2760 : 2761 : Version number bumped to 110.39.1. NEW BOOTFILES! 2762 : 2763 : Entrypoints: non-zero offset into a code object where execution should begin. 2764 : 2765 : - Added the notion of an entrypoint to CodeObj. 2766 : - Added reading/writing of entrypoint info to Binfile. 2767 : - Made runtime system bootloader aware of entrypoints. 2768 : - Use the address of the label of the first function given to mlriscGen 2769 : as the entrypoint. This address is currently always 0, but it will 2770 : not be 0 once we turn on block placement. 2771 : - Removed the linkage cluster code (which was The Other Way(tm) of dealing 2772 : with entry points) from mlriscGen. 2773 : 2774 : ---------------------------------------------------------------------- 2775 : leunga 1127 Name: Allen Leung 2776 : Date: 2002/03/07 20:45:00 EST 2777 : Tag: leunga-20020307-x86-cmov 2778 : Description: 2779 : 2780 : Bug fixes for CMOVcc on x86. 2781 : 2782 : 1. Added machine code generation for CMOVcc 2783 : 2. CMOVcc is now generated in preference over SETcc on PentiumPro or above. 2784 : 3. CMOVcc cannot have an immediate operand as argument. 2785 : 2786 : ---------------------------------------------------------------------- 2787 : blume 1126 Name: Matthias Blume 2788 : Date: 2002/03/07 16:15:00 EST 2789 : Tag: blume-20020307-controls 2790 : Description: 2791 : 2792 : This is a very large but mostly boring patch which makes (almost) 2793 : every tuneable compiler knob (i.e., pretty much everything under 2794 : Control.* plus a few other things) configurable via both the command 2795 : line and environment variables in the style CM did its configuration 2796 : until now. 2797 : 2798 : Try starting sml with '-h' (or, if you are brave, '-H') 2799 : 2800 : To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which 2801 : implements the underlying generic mechanism. 2802 : 2803 : The interface to some of the existing such facilities has changed somewhat. 2804 : For example, the MLRiscControl module now provides mkFoo instead of getFoo. 2805 : (The getFoo interface is still there for backward-compatibility, but its 2806 : use is deprecated.) 2807 : 2808 : The ml-build script passes -Cxxx=yyy command-line arguments through so 2809 : that one can now twiddle the compiler settings when using this "batch" 2810 : compiler. 2811 : 2812 : TODO items: 2813 : 2814 : We should go through and throw out all controls that are no longer 2815 : connected to anything. Moreover, we should go through and provide 2816 : meaningful (and correct!) documentation strings for those controls 2817 : that still are connected. 2818 : 2819 : Currently, multiple calls to Controls.new are accepted (only the first 2820 : has any effect). Eventually we should make sure that every control 2821 : is being made (via Controls.new) exactly once. Future access can then 2822 : be done using Controls.acc. 2823 : 2824 : Finally, it would probably be a good idea to use the getter-setter 2825 : interface to controls rather than ref cells. For the time being, both 2826 : styles are provided by the Controls module, but getter-setter pairs are 2827 : better if thread-safety is of any concern because they can be wrapped. 2828 : 2829 : ***************************************** 2830 : 2831 : One bug fix: The function blockPlacement in three of the MLRISC 2832 : backpatch files used to be hard-wired to one of two possibilities at 2833 : link time (according to the value of the placementFlag). But (I 2834 : think) it should rather sense the flag every time. 2835 : 2836 : ***************************************** 2837 : 2838 : Other assorted changes (by other people who did not supply a HISTORY entry): 2839 : 2840 : 1. the cross-module inliner now works much better (Monnier) 2841 : 2. representation of weights, frequencies, and probabilities in MLRISC 2842 : changed in preparation of using those for weighted block placement 2843 : (Reppy, George) 2844 : 2845 : ---------------------------------------------------------------------- 2846 : george 1116 Name: Lal George 2847 : george 1124 Date: 2002/03/07 14:44:24 EST 2002 2848 : Tag: george-20020307-weighted-block-placement 2849 : 2850 : Tested the weighted block placement optimization on all architectures 2851 : (except the hppa) using AMPL to generate the block and edge frequencies. 2852 : Changes were required in the machine properties to correctly 2853 : categorize trap instructions. There is an MLRISC flag 2854 : "weighted-block-placement" that can be used to enable weighted block 2855 : placement, but this will be ineffective without block/edge 2856 : frequencies (coming soon). 2857 : 2858 : 2859 : ---------------------------------------------------------------------- 2860 : Name: Lal George 2861 : george 1116 Date: 2002/03/05 17:24:48 EST 2862 : Tag: george-20020305-linkage-cluster 2863 : 2864 : george 1124 In order to support the block placement optimization, a new cluster 2865 : is generated as the very first cluster (called the linkage cluster). 2866 : It contains a single jump to the 'real' entry point for the compilation 2867 : unit. Block placement has no effect on the linkage cluster itself, but 2868 : all the other clusters have full freedom in the manner in which they 2869 : reorder blocks or functions. 2870 : george 1116 2871 : On the x86 the typical linkage code that is generated is: 2872 : ---------------------- 2873 : .align 2 2874 : L0: 2875 : addlL1-L0, 72(%esp) 2876 : george 1124 jmp L1 2877 : george 1116 2878 : 2879 : .align 2 2880 : L1: 2881 : ---------------------- 2882 : 2883 : 72(%esp) is the memory location for the stdlink register. This 2884 : must contain the address of the CPS function being called. In the 2885 : george 1124 above example, it contains the address of L0; before 2886 : george 1116 calling L1 (the real entry point for the compilation unit), it 2887 : must contain the address for L1, and hence 2888 : 2889 : addl $L1-L0, 72(%esp) 2890 : 2891 : george 1124 I have tested this on all architectures except the hppa.The increase 2892 : in code size is of course negligible 2893 : george 1116 2894 : ---------------------------------------------------------------------- 2895 : leunga 1115 Name: Allen Leung 2896 : Date: 2002/03/03 13:20:00 EST 2897 : Tag: leunga-20020303-mlrisc-tools 2898 : 2899 : Added #[ ... ] expressions to mlrisc tools 2900 : 2901 : ---------------------------------------------------------------------- 2902 : blume 1096 Name: Matthias Blume 2903 : blume 1098 Date: 2002/02/27 12:29:00 EST 2904 : Tag: blume-20020227-cdebug 2905 : Description: 2906 : 2907 : - made types in structure C and C_Debug to be equal 2908 : - got rid of code duplication (c-int.sml vs. c-int-debug.sml) 2909 : - there no longer is a C_Int_Debug (C_Debug is directly derived from C) 2910 : 2911 : ---------------------------------------------------------------------- 2912 : Name: Matthias Blume 2913 : blume 1096 Date: 2002/02/26 12:00:00 EST 2914 : Tag: blume-20020226-ffi 2915 : Description: 2916 : 2917 : 1. Fixed a minor bug in CM's "noweb" tool: 2918 : If numbering is turned off, then truly don't number (i.e., do not 2919 : supply the -L option to noweb). The previous behavior was to supply 2920 : -L'' -- which caused noweb to use the "default" line numbering scheme. 2921 : Thanks to Chris Richards for pointing this out (and supplying the fix). 2922 : 2923 : 2. Once again, I reworked some aspects of the FFI: 2924 : 2925 : A. The incomplete/complete type business: 2926 : 2927 : - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are 2928 : gone! 2929 : - ML types representing an incomplete type are now *equal* to 2930 : ML types representing their corresponding complete types (just like 2931 : in C). This is still safe because ml-nlffigen will not generate 2932 : RTTI for incomplete types, nor will it generate functions that 2933 : require access to such RTTI. But when ML code generated from both 2934 : incomplete and complete versions of the C type meet, the ML types 2935 : are trivially interoperable. 2936 : 2937 : NOTE: These changes restore the full generality of the translation 2938 : (which was previously lost when I eliminated functorization)! 2939 : 2940 : B. Enum types: 2941 : 2942 : - Structure C now has a type constructor "enum" that is similar to 2943 : how the "su" constructor works. However, "enum" is not a phantom 2944 : type because each "T enum" has values (and is isomorphic to 2945 : MLRep.Signed.int). 2946 : - There are generic access operations for enum objects (using 2947 : MLRep.Signed.int). 2948 : - ml-nlffigen will generate a structure E_foo for each "enum foo". 2949 : * The structure contains the definition of type "mlrep" (the ML-side 2950 : representation type of the enum). Normally, mlrep is the same 2951 : as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec", 2952 : then mlrep will be defined as a datatype -- thus facilitating 2953 : pattern matching on mlrep values. 2954 : ("-ec" will be suppressed if there are duplicate values in an 2955 : enumeration.) 2956 : * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep 2957 : will be generated for each C enum constant xxx. 2958 : * Conversion functions m2i and i2m convert between mlrep and 2959 : MLRep.Signed.int. (Without "-ec", these functions are identities.) 2960 : * Coversion functions c and ml convert between mlrep and "tag enum". 2961 : * Access functions (get/set) fetch and store mlrep values. 2962 : - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed 2963 : enumerations are merged into one single enumeration represented by 2964 : structure E_'. 2965 : 2966 : ---------------------------------------------------------------------- 2967 : leunga 1080 Name: Allen Leung 2968 : leunga 1094 Date: 2002/02/25 04:45:00 EST 2969 : Tag: leunga-20020225-cps-spill 2970 : 2971 : This is a new implementation of the CPS spill phase. 2972 : The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml 2973 : In case of problems, replace it with the old file spill.sml 2974 : 2975 : The current compiler runs into some serious performance problems when 2976 : constructing a large record. This can happen when we try to compile a 2977 : structure with many items. Even a very simple structure like the following 2978 : makes the compiler slow down. 2979 : 2980 : structure Foo = struct 2981 : val x_1 = 0w1 : Word32.int 2982 : val x_2 = 0w2 : Word32.int 2983 : val x_3 = 0w3 : Word32.int 2984 : ... 2985 : val x_N = 0wN : Word32.int 2986 : end 2987 : 2988 : The following table shows the compile time, from N=1000 to N=4000, 2989 : with the old compiler: 2990 : 2991 : N 2992 : 1000 CPS 100 spill 0.04u 0.00s 0.00g 2993 : MLRISC ra 0.06u 0.00s 0.05g 2994 : (spills = 0 reloads = 0) 2995 : TOTAL 0.63u 0.07s 0.21g 2996 : 2997 : 1100 CPS 100 spill 8.25u 0.32s 0.64g 2998 : MLRISC ra 5.68u 0.59s 3.93g 2999 : (spills = 0 reloads = 0) 3000 : TOTAL 14.71u 0.99s 4.81g 3001 : 3002 : 1500 CPS 100 spill 58.55u 2.34s 1.74g 3003 : MLRISC ra 5.54u 0.65s 3.91g 3004 : (spills = 543 reloads = 1082) 3005 : TOTAL 65.40u 3.13s 6.00g 3006 : 3007 : 2000 CPS 100 spill 126.69u 4.84s 3.08g 3008 : MLRISC ra 0.80u 0.10s 0.55g 3009 : (spills = 42 reloads = 84) 3010 : TOTAL 129.42u 5.10s 4.13g 3011 : 3012 : 3000 CPS 100 spill 675.59u 19.03s 11.64g 3013 : MLRISC ra 2.69u 0.27s 1.38g 3014 : (spills = 62 reloads = 124) 3015 : TOTAL 682.48u 19.61s 13.99g 3016 : 3017 : 4000 CPS 100 spill 2362.82u 56.28s 43.60g 3018 : MLRISC ra 4.96u 0.27s 2.72g 3019 : (spills = 85 reloads = 170) 3020 : TOTAL 2375.26u 57.21s 48.00g 3021 : 3022 : As you can see the old cps spill module suffers from some serious 3023 : leunga 1115 performance problem. But since I cannot decipher the old code fully, 3024 : instead of patching the problems up, I'm reimplementing it 3025 : leunga 1094 with a different algorithm. The new code is more modular, 3026 : smaller when compiled, and substantially faster 3027 : (O(n log n) time and O(n) space). Timing of the new spill module: 3028 : 3029 : 4000 CPS 100 spill 0.02u 0.00s 0.00g 3030 : MLRISC ra 0.25u 0.02s 0.15g 3031 : (spills=1 reloads=3) 3032 : TOTAL 7.74u 0.34s 1.62g 3033 : 3034 : Implementation details: 3035 : 3036 : As far as I can tell, the purpose of the CPS spill module is to make sure the 3037 : number of live variables at any program point (the bandwidth) 3038 : does not exceed a certain limit, which is determined by the 3039 : size of the spill area. 3040 : 3041 : When the bandwidth is too large, we decrease the register pressure by 3042 : packing live variables into spill records. How we achieve this is 3043 : completely different than what we did in the old code. 3044 : 3045 : First, there is something about the MLRiscGen code generator 3046 : that we should be aware of: 3047 : 3048 : o MLRiscGen performs code motion! 3049 : 3050 : In particular, it will move floating point computations and 3051 : address computations involving only the heap pointer to 3052 : their use sites (if there is only a single use). 3053 : What this means is that if we have a CPS record construction 3054 : statement 3055 : 3056 : RECORD(k,vl,w,e) 3057 : 3058 : we should never count the new record address w as live if w 3059 : has only one use (which is often the case). 3060 : 3061 : We should do something similar to floating point, but the transformation 3062 : there is much more complex, so I won't deal with that. 3063 : 3064 : Secondly, there are now two new cps primops at our disposal: 3065 : 3066 : 1. rawrecord of record_kind option 3067 : This pure operator allocates some uninitialized storage from the heap. 3068 : There are two forms: 3069 : 3070 : rawrecord NONE [INT n] allocates a tagless record of length n 3071 : rawrecord (SOME rk) [INT n] allocates a tagged record of length n 3072 : and initializes the tag. 3073 : 3074 : 2. rawupdate of cty 3075 : rawupdate cty (v,i,x) 3076 : Assigns to x to the ith component of record v. 3077 : The storelist is not updated. 3078 : 3079 : We use these new primops for both spilling and increment record construction. 3080 : 3081 : 1. Spilling. 3082 : 3083 : This is implemented with a linear scan algorithm (but generalized 3084 : to trees). The algorithm will create a single spill record at the 3085 : beginning of the cps function and use rawupdate to spill to it, 3086 : and SELECT or SELp to reload from it. So both spills and reloads 3087 : are fine-grain operations. In contrast, in the old algorithm 3088 : "spills" have to be bundled together in records. 3089 : 3090 : Ideally, we should sink the spill record construction to where 3091 : it is needed. We can even split the spill record into multiple ones 3092 : at the places where they are needed. But CPS is not a good 3093 : representation for global code motion, so I'll keep it simple and 3094 : am not attempting this. 3095 : 3096 : 2. Incremental record construction (aka record splitting). 3097 : 3098 : Long records with many component values which are simulatenously live 3099 : (recall that single use record addresses are not considered to 3100 : be live) are constructed with rawrecord and rawupdate. 3101 : We allocate space on the heap with rawrecord first, then gradually 3102 : fill it in with rawupdate. This is the technique suggested to me 3103 : by Matthias. 3104 : 3105 : Some restrictions on when this is applicable: 3106 : 1. It is not a VECTOR record. The code generator currently does not handle 3107 : this case. VECTOR record uses double indirection like arrays. 3108 : 2. All the record component values are defined in the same "basic block" 3109 : as the record constructor. This is to prevent speculative 3110 : record construction. 3111 : 3112 : ---------------------------------------------------------------------- 3113 : Name: Allen Leung 3114 : leunga 1086 Date: 2002/02/22 01:02:00 EST 3115 : Tag: leunga-20020222-mlrisc-tools 3116 : 3117 : Minor bug fixes in the parser and rewriter 3118 : 3119 : ---------------------------------------------------------------------- 3120 : Name: Allen Leung 3121 : leunga 1085 Date: 2002/02/21 20:20:00 EST 3122 : Tag: leunga-20020221-peephole 3123 : 3124 : Regenerated the peephole files. Some contained typos in the specification 3125 : and some didn't compile because of pretty printing bugs in the old version 3126 : of 'nowhere'. 3127 : 3128 : ---------------------------------------------------------------------- 3129 : Name: Allen Leung 3130 : leunga 1079 Date: 2002/02/19 20:20:00 EST 3131 : Tag: leunga-20020219-mlrisc-tools 3132 : Description: 3133 : 3134 : Minor bug fixes to the mlrisc-tools library: 3135 : 3136 : 1. Fixed up parsing colon suffixed keywords 3137 : 2. Added the ability to shut the error messages up 3138 : 3. Reimplemented the pretty printer and fixed up/improved 3139 : the pretty printing of handle and -> types. 3140 : 4. Fixed up generation of literal symbols in the nowhere tool. 3141 : 5. Added some SML keywords to to sml.sty 3142 : 3143 : ---------------------------------------------------------------------- 3144 : Name: Matthias Blume 3145 : blume 1078 Date: 2002/02/19 16:20:00 EST 3146 : Tag: blume-20020219-cmffi 3147 : Description: 3148 : 3149 : A wild mix of changes, some minor, some major: 3150 : 3151 : * All C FFI-related libraries are now anchored under$c: 3152 : $/c.cm -->$c/c.cm 3153 : $/c-int.cm -->$c/internals/c-int.cm 3154 : $/memory.cm -->$c/memory/memory.cm 3155 : 3156 : * "make" tool (in CM) now treats its argument pathname slightly 3157 : differently: 3158 : 1. If the native expansion is an absolute name, then before invoking 3159 : the "make" command on it, CM will apply OS.Path.mkRelative 3160 : (with relativeTo = OS.FileSys.getDir()) to it. 3161 : 2. The argument will be passed through to subsequent phases of CM 3162 : processing without "going native". In particular, if the argument 3163 : was an anchored path, then "make" will not lose track of that anchor. 3164 : 3165 : * Compiler backends now "know" their respective C calling conventions 3166 : instead of having to be told about it by ml-nlffigen. This relieves 3167 : ml-nlffigen from one of its burdens. 3168 : 3169 : * The X86Backend has been split into X86CCallBackend and X86StdCallBackend. 3170 : 3171 : * Export C_DEBUG and C_Debug from $c/c.cm. 3172 : 3173 : * C type encoding in ml-nlffi-lib has been improved to model the conceptual 3174 : subtyping relationship between incomplete pointers and their complete 3175 : counterparts. For this, ('t, 'c) ptr has been changed to 'o ptr -- 3176 : with the convention of instantiating 'o with ('t, 'c) obj whenever 3177 : the pointer target type is complete. In the incomplete case, 'o 3178 : will be instantiated with some "'c iobj" -- a type obtained by 3179 : using one of the functors PointerToIncompleteType or PointerToCompleteType. 3180 : 3181 : Operations that work on both incomplete and complete pointer types are 3182 : typed as taking an 'o ptr while operations that require the target to 3183 : be known are typed as taking some ('t, 'c) obj ptr. 3184 : 3185 : voidptr is now a bit "more concrete", namely "type voidptr = void ptr'" 3186 : where void is an eqtype without any values. This makes it possible 3187 : to work on voidptr values using functions meant to operate on light 3188 : incomplete pointers. 3189 : 3190 : * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has 3191 : been vastly simplified. 3192 : 3193 : ---------------------------------------------------------------------- 3194 : Name: Matthias Blume 3195 : blume 1076 Date: 2002/02/19 10:48:00 EST 3196 : Tag: blume-20020219-pqfix 3197 : Description: 3198 : 3199 : Applied Chris Okasaki's bug fix for priority queues. 3200 : 3201 : ---------------------------------------------------------------------- 3202 : Name: Matthias Blume 3203 : blume 1073 Date: 2002/02/15 17:05:00 3204 : blume 1069 Tag: Release_110_39 3205 : Description: 3206 : 3207 : blume 1073 Last-minute retagging is becoming a tradition... :-( 3208 : 3209 : This is the working release 110.39. 3210 : 3211 : ---------------------------------------------------------------------- 3212 : Name: Matthias Blume 3213 : Date: 2002/02/15 16:00:00 EST 3214 : Tag: Release_110_39-orig 3215 : Description: 3216 : 3217 : blume 1069 Working release 110.39. New bootfiles. 3218 : 3219 : blume 1073 (Update: There was a small bug in the installer so it wouldn't work 3220 : with all shells. So I retagged. -Matthias) 3221 : 3222 : blume 1069 ---------------------------------------------------------------------- 3223 : Name: Matthias Blume 3224 : blume 1068 Date: 2002/02/15 14:17:00 EST 3225 : Tag: blume-20020215-showbindings 3226 : Description: 3227 : 3228 : Added EnvRef.listBoundSymbols and CM.State.showBindings. Especially 3229 : the latter can be useful for exploring what bindings are available at 3230 : the interactive prompt. (The first function returns only the list 3231 : of symbols that are really bound, the second prints those but also the 3232 : ones that CM's autoloading mechanism knows about.) 3233 : 3234 : ---------------------------------------------------------------------- 3235 : Name: Matthias Blume 3236 : blume 1067 Date: 2002/02/15 12:08:00 EST 3237 : Tag: blume-20020215-iptrs 3238 : Description: 3239 : 3240 : Two improvements to ml-nlffigen: 3241 : 3242 : 1. Write files only if they do not exist or if their current contents 3243 : do not coincide with what's being written. (That is, avoid messing 3244 : with the time stamps unless absolutely necessary.) 3245 : 3246 : 2. Implement a "repository" mechanism for generated files related 3247 : to "incomplete pointer types". See the README file for details. 3248 : 3249 : ---------------------------------------------------------------------- 3250 : Name: Matthias Blume 3251 : blume 1066 Date: 2002/02/14 11:50:00 EST 3252 : Tag: blume-20020214-quote 3253 : Description: 3254 : 3255 : Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm). This is required 3256 : because of the new and improved tag generation scheme. (Thanks to Allen 3257 : Leung for pointing it out.) 3258 : 3259 : ---------------------------------------------------------------------- 3260 : george 1065 Name: Lal George 3261 : Date: 2002/02/14 09:55:27 EST 2002 3262 : Tag: george-20020214-isabelle-bug 3263 : Description: 3264 : 3265 : Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation 3266 : of Isabelle on the x86. 3267 : 3268 : From Allen: 3269 : ----------- 3270 : I've found the problem: 3271 : 3272 : in ra-core.sml, I use the counter "blocked" to keep track of the 3273 : true number of elements in the freeze queue. When the counter goes 3274 : to zero, I skip examining the queue. But I've messed up the 3275 : bookkeeping in combine(): 3276 : 3277 : else (); 3278 : case !ucol of 3279 : PSEUDO => (if !cntv > 0 then 3280 : (if !cntu > 0 then blocked := !blocked - 1 else (); 3281 : ^^^^^^^^^^^^^^^^^^^^^^^ 3282 : moveu := mergeMoveList(!movev, !moveu) 3283 : ) 3284 : else (); 3285 : 3286 : combine() is called to coalesce two nodes u and v. 3287 : I think I was thinking that if the move counts of u and v are both 3288 : greater than zero then after they are coalesced then one node is 3289 : removed from the freeze queue. Apparently I was thinking that 3290 : both u and v are of low degree, but that's clearly not necessarily true. 3291 : 3292 : 3293 : 02/12/2002: 3294 : Here's the patch. HOL now compiles. 3295 : 3296 : I don't know how this impact on performance (compile 3297 : time or runtime). This bug caused the RA (especially on the x86) 3298 : to go thru the potential spill phase when there are still nodes on the 3299 : freeze queue. 3300 : 3301 : 3302 : 3303 : 3304 : ---------------------------------------------------------------------- 3305 : blume 1058 Name: Matthias Blume 3306 : blume 1064 Date: 2002/02/13 22:40:00 EST 3307 : Tag: blume-20020213-fptr-rtti 3308 : Description: 3309 : 3310 : Fixed a bug in ml-nlffigen that was introduced with one of the previous 3311 : updates. 3312 : 3313 : ---------------------------------------------------------------------- 3314 : Name: Matthias Blume 3315 : blume 1063 Date: 2002/02/13 16:41:00 EST 3316 : Tag: blume-20020213-cmlpq 3317 : Description: 3318 : 3319 : Added new priority queue export symbols (which have just been added to 3320 : smlnj-lib.cm) to CML's version of smlnj-lib.cm. (Otherwise CML would 3321 : not compile and the installer would choke.) 3322 : 3323 : ---------------------------------------------------------------------- 3324 : Name: Matthias Blume 3325 : blume 1062 Date: 2002/02/13 16:15:00 EST 3326 : Tag: blume-20020213-various 3327 : Description: 3328 : 3329 : 1. More tweaks to ml-nlffigen: 3330 : 3331 : - better internal datastructures (resulting in slight speedup) 3332 : - "-match" option requires exact match 3333 : - "localized" gensym counters (untagged structs/unions nested within 3334 : other structs/unions or within typedefs get a fresh counter; their 3335 : tag will be prefixed by a concatenation of their parents' tags) 3336 : - bug fixes (related to calculation of transitive closure of types 3337 : to be included in the output) 3338 : 3339 : 2. Minor Basis updates: 3340 : 3341 : - added implementations for List.collate and Option.app 3342 : 3343 : ---------------------------------------------------------------------- 3344 : Name: Matthias Blume 3345 : blume 1060 Date: 2002/02/11 15:55:00 EST 3346 : Tag: blume-20020211-gensym 3347 : Description: 3348 : 3349 : Added a "-gensym" option to command line of ml-nlffigen. This can be 3350 : used to specify a "stem" -- a string that is inserted in all "gensym'd" 3351 : names (ML structure names that correspond to unnamed C structs, unions, 3352 : and enums), so that separate runs of ml-nlffigen do not clash. 3353 : 3354 : ---------------------------------------------------------------------- 3355 : Name: Matthias Blume 3356 : blume 1059 Date: 2002/02/11 12:05:00 EST 3357 : Tag: blume-20020211-gensml 3358 : Description: 3359 : 3360 : A quick fix for a problem with GenSML (in the pgraph-util library): 3361 : Make generation of toplevel "local" optional. (Strictly speaking, 3362 : signature definitions within "local" are not legal SML.) 3363 : 3364 : Other than that: updates to INSTALL and cm/TODO. 3365 : 3366 : ---------------------------------------------------------------------- 3367 : Name: Matthias Blume 3368 : blume 1058 Date: 2002/02/08 15:00:00 EST 3369 : Tag: blume-20020208-uniquepid 3370 : Description: 3371 : 3372 : 0. Version number has been bumped to 110.38.1. NEW BOOTFILES!!! 3373 : 3374 : 1. The installer (config/install.sh) has gotten smarter: 3375 : 3376 : - Configuration options are a bit easier to specify now 3377 : (in config/targets). 3378 : - Bug in recognizing .tar.bz2 files fixed. 3379 : - Installer automatically resolves dependencies between 3380 : configuration options (e.g., if you ask for eXene, you will 3381 : also get cml -- regardless whether you asked for it or not). 3382 : - Installer can run in "quieter mode" by setting the environment 3383 : variable INSTALL_QUIETLY to "true". "Quieter" does not mean 3384 : "completely silent", though. 3385 : - Build HashCons library as part of smlnj-lib. 3386 : 3387 : 2. A new scheme for assigning persistent identifiers to compilation 3388 : units (and, by extension, to types etc.) has been put into place. 3389 : This fixes a long-standing bug where types and even dynamic values 3390 : can get internally confused, thereby compromising type safety 3391 : (abstraction) and dynamic correctness. See 3392 : 3393 : http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz 3394 : 3395 : for an example of how things could go wrong until now. 3396 : 3397 : The downside of the new scheme is that pids are not quite as 3398 : persistent as they used to be: CM will generate a fresh pid 3399 : for every compilation unit that it thinks it sees for the first 3400 : time. That means that if you compile starting from a clean, fresh 3401 : source tree at two different times, you end up with different 3402 : binaries. 3403 : 3404 : Cutoff recompilation, however, has not been compromised because 3405 : CM keeps pid information in special caches between runs. 3406 : 3407 : ---------------------------------------------------------------------- 3408 : george 1053 Name: Lal George 3409 : george 1055 Date: 2002/02/07 15:34:13 EST 2002 3410 : Tag: 3411 : Description: 3412 : 3413 : Compilers that generate assembly code may produce global labels 3414 : whose value is resolved at link time. The various peephole optimization 3415 : modules did not take this in account. 3416 : 3417 : TODO. The Labels.addrOf function should really return an option 3418 : type so that clients are forced to deal with this issue, rather 3419 : than an exception being raised. 3420 : 3421 : ---------------------------------------------------------------------- 3422 : Name: Lal George 3423 : george 1053 Date: 2002/02/06 13:55:02 EST 3424 : Tag: george-20020206-ra-breakup 3425 : Description: 3426 : 3427 : 1. A bug fix from Allen. 3428 : 3429 : A typo causes extra fstp %st(0)'s to be generated at compensation 3430 : edges, which might cause stack underflow traps at runtime. This 3431 : occurs in fft where there are extraneous fstps right before the 'into' 3432 : trap instruction (in this case they are harmless since none of the 3433 : integers overflow.) 3434 : 3435 : george 1054 2. Pulled out various utility modules that were embedded in the modules 3436 : george 1053 of the register allocator. I need these modules for other purposes, but 3437 : george 1054 they are not complete enough to put into a library (just yet). 3438 : george 1053 ---------------------------------------------------------------------- 3439 : blume 1049 Name: Matthias Blume 3440 : Date: 2002/01/31 16:05:00 EST 3441 : Tag: blume-20020131-sparc-ccalls 3442 : Description: 3443 : 3444 : 1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes) 3445 : of extra stack space by mistake. Fixed. 3446 : 3447 : 2. Bug in logic of handling of command-line options in ml-nlffigen fixed. 3448 : 3449 : ---------------------------------------------------------------------- 3450 : leunga 1048 Name: Allen Leung 3451 : Date: 2002/01/30 3452 : Tag: leunga-20020130-nowhere-bug-fix 3453 : Description: 3454 : 3455 : MLRISC bug fixes: 3456 : 1. Fixed a bindings computation bug in the 'nowhere' program generator tool. 3457 : 2. MachineInt.fromString was negating its value. 3458 : 3459 : ---------------------------------------------------------------------- 3460 : blume 1039 Name: Matthias Blume 3461 : blume 1047 Date: 2002/01/29 3462 : Tag: blume-20020129-INSTALL 3463 : Description: 3464 : 3465 : - Added somewhat detailed installation instructions (file INSTALL). 3466 : - Fixed curl-detection bug in config/install.sh. 3467 : - It is now possible to select the URL getter using the URLGETTER 3468 : environment variable: 3469 : 3470 : not set / "unknown" --> automatic detection (script tries wget, 3471 : curl, and lynx) 3472 : "wget" / "curl" / "lynx" --> use the specified program (script "knows" 3473 : how to properly invoke them) 3474 : other --> use$URLGETTER directly, it must take 3475 : precisely two command-line arguments 3476 : (source URL and destination file name) 3477 : 3478 : ---------------------------------------------------------------------- 3479 : Name: Matthias Blume 3480 : blume 1044 Date: 2002/01/28 3481 : blume 1045 Tag: blume-20020128-sparc-ccalls 3482 : Description: 3483 : 3484 : - Fixed problem with calculation of "used" registers in sparc-c-calls. 3485 : - Make use of the allocParam argument in sparc-c-calls. 3486 : 3487 : ---------------------------------------------------------------------- 3488 : Name: Matthias Blume 3489 : Date: 2002/01/28 3490 : blume 1044 Tag: blume-20020128-allocParam 3491 : Description: 3492 : 3493 : John Reppy: Changes c-calls API to accept client-callback for 3494 : allocating extra stack space. 3495 : me: Corresponding changes to mlriscGen (using a dummy argument that 3496 : does not change the current behavior). 3497 : 3498 : ---------------------------------------------------------------------- 3499 : Name: Matthias Blume 3500 : blume 1040 Date: 2002/01/28 12:00:00 3501 : Tag: Release_110_38 3502 : Description: 3503 : 3504 : This time for real!!! 3505 : 3506 : ---------------------------------------------------------------------- 3507 : Name: Matthias Blume 3508 : blume 1039 Date: 2002/01/28 10:56:00 EST 3509 : Tag: blume-20020128-retraction 3510 : Description: 3511 : 3512 : 0. Retracted earlier 110.38. (The Release_110_38 tag has been replaced 3513 : with blume-Release_110_38-retracted.) 3514 : 3515 : 1. Fixed a problem with incorrect rounding modes in real64.sml. 3516 : (Thanks to Andrew Mccreight .) 3517 : 3518 : 2. A bug in ml-nlffigen related to the handling of unnamed structs, unions, 3519 : and enums fixed. The naming of corresponding ML identifiers should 3520 : now be consistent again. 3521 : 3522 : ---------------------------------------------------------------------- 3523 : leunga 1038 Name: Allen Leung 3524 : Date: 2002/01/27 3525 : Tag: leunga-20020127-nowhere 3526 : Description: 3527 : 3528 : Added a target called nowhere in the configuration scripts. 3529 : Enabling this will build the MLRISC 'nowhere' tool (for translating 3530 : programs with where-clauses into legal SML code) during installation. 3531 : 3532 : ---------------------------------------------------------------------- 3533 : blume 1036 Name: Matthias Blume 3534 : leunga 1038 Date: 2002/01/25 21:27:00 EST 3535 : blume 1039 Tag: blume-Release_110_38-retracted 3536 : blume 1037 Description: 3537 : 3538 : Call it a (working) release! Version is 110.38. Bootfiles are ready. 3539 : 3540 : README will be added later. 3541 : 3542 : blume 1039 !!! NOTE: Re-tagged as blume-Release_110_38-retracted. Original tag 3543 : (Release_110_38) removed. Reason: Last-minute bug fixes. 3544 : 3545 : blume 1037 ---------------------------------------------------------------------- 3546 : Name: Matthias Blume 3547 : blume 1036 Date: 2002/01/25 3548 : Tag: blume-20020125-ffi 3549 : Description: 3550 : 3551 : A large number of tweaks and improvements to ml-nlffi-lib and 3552 : ml-nlffigen: 3553 : 3554 : - ML represenation types have been streamlined 3555 : - getter and setter functions work with concrete values, not abstract 3556 : ones where possible 3557 : - ml-nlffigen command line more flexible (see README file there) 3558 : - some bugs have been fixed (hopefully) 3559 : 3560 : ---------------------------------------------------------------------- 3561 : george 1034 Name: Lal George 3562 : Date: 2002/01/24 3563 : Tag: george-20020124-risc-ra-interface 3564 : Description: 3565 : 3566 : There is a dramatic simplification in the interface to the 3567 : register allocator for RISC architectures as a result of making 3568 : parallel copy instructions explicit. 3569 : 3570 : ---------------------------------------------------------------------- 3571 : blume 1028 Name: Matthias Blume 3572 : blume 1032 Date: 2002/01/22 3573 : Tag: blume-20020122-x86-ccalls 3574 : Description: 3575 : 3576 : Bug fix for c-calls on x86 (having to do with how char- and 3577 : short-arguments are being handled). 3578 : 3579 : ---------------------------------------------------------------------- 3580 : Name: Matthias Blume 3581 : blume 1031 Date: 2002/01/21 3582 : Tag: blume-20020121-ff 3583 : Description: 3584 : 3585 : Another day of fiddling with the FFI... 3586 : 3587 : 1. Bug fix/workaround: CKIT does not complain about negative array 3588 : dimensions, so ml-nlffigen has to guard itself against this possibility. 3589 : (Otherwise a negative dimension would send it into an infinite loop.) 3590 : 3591 : 2. Some of the abstract types (light objects, light pointers, most "base" 3592 : types) in structure C are now eqtypes. 3593 : 3594 : 3. Added constructors and test functions for NULL function pointers. 3595 : 3596 : ---------------------------------------------------------------------- 3597 : Name: Matthias Blume 3598 : blume 1028 Date: 2002/01/18 3599 : blume 1030 Tag: blume-20020118-ready-for-new-release 3600 : Description: 3601 : 3602 : Made config/srcarchiveurl point to a new place. (Will provide boot 3603 : files shortly.) 3604 : 3605 : Maybe we christen this to be 110.38? 3606 : 3607 : ---------------------------------------------------------------------- 3608 : Name: Matthias Blume 3609 : Date: 2002/01/18 3610 : blume 1029 Tag: blume-20020118-more-ffifiddle 3611 : Description: 3612 : 3613 : Today's FFI fiddling: 3614 : 3615 : - Provided a structure CGetSet with "convenient" versions of C.Get.* and 3616 : C.Set.* that use concrete (MLRep.*) arguments and results instead 3617 : of abstract ones. 3618 : 3619 : - Provided word-style bit operations etc. for "int" representation 3620 : types in MLRep.SBitops where ranges over Char, Int, Short, 3621 : and Long. 3622 : 3623 : ---------------------------------------------------------------------- 3624 : Name: Matthias Blume 3625 : Date: 2002/01/18 3626 : blume 1028 Tag: blume-20020118-use-x86-fp 3627 : Description: 3628 : 3629 : Now that x86-fast-fp seems to be working, I turned it back on again 3630 : by default. (Seems to work fine now, even with the FFI.) 3631 : 3632 : Other than that, I added some documentation about the FFI to 3633 : src/ml-nlffigen/README and updated the FFI test examples in 3634 : src/ml-nlffi-lib/Tests/*. 3635 : 3636 : ---------------------------------------------------------------------- 3637 : leunga 1025 Name: Allen Leung 3638 : Date: 2002/01/17 3639 : Tag: leunga-20020117-x86-fast-fp-call 3640 : Description: 3641 : 3642 : 1. Fixed a problem with handling return fp values when x86's fast fp 3643 : mode is turned on. 3644 : 3645 : 2. Minor pretty printing fix for cellset. Print %st(0) as %st(0) instead 3646 : of %f32. 3647 : 3648 : 3. Added a constructor INT32lit to the ast of MLRISC tools. 3649 : 3650 : ---------------------------------------------------------------------- 3651 : blume 1021 Name: Matthias Blume 3652 : Date: 2002/01/16 3653 : Tag: blume-20020116-ffifiddle 3654 : Description: 3655 : 3656 : More fiddling with the FFI interface: 3657 : 3658 : - Make constness 'c instead of rw wherever possible. This eliminates 3659 : the need for certain explicit coercions. (However, due to ML's 3660 : value polymorphism, there will still be many cases where explicit 3661 : coercions are necessary. Phantom types are not the whole answer 3662 : to modeling a subtyping relationship in ML.) 3663 : 3664 : - ro/rw coersions for pointers added. (Avoids the detour through */&.) 3665 : 3666 : - "printf" test example added to src/ml-nlffi-lib/Tests. (Demonstrates 3667 : clumsy workaround for varargs problem.) 3668 : 3669 : ---------------------------------------------------------------------- 3670 : george 1016 Name: Lal George 3671 : Date: 2002/01/15 3672 : Tag: 3673 : Description: 3674 : 3675 : 1. Since COPY instructions are no longer native to the architecture, 3676 : a generic functor can be used to implement the expandCopies function. 3677 : 3678 : 2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a 3679 : TEXT segment. 3680 : 3681 : ---------------------------------------------------------------------- 3682 : blume 1011 Name: Matthias Blume 3683 : blume 1015 Date: 2002/01/15 3684 : Tag: blume-20020115-ffiupdates 3685 : Description: 3686 : 3687 : 1. Fix for bug resulting in single-precision float values being returned 3688 : incorrectly from FFI calls. 3689 : 3690 : 2. Small modifications to C FFI API: 3691 : 3692 : - memory-allocation routines return straight objects (no options) 3693 : and raise an exception in out-of-memory situations 3694 : - unsafe extensions to cast between function pointers and pointers 3695 : from/to ints 3696 : - added structure C_Debug as an alternative to structure C where 3697 : pointer-dereferencing (|*| and |*!) always check for null-pointers 3698 : - added open_lib' to DynLinkage; open_lib' works like open_lib 3699 : but also takes a (possibly empty) list of existing library handles 3700 : that the current library depends on 3701 : 3702 : ---------------------------------------------------------------------- 3703 : Name: Matthias Blume 3704 : blume 1011 Date: 2002/01/10 3705 : Tag: blume-20020110-newffigen 3706 : Description: 3707 : 3708 : 1. Updates to portable graph code. 3709 : 3710 : 2. Major update to ml-nlffigen and ml-nlffi-lib. Things are much 3711 : more scalable now so that even huge interfaces such as the one 3712 : for GTK compile in finite time and space. :-) 3713 : See src/ml-nlffigen/README for details on what's new. 3714 : 3715 : ---------------------------------------------------------------------- 3716 : george 1003 Name: Lal George 3717 : george 1009 Date: 2001/01/09 14:31:35 EST 2002 3718 : Tag: george-20011206-rm-native-copy 3719 : Description: 3720 : 3721 : Removed the native COPY and FCOPY instructions 3722 : from all the architectures and replaced it with the 3723 : explicit COPY instruction from the previous commit. 3724 : 3725 : It is now possible to simplify many of the optimizations 3726 : modules that manipulate copies. This has not been 3727 : done in this change. 3728 : 3729 : ---------------------------------------------------------------------- 3730 : Name: Lal George 3731 : george 1003 Date: 2001/12/06 16:50:13 EST 2001 3732 : Tag: george-20011206-mlrisc-instruction 3733 : Description: 3734 : george 824 3735 : george 1003 Changed the representation of instructions from being fully abstract 3736 : to being partially concrete. That is to say: 3737 : 3738 : from 3739 : type instruction 3740 : 3741 : to 3742 : type instr (* machine instruction *) 3743 : 3744 : datatype instruction = 3745 : LIVE of {regs: C.cellset, spilled: C.cellset} 3746 : | KILL of {regs: C.cellset, spilled: C.cellset} 3747 : | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list} 3748 : | ANNOTATION of {i: instruction, a: Annotations.annotation} 3749 : | INSTR of instr 3750 : 3751 : This makes the handling of certain special instructions that appear on 3752 : all architectures easier and uniform. 3753 : 3754 : LIVE and KILL say that a list of registers are live or killed at the 3755 : program point where they appear. No spill code is generated when an 3756 : element of the 'regs' field is spilled, but the register is moved to 3757 : the 'spilled' (which is present, more for debugging than anything else). 3758 : 3759 : LIVE replaces the (now deprecated) DEFFREG instruction on the alpha. 3760 : We used to generate: 3761 : 3762 : DEFFREG f1 3763 : f1 := f2 + f3 3764 : trapb 3765 : 3766 : but now generate: 3767 : 3768 : f1 := f2 + f3 3769 : trapb 3770 : LIVE {regs=[f1,f2,f3], spilled=[]} 3771 : 3772 : Furthermore, the DEFFREG (hack) required that all floating point instruction 3773 : use all registers mentioned in the instruction. Therefore f1 := f2 + f3, 3774 : defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting 3775 : in a cleaner alpha implementation. (Hopefully, intel will not get rid of 3776 : this architecture). 3777 : 3778 : COPYXXX is intended to replace the parallel COPY and FCOPY available on 3779 : all the architectures. This will result in further simplification of the 3780 : register allocator that must be aware of them for coalescing purposes, and 3781 : will also simplify certain aspects of the machine description that provides 3782 : callbacks related to parallel copies. 3783 : 3784 : ANNOTATION should be obvious, and now INSTR represents the honest to God 3785 : machine instruction set! 3786 : 3787 : The /instructions/Instr.sml files define certain utility 3788 : functions for making porting easier -- essentially converting upper case 3789 : to lower case. All machine instructions (of type instr) are in upper case, 3790 : and the lower case form generates an MLRISC instruction. For example on 3791 : the alpha we have: 3792 : 3793 : datatype instr = 3794 : LDA of {r:cell, b:cell, d:operand} 3795 : | ... 3796 : 3797 : val lda : {r:cell, b:cell, d:operand} -> instruction 3798 : ... 3799 : 3800 : where lda is just (INSTR o LDA), etc. 3801 : 3802 : leunga 815 ---------------------------------------------------------------------- 3803 : blume 986 Name: Matthias Blume 3804 : blume 993 Date: 2001/11/22 21:40:00 EST 3805 : Tag: Release_110_37 3806 : Description: 3807 : 3808 : Release 110.37. This time for real. 3809 : 3810 : ---------------------------------------------------------------------- 3811 : Name: Matthias Blume 3812 : blume 988 Date: 2001/11/21 16:35:00 EST 3813 : Tag: blume-20011121-foot-in-mouth 3814 : Description: 3815 : 3816 : Removed the "Release_110_37" tag because of a serious bug. 3817 : This will be re-tagged once the bug is fixed. 3818 : 3819 : ---------------------------------------------------------------------- 3820 : Name: Matthias Blume 3821 : blume 987 Date: 2001/11/21 16:14:00 EST 3822 : Tag: blume-20011121-forgottenfile 3823 : Description: 3824 : 3825 : Forgot to add a file. (Just a .tex-file -- part of 3826 : the CM manual source.) 3827 : 3828 : ---------------------------------------------------------------------- 3829 : Name: Matthias Blume 3830 : blume 986 Date: 2001/11/21 16:10:00 EST 3831 : blume 988 Tag: blume-20011121-invalid_110_37 3832 : blume 986 Description: 3833 : 3834 : blume 988 Note: I removed the original tag "Release_110_37" from this commit 3835 : because we found a serious bug in all non-x86 backends. 3836 : - Matthias 3837 : 3838 : blume 986 1. Modifications to the SML/NJ code generator and to the runtime system 3839 : so that code object name strings are directly inserted into code 3840 : objects at code generation time. The only business the runtime system 3841 : has with this is now to read the name strings on occasions. 3842 : (The encoding of the name string has also changed somewhat.) 3843 : 3844 : 2. CM now implements a simple "set calculus" for specifying export lists. 3845 : In particular, it is now possible to refer to the export lists of 3846 : other libraries/groups/sources and form unions as well as differences. 3847 : See the latest CM manual for details. 3848 : 3849 : 3. An separate notion of "proxy" libraries has again be eliminated from 3850 : CM's model. (Proxy libraries are now simply a special case of using 3851 : the export list calculus.) 3852 : 3853 : 4. Some of the existing libraries now take advantage of the new set 3854 : calculus. 3855 : (Notice that not all libraries have been converted because some 3856 : of the existing .cm-files are supposed to be backward compatible 3857 : with 110.0.x.) 3858 : 3859 : 5. Some cleanup in stand-alone programs. (Don't use "exnMessage" -- use 3860 : "General.exnMessage"! The former relies on a certain hook to be 3861 : initialized, and that often does not happen in the stand-alone case.) 3862 : 3863 : ---------------------------------------------------------------------- 3864 : george 984 Name: Lal George 3865 : Date: 2001/11/21 13:56:18 EST 3866 : Tag: george-2001121-pseudo-ops 3867 : Description: 3868 : 3869 : Implemented a complete redesign of MLRISC pseudo-ops. Now there 3870 : ought to never be any question of incompatabilities with 3871 : pseudo-op syntax expected by host assemblers. 3872 : 3873 : For now, only modules supporting GAS syntax are implemented 3874 : but more should follow, such as MASM, and vendor assembler 3875 : syntax, e.g. IBM as, Sun as, etc. 3876 : 3877 : ---------------------------------------------------------------------- 3878 : blume 977 Name: Matthias Blume 3879 : Date: 2001/11/14 11:52:00 EST 3880 : Tag: blume-20011114-srcname 3881 : Description: 3882 : 3883 : 1. Routed the name of the current source file to mlriscgen where it 3884 : should be directly emitted into the code object. (This last part 3885 : is yet to be done.) 3886 : 3887 : 2. Some cleanup of the pgraph code to make it match the proposal that 3888 : I put out the other day. (The proposal notwithstanding, things are 3889 : still in flux here.) 3890 : 3891 : ---------------------------------------------------------------------- 3892 : george 976 Name: Lal George 3893 : Date: 2001/11/14 09:44:04 EST 3894 : Tag: 3895 : Description: 3896 : 3897 : Fix for a backpatching bug reported by Allen. 3898 : 3899 : Because the boundary between short and long span-dependent 3900 : instructions is +/- 128, there are an astounding number of 3901 : span-dependent instructions whose size is over estimated. 3902 : 3903 : Allen came up with the idea of letting the size of span 3904 : dependent instructions be non-monotonic, for a maxIter 3905 : number of times, after which the size must be monotonically 3906 : increasing. 3907 : 3908 : This table shows the number of span-dependent instructions 3909 : whose size was over-estimated as a function of maxIter, for the 3910 : file Parse/parse/ml.grm.sml: 3911 : 3912 : maxIter # of instructions: 3913 : 10 687 3914 : 20 438 3915 : 30 198 3916 : 40 0 3917 : 3918 : In compiling the compiler, there is no significant difference in 3919 : compilation speed between maxIter=10 and maxIter=40. Actually, 3920 : my measurements showed that maxIter=40 was a tad faster than 3921 : maxIter=10! Also 96% of the files in the compiler reach a fix 3922 : point within 13 iterations, so fixing maxIter at 40, while high, 3923 : is okay. 3924 : 3925 : ---------------------------------------------------------------------- 3926 : blume 975 Name: Matthias Blume 3927 : Date: 2001/10/31 15:25:00 EST 3928 : Tag: blume-20011031-pgraph 3929 : Description: 3930 : 3931 : CKIT: 3932 : * Changed the "Function" constructor of type Ast.ctype to carry optional 3933 : argument identifiers. 3934 : * Changed the return type of TypeUtil.getFunction accordingly. 3935 : * Type equality ignores the argument names. 3936 : * TypeUtil.composite tries to preserve argument names but gives up quickly 3937 : if there is a mismatch. 3938 : 3939 : installation script: 3940 : * attempts to use "curl" if available (unless "wget" is available as well) 3941 : 3942 : CM: 3943 : * has an experimental implementation of "portable graphs" which I will 3944 : soon propose as an implementation-independent library format 3945 : * there are also new libraries $/pgraph.cm and$/pgraph-util.cm 3946 : 3947 : NLFFI-LIB: 3948 : * some cleanup (all cosmetic) 3949 : 3950 : NLFFIGEN: 3951 : * temporarily disabled the mechanism that suppresses ML output for 3952 : C definitions whose identifiers start with an underscore character 3953 : * generate val bindings for enum constants 3954 : * user can request that only one style (light or heavy) is being used; 3955 : default is to use both (command-line arguments: -heavy and -light) 3956 : * fixed bug in handling of function types involving incomplete pointers 3957 : * generate ML entry points that take record arguments (i.e., using 3958 : named arguments) for C functions that have a prototype with named 3959 : arguments 3960 : (see changes to CKIT) 3961 : 3962 : ---------------------------------------------------------------------- 3963 : leunga 972 Name: Allen Leung 3964 : Date: 2001/10/27 20:34:00 EDT 3965 : Tag: leunga-20011027-x86-fast-fp-call 3966 : Description: 3967 : 3968 : Fixed the bug described in blume-20010920-slowfp. 3969 : 3970 : The fix involves 3971 : 1. generating FCOPYs in FSTP in ia32-svid 3972 : 2. marking a CALL with the appropriate annotation 3973 : 3974 : ---------------------------------------------------------------------- 3975 : blume 961 Name: Matthias Blume 3976 : Date: 2001/10/16 11:32:00 EDT 3977 : Tag: blume-20011016-netbsd 3978 : Description: 3979 : 3980 : Underscore patch from Chris Richards (fixing problem with compiling 3981 : runtime system under recent NetBSD). 3982 : 3983 : ---------------------------------------------------------------------- 3984 : george 959 Name: Allen Leung 3985 : Date: 2001/10/12 17:18:32 EDT 2001 3986 : Tag: leung-20011012-x86-printflowgraph 3987 : Description: 3988 : 3989 : X86RA now uses a valid (instead of dummy) PrintFlowgraph module. 3990 : 3991 : ---------------------------------------------------------------------- 3992 : george 952 Name: Lal George 3993 : george 958 Date: 2001/10/11 23:51:34 EDT 3994 : Tag: george-20011011-too-many-instrs 3995 : Description: 3996 : 3997 : The representation of a program point never expected to see more 3998 : than 65536 instructions in a basic block! 3999 : 4000 : ---------------------------------------------------------------------- 4001 : Name: Lal George 4002 : george 952 Date: 2001/10/09 09:41:37 EDT 4003 : Tag: george-20011008-mlrisc-labels 4004 : Description: 4005 : 4006 : Changed the machine description files to support printing of 4007 : local and global labels in assembly code, based on host assembler 4008 : conventions. 4009 : 4010 : ---------------------------------------------------------------------- 4011 : blume 916 Name: Matthias Blume 4012 : blume 939 Date: 2001/09/25 15:25:00 EDT 4013 : Tag: blume-20010925-exninfo 4014 : Description: 4015 : 4016 : I provided a non-hook implementation of exnName (at the toplevel) and 4017 : made the "dummy" implementation of exnMessage (at the toplevel) more 4018 : useful: if nothing gets "hooked in", then at least you are going to 4019 : see the exception name and a message indicating why you don't see more. 4020 : 4021 : [For the time being, programs that need exnMessage and want to use 4022 : ml-build should either use General.exnMessage (strongly recommended) or 4023 : refer to structure General at some other point so that CM sees a 4024 : static dependency.] 4025 : 4026 : [Similar remarks go for "print" and "use": If you want to use their 4027 : functionality in stand-alone programs generated by ml-build, then use 4028 : TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).] 4029 : 4030 : ---------------------------------------------------------------------- 4031 : Name: Matthias Blume 4032 : blume 938 Date: 2001/09/20 17:28:00 EDT 4033 : Tag: blume-20010920-slowfp 4034 : Description: 4035 : 4036 : Allen says that x86-fast-fp is not safe yet, so I turned it off again... 4037 : 4038 : ---------------------------------------------------------------------- 4039 : Name: Matthias Blume 4040 : blume 937 Date: 2001/09/20 17:20:00 EDT 4041 : Tag: blume-20010920-canonicalpaths 4042 : Description: 4043 : 4044 : 0. Updated the BOOT file (something that I forgot to do earlier). 4045 : 4046 : 1. Small internal change to CM so that it avoids "/../" in filenames 4047 : as much as possible (but only where it is safe). 4048 : 4049 : 2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so 4050 : that arguments that contain delimiters are passed through correctly. 4051 : This change also means that all "special" arguments of the form 4052 : @SMLxxx... must come first. 4053 : 4054 : 3. Changed install script to put relative anchor names for tool commands 4055 : into pathconfig. 4056 : 4057 : ---------------------------------------------------------------------- 4058 : Name: Matthias Blume 4059 : blume 932 Date: 2001/09/18 15:35:00 EDT 4060 : Tag: blume-20010918-readme11036 4061 : Description: 4062 : 4063 : Added README files. 4064 : 4065 : ---------------------------------------------------------------------- 4066 : Name: Matthias Blume 4067 : blume 929 Date: 2001/09/18 11:45:00 EDT 4068 : Tag: Release_110_36 (retag) 4069 : Description: 4070 : 4071 : Fixed mistake in config/preloads. Retagged as 110.36. 4072 : 4073 : ---------------------------------------------------------------------- 4074 : Name: Matthias Blume 4075 : blume 928 Date: 2001/09/18 09:40:00 EDT 4076 : blume 929 Tag: Release_110_36_orig (tag changed) 4077 : blume 928 Description: 4078 : 4079 : New version (110.36). New bootfiles. 4080 : 4081 : ---------------------------------------------------------------------- 4082 : Name: Matthias Blume 4083 : blume 927 Date: 2001/09/14 16:15:00 EDT 4084 : Tag: blume-20010914-x86fastfp 4085 : Description: 4086 : 4087 : John committed some changes that Allen made, in particular a (hopefully) 4088 : correctly working version of the x86-fp module. 4089 : 4090 : I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp" 4091 : flag to "true". Everything seems to compile to a fixpoint ok, and 4092 : "mandelbrot" speeds up by about 15%. 4093 : 4094 : ---------------------------------------------------------------------- 4095 : Name: Matthias Blume 4096 : blume 923 Date: 2001/09/13 11:20:00 EDT 4097 : Tag: blume-20010913-minimal 4098 : Description: 4099 : 4100 : 1. Stefan Monnier's patch to fix a miscompilation problem that 4101 : was brought to light by John Reppy's work on Moby. 4102 : 4103 : 2. Implemented a minimal "structure Compiler" that contains just 4104 : "version" and "architecture". The minimal version will be 4105 : available when the full version is not. This is for backward- 4106 : compatibility with code that wants to test Compiler.version. 4107 : 4108 : ---------------------------------------------------------------------- 4109 : Name: Matthias Blume 4110 : blume 920 Date: 2001/08/28 14:03:00 EDT 4111 : Tag: blume-20010828-ml-lex 4112 : Description: 4113 : 4114 : Fix for bug 1581, received from Neophytos Michael. 4115 : 4116 : ---------------------------------------------------------------------- 4117 : Name: Matthias Blume 4118 : blume 918 Date: 2001/08/27 11:20:00 EDT 4119 : Tag: blume-20010827-readme11035 4120 : Description: 4121 : 4122 : Fleshed out the README file for 110.35. 4123 : 4124 : ---------------------------------------------------------------------- 4125 : Name: Matthias Blume 4126 : blume 916 Date: 2001/08/24 17:10:00 EDT 4127 : Tag: Release_110_35 4128 : Description: 4129 : 4130 : New version number (110.35). New bootfiles. 4131 : 4132 : ---------------------------------------------------------------------- 4133 : george 915 Name: Lal George 4134 : george 909 Date: 2001/08/24 13:47:18 EDT 2001 4135 : Tag: george-20010824-MLRISC-graphs 4136 : Description: 4137 : 4138 : removed clusters from MLRISC completely and replaced with graphs. 4139 : 4140 : ---------------------------------------------------------------------- 4141 : Name: Matthias Blume 4142 : blume 905 Date: 2001/08/23 17:50:00 EDT 4143 : Tag: blume-20010823-toplevel 4144 : Description: 4145 : 4146 : - some reorganization of the code that implements various kinds of 4147 : environments in the compiler (static, dynamic, symbolic, combined) 4148 : - re-implemented the EnvRef module so that evalStream works properly 4149 : (if the stream contains references to "use", "CM.make", etc.) 4150 : - cleaned up evalloop.sml and interact.sml (but they need more cleaning) 4151 : 4152 : ---------------------------------------------------------------------- 4153 : Name: Matthias Blume 4154 : blume 904 Date: 2001/08/20 15:50 EDT 4155 : Tag: blume20010820-slipup 4156 : Description: 4157 : 4158 : I forgot to commit a few files. Here they are... 4159 : 4160 : ---------------------------------------------------------------------- 4161 : Name: Matthias Blume 4162 : blume 903 Date: 2001/08/20 15:35:00 EDT 4163 : Tag: blume-20010820-debugprof 4164 : Description: 4165 : 4166 : !!!! NEW BOOTFILES !!!! 4167 : 4168 : This is another round of reorganizing the compiler sources. This 4169 : time the main goal was to factor out all the "instrumentation" 4170 : passes (for profiling and backtracing) into their own library. 4171 : The difficulty was to do it in such a way that it does not depend 4172 : on elaborate.cm but only on elabdata.cm. 4173 : 4174 : Therefore there have been further changes to both elaborate.cm and 4175 : elabdata.cm -- more "generic" things have been moved from the former 4176 : to the latter. As a result, I was forced to split the assignment 4177 : of numbers indicating "primtyc"s into two portions: SML-generic and 4178 : SML/NJ-specific. Since it would have been awkward to maintain, 4179 : I bit the bullet and actually _changed_ the mapping between these 4180 : numbers and primtycs. The bottom line of this is that you need 4181 : a new set of bin- and bootfiles. 4182 : 4183 : I have built new bootfiles for all architectures, so doing a fresh 4184 : checkout and config/install.sh should be all you need. 4185 : 4186 : The newly created library's name is 4187 : 4188 :$smlnj/viscomp/debugprof.cm 4189 : 4190 : and its sources live under 4191 : 4192 : src/compiler/DebugProf 4193 : 4194 : ---------------------------------------------------------------------- 4195 : Name: Matthias Blume 4196 : blume 902 Date: 2001/08/15 17:15:00 EDT 4197 : Tag: blume-20010815-compreorg 4198 : Description: 4199 : 4200 : This is a first cut at reorganizing the CM libraries that make up the 4201 : core of the compiler. The idea is to separate out pieces that could 4202 : be used independently by tools, e.g., the parser, the typechecker, etc. 4203 : 4204 : The current status is a step in this direction, but it is not quite 4205 : satisfactory yet. Expect more changes in the future. 4206 : 4207 : Here is the current (new) organization... 4208 : 4209 : What used to be $smlnj/viscomp/core.cm is now divided into 4210 : six CM libraries: 4211 : 4212 :$smlnj/viscomp/basics.cm 4213 : /parser.cm 4214 : /elabdata.cm 4215 : /elaborate.cm 4216 : /execute.cm 4217 : /core.cm 4218 : 4219 : The CM files for these libraries live under src/system/smlnj/viscomp. 4220 : All these libraries are proxy libraries that contain precisely 4221 : one CM library component. Here are the locations of the components 4222 : (all within the src/compiler tree): 4223 : 4224 : Basics/basics.cm 4225 : Parse/parser.cm 4226 : ElabData/elabdata.cm 4227 : Elaborator/elaborate.cm 4228 : Execution/execute.cm 4229 : core.cm 4230 : 4231 : [This organization is the same that has been used already 4232 : for a while for the architecture-specific parts of the visible 4233 : compiler and for the old version of core.cm.] 4234 : 4235 : As you will notice, many source files have been moved from their 4236 : respective original locations to a new home in one of the above 4237 : subtrees. 4238 : 4239 : The division of labor between the new libraries is the following: 4240 : 4241 : basics.cm: 4242 : - Simple, basic definitions that pertain to many (or all) of 4243 : the other libraries. 4244 : parser.cm: 4245 : - The SML parser, producing output of type Ast.dec. 4246 : - The type family for Ast is also defined and exported here. 4247 : elabdata.cm: 4248 : - The datatypes that describe input and output of the elaborator. 4249 : This includes types, absyn, and static environments. 4250 : elaborator.cm: 4251 : - The SML/NJ type checker and elaborator. 4252 : This maps an Ast.dec (with a given static environment) to 4253 : an Absyn.dec (with a new static environment). 4254 : - This libraries implements certain modules that used to be 4255 : structures as functors (to remove dependencies on FLINT). 4256 : execute.cm: 4257 : - Everything having to do with executing binary code objects. 4258 : - Dynamic environments. 4259 : core.cm: 4260 : - SML/NJ-specific instantiations of the elaborator and MLRISC. 4261 : - Top-level modules. 4262 : - FLINT (this should eventually become its own library) 4263 : 4264 : Notes: 4265 : 4266 : I am not 100% happy with the way I separated the elaborator (and its 4267 : data structures) from FLINT. Two instances of the same problem: 4268 : 4269 : 1. Data structures contain certain fields that carry FLINT-specific 4270 : information. I hacked around this using exn and the property list 4271 : module from smlnj-lib. But the fact that there are middle-end 4272 : specific fields around at all is a bit annoying. 4273 : 4274 : 2. The elaborator calculates certain FLINT-related information. I tried 4275 : to make this as abstract as I could using functorization, but, again, 4276 : the fact that the elaborator has to perform calculations on behalf 4277 : of the middle-end at all is not nice. 4278 : 4279 : 3. Having to used exn and property lists is unfortunate because it 4280 : weakens type checking. The other alternative (parameterizing 4281 : nearly *everything*) is not appealing, though. 4282 : 4283 : I removed the "rebinding =" warning hack because due to the new organization 4284 : it was awkward to maintain it. As a result, the compiler now issues some of 4285 : these warnings when compiling init.cmi during bootstrap compilation. On 4286 : the plus side, you also get a warning when you do, for example: 4287 : val op = = Int32.+ 4288 : which was not the case up to now. 4289 : 4290 : I placed "assign" and "deref" into the _Core structure so that the 4291 : code that deals with the "lazy" keyword can find them there. This