| 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 : |
|
|
Name:
|
| 11 : |
|
|
Date:
|
| 12 : |
|
|
Tag: <post-commit CVS tag>
|
| 13 : |
|
|
Description:
|
| 14 : |
george |
705 |
|
| 15 : |
leunga |
591 |
----------------------------------------------------------------------
|
| 16 : |
blume |
715 |
Name: Matthias Blume
|
| 17 : |
blume |
716 |
Date: 2000/11/04 14:45:00
|
| 18 : |
|
|
Tag: blume-20001104-mlbuild
|
| 19 : |
|
|
Description:
|
| 20 : |
|
|
|
| 21 : |
|
|
- Made ml-build faster on startup.
|
| 22 : |
|
|
- Documentation fixes.
|
| 23 : |
|
|
|
| 24 : |
|
|
----------------------------------------------------------------------
|
| 25 : |
|
|
Name: Matthias Blume
|
| 26 : |
blume |
715 |
Date: 2000/11/02 17:00:00 JST
|
| 27 : |
|
|
Tag: blume-20001102-condcomp
|
| 28 : |
|
|
Description:
|
| 29 : |
|
|
|
| 30 : |
|
|
- Small tweaks to pickler -- new BOOTFILES!
|
| 31 : |
|
|
- Version bumped to 110.29.2.
|
| 32 : |
|
|
- Added conditional compilation facility to init.cmi (see comment there).
|
| 33 : |
|
|
----------------------------------------------------------------------
|
| 34 : |
leunga |
713 |
Name: Allen Leung
|
| 35 : |
|
|
Date: 2000/10/23 19:31:00
|
| 36 : |
|
|
Tag: leunga-20001023-demo-ra
|
| 37 : |
|
|
|
| 38 : |
|
|
1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
|
| 39 : |
|
|
2. Test programs for the graph library updated
|
| 40 : |
|
|
3. Some new MLRISC demo programs added
|
| 41 : |
|
|
|
| 42 : |
|
|
----------------------------------------------------------------------
|
| 43 : |
blume |
710 |
Name: Matthias Blume
|
| 44 : |
blume |
711 |
Date: 2000/08/31 22:15:00 JST
|
| 45 : |
|
|
Tag: blume-20001017-errmsg
|
| 46 : |
|
|
Description:
|
| 47 : |
|
|
|
| 48 : |
|
|
More error message grief: Where there used to be no messages, there
|
| 49 : |
|
|
now were some that had bogus error regions. Fixed.
|
| 50 : |
|
|
|
| 51 : |
|
|
----------------------------------------------------------------------
|
| 52 : |
|
|
Name: Matthias Blume
|
| 53 : |
blume |
710 |
Date: 2000/08/31 17:30:00 JST
|
| 54 : |
|
|
Tag: blume-20001017-v110p29p1
|
| 55 : |
|
|
Description:
|
| 56 : |
|
|
|
| 57 : |
|
|
I made a version 110.29.1 with new bootfiles.
|
| 58 : |
|
|
|
| 59 : |
|
|
Changes: Modified pickler/unpickler for faster and leaner unpickling.
|
| 60 : |
|
|
CM documentation changes and a small bugfix in CM's error reporting.
|
| 61 : |
|
|
|
| 62 : |
|
|
----------------------------------------------------------------------
|
| 63 : |
george |
705 |
Name: Lal George
|
| 64 : |
|
|
Date: 2000/09/27 14:42:35 EDT
|
| 65 : |
|
|
Tag: george-20000927-nodestatus
|
| 66 : |
|
|
Description:
|
| 67 : |
|
|
|
| 68 : |
|
|
Changed the type of the nodestatus, so that:
|
| 69 : |
|
|
|
| 70 : |
|
|
SPILLED(~1) is now SPILLED
|
| 71 : |
|
|
SPILLED(m) where m>=0 is now MEMREG(m)
|
| 72 : |
|
|
SPILLED(s) where s<~1 is now SPILL_LOC(~s)
|
| 73 : |
|
|
|
| 74 : |
|
|
----------------------------------------------------------------------
|
| 75 : |
blume |
697 |
Name: Matthias Blume
|
| 76 : |
blume |
703 |
Date: 2000/09/07 14:45:00 JST
|
| 77 : |
|
|
Tag: blume-20000907-cmerrmsg
|
| 78 : |
|
|
Description:
|
| 79 : |
|
|
|
| 80 : |
|
|
Small tweak to CM to avoid getting ML syntax error messages twice.
|
| 81 : |
|
|
|
| 82 : |
|
|
----------------------------------------------------------------------
|
| 83 : |
|
|
Name: Matthias Blume
|
| 84 : |
blume |
701 |
Date: 2000/08/31 18:00:00 JST
|
| 85 : |
|
|
Tag: blume-20000831-cvsbootfiles
|
| 86 : |
|
|
Description:
|
| 87 : |
|
|
|
| 88 : |
|
|
New URL for boot files (because the 110.29 files on the BL server do
|
| 89 : |
|
|
now work correctly with my updated install scripts for yacc and lex).
|
| 90 : |
|
|
|
| 91 : |
|
|
----------------------------------------------------------------------
|
| 92 : |
|
|
Name: Matthias Blume
|
| 93 : |
blume |
697 |
Date: 2000/08/08 12:33:00 JST
|
| 94 : |
|
|
Tag: blume-20000808-manual
|
| 95 : |
|
|
Description:
|
| 96 : |
|
|
|
| 97 : |
|
|
Tiny update to CM manual.
|
| 98 : |
|
|
|
| 99 : |
|
|
----------------------------------------------------------------------
|
| 100 : |
leunga |
695 |
Name: Allen Leung
|
| 101 : |
|
|
Date: 2000/08/7 19:31:00
|
| 102 : |
|
|
Tag: leunga-20000807-a-whole-bunch-of-stuff
|
| 103 : |
|
|
|
| 104 : |
|
|
Moby, C--, SSA, x86, machine descriptions etc. Should only affect C--
|
| 105 : |
|
|
and Mobdy.
|
| 106 : |
|
|
|
| 107 : |
|
|
1. x86
|
| 108 : |
|
|
|
| 109 : |
|
|
a. Fixes to peephole module by John and Dan.
|
| 110 : |
|
|
b. Assembly fix to SETcc by Allen.
|
| 111 : |
|
|
c. Fix to c-call by John.
|
| 112 : |
|
|
d. Fix to spilling by John. (This one deals with the missing FSTPT case)
|
| 113 : |
|
|
e. Instruction selection optimization to SETcc as suggested by John.
|
| 114 : |
|
|
|
| 115 : |
|
|
For example,
|
| 116 : |
|
|
|
| 117 : |
|
|
MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
|
| 118 : |
|
|
|
| 119 : |
|
|
should generate:
|
| 120 : |
|
|
|
| 121 : |
|
|
MOVL a, x
|
| 122 : |
|
|
SUBL b, x
|
| 123 : |
|
|
SHRL 31, x
|
| 124 : |
|
|
|
| 125 : |
|
|
2. IR stuff
|
| 126 : |
|
|
|
| 127 : |
|
|
A bunch of new DJ-graph related algorithms added. These
|
| 128 : |
|
|
speed up SSA construction.
|
| 129 : |
|
|
|
| 130 : |
|
|
3. SSA + Scheduling
|
| 131 : |
|
|
|
| 132 : |
|
|
Added code for SSA and scheduling to the repository
|
| 133 : |
|
|
|
| 134 : |
|
|
----------------------------------------------------------------------
|
| 135 : |
george |
694 |
Name: Lal George
|
| 136 : |
|
|
Date: 2000/07/27 11:53:14 EDT
|
| 137 : |
|
|
|
| 138 : |
|
|
Tag: lal-20000727-linux-ppc
|
| 139 : |
|
|
Description:
|
| 140 : |
|
|
|
| 141 : |
|
|
Made changes to support Linux PPC.
|
| 142 : |
|
|
p.s. I have confirmation that the 110.29 boot files work fine.
|
| 143 : |
|
|
|
| 144 : |
|
|
----------------------------------------------------------------------
|
| 145 : |
blume |
684 |
Name: Matthias Blume
|
| 146 : |
blume |
692 |
Date: 2000/07/27 17:40:00 JST
|
| 147 : |
|
|
Tag: blume-20000727-scripts
|
| 148 : |
|
|
Description:
|
| 149 : |
|
|
|
| 150 : |
|
|
!!!! WARNING !!!!
|
| 151 : |
|
|
You must recompile the runtime system!
|
| 152 : |
|
|
!!!! WARNING !!!!
|
| 153 : |
|
|
|
| 154 : |
|
|
This is basically another round of script-enhancements:
|
| 155 : |
|
|
|
| 156 : |
|
|
1. sml, ml-build, and ml-makedepend accept options -D and -U to define
|
| 157 : |
|
|
and undefine CM preprocessor symbols.
|
| 158 : |
|
|
|
| 159 : |
|
|
2. ml-build avoids generating a new heap image if it finds that the
|
| 160 : |
|
|
existing one is still ok. (The condition is that no ML file had to
|
| 161 : |
|
|
be recompiled and all ML files are found to be older that the heap
|
| 162 : |
|
|
file.)
|
| 163 : |
|
|
|
| 164 : |
|
|
To make this work smoothly, I also hacked the runtime system as
|
| 165 : |
|
|
well as SMLofNJ.SysInfo to get access to the heap image suffix
|
| 166 : |
|
|
(.sparc-solaris, ...) that is currently being used.
|
| 167 : |
|
|
|
| 168 : |
|
|
Moreover, the signature of CM.mk_standalone has changed. See the
|
| 169 : |
|
|
CM manual.
|
| 170 : |
|
|
|
| 171 : |
|
|
3. ml-makedepend accepts additional options -n, -a, and -o. (See the
|
| 172 : |
|
|
CM manual for details.)
|
| 173 : |
|
|
|
| 174 : |
|
|
4. More CM manual updates:
|
| 175 : |
|
|
- all of the above has been documented.
|
| 176 : |
|
|
- there is now a section describing the (CM-related) command line
|
| 177 : |
|
|
arguments that are accepted by the "sml" command
|
| 178 : |
|
|
|
| 179 : |
|
|
----------------------------------------------------------------------
|
| 180 : |
|
|
Name: Matthias Blume
|
| 181 : |
blume |
691 |
Date: 2000/07/25 16:20:00 JST
|
| 182 : |
|
|
Tag: blume-20000725-makedepend
|
| 183 : |
|
|
Description:
|
| 184 : |
|
|
|
| 185 : |
|
|
Added a script called ml-makedepend. This can be used in makefiles
|
| 186 : |
|
|
for Unix' make in a way very similar to the "makedepend" command for
|
| 187 : |
|
|
C.
|
| 188 : |
|
|
|
| 189 : |
|
|
The script internally uses function CM.sources.
|
| 190 : |
|
|
|
| 191 : |
|
|
Synopsis:
|
| 192 : |
|
|
|
| 193 : |
|
|
ml-makedepend [-f makefile] cmfile targetname
|
| 194 : |
|
|
|
| 195 : |
|
|
The default for the makefile is "makefile" (or "Makefile" should
|
| 196 : |
|
|
"makefile" not exist).
|
| 197 : |
|
|
|
| 198 : |
|
|
ml-makedepend adds a cmfile/targetname-specific section to this
|
| 199 : |
|
|
makefile (after removing the previous version of this section). The
|
| 200 : |
|
|
section contains a single dependency specification with targetname on
|
| 201 : |
|
|
the LHS (targetname is an arbitrary name), and a list of files derived
|
| 202 : |
|
|
from the cmfile on the RHS. Some of the files on the RHS are
|
| 203 : |
|
|
ARCH/OPSYS-specific. Therefore, ml-makedepend inserts references to
|
| 204 : |
|
|
"make" variables $(ARCH) and $(OPSYS) in place of the corresponding
|
| 205 : |
|
|
path names. The makefile writer is responsible for making sure that
|
| 206 : |
|
|
these variables have correct at the time "make" is invoked.
|
| 207 : |
|
|
|
| 208 : |
|
|
----------------------------------------------------------------------
|
| 209 : |
|
|
Name: Matthias Blume
|
| 210 : |
blume |
690 |
Date: 2000/07/22 23:30:00 JST
|
| 211 : |
|
|
Tag: blume-20000722-urlupdate
|
| 212 : |
|
|
Description:
|
| 213 : |
|
|
|
| 214 : |
|
|
Changed BOOT and config/srcarchiveurl to point to BL server:
|
| 215 : |
|
|
|
| 216 : |
|
|
ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
|
| 217 : |
|
|
|
| 218 : |
|
|
----------------------------------------------------------------------
|
| 219 : |
|
|
Name: Matthias Blume
|
| 220 : |
blume |
686 |
Date: 2000/07/18 18:00:00 JST
|
| 221 : |
|
|
Tag: blume-20000718-Version_110_29
|
| 222 : |
|
|
Description:
|
| 223 : |
|
|
|
| 224 : |
|
|
1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
|
| 225 : |
|
|
|
| 226 : |
|
|
2. Updated config/version to 110.29
|
| 227 : |
|
|
|
| 228 : |
|
|
3. Updated config/srcarchiveurl
|
| 229 : |
|
|
|
| 230 : |
|
|
3. New boot files!
|
| 231 : |
|
|
ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
|
| 232 : |
|
|
|
| 233 : |
|
|
----------------------------------------------------------------------
|
| 234 : |
|
|
Name: Matthias Blume
|
| 235 : |
blume |
684 |
Date: 2000/07/11 13:58:00 JST
|
| 236 : |
|
|
Tag: blume-20000711-doctypo
|
| 237 : |
|
|
Description:
|
| 238 : |
|
|
|
| 239 : |
|
|
Fixed a few typos in CM manual.
|
| 240 : |
|
|
|
| 241 : |
|
|
----------------------------------------------------------------------
|
| 242 : |
leunga |
683 |
Name: Allen Leung
|
| 243 : |
|
|
Date: 2000/06/15 00:38:00
|
| 244 : |
|
|
Tag: leunga-20000704-sparc-x86
|
| 245 : |
|
|
|
| 246 : |
|
|
1. x86 peephole improvement sp += k; sp -= k => nop [from John]
|
| 247 : |
|
|
2. fix to x86 RET bug [found by Dan Grossman]
|
| 248 : |
|
|
3. sparc assembly bug fix for ticc instructions [found by Fermin]
|
| 249 : |
|
|
|
| 250 : |
|
|
Affects c-- and moby only
|
| 251 : |
|
|
|
| 252 : |
|
|
----------------------------------------------------------------------
|
| 253 : |
blume |
680 |
Name: Matthias Blume
|
| 254 : |
blume |
682 |
Date: 2000/07/04 15:26:00
|
| 255 : |
|
|
Tag: blume-20000704-trigger
|
| 256 : |
|
|
Description:
|
| 257 : |
|
|
|
| 258 : |
|
|
1. Improvements to CM manual.
|
| 259 : |
|
|
2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
|
| 260 : |
|
|
of getting a back-trace. The function, when called, raises an
|
| 261 : |
|
|
internal exception which explicitly carries the full back-trace history,
|
| 262 : |
|
|
so it is unaffected by any intervening handle-raise pairs ("trivial"
|
| 263 : |
|
|
or not). The interactive loop will print that history once it arrives
|
| 264 : |
|
|
at top level.
|
| 265 : |
|
|
Short of having all exceptions implicitly carry the full history, the
|
| 266 : |
|
|
recommended way of using this facility is:
|
| 267 : |
|
|
- compile your program with instrumentation "on"
|
| 268 : |
|
|
- run it, when it raises an exception, look at the history
|
| 269 : |
|
|
- if the history is "cut off" because of some handler, go and modify
|
| 270 : |
|
|
your program so that it explicitly calls BTrace.trigger
|
| 271 : |
|
|
- recompile (still instrumented), and rerun; look at the full history
|
| 272 : |
|
|
|
| 273 : |
|
|
----------------------------------------------------------------------
|
| 274 : |
|
|
Name: Matthias Blume
|
| 275 : |
|
|
Date: 2000/07/03 15:36:00 JST
|
| 276 : |
blume |
680 |
Tag: blume-20000702-manual
|
| 277 : |
|
|
Description:
|
| 278 : |
|
|
|
| 279 : |
|
|
Small corrections and updates to CM manual.
|
| 280 : |
|
|
|
| 281 : |
|
|
----------------------------------------------------------------------
|
| 282 : |
blume |
682 |
Name: Matthias Blume
|
| 283 : |
blume |
679 |
Date: 2000/06/29 16:04:00 JST
|
| 284 : |
|
|
Tag: blume-20000629-yacctool
|
| 285 : |
|
|
Description:
|
| 286 : |
|
|
|
| 287 : |
|
|
Changes:
|
| 288 : |
|
|
|
| 289 : |
|
|
1. Class "mlyacc" now takes separate arguments to pass options to
|
| 290 : |
|
|
generated .sml- and .sig-files independently.
|
| 291 : |
|
|
2. Corresponding CM manual updates.
|
| 292 : |
|
|
3. BTrace module now also reports call sites. (However, for loop clusters
|
| 293 : |
|
|
it only shows from where the cluster was entered.) There are associated
|
| 294 : |
|
|
modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
|
| 295 : |
|
|
|
| 296 : |
|
|
----------------------------------------------------------------------
|
| 297 : |
blume |
675 |
Name: Matthias Blume
|
| 298 : |
blume |
678 |
Date: 2000/06/27 16:51:00 JST
|
| 299 : |
|
|
Tag: blume-20000627-noweb
|
| 300 : |
|
|
Description:
|
| 301 : |
|
|
|
| 302 : |
|
|
Changes:
|
| 303 : |
|
|
|
| 304 : |
|
|
1. Implemented "subdir" and "witness" options for noweb tool.
|
| 305 : |
|
|
This caused some slight internal changes in CM's tool implementation.
|
| 306 : |
|
|
2. Fixed bug in "tool plugin" mechanism. This is essentially cleaning
|
| 307 : |
|
|
some remaining issues from earlier path anchor changes.
|
| 308 : |
|
|
3. Updated CM manual accordingly.
|
| 309 : |
|
|
|
| 310 : |
|
|
4. Changed implementation of back-tracing so that I now consider it
|
| 311 : |
|
|
ready for prime-time.
|
| 312 : |
|
|
|
| 313 : |
|
|
In particular, you don't have to explicitly trigger the back-trace
|
| 314 : |
|
|
anymore. Instead, if you are running BTrace-instrumented code and
|
| 315 : |
|
|
there is an uncaught exception (regardless of whether or not it was
|
| 316 : |
|
|
raised in instrumented code), the top-level evalloop will print
|
| 317 : |
|
|
the back-trace.
|
| 318 : |
|
|
|
| 319 : |
|
|
Features:
|
| 320 : |
|
|
|
| 321 : |
|
|
- Instrumented and uninstrumented code work together seemlessly.
|
| 322 : |
|
|
(Of course, uninstrumented code is never mentioned in actual
|
| 323 : |
|
|
back-traces.)
|
| 324 : |
|
|
|
| 325 : |
|
|
- Asymptotic time- and space-complexity of instrumented code is
|
| 326 : |
|
|
equal to that of uninstrumented code. (This means that
|
| 327 : |
|
|
tail-recursion is preserved by the instrumentation phase.)
|
| 328 : |
|
|
|
| 329 : |
|
|
- Modules whose code has been instrumented in different sessions
|
| 330 : |
|
|
work together without problem.
|
| 331 : |
|
|
|
| 332 : |
|
|
- There is no penalty whatsoever on uninstrumented code.
|
| 333 : |
|
|
|
| 334 : |
|
|
- There is no penalty on "raise" expressions, even in
|
| 335 : |
|
|
instrumented code.
|
| 336 : |
|
|
|
| 337 : |
|
|
A potential bug (or perhaps it is a feature, too):
|
| 338 : |
|
|
|
| 339 : |
|
|
A back-trace reaches no further than the outermost instrumented
|
| 340 : |
|
|
non-trivial "raise". Here, a "trivial" raise is one that is the
|
| 341 : |
|
|
sole RHS of a "handle" rule. Thus, back-traces reach trough
|
| 342 : |
|
|
|
| 343 : |
|
|
<exp> handle e => raise e
|
| 344 : |
|
|
|
| 345 : |
|
|
and even
|
| 346 : |
|
|
|
| 347 : |
|
|
<exp> handle Foo => raise Bar
|
| 348 : |
|
|
|
| 349 : |
|
|
and, of course, through
|
| 350 : |
|
|
|
| 351 : |
|
|
<exp> handle Foo => ...
|
| 352 : |
|
|
|
| 353 : |
|
|
if the exception was not Foo.
|
| 354 : |
|
|
|
| 355 : |
|
|
Back-traces always reach right through any un-instrumented code
|
| 356 : |
|
|
including any of its "handle" expressions, trivial or not.
|
| 357 : |
|
|
|
| 358 : |
|
|
To try this out, do the following:
|
| 359 : |
|
|
|
| 360 : |
|
|
- Erase all existing binfiles for your program.
|
| 361 : |
|
|
(You may keep binfiles for those modules where you think you
|
| 362 : |
|
|
definitely don't need back-tracing.)
|
| 363 : |
|
|
- Turn on back-trace instrumentation:
|
| 364 : |
|
|
SMLofNJ.Internals.BTrace.mode (SOME true);
|
| 365 : |
|
|
- Recompile your program. (I.e., run "CM.make" or "use".)
|
| 366 : |
|
|
- You may now turn instrumentation off again (if you want):
|
| 367 : |
|
|
SMLofNJ.Internals.BTrace.mode (SOME false);
|
| 368 : |
|
|
- Run your program as usual. If it raises an exception that
|
| 369 : |
|
|
reaches the interactive toplevel, then a back-trace will
|
| 370 : |
|
|
automatically be printed. After that, the toplevel loop
|
| 371 : |
|
|
will print the exception history as usual.
|
| 372 : |
|
|
|
| 373 : |
|
|
----------------------------------------------------------------------
|
| 374 : |
|
|
Name: Matthias Blume
|
| 375 : |
blume |
677 |
Date: 2000/06/26 09:56:46 JST
|
| 376 : |
|
|
Tag: blume-20000626-setup
|
| 377 : |
|
|
Description:
|
| 378 : |
|
|
|
| 379 : |
|
|
CM: - setup-parameter to "sml" added; this can be used to run arbitrary
|
| 380 : |
|
|
ML code before and after compiling a file (e.g., to set compiler
|
| 381 : |
|
|
flags)
|
| 382 : |
|
|
|
| 383 : |
|
|
Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
|
| 384 : |
|
|
- associated changes to btrace.sml (BTrace instrumentation pass)
|
| 385 : |
|
|
- cleaner implementation of btimp.sml (BTrace tracing and report
|
| 386 : |
|
|
module)
|
| 387 : |
|
|
|
| 388 : |
|
|
CM manual: * new path encoding documented
|
| 389 : |
|
|
* description of setup-parameter to "sml" added
|
| 390 : |
|
|
|
| 391 : |
|
|
The biggest user-visible change to back-tracing is that it is no
|
| 392 : |
|
|
longer necessary to compile all traced modules within the same
|
| 393 : |
|
|
session. (This was a real limitation.)
|
| 394 : |
|
|
|
| 395 : |
|
|
----------------------------------------------------------------------
|
| 396 : |
|
|
Name: Matthias Blume
|
| 397 : |
blume |
676 |
Date: 2000/06/24 12:40:00 JST
|
| 398 : |
|
|
Tag: blume-20000624-startup
|
| 399 : |
|
|
Description:
|
| 400 : |
|
|
|
| 401 : |
|
|
Fixes startup slowdown problem. (I was calling SrcPath.sync a _tad_
|
| 402 : |
|
|
bit too often -- to put it mildly. :)
|
| 403 : |
|
|
|
| 404 : |
|
|
----------------------------------------------------------------------
|
| 405 : |
|
|
Name: Matthias Blume
|
| 406 : |
blume |
675 |
Date: 2000/06/23 18:20:00 JST
|
| 407 : |
|
|
Tag: blume-20000623-btrace
|
| 408 : |
|
|
Description:
|
| 409 : |
|
|
|
| 410 : |
|
|
This updates adds a backtrace facility to aid programmers in debugging
|
| 411 : |
|
|
their programs. This involves the following changes:
|
| 412 : |
|
|
|
| 413 : |
|
|
1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
|
| 414 : |
|
|
keeping track of the current call stack. When programs are compiled
|
| 415 : |
|
|
in a special mode, the compiler will insert calls to these hooks
|
| 416 : |
|
|
into the user program.
|
| 417 : |
|
|
"Hook" means that it is possible for different implementations of
|
| 418 : |
|
|
back-tracing to register themselves (at different times).
|
| 419 : |
|
|
|
| 420 : |
|
|
2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
|
| 421 : |
|
|
as an Absyn.dec->Absyn.dec rewrite. Normally this phase is turned off.
|
| 422 : |
|
|
It can be turned on using this call:
|
| 423 : |
|
|
SMLofNJ.Internals.BTrace.mode (SOME true);
|
| 424 : |
|
|
Turning it off again:
|
| 425 : |
|
|
SMLofNJ.Internals.BTrace.mode (SOME false);
|
| 426 : |
|
|
Querying the current status:
|
| 427 : |
|
|
SMLofNJ.Internals.BTrace.mode NONE;
|
| 428 : |
|
|
Annotated programs are about twice as big as normal ones, and they
|
| 429 : |
|
|
run a factor of 2 to 4 slower with a dummy back-trace plugin (one
|
| 430 : |
|
|
where all hooks do nothing). The slowdown with a plugin that is
|
| 431 : |
|
|
actually useful (such as the one supplied by default) is even greater,
|
| 432 : |
|
|
but in the case of the default plugin it is still only an constant
|
| 433 : |
|
|
factor (amortized).
|
| 434 : |
|
|
|
| 435 : |
|
|
3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
|
| 436 : |
|
|
with a sub-structure BTrace for controlling back-tracing. In particular,
|
| 437 : |
|
|
the above-mentioned function "mode" controls whether the annotation
|
| 438 : |
|
|
phase is invoked by the compiler. Another important function is
|
| 439 : |
|
|
"trigger": when called it aborts the current execution and causes
|
| 440 : |
|
|
the top-level loop to print a full back-trace.
|
| 441 : |
|
|
|
| 442 : |
|
|
4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
|
| 443 : |
|
|
for back-tracing. It keeps track of the dynamic call stack and in
|
| 444 : |
|
|
addition to that it keeps a partial history at each "level" of that
|
| 445 : |
|
|
stack. For example, if a tail-calls b, b tail-calls c, and c tail-calls
|
| 446 : |
|
|
d and b (at separate times, dynamically), then the report will show:
|
| 447 : |
|
|
|
| 448 : |
|
|
GOTO d
|
| 449 : |
|
|
/c
|
| 450 : |
|
|
GOTO \b
|
| 451 : |
|
|
CALL a
|
| 452 : |
|
|
|
| 453 : |
|
|
This shows that there was an initial non-tail call of a, then a
|
| 454 : |
|
|
tail-call to b or c, looping behavior in a cluster of functions that
|
| 455 : |
|
|
consist of b and c, and then a goto from that cluster (i.e., either from
|
| 456 : |
|
|
b or from c) to d.
|
| 457 : |
|
|
|
| 458 : |
|
|
Note that (depending on the user program) the amount of information
|
| 459 : |
|
|
that the back-trace module has to keep track of at each level is bounded
|
| 460 : |
|
|
by a constant. Thus, the whole implementation has the same asymptotical
|
| 461 : |
|
|
complexity as the original program (both in space and in time).
|
| 462 : |
|
|
|
| 463 : |
|
|
5. compiler/TopLevel/interact/evalloop.sml has been modified to
|
| 464 : |
|
|
handle the special exception SMLofNJ.Internals.BTrace.BTrace
|
| 465 : |
|
|
which is raised by the "trigger" function mentioned above.
|
| 466 : |
|
|
|
| 467 : |
|
|
Notes on usage:
|
| 468 : |
|
|
|
| 469 : |
|
|
- Annotated code works well together with unannotated code:
|
| 470 : |
|
|
Unannotated calls simply do not show up at all in the backtrace.
|
| 471 : |
|
|
|
| 472 : |
|
|
- It is not a good idea to let modules that were annotated during
|
| 473 : |
|
|
different sessions run at the same time. This is because the compiler
|
| 474 : |
|
|
chooses small integers to identify individual functions, and there
|
| 475 : |
|
|
will be clashes if different modules were compiled in separate sessions.
|
| 476 : |
|
|
(Nothing will crash, and you will even be told about the clashes, but
|
| 477 : |
|
|
back-trace information will in general not be useful.)
|
| 478 : |
|
|
|
| 479 : |
|
|
- Back-tracing can be confused by callcc and capture.
|
| 480 : |
|
|
|
| 481 : |
|
|
- The only way of getting a back-trace right now is to explicitly
|
| 482 : |
|
|
invoke the "trigger" function from your user program. Eventually, we
|
| 483 : |
|
|
should make every exception carry back-trace information (if
|
| 484 : |
|
|
available). But since this creates more overhead at "raise"-time
|
| 485 : |
|
|
(similar to the current exnHistory overhead), I have not yet
|
| 486 : |
|
|
implemented this. (The implementation will be rather easy.) With
|
| 487 : |
|
|
exceptions carrying back-trace information, this facility will be even
|
| 488 : |
|
|
more useful because users don't need to modify their programs...
|
| 489 : |
|
|
|
| 490 : |
|
|
- While it is possible to compile the compiler with back-trace
|
| 491 : |
|
|
annotations turned on (I did it to get some confidence in
|
| 492 : |
|
|
correctness), you must make absolutely sure that core.sml and
|
| 493 : |
|
|
btimp.sml are compiled WITHOUT annotation! (core.sml cannot actually
|
| 494 : |
|
|
be compiled with annotation because there is no core access yet, but
|
| 495 : |
|
|
if you compile btimp.sml with annotation, then the system will go into
|
| 496 : |
|
|
an infinite recursion and crash.)
|
| 497 : |
|
|
Since CM currently does not know about BTrace, the only way to turn
|
| 498 : |
|
|
annotations on and off for different modules of the compiler is to
|
| 499 : |
|
|
interrupt CMB.make, change the settings, and re-invoke it. Of course,
|
| 500 : |
|
|
this is awkward and clumsy.
|
| 501 : |
|
|
|
| 502 : |
|
|
Sample sessions:
|
| 503 : |
|
|
|
| 504 : |
|
|
Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
|
| 505 : |
|
|
- SMLofNJ.Internals.BTrace.mode (SOME true);
|
| 506 : |
|
|
[autoloading]
|
| 507 : |
|
|
[autoloading done]
|
| 508 : |
|
|
val it = false : bool
|
| 509 : |
|
|
- structure X = struct
|
| 510 : |
|
|
- fun main n = let
|
| 511 : |
|
|
- fun a (x, 0) = d x
|
| 512 : |
|
|
- | a (x, n) = b (x, n - 1)
|
| 513 : |
|
|
- and b (x, n) = c (x, n)
|
| 514 : |
|
|
- and c (x, n) = a (x, n)
|
| 515 : |
|
|
- and d x = e (x, 3)
|
| 516 : |
|
|
- and e (x, 0) = f x
|
| 517 : |
|
|
- | e (x, n) = e (x, n - 1)
|
| 518 : |
|
|
- and f 0 = SMLofNJ.Internals.BTrace.trigger ()
|
| 519 : |
|
|
- | f n = n * g (n - 1)
|
| 520 : |
|
|
- and g n = a (n, 3)
|
| 521 : |
|
|
- in
|
| 522 : |
|
|
- f n
|
| 523 : |
|
|
- end
|
| 524 : |
|
|
- end;
|
| 525 : |
|
|
structure X : sig val main : int -> int end
|
| 526 : |
|
|
- X.main 3;
|
| 527 : |
|
|
*** BACK-TRACE ***
|
| 528 : |
|
|
GOTO stdIn:4.2-13.20: X.main[2].f
|
| 529 : |
|
|
GOTO-( stdIn:4.2-13.20: X.main[2].e
|
| 530 : |
|
|
GOTO stdIn:4.2-13.20: X.main[2].d
|
| 531 : |
|
|
/ stdIn:4.2-13.20: X.main[2].a
|
| 532 : |
|
|
| stdIn:4.2-13.20: X.main[2].b
|
| 533 : |
|
|
GOTO-\ stdIn:4.2-13.20: X.main[2].c
|
| 534 : |
|
|
CALL stdIn:4.2-13.20: X.main[2].g
|
| 535 : |
|
|
GOTO stdIn:4.2-13.20: X.main[2].f
|
| 536 : |
|
|
GOTO-( stdIn:4.2-13.20: X.main[2].e
|
| 537 : |
|
|
GOTO stdIn:4.2-13.20: X.main[2].d
|
| 538 : |
|
|
/ stdIn:4.2-13.20: X.main[2].a
|
| 539 : |
|
|
| stdIn:4.2-13.20: X.main[2].b
|
| 540 : |
|
|
GOTO-\ stdIn:4.2-13.20: X.main[2].c
|
| 541 : |
|
|
CALL stdIn:4.2-13.20: X.main[2].g
|
| 542 : |
|
|
GOTO stdIn:4.2-13.20: X.main[2].f
|
| 543 : |
|
|
GOTO-( stdIn:4.2-13.20: X.main[2].e
|
| 544 : |
|
|
GOTO stdIn:4.2-13.20: X.main[2].d
|
| 545 : |
|
|
/ stdIn:4.2-13.20: X.main[2].a
|
| 546 : |
|
|
| stdIn:4.2-13.20: X.main[2].b
|
| 547 : |
|
|
GOTO-\ stdIn:4.2-13.20: X.main[2].c
|
| 548 : |
|
|
CALL stdIn:4.2-13.20: X.main[2].g
|
| 549 : |
|
|
GOTO stdIn:4.2-13.20: X.main[2].f
|
| 550 : |
|
|
CALL stdIn:2.15-17.4: X.main[2]
|
| 551 : |
|
|
-
|
| 552 : |
|
|
|
| 553 : |
|
|
(Note that because of a FLINt bug the above code currently does not
|
| 554 : |
|
|
compile without BTrace turned on.)
|
| 555 : |
|
|
|
| 556 : |
|
|
Here is another example, using my modified Tiger compiler:
|
| 557 : |
|
|
|
| 558 : |
|
|
Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
|
| 559 : |
|
|
- SMLofNJ.Internals.BTrace.mode (SOME true);
|
| 560 : |
|
|
[autoloading]
|
| 561 : |
|
|
[autoloading done]
|
| 562 : |
|
|
val it = false : bool
|
| 563 : |
|
|
- CM.make "sources.cm";
|
| 564 : |
|
|
[autoloading]
|
| 565 : |
|
|
...
|
| 566 : |
|
|
[autoloading done]
|
| 567 : |
|
|
[scanning sources.cm]
|
| 568 : |
|
|
[parsing (sources.cm):parse.sml]
|
| 569 : |
|
|
[creating directory CM/SKEL ...]
|
| 570 : |
|
|
[parsing (sources.cm):tiger.lex.sml]
|
| 571 : |
|
|
...
|
| 572 : |
|
|
[wrote CM/sparc-unix/semant.sml]
|
| 573 : |
|
|
[compiling (sources.cm):main.sml]
|
| 574 : |
|
|
[wrote CM/sparc-unix/main.sml]
|
| 575 : |
|
|
[New bindings added.]
|
| 576 : |
|
|
val it = true : bool
|
| 577 : |
|
|
- Main.compile ("../testcases/merge.tig", "foo.out");
|
| 578 : |
|
|
*** BACK-TRACE ***
|
| 579 : |
|
|
CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
|
| 580 : |
|
|
CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
|
| 581 : |
|
|
CALL lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
|
| 582 : |
|
|
GOTO lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
|
| 583 : |
|
|
CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
|
| 584 : |
|
|
CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
|
| 585 : |
|
|
CALL lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
|
| 586 : |
|
|
/ lib/semant.sml:411.65-543.8: SemantFun[2].transDec
|
| 587 : |
|
|
CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
|
| 588 : |
|
|
CALL lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
|
| 589 : |
|
|
CALL lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
|
| 590 : |
|
|
CALL main.sml:1.18-118.4: Main.compile[2]
|
| 591 : |
|
|
-
|
| 592 : |
|
|
|
| 593 : |
|
|
----------------------------------------------------------------------
|
| 594 : |
blume |
673 |
Name: Matthias Blumen
|
| 595 : |
|
|
Date: 2000/06/21 18:00:00 JST
|
| 596 : |
|
|
Tag: blume-20000621-manual
|
| 597 : |
|
|
Description:
|
| 598 : |
|
|
|
| 599 : |
|
|
CM manual update: Path environments documented.
|
| 600 : |
|
|
|
| 601 : |
|
|
----------------------------------------------------------------------
|
| 602 : |
blume |
666 |
Name: Matthias Blume
|
| 603 : |
blume |
672 |
Date: 2000/06/19 13:40:00
|
| 604 : |
|
|
Tag: blume-20000619-manual
|
| 605 : |
|
|
Description:
|
| 606 : |
|
|
|
| 607 : |
|
|
CM manual and system/README update. This only covers the fact that
|
| 608 : |
|
|
there are no more implicit anchors. (Path environments and the "bind"
|
| 609 : |
|
|
option to "cm" have yet to be documented.)
|
| 610 : |
|
|
|
| 611 : |
|
|
----------------------------------------------------------------------
|
| 612 : |
|
|
Name: Matthias Blume
|
| 613 : |
blume |
671 |
Date: 2000/06/19 11:05:00 JST
|
| 614 : |
|
|
Tag: blume-20000619-chdir-bugfix
|
| 615 : |
|
|
Description:
|
| 616 : |
|
|
|
| 617 : |
|
|
Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
|
| 618 : |
|
|
|
| 619 : |
|
|
----------------------------------------------------------------------
|
| 620 : |
|
|
Name: Matthias Blume
|
| 621 : |
blume |
670 |
Date: 2000/06/18 22:00:10 JST
|
| 622 : |
|
|
Tag: blume-20000618-implicit-anchors-really-gone
|
| 623 : |
|
|
Description:
|
| 624 : |
|
|
|
| 625 : |
|
|
I updates the previous HISTORY entry where I forgot to mention that
|
| 626 : |
|
|
implicit anchors are no longer with us.
|
| 627 : |
|
|
|
| 628 : |
|
|
The current update also gets rid of the (now useless) controller
|
| 629 : |
|
|
CM.Control.implicit_anchors.
|
| 630 : |
|
|
|
| 631 : |
|
|
----------------------------------------------------------------------
|
| 632 : |
|
|
Name: Matthias Blume
|
| 633 : |
blume |
666 |
Date: 2000/06/16 17:30:00 JST
|
| 634 : |
|
|
Tag: blume-20000616-anchorenv
|
| 635 : |
|
|
Description:
|
| 636 : |
|
|
|
| 637 : |
|
|
This patch implements the long anticipated (just kidding :) "anchor
|
| 638 : |
|
|
environment" mechanism. In the course of doing this, I also
|
| 639 : |
|
|
re-implemented CM's internal "SrcPath" module from scratch. The new
|
| 640 : |
|
|
one should be more robust in certain boundary cases. In any case, it
|
| 641 : |
|
|
is a lot cleaner than its predecessor (IMHO).
|
| 642 : |
|
|
|
| 643 : |
|
|
This time, although there is yet another boot file format change, I
|
| 644 : |
|
|
kept the unpickler backward-compatible. As a result, no new bootfiles
|
| 645 : |
|
|
are necessary and bootstrapping is straightforward. (You cannot read
|
| 646 : |
|
|
new bootfiles into an old system, but the other way around is no
|
| 647 : |
|
|
problem.)
|
| 648 : |
|
|
|
| 649 : |
|
|
Visible changes:
|
| 650 : |
|
|
|
| 651 : |
blume |
670 |
** 0. Implicit path anchors (without the leading $-symbol) are no
|
| 652 : |
|
|
longer recognized at all. This means that such path names are not
|
| 653 : |
|
|
illegal either. For example, the name basis.cm simply refers to a
|
| 654 : |
|
|
local file called "basis.cm" (i.e, the name is an ordinary path
|
| 655 : |
|
|
relative to .cm-files directory). Or, to put it differently, only
|
| 656 : |
|
|
names that start with $ are anchored paths.
|
| 657 : |
blume |
666 |
|
| 658 : |
blume |
670 |
** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
|
| 659 : |
|
|
vanished.
|
| 660 : |
|
|
|
| 661 : |
|
|
John (Reppy) had critizised this as soon as I originally proposed and
|
| 662 : |
blume |
666 |
implemented it, but at that time I did not really deeply believe
|
| 663 : |
|
|
him. :) Now I came full-circle because I need the $<singlearc> syntax
|
| 664 : |
|
|
in another place where it cannot be seen as an abbreviation for
|
| 665 : |
|
|
$/<singlearc>. To avoid the confusion, $<singlearc> now means what it
|
| 666 : |
|
|
seems to mean (i.e., it "expands" into the corresponding anchor
|
| 667 : |
|
|
value).
|
| 668 : |
|
|
|
| 669 : |
|
|
However, when paths are used as members in CM description files, it
|
| 670 : |
|
|
continues to be true that there must be at least another arc after the
|
| 671 : |
|
|
anchor. This is now enforced separately during semantic analysis
|
| 672 : |
blume |
670 |
(i.e., from a lexical/syntactical point of view, the notation is ok.)
|
| 673 : |
blume |
666 |
|
| 674 : |
|
|
** 2. The "cm" class now accepts an option "bind". The option's value
|
| 675 : |
|
|
is a sub-option list of precisely two items -- one labeled "anchor"
|
| 676 : |
|
|
and the other one labeled "value". As you might expect, "anchor" is
|
| 677 : |
|
|
used to specify an anchor name to be bound, and "value" specifies what
|
| 678 : |
|
|
the anchor is being bound to.
|
| 679 : |
|
|
|
| 680 : |
|
|
The value must be a directory name and can be given in either standard
|
| 681 : |
|
|
syntax (including the possibility that it is itself an anchored path)
|
| 682 : |
|
|
or native syntax.
|
| 683 : |
|
|
|
| 684 : |
|
|
Examples:
|
| 685 : |
|
|
|
| 686 : |
|
|
foo.cm (bind:(anchor:bar value:$mystuff/bar))
|
| 687 : |
|
|
lib.cm (bind:(anchor:a value:"H:\\x\\y\\z")) (* only works under windows *)
|
| 688 : |
|
|
|
| 689 : |
|
|
and so on.
|
| 690 : |
|
|
|
| 691 : |
|
|
The meaning of this is that the .cm-file will be processed with an
|
| 692 : |
|
|
augmented anchor environment where the given anchor(s) is/are bound to
|
| 693 : |
|
|
the given values(s).
|
| 694 : |
|
|
|
| 695 : |
|
|
The rationale for having this feature is this: Suppose you are trying
|
| 696 : |
|
|
to use two different (already stable) libraries a.cm and b.cm (that
|
| 697 : |
|
|
you perhaps didn't write yourself). Further, suppose each of these
|
| 698 : |
|
|
two libraries internally uses its own auxiliary library $aux/lib.cm.
|
| 699 : |
|
|
Normally you would now have a problem because the anchor "lib" can not
|
| 700 : |
|
|
be bound to more than one value globally. Therefore, the project that
|
| 701 : |
|
|
uses both a.cm and b.cm must locally redirect the anchor to some other
|
| 702 : |
|
|
place:
|
| 703 : |
|
|
|
| 704 : |
|
|
a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
|
| 705 : |
|
|
b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
|
| 706 : |
|
|
|
| 707 : |
|
|
This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
|
| 708 : |
|
|
/usr/lib/smlnj/b-stuff/aux.cm, respectively.
|
| 709 : |
|
|
|
| 710 : |
|
|
Hard-wiring path names is a bit inflexible (and CM will verbosely warn
|
| 711 : |
|
|
you when you do so at the time of CM.stabilize). Therefore, you can
|
| 712 : |
|
|
also use an anchored path as the value:
|
| 713 : |
|
|
|
| 714 : |
|
|
a.cm (bind:(anchor:lib value:$a-lib))
|
| 715 : |
|
|
b.cm (bind:(anchor:lib value:$b-lib))
|
| 716 : |
|
|
|
| 717 : |
|
|
Now you can globally configure (using the usual CM.Anchor.anchor or
|
| 718 : |
|
|
pathconfig machinery) bindings for "a-lib" and "b-lib". Since "lib"
|
| 719 : |
|
|
itself is always locally bound, setting it globally is no longer
|
| 720 : |
|
|
meaningful or necessary (but it does not hurt either). In fact, "lib"
|
| 721 : |
|
|
can still be used as a global anchor for separate purposes. As a
|
| 722 : |
|
|
matter of fact, one can locally define "lib" in terms of a global
|
| 723 : |
|
|
"lib":
|
| 724 : |
|
|
|
| 725 : |
|
|
a.cm (bind:(anchor:lib value:$lib/a))
|
| 726 : |
|
|
b.cm (bind:(anchor:lib value:$lib/b))
|
| 727 : |
|
|
|
| 728 : |
|
|
** 3: The encoding of path names has changed. This affects the way
|
| 729 : |
|
|
path names are shown in CM's progress report and also the internal
|
| 730 : |
|
|
protocol encoding used for parallel make.
|
| 731 : |
|
|
|
| 732 : |
|
|
The encoding now uses one or more ':'-separated segments. Each
|
| 733 : |
|
|
segments corresponds to a file that has been specified relative to the
|
| 734 : |
|
|
file given by its preceding segment. The first segment is either
|
| 735 : |
|
|
relative to the CWD, absolute, or anchored. Each segment itself is
|
| 736 : |
|
|
basically a Unix pathname; all segments but the first are relative.
|
| 737 : |
|
|
|
| 738 : |
|
|
Example:
|
| 739 : |
|
|
|
| 740 : |
|
|
$foo/bar/baz.cm:a/b/c.sml
|
| 741 : |
|
|
|
| 742 : |
|
|
This path denotes the file bar/a/b/c.sml relative to the directory
|
| 743 : |
|
|
denoted by anchor "foo". Notice that the encoding also includes
|
| 744 : |
|
|
baz.cm which is the .cm-file that listed a/b/c.sml. As usual, such
|
| 745 : |
|
|
paths are resolved relative to the .cm-files directory, so baz.cm must
|
| 746 : |
|
|
be ignored to get the "real" pathname.
|
| 747 : |
|
|
|
| 748 : |
|
|
To make this fact more obvious, CM puts the names of such "virtual
|
| 749 : |
|
|
arcs" into parentheses when they appear in progress reports. (No
|
| 750 : |
|
|
parentheses will appear in the internal protocol encoding.) Thus,
|
| 751 : |
|
|
what you really see is:
|
| 752 : |
|
|
|
| 753 : |
|
|
$foo/bar/(baz.cm):a/b/c.sml
|
| 754 : |
|
|
|
| 755 : |
|
|
I find this notation to be much more informative than before.
|
| 756 : |
|
|
|
| 757 : |
|
|
Another new feature of the encoding is that special characters
|
| 758 : |
|
|
including parentheses, colons, (back)slashes, and white space are
|
| 759 : |
|
|
written as \ddd (where ddd is the decimal encoding of the character).
|
| 760 : |
|
|
|
| 761 : |
blume |
670 |
*** The CM manual still needs to be updated.
|
| 762 : |
|
|
|
| 763 : |
blume |
666 |
----------------------------------------------------------------------
|
| 764 : |
leunga |
660 |
Name: Allen Leung
|
| 765 : |
|
|
Date: 2000/06/15 00:38:00
|
| 766 : |
|
|
Tag: leunga-20000615-x86-peephole
|
| 767 : |
|
|
|
| 768 : |
|
|
x86 Peephole fix by Fermin. Affects c-- and moby only.
|
| 769 : |
|
|
|
| 770 : |
|
|
----------------------------------------------------------------------
|
| 771 : |
blume |
658 |
Name: Matthias Blume
|
| 772 : |
|
|
Date: 2000/06/12 11:40:00
|
| 773 : |
|
|
Tag: blume-20000612-parmakefix
|
| 774 : |
|
|
Description:
|
| 775 : |
|
|
|
| 776 : |
|
|
More cleanup after changing the file naming scheme: This time I
|
| 777 : |
|
|
repaired the parallel make mechanism for CMB.make which I broke earlier.
|
| 778 : |
|
|
|
| 779 : |
|
|
----------------------------------------------------------------------
|
| 780 : |
leunga |
657 |
Name: Allen Leung
|
| 781 : |
|
|
Date: 2000/06/09 01:25:00
|
| 782 : |
|
|
Tag: leunga-20000609-various
|
| 783 : |
|
|
|
| 784 : |
|
|
None of these things should affect normal SML/NJ operations
|
| 785 : |
|
|
|
| 786 : |
|
|
1. Peephole improvements provided by Fermin (c--)
|
| 787 : |
|
|
2. New annotation DEFUSE for adding extra dependence (moby)
|
| 788 : |
|
|
3. New X86 LOCK instructions (moby)
|
| 789 : |
|
|
4. New machine description language for reservation tables (scheduling)
|
| 790 : |
|
|
5. Fixes to various optimization/analysis modules (branch chaining, dominator
|
| 791 : |
|
|
trees etc.)
|
| 792 : |
|
|
6. I've changed the CM files so that they can work with versions
|
| 793 : |
|
|
110.0.6, 110.25 and 110.28
|
| 794 : |
|
|
|
| 795 : |
|
|
----------------------------------------------------------------------
|
| 796 : |
blume |
652 |
Name: Matthias Blume
|
| 797 : |
blume |
656 |
Date: 2000/06/09 12:40:00
|
| 798 : |
|
|
Tag: blume-20000609-log
|
| 799 : |
|
|
Description:
|
| 800 : |
|
|
|
| 801 : |
|
|
- Removed all(?) remaining RCS Log entries from sources.
|
| 802 : |
|
|
|
| 803 : |
|
|
- Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
|
| 804 : |
|
|
anchored paths).
|
| 805 : |
|
|
|
| 806 : |
|
|
----------------------------------------------------------------------
|
| 807 : |
|
|
Name: Matthias Blume
|
| 808 : |
blume |
653 |
Date: 2000/06/07 17:00:00 JST
|
| 809 : |
|
|
Tag: blume-20000607-no-implicit-anchors
|
| 810 : |
|
|
Description:
|
| 811 : |
|
|
|
| 812 : |
|
|
1. This update changes the default setting for
|
| 813 : |
|
|
CM.Control.implicit_anchors from true to false. This means that
|
| 814 : |
|
|
implicit anchors are no longer permitted by default. I also tried to
|
| 815 : |
|
|
make sure that nothing else still relies on implicit anchors.
|
| 816 : |
|
|
(This is the next step on the schedule towards a CM that does not even
|
| 817 : |
|
|
have the notion of implicit anchors anymore.)
|
| 818 : |
|
|
|
| 819 : |
|
|
2. More CM manual updates.
|
| 820 : |
|
|
|
| 821 : |
|
|
3. I managed to track down and fix the pickling bug I mentioned last
|
| 822 : |
|
|
time. Because of the previously existing workaround, this entails no
|
| 823 : |
|
|
immediate practical changes.
|
| 824 : |
|
|
|
| 825 : |
|
|
----------------------------------------------------------------------
|
| 826 : |
|
|
Name: Matthias Blume
|
| 827 : |
blume |
652 |
Date: 2000/06/06 11:15:00 JST
|
| 828 : |
|
|
Tag: blume-20000606-lazierpickle
|
| 829 : |
|
|
Description:
|
| 830 : |
|
|
|
| 831 : |
|
|
!!!! NEW BOOT FILES !!!!
|
| 832 : |
|
|
|
| 833 : |
|
|
* The main purpose of this update is to make library pickles lazier in
|
| 834 : |
|
|
order to reduce the initial space penalty for autoloading a library.
|
| 835 : |
|
|
As a result, it is now possible to have $smlnj/compiler.cm
|
| 836 : |
|
|
pre-registered. This should take care of the many complaints or
|
| 837 : |
|
|
inquiries about missing structure Compiler. This required changes to
|
| 838 : |
|
|
CM's internal data structures and small tweaks to some algorithms.
|
| 839 : |
|
|
|
| 840 : |
|
|
As a neat additional effect, it is no longer necessary (for the sake
|
| 841 : |
|
|
of lean heap image files) to distinguish between a "minimal" CM and a
|
| 842 : |
|
|
"full" CM. Now, there is only one CM (i.e., the "full" version:
|
| 843 : |
|
|
$smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
|
| 844 : |
|
|
interactive top level. ($smlnj/cm/minimal.cm is gone.)
|
| 845 : |
|
|
|
| 846 : |
|
|
To make the life of compiler-hackers easier, "makeml" now also
|
| 847 : |
|
|
pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm). In other
|
| 848 : |
|
|
words, after you bootstrap a new sml for the first time, you will not
|
| 849 : |
|
|
have to autoload $smlnj/cmb.cm again afterwards. (The first time
|
| 850 : |
|
|
around you will still have to do it, though.)
|
| 851 : |
|
|
|
| 852 : |
|
|
* A second change consists of major updates to the CM manual. There
|
| 853 : |
|
|
are now several appendices with summary information and also a full
|
| 854 : |
|
|
specification of the CM description file syntax.
|
| 855 : |
|
|
|
| 856 : |
|
|
* In directory src/system I added the script "allcross". This script
|
| 857 : |
|
|
invokes sml and cross-compiles the compiler for all supported
|
| 858 : |
|
|
architectures. (Useful when providing a new set of boot files.)
|
| 859 : |
|
|
|
| 860 : |
|
|
* There seems to be a latent bug in my "lazy pickles" mechanism. I
|
| 861 : |
|
|
added a small tweak to pickle-util.sml to work around this problem,
|
| 862 : |
|
|
but it is not a proper fix yet. I will investigate further. (The
|
| 863 : |
|
|
effect of the bug was an inflation of library pickle size.)
|
| 864 : |
|
|
|
| 865 : |
|
|
* Version number increased to 110.28.1 (to avoid compatibility problems).
|
| 866 : |
|
|
|
| 867 : |
|
|
----------------------------------------------------------------------
|
| 868 : |
leunga |
646 |
Name: Allen Leung
|
| 869 : |
leunga |
648 |
Date: 2000/05/25 17:28 EDT
|
| 870 : |
|
|
Tag: leunga-20000525-ra
|
| 871 : |
|
|
Description:
|
| 872 : |
|
|
|
| 873 : |
|
|
Fixed a bug in freezing phase of the register allocator.
|
| 874 : |
|
|
|
| 875 : |
|
|
----------------------------------------------------------------------
|
| 876 : |
|
|
Name: Allen Leung
|
| 877 : |
leunga |
647 |
Date: 2000/05/15 22:53 EDT
|
| 878 : |
leunga |
646 |
Tag: leunga-20000515-alpha-x86-ra
|
| 879 : |
|
|
Description:
|
| 880 : |
|
|
|
| 881 : |
|
|
1. Alpha
|
| 882 : |
|
|
|
| 883 : |
|
|
Slight cleanup. Removed the instruction SGNXL
|
| 884 : |
|
|
|
| 885 : |
|
|
2. X86
|
| 886 : |
|
|
|
| 887 : |
|
|
Added the following instructions to the instruction set:
|
| 888 : |
|
|
|
| 889 : |
|
|
ROLx, RORx,
|
| 890 : |
|
|
BTx, BTSx, BTLx, BTRx,
|
| 891 : |
|
|
XCHGx, and variants with the LOCK prefix
|
| 892 : |
|
|
|
| 893 : |
|
|
3. Register Allocation
|
| 894 : |
|
|
|
| 895 : |
|
|
The module ra-rewrite-with-renaming has been improved.
|
| 896 : |
|
|
|
| 897 : |
leunga |
647 |
These have no effect on SML/NJ.
|
| 898 : |
|
|
|
| 899 : |
leunga |
646 |
----------------------------------------------------------------------
|
| 900 : |
blume |
642 |
Name: Matthias Blume
|
| 901 : |
blume |
645 |
Date: 2000/05/15 16:20:00 JST
|
| 902 : |
|
|
Tag: blume-20000515-lightrebuild
|
| 903 : |
|
|
Description:
|
| 904 : |
|
|
|
| 905 : |
|
|
1. I added an alternative to "-rebuild" to "makeml". The difference is
|
| 906 : |
|
|
that prior to calling CMB.make' the CM-variable "LIGHT" will be
|
| 907 : |
|
|
defined. In effect, the command will not build any cross-compiler
|
| 908 : |
|
|
backends and therefore finish more quickly.
|
| 909 : |
|
|
|
| 910 : |
|
|
The "fixpt" script also takes a "-light" switch to be able to use
|
| 911 : |
|
|
this new facility while compiling for a fixpoint.
|
| 912 : |
|
|
|
| 913 : |
|
|
2. I replaced all mentions of anchored paths in group owner specifications
|
| 914 : |
|
|
with simple relative paths (usually starting with "..").
|
| 915 : |
|
|
The rationale is that a library's internal workings should not be
|
| 916 : |
|
|
compromised by the lack of some anchor. (An anchor is necessary
|
| 917 : |
|
|
for someone who wants to refer to the library by an anchored path,
|
| 918 : |
|
|
but it should not be necessary to build the same library in the first
|
| 919 : |
|
|
place.)
|
| 920 : |
|
|
|
| 921 : |
|
|
3. I changed the way CM's tool mechanism determines the shell command
|
| 922 : |
|
|
string used for things like ml-yacc etc. so that it does not break
|
| 923 : |
|
|
when CM.Control.implicit_anchors is turned off.
|
| 924 : |
|
|
|
| 925 : |
|
|
----------------------------------------------------------------------
|
| 926 : |
|
|
Name: Matthias Blume
|
| 927 : |
blume |
644 |
Date: 2000/05/12 18:20:00 JST
|
| 928 : |
|
|
Tag: blume-20000512-ml-build
|
| 929 : |
|
|
Description:
|
| 930 : |
|
|
|
| 931 : |
|
|
Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
|
| 932 : |
|
|
getting installed properly (by config/install.sh).
|
| 933 : |
|
|
|
| 934 : |
|
|
----------------------------------------------------------------------
|
| 935 : |
|
|
Name: Matthias Blume
|
| 936 : |
blume |
643 |
Date: 2000/05/12 17:30:00 JST
|
| 937 : |
|
|
Tag: blume-20000512-anchors
|
| 938 : |
|
|
Description:
|
| 939 : |
|
|
|
| 940 : |
|
|
!!! NEW BOOT FILES !!!
|
| 941 : |
|
|
|
| 942 : |
|
|
This change is in preparation of fading out support for "implicitly
|
| 943 : |
|
|
anchored path names". I went through all sources and used the
|
| 944 : |
|
|
explicit (and relatively new) $-notation. See system/README and the
|
| 945 : |
|
|
CM manual for more info on this.
|
| 946 : |
|
|
|
| 947 : |
|
|
I also modified the anchoring scheme for some things such as "smlnj",
|
| 948 : |
|
|
"MLRISC", "cm", etc. to take advantage of the fact that explicit
|
| 949 : |
|
|
anchors are more expressive: anchor name and first arc do not have to
|
| 950 : |
|
|
coincide. This entails the following user-visible change:
|
| 951 : |
|
|
|
| 952 : |
|
|
You have to write $smlnj/foo/bar instead of smlnj/foo/bar. In
|
| 953 : |
|
|
particular, when you fire up sml with a command-line argument, say,
|
| 954 : |
|
|
e.g.:
|
| 955 : |
|
|
|
| 956 : |
|
|
sml '$smlnj/cmb.cm'
|
| 957 : |
|
|
|
| 958 : |
|
|
At the ML toplevel prompt:
|
| 959 : |
|
|
|
| 960 : |
|
|
CM.autoload "$smlnj/cmb.cm";
|
| 961 : |
|
|
|
| 962 : |
|
|
There is also a new controller in CM.Control that can be used to turn
|
| 963 : |
|
|
off all remaining support for implicit anchors by saying:
|
| 964 : |
|
|
|
| 965 : |
|
|
CM.autoload "$smlnj/
|
| 966 : |
|
|
#set CM.Control.implicit_anchors false;
|
| 967 : |
|
|
|
| 968 : |
|
|
This causes CM to reject implicitly anchored paths. This is (for the
|
| 969 : |
|
|
time being) less permissive than the "final" version where there will
|
| 970 : |
|
|
be no more such implicit anchors and relative paths will be just that:
|
| 971 : |
|
|
relative.
|
| 972 : |
|
|
|
| 973 : |
|
|
The next step (version after next version?) will be to make the
|
| 974 : |
|
|
default for CM.Control.implicit_anchors false. After the dust has
|
| 975 : |
|
|
settled, I can then produce the "final" version of this...
|
| 976 : |
|
|
|
| 977 : |
|
|
Note: Since bootstrapping is a bit tricky, I provided new boot files.
|
| 978 : |
|
|
|
| 979 : |
|
|
----------------------------------------------------------------------
|
| 980 : |
|
|
Name: Matthias Blume
|
| 981 : |
blume |
642 |
Date: 2000/05/11 16:30:00 JST
|
| 982 : |
|
|
Tag: blume-20000511-sources
|
| 983 : |
|
|
Description:
|
| 984 : |
|
|
|
| 985 : |
|
|
The main change is that I added function CM.sources as a generalized
|
| 986 : |
|
|
version of the earlier CM.makedepend. This entails the following
|
| 987 : |
|
|
additional changes:
|
| 988 : |
|
|
|
| 989 : |
|
|
- CM.makedepend has been dropped.
|
| 990 : |
|
|
|
| 991 : |
|
|
- CM manual has been updated.
|
| 992 : |
|
|
|
| 993 : |
|
|
- TOOLS signature and API have been changed.
|
| 994 : |
|
|
|
| 995 : |
|
|
----------------------------------------------------------------------
|
| 996 : |
leunga |
640 |
Name: Allen Leung
|
| 997 : |
leunga |
641 |
Date: 2000/05/10 21:17 EDT
|
| 998 : |
|
|
Tag: leunga-20000510-moby-c--ssa
|
| 999 : |
|
|
Description:
|
| 1000 : |
|
|
|
| 1001 : |
|
|
Various bug fixes and new features for C--, Moby and MLRISC optimizations.
|
| 1002 : |
|
|
None of these affect SML/NJ.
|
| 1003 : |
|
|
|
| 1004 : |
|
|
1. Register Allocation
|
| 1005 : |
|
|
|
| 1006 : |
|
|
a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
|
| 1007 : |
|
|
This module tries to remove local (i.e. basic block level) redundancies
|
| 1008 : |
|
|
during spilling.
|
| 1009 : |
|
|
|
| 1010 : |
|
|
b. A new framework for performing region based register allocation.
|
| 1011 : |
|
|
Not yet entirely functional.
|
| 1012 : |
|
|
|
| 1013 : |
|
|
2. X86
|
| 1014 : |
|
|
|
| 1015 : |
|
|
a. DefUse for POP was missing the stack pointer [found by Lal]
|
| 1016 : |
|
|
b. Reload for CALL was incorrect in X86Spill [found by John]
|
| 1017 : |
|
|
c. Various fixes in X86Spill so that it can be used correctly for
|
| 1018 : |
|
|
the new spilling module.
|
| 1019 : |
|
|
|
| 1020 : |
|
|
3. SSA/IR
|
| 1021 : |
|
|
|
| 1022 : |
|
|
a. New module ir/dj-dataflow.sml implements elimination based
|
| 1023 : |
|
|
data flow analysis.
|
| 1024 : |
|
|
|
| 1025 : |
|
|
4. MLRiscGen
|
| 1026 : |
|
|
|
| 1027 : |
|
|
a. Fix for gc type annotation
|
| 1028 : |
|
|
|
| 1029 : |
|
|
5. MDGen
|
| 1030 : |
|
|
|
| 1031 : |
|
|
Various fixes for machine description -> ml code translation. For ssa
|
| 1032 : |
|
|
only.
|
| 1033 : |
|
|
|
| 1034 : |
|
|
----------------------------------------------------------------------
|
| 1035 : |
|
|
Name: Allen Leung
|
| 1036 : |
leunga |
640 |
Date: 2000/05/08 22:17 EDT
|
| 1037 : |
|
|
Tag: leunga-20000508-labexp
|
| 1038 : |
|
|
Description:
|
| 1039 : |
|
|
|
| 1040 : |
|
|
Fermin has found a few assembly problems with constant expressions
|
| 1041 : |
|
|
generated in LabelExp. Mostly, the problems involve extra parentheses,
|
| 1042 : |
|
|
which choke on dumb assemblers. This is his fix.
|
| 1043 : |
|
|
|
| 1044 : |
|
|
----------------------------------------------------------------------
|
| 1045 : |
dbm |
635 |
Name: Dave MacQueen
|
| 1046 : |
|
|
Date: 2000/04/09 14:00 EDT
|
| 1047 : |
|
|
Tag: dbm-20000502-Version_110_28
|
| 1048 : |
|
|
Description:
|
| 1049 : |
|
|
|
| 1050 : |
|
|
1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
|
| 1051 : |
|
|
|
| 1052 : |
|
|
2. Updated config/version to 110.28
|
| 1053 : |
|
|
|
| 1054 : |
|
|
3. Updated config/srcarchiveurl
|
| 1055 : |
|
|
|
| 1056 : |
|
|
3. New boot files!
|
| 1057 : |
|
|
ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
|
| 1058 : |
|
|
|
| 1059 : |
|
|
----------------------------------------------------------------------
|
| 1060 : |
blume |
634 |
Name: Matthias Blume
|
| 1061 : |
|
|
Date: 2000/05/01 19:05:00 JST
|
| 1062 : |
|
|
Tag: blume-20000501-noweb
|
| 1063 : |
|
|
Description:
|
| 1064 : |
|
|
|
| 1065 : |
|
|
A new noweb tool has been added. The existing system is entirely
|
| 1066 : |
|
|
unaffected by this, but some CM users have asked for renewed noweb
|
| 1067 : |
|
|
support. Everything is documented in the CM manual.
|
| 1068 : |
|
|
|
| 1069 : |
|
|
New (plugin) libraries:
|
| 1070 : |
|
|
|
| 1071 : |
|
|
noweb-tool.cm
|
| 1072 : |
|
|
nw-ext.cm
|
| 1073 : |
|
|
|
| 1074 : |
|
|
----------------------------------------------------------------------
|
| 1075 : |
dbm |
633 |
Name: Dave MacQueen
|
| 1076 : |
|
|
Date: 2000/04/30 12:40PM EDT
|
| 1077 : |
|
|
Tag: dbm-20000430-bug_fixes
|
| 1078 : |
|
|
Description:
|
| 1079 : |
|
|
|
| 1080 : |
|
|
1. Fix for bug 1498
|
| 1081 : |
|
|
smlnj/src/system/Basis/Implementation/Unsafe/object.sig
|
| 1082 : |
|
|
smlnj/src/system/Basis/Implementation/Unsafe/object.sml
|
| 1083 : |
|
|
added toRealArray function
|
| 1084 : |
|
|
smlnj/src/compiler/MiscUtil/print/ppobj.sml
|
| 1085 : |
|
|
added check for tag Obj.RealArray to array printing case in ppObj
|
| 1086 : |
|
|
|
| 1087 : |
|
|
2. Fix for bug 1510
|
| 1088 : |
|
|
smlnj/src/compiler/Semant/types/typesutil.sml
|
| 1089 : |
|
|
fixed definition of dummyargs (used by equalTycon) so that
|
| 1090 : |
|
|
dummy args are distinct types
|
| 1091 : |
|
|
|
| 1092 : |
|
|
----------------------------------------------------------------------
|
| 1093 : |
blume |
629 |
Name: Matthias Blume
|
| 1094 : |
blume |
632 |
Date: 2000/04/30 01:00:00 JST
|
| 1095 : |
|
|
Tag: blume-20000430-versions
|
| 1096 : |
|
|
Description:
|
| 1097 : |
|
|
|
| 1098 : |
|
|
1. CM version numbering added. This is an implementation of Lal's
|
| 1099 : |
|
|
proposal for adding version numbers and version checking to .cm
|
| 1100 : |
|
|
files. Lal said that his proposal was just that -- a proposal.
|
| 1101 : |
|
|
For the time being I went ahead and implemented it so that people
|
| 1102 : |
|
|
can comment on it. Everything is completely backward-compatible
|
| 1103 : |
|
|
(except for the stable library format, i.e., new bootfiles!).
|
| 1104 : |
|
|
|
| 1105 : |
|
|
As usual, see the CM manual for details.
|
| 1106 : |
|
|
|
| 1107 : |
|
|
2. An alternative syntax for anchored paths has been implemented.
|
| 1108 : |
|
|
Dave has recently voiced the same concerns that I had when I did
|
| 1109 : |
|
|
this, so there should be some support. My take is that eventually
|
| 1110 : |
|
|
I will let support for the current syntax (where anchors are
|
| 1111 : |
|
|
"implicit") fade out in favor of the new, explicit syntax.
|
| 1112 : |
|
|
In order to be backward-compatible, both old and new syntax are
|
| 1113 : |
|
|
currently supported.
|
| 1114 : |
|
|
|
| 1115 : |
|
|
Again, see the CM manual for details.
|
| 1116 : |
|
|
|
| 1117 : |
|
|
3. Parallel make is trying to be slightly smarter: When the master
|
| 1118 : |
|
|
process finds a "bottleneck", i.e., when there is only one
|
| 1119 : |
|
|
compilation unit that can be compiled and everybody else is
|
| 1120 : |
|
|
waiting on it, then it will simply compile it directly instead
|
| 1121 : |
|
|
of clumsily telling one of the slaves to do it.
|
| 1122 : |
|
|
|
| 1123 : |
|
|
4. Support for "unsharing" added. This is necessary in order to be
|
| 1124 : |
|
|
able to have two different versions of the same library running
|
| 1125 : |
|
|
at the same time (e.g., for trying out a new MLRISC while still
|
| 1126 : |
|
|
having the old MLRISC linked into the current compiler, etc.)
|
| 1127 : |
|
|
See the CM manual.
|
| 1128 : |
|
|
|
| 1129 : |
|
|
5. Simple "makedepend" functionality added for generating Makefile
|
| 1130 : |
|
|
dependency information. (This is rather crude at the moment.
|
| 1131 : |
|
|
Expect some changes here in the future.)
|
| 1132 : |
|
|
|
| 1133 : |
|
|
6. ".fun" added as a recognized suffix for ML files. Also documented
|
| 1134 : |
|
|
explicitly in the manual that the fallback behavior (unknown suffix
|
| 1135 : |
|
|
-> ML file) is not an official feature!
|
| 1136 : |
|
|
|
| 1137 : |
|
|
7. Small changes to the pickler for stable libraries.
|
| 1138 : |
|
|
|
| 1139 : |
|
|
8. Several internal changes to CM (for cleanup/improvement).
|
| 1140 : |
|
|
|
| 1141 : |
|
|
|
| 1142 : |
|
|
!!!! NEW BINFILES !!!!
|
| 1143 : |
|
|
|
| 1144 : |
|
|
----------------------------------------------------------------------
|
| 1145 : |
|
|
Name: Matthias Blume
|
| 1146 : |
blume |
631 |
Date: 2000/04/28 17:30:00 JST
|
| 1147 : |
|
|
Tag: blume-20000428-pathconfig
|
| 1148 : |
|
|
Description:
|
| 1149 : |
|
|
|
| 1150 : |
|
|
1. I changed config/install.sh to remove duplicate entries from the
|
| 1151 : |
|
|
lib/pathconfig file at the end. Moreover, the final version of
|
| 1152 : |
|
|
lib/pathconfig is sorted alphabetically. The same (sorting) is done
|
| 1153 : |
|
|
in src/system/installml.
|
| 1154 : |
|
|
|
| 1155 : |
|
|
2. The config/install.sh script now consistently uses relative
|
| 1156 : |
|
|
pathnames in lib/pathconfig whenever the anchor is in the lib
|
| 1157 : |
|
|
directory. (So far this was true for the libraries that come
|
| 1158 : |
|
|
pre-compiled and bundled as part of the bootfiles but not for
|
| 1159 : |
|
|
libraries that are compiled by the script itself.)
|
| 1160 : |
|
|
|
| 1161 : |
|
|
----------------------------------------------------------------------
|
| 1162 : |
|
|
Name: Matthias Blume
|
| 1163 : |
blume |
629 |
Date: 2000/04/26 13:10:00 JST
|
| 1164 : |
|
|
Tag: blume-20000426-fun_suffix
|
| 1165 : |
|
|
Description:
|
| 1166 : |
|
|
|
| 1167 : |
|
|
Added ".fun" as a recognized file name suffix (for ML code).
|
| 1168 : |
|
|
|
| 1169 : |
|
|
----------------------------------------------------------------------
|
| 1170 : |
leunga |
628 |
Name: Allen Leung
|
| 1171 : |
|
|
Date: 2000/04/25 17:00:00 EST
|
| 1172 : |
|
|
Tag: leunga-20000425-alpha-ra
|
| 1173 : |
|
|
Description:
|
| 1174 : |
|
|
|
| 1175 : |
|
|
1. Alpha
|
| 1176 : |
|
|
|
| 1177 : |
|
|
PSEUDOARITH was missing in AlphaRewrite. This causes an endless loop
|
| 1178 : |
|
|
in C--.
|
| 1179 : |
|
|
|
| 1180 : |
|
|
2. RA
|
| 1181 : |
|
|
|
| 1182 : |
|
|
Added a flag "ra-dump-size" to print out the size of the flowgraph
|
| 1183 : |
|
|
and the interference graph.
|
| 1184 : |
|
|
|
| 1185 : |
|
|
----------------------------------------------------------------------
|
| 1186 : |
dbm |
627 |
Name: Dave MacQueen
|
| 1187 : |
|
|
Date: 2000/04/25/
|
| 1188 : |
|
|
Tag: dbm-20000425-mlyacc_doc_examples
|
| 1189 : |
|
|
Description:
|
| 1190 : |
|
|
Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
|
| 1191 : |
|
|
Updated all three examples in src/ml-yacc/examples to run
|
| 1192 : |
|
|
under 110.* using CM.make.
|
| 1193 : |
|
|
|
| 1194 : |
|
|
----------------------------------------------------------------------
|
| 1195 : |
leunga |
624 |
Name: Allen Leung
|
| 1196 : |
|
|
Date: 2000/04/20 23:04:00 EST
|
| 1197 : |
|
|
Tag: leunga-20000420-ssa-c---stuff
|
| 1198 : |
|
|
Description:
|
| 1199 : |
|
|
|
| 1200 : |
|
|
This update synchronizes my repository with Yale's. Most of these
|
| 1201 : |
|
|
changes, however, do not affect SML/NJ at all (the RA is an exception).
|
| 1202 : |
|
|
|
| 1203 : |
|
|
1. Register Allocator
|
| 1204 : |
|
|
|
| 1205 : |
|
|
a. An improvement in the interference graph construction:
|
| 1206 : |
|
|
Given a copy
|
| 1207 : |
|
|
|
| 1208 : |
|
|
s <- t
|
| 1209 : |
|
|
|
| 1210 : |
leunga |
628 |
no interference edge between s and t is added for this definition of s.
|
| 1211 : |
leunga |
624 |
|
| 1212 : |
|
|
b. I've added two new spill heuristic modules that Fermin and I developed
|
| 1213 : |
|
|
(in the new library RA.cm). These are unused in SML/NJ but maybe
|
| 1214 : |
|
|
useful for others (Moby?)
|
| 1215 : |
|
|
|
| 1216 : |
|
|
2. X86
|
| 1217 : |
|
|
|
| 1218 : |
|
|
a. Various fixes in the backend provided by Fermin [C--] and Lal.
|
| 1219 : |
|
|
|
| 1220 : |
|
|
3. Alpha
|
| 1221 : |
|
|
|
| 1222 : |
|
|
a. Added the BSR instruction and code generation that goes with it [C--]
|
| 1223 : |
|
|
b. Other fixes too numerous to recount provided by Fermin [C--]
|
| 1224 : |
|
|
|
| 1225 : |
|
|
4. Regmaps
|
| 1226 : |
|
|
|
| 1227 : |
|
|
a. The regmaps are not initialized with the identity physical bindings
|
| 1228 : |
|
|
at creation time. This is unneeded.
|
| 1229 : |
|
|
|
| 1230 : |
|
|
5. MLRISC Optimizations
|
| 1231 : |
|
|
|
| 1232 : |
|
|
a. The DJ-Graph module can now compute the iterated dominance frontiers
|
| 1233 : |
|
|
intersects with liveness incrementally in linear time! Woohoo!
|
| 1234 : |
|
|
This is now used in my new SSA construction algorithm.
|
| 1235 : |
|
|
|
| 1236 : |
|
|
b. THe branch reorganization module is now smarter about linear chains of
|
| 1237 : |
|
|
basic blocks.
|
| 1238 : |
|
|
|
| 1239 : |
|
|
|
| 1240 : |
|
|
----------------------------------------------------------------------
|
| 1241 : |
blume |
615 |
Name: Matthias Blume
|
| 1242 : |
|
|
Date: 2000/04/12 13:52:00 JST
|
| 1243 : |
|
|
Tag: blume_main_v110p27_1
|
| 1244 : |
|
|
Description:
|
| 1245 : |
|
|
|
| 1246 : |
|
|
Changed install.sh script to handle archive files without version number
|
| 1247 : |
|
|
and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
|
| 1248 : |
|
|
name of the boot file archive.
|
| 1249 : |
|
|
|
| 1250 : |
|
|
----------------------------------------------------------------------
|
| 1251 : |
dbm |
607 |
Name: Dave MacQueen
|
| 1252 : |
|
|
Date: 2000/04/09 14:00 EDT
|
| 1253 : |
|
|
Tag: dbm-20000410-Version_110_27
|
| 1254 : |
|
|
Description:
|
| 1255 : |
|
|
|
| 1256 : |
|
|
1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
|
| 1257 : |
|
|
|
| 1258 : |
|
|
2. Updated src/config/version to 110.27
|
| 1259 : |
|
|
|
| 1260 : |
|
|
3. New boot files!
|
| 1261 : |
|
|
|
| 1262 : |
|
|
----------------------------------------------------------------------
|
| 1263 : |
leunga |
606 |
Name: Allen Leung
|
| 1264 : |
|
|
Date: 2000/04/09 19:09:00 EST
|
| 1265 : |
|
|
Tag: leunga-20000409-misc
|
| 1266 : |
|
|
Description:
|
| 1267 : |
|
|
|
| 1268 : |
|
|
1. Yet another fix for x86 assembly for idivl, imull, mull and friends.
|
| 1269 : |
|
|
|
| 1270 : |
|
|
2. Miscellaneous improvements to MLRISC (unused in sml/nj)
|
| 1271 : |
|
|
|
| 1272 : |
|
|
----------------------------------------------------------------------
|
| 1273 : |
monnier |
605 |
Name: Stefan
|
| 1274 : |
|
|
Date: 2000/04/07 10:00:00 EDT
|
| 1275 : |
|
|
Tag: monnier-20000406-branch-handling
|
| 1276 : |
|
|
Description:
|
| 1277 : |
|
|
|
| 1278 : |
|
|
Improved handling of branches (mostly those generated from
|
| 1279 : |
|
|
polymorphic equality), removed switchoff and changed the
|
| 1280 : |
|
|
default optimization settings (more cpsopt and less flintopt).
|
| 1281 : |
|
|
|
| 1282 : |
|
|
----------------------------------------------------------------------
|
| 1283 : |
leunga |
601 |
Name: Allen Leung
|
| 1284 : |
leunga |
602 |
Date: 2000/04/06 01:30:00 EST
|
| 1285 : |
|
|
Tag: leunga-20000406-peephole-x86-SSA-2
|
| 1286 : |
|
|
Description:
|
| 1287 : |
|
|
|
| 1288 : |
|
|
Forgot a few files.
|
| 1289 : |
|
|
|
| 1290 : |
|
|
----------------------------------------------------------------------
|
| 1291 : |
|
|
Name: Allen Leung
|
| 1292 : |
leunga |
601 |
Date: 2000/04/06 00:36:00 EST
|
| 1293 : |
|
|
Tag: leunga-20000406-peephole-x86-SSA
|
| 1294 : |
|
|
Description:
|
| 1295 : |
|
|
|
| 1296 : |
|
|
1. New Peephole code
|
| 1297 : |
|
|
|
| 1298 : |
|
|
2. Minor improvement to X86 instruction selection
|
| 1299 : |
|
|
|
| 1300 : |
|
|
3. Various fixes to SSA and machine description -> code translator
|
| 1301 : |
|
|
|
| 1302 : |
|
|
----------------------------------------------------------------------
|
| 1303 : |
blume |
595 |
Name: Matthias Blume
|
| 1304 : |
|
|
Date: 2000/04/05 12:30:00 JST
|
| 1305 : |
|
|
Tag: blume_main_v110p26p2_3
|
| 1306 : |
|
|
Description:
|
| 1307 : |
|
|
|
| 1308 : |
|
|
This update just merges three minor cosmetic updates to CM's sources
|
| 1309 : |
|
|
to get ready for the 110.27 code freeze on Friday. No functionality
|
| 1310 : |
|
|
has changed.
|
| 1311 : |
|
|
|
| 1312 : |
|
|
----------------------------------------------------------------------
|
| 1313 : |
leunga |
593 |
Name: Allen Leung
|
| 1314 : |
leunga |
594 |
Date: 2000/04/04 19:39:00 EST
|
| 1315 : |
|
|
Tag: leunga-20000404-x86-asm
|
| 1316 : |
|
|
Description:
|
| 1317 : |
|
|
|
| 1318 : |
|
|
1. Fixed a problem in X86 assembly.
|
| 1319 : |
|
|
|
| 1320 : |
|
|
Things like
|
| 1321 : |
|
|
|
| 1322 : |
|
|
jmp %eax
|
| 1323 : |
|
|
jmp (%eax)
|
| 1324 : |
|
|
|
| 1325 : |
|
|
should be output as
|
| 1326 : |
|
|
|
| 1327 : |
|
|
jmp *%eax
|
| 1328 : |
|
|
jmp *(%eax)
|
| 1329 : |
|
|
|
| 1330 : |
|
|
2. Assembly output
|
| 1331 : |
|
|
|
| 1332 : |
|
|
Added a new flag
|
| 1333 : |
|
|
|
| 1334 : |
|
|
"asm-indent-copies" (default to false)
|
| 1335 : |
|
|
|
| 1336 : |
|
|
When this flag is on, parallel copies will be indented an extra level.
|
| 1337 : |
|
|
|
| 1338 : |
|
|
----------------------------------------------------------------------
|
| 1339 : |
|
|
Name: Allen Leung
|
| 1340 : |
leunga |
593 |
Date: 2000/04/04 03:18:00 EST
|
| 1341 : |
|
|
Tag: leunga-20000404-C--Moby
|
| 1342 : |
|
|
Description:
|
| 1343 : |
|
|
|
| 1344 : |
|
|
All of these fixes are related to C--, Moby, and my own optimization
|
| 1345 : |
|
|
stuff; so they shouldn't affect SML/NJ.
|
| 1346 : |
|
|
|
| 1347 : |
|
|
1. X86
|
| 1348 : |
|
|
|
| 1349 : |
|
|
Various fixes related floating point, and extensions.
|
| 1350 : |
|
|
|
| 1351 : |
|
|
2. Alpha
|
| 1352 : |
|
|
|
| 1353 : |
|
|
Some extra patterns related to loads with signed/zero extension
|
| 1354 : |
|
|
provided by Fermin.
|
| 1355 : |
|
|
|
| 1356 : |
|
|
3. Assembly
|
| 1357 : |
|
|
|
| 1358 : |
monnier |
659 |
When generating assembly, resolve the value of client defined constants,
|
| 1359 : |
leunga |
593 |
instead of generating symbolic values. This is controlled by the
|
| 1360 : |
|
|
new flag "asm-resolve-constants", which is default to true.
|
| 1361 : |
|
|
|
| 1362 : |
|
|
4. Machine Descriptions
|
| 1363 : |
|
|
|
| 1364 : |
|
|
a. The precedence parser was slightly broken when parsing infixr symbols.
|
| 1365 : |
|
|
b. The type generalizing code had the bound variables reversed, resulting
|
| 1366 : |
|
|
in a problem during arity raising.
|
| 1367 : |
|
|
c. Various fixes in machine descriptions.
|
| 1368 : |
|
|
|
| 1369 : |
|
|
----------------------------------------------------------------------
|
| 1370 : |
blume |
592 |
Name: Matthias Blume
|
| 1371 : |
|
|
Date: 2000/04/03 16:05:00 JST
|
| 1372 : |
|
|
Tag: blume_main_v110p26p2_2
|
| 1373 : |
|
|
Description:
|
| 1374 : |
|
|
|
| 1375 : |
|
|
I eliminated coreEnv from compInfo. Access to the "Core" structure is
|
| 1376 : |
|
|
now done via the ordinary static environment that is context to each
|
| 1377 : |
|
|
compilation unit.
|
| 1378 : |
|
|
|
| 1379 : |
|
|
To this end, I arranged that instead of "structure Core" as "structure
|
| 1380 : |
|
|
_Core" is bound in the pervasive environment. Core access is done via
|
| 1381 : |
monnier |
659 |
_Core (which can never be accidentally rebound because _Core is not a
|
| 1382 : |
blume |
592 |
legal surface-syntax symbol).
|
| 1383 : |
|
|
|
| 1384 : |
|
|
The current solution is much cleaner because the core environment is
|
| 1385 : |
|
|
now simply part of the pervasive environment which is part of every
|
| 1386 : |
|
|
compilation unit's context anyway. In particular, this eliminates all
|
| 1387 : |
|
|
special-case handling that was necessary until now in order to deal
|
| 1388 : |
|
|
with dynamic and symbolic parts of the core environment.
|
| 1389 : |
|
|
|
| 1390 : |
|
|
Remaining hackery (to bind the "magic" symbol _Core) is localized in the
|
| 1391 : |
monnier |
659 |
compilation manager's bootstrap compiler (actually: in the "init group"
|
| 1392 : |
blume |
592 |
handling). See the comments in src/system/smlnj/init/init.cmi for
|
| 1393 : |
|
|
more details.
|
| 1394 : |
|
|
|
| 1395 : |
|
|
I also tried to track down all mentions of "Core" (as string argument
|
| 1396 : |
|
|
to Symbol.strSymbol) in the compiler and replaced them with a
|
| 1397 : |
|
|
reference to the new CoreSym.coreSym. Seems cleaner since the actual
|
| 1398 : |
|
|
name appears in one place only.
|
| 1399 : |
|
|
|
| 1400 : |
|
|
Binfile and bootfile format have not changed, but the switchover from
|
| 1401 : |
|
|
the old "init.cmi" to the new one is a bit tricky, so I supplied new
|
| 1402 : |
|
|
bootfiles anyway.
|
| 1403 : |
|
|
|
| 1404 : |
|
|
----------------------------------------------------------------------
|
| 1405 : |
leunga |
591 |
Name: Allen Leung
|
| 1406 : |
|
|
Date: 2000/04/02 21:17:00 EST
|
| 1407 : |
|
|
Tag: leunga-20000402-mltree
|
| 1408 : |
|
|
Description:
|
| 1409 : |
leunga |
585 |
|
| 1410 : |
leunga |
591 |
1. Renamed the constructor CALL in MLTREE by popular demand.
|
| 1411 : |
|
|
2. Added a bunch of files from my repository. These are currently
|
| 1412 : |
|
|
used by other non-SMLNJ backends.
|
| 1413 : |
|
|
|
| 1414 : |
leunga |
576 |
----------------------------------------------------------------------
|
| 1415 : |
leunga |
590 |
Name: Allen Leung
|
| 1416 : |
|
|
Date: 2000/03/31 21:15:00 EST
|
| 1417 : |
|
|
Tag: leunga-20000331-aliasing
|
| 1418 : |
|
|
Description:
|
| 1419 : |
|
|
|
| 1420 : |
|
|
This update contains a rewritten (and hopefully more correct) module
|
| 1421 : |
|
|
for extracting aliasing information from CPS.
|
| 1422 : |
|
|
|
| 1423 : |
|
|
To turn on this feature:
|
| 1424 : |
|
|
|
| 1425 : |
|
|
Compiler.Control.CG.memDisambiguate := true
|
| 1426 : |
|
|
|
| 1427 : |
|
|
To pretty print the region information with assembly
|
| 1428 : |
|
|
|
| 1429 : |
|
|
Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
|
| 1430 : |
|
|
|
| 1431 : |
|
|
To control how many levels of aliasing information are printed, use:
|
| 1432 : |
|
|
|
| 1433 : |
|
|
Compiler.Control.MLRISC.getInt "points-to-show-level" := n
|
| 1434 : |
|
|
|
| 1435 : |
|
|
The default of n is 3.
|
| 1436 : |
|
|
|
| 1437 : |
|
|
----------------------------------------------------------------------
|
| 1438 : |
dbm |
589 |
Name: David MacQueen
|
| 1439 : |
|
|
Date: 2000/03/31 11:15:00 EST
|
| 1440 : |
|
|
Tag: dbm-20000331-runtime_fix
|
| 1441 : |
|
|
Description:
|
| 1442 : |
|
|
|
| 1443 : |
|
|
This update contains:
|
| 1444 : |
|
|
|
| 1445 : |
|
|
1. runtime/c-lib/c-libraries.c
|
| 1446 : |
|
|
includes added in revision 1.2 caused compilation errors on hppa-hpux
|
| 1447 : |
|
|
|
| 1448 : |
|
|
2. fix for bug 1556
|
| 1449 : |
|
|
system/Basis/Implementation/NJ/internal-signals.sml
|
| 1450 : |
|
|
|
| 1451 : |
|
|
----------------------------------------------------------------------
|
| 1452 : |
blume |
587 |
Name: Matthias Blume
|
| 1453 : |
blume |
588 |
Date: 2000/03/31 18:00:00 JST
|
| 1454 : |
|
|
Tag: blume_main_v110p26p2_1
|
| 1455 : |
|
|
Description:
|
| 1456 : |
|
|
|
| 1457 : |
|
|
This update contains:
|
| 1458 : |
|
|
|
| 1459 : |
|
|
1. A small change to CM's handling of stable libraries:
|
| 1460 : |
|
|
CM now maintains one "global" modmap that is used for all stable
|
| 1461 : |
|
|
libraries. The use of such a global modmap maximizes sharing and
|
| 1462 : |
|
|
minimizes the need for re-traversing parts of environments during
|
| 1463 : |
|
|
modmap construction. (However, this has minor impact since modmap
|
| 1464 : |
|
|
construction seems to account for just one percent or less of total
|
| 1465 : |
|
|
compile time.)
|
| 1466 : |
|
|
|
| 1467 : |
|
|
2. I added a "genmap" phase to the statistics. This is where I got the
|
| 1468 : |
|
|
"one percent" number (see above).
|
| 1469 : |
|
|
|
| 1470 : |
|
|
3. CM's new tool parameter mechanism just became _even_ better. :)
|
| 1471 : |
|
|
- The parser understands named parameters and recursive options.
|
| 1472 : |
|
|
- The "make" and "shell" tools use these new features.
|
| 1473 : |
|
|
(This makes it a lot easier to cascade these tools.)
|
| 1474 : |
|
|
- There is a small syntax change: named parameters use a
|
| 1475 : |
|
|
|
| 1476 : |
|
|
<name> : ( <option> ... ) or
|
| 1477 : |
|
|
<name> : <string>
|
| 1478 : |
|
|
|
| 1479 : |
|
|
syntax. Previously, named parameters were implemented in an
|
| 1480 : |
|
|
ad-hoc fashion by each tool individually (by parsing strings)
|
| 1481 : |
|
|
and had the form
|
| 1482 : |
|
|
|
| 1483 : |
|
|
<name>=<string>
|
| 1484 : |
|
|
|
| 1485 : |
|
|
See the CM manual for a full description of these issues.
|
| 1486 : |
|
|
|
| 1487 : |
|
|
----------------------------------------------------------------------
|
| 1488 : |
|
|
Name: Matthias Blume
|
| 1489 : |
blume |
587 |
Date: 2000/03/30 18:00:00 JST
|
| 1490 : |
|
|
Tag: blume_main_v110p26p2_0
|
| 1491 : |
|
|
Description:
|
| 1492 : |
|
|
|
| 1493 : |
|
|
!!!!! WARNING !!!!!!
|
| 1494 : |
|
|
!! New binfiles !!
|
| 1495 : |
|
|
!!!!!!!!!!!!!!!!!!!!
|
| 1496 : |
|
|
|
| 1497 : |
|
|
This update contains:
|
| 1498 : |
|
|
|
| 1499 : |
|
|
1. Moderate changes to CM:
|
| 1500 : |
|
|
|
| 1501 : |
|
|
- Changes to CM's tools mechanism. In particular, it is now possible
|
| 1502 : |
|
|
to have tools that accept additional "command line" parameters
|
| 1503 : |
|
|
(specified in the .cm file at each instance where the tool's class is
|
| 1504 : |
|
|
used).
|
| 1505 : |
|
|
|
| 1506 : |
monnier |
659 |
This was done to accommodate the new "make" and "shell" tools which
|
| 1507 : |
|
|
facilitate fairly seamless hookup to portions of code managed using
|
| 1508 : |
blume |
587 |
Makefiles or Shell scripts.
|
| 1509 : |
|
|
|
| 1510 : |
|
|
There are no classes "shared" or "private" anymore. Instead, the
|
| 1511 : |
|
|
sharing annotation is now a parameter to the "sml" class.
|
| 1512 : |
|
|
|
| 1513 : |
|
|
There is a bit of generic machinery for implementing one's own
|
| 1514 : |
|
|
tools that accept command-line parameters. However, I am not yet fully
|
| 1515 : |
|
|
satisfied with that part, so expect changes here in the future.
|
| 1516 : |
|
|
|
| 1517 : |
|
|
All existing tools are described in the CM manual.
|
| 1518 : |
|
|
|
| 1519 : |
monnier |
659 |
- Slightly better error handling. (CM now suppresses many followup
|
| 1520 : |
blume |
587 |
error messages that tended to be more annoying than helpful.)
|
| 1521 : |
|
|
|
| 1522 : |
|
|
2. Major changes to the compiler's static environment data structures.
|
| 1523 : |
|
|
|
| 1524 : |
|
|
- no CMStaticEnv anymore.
|
| 1525 : |
|
|
- no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
|
| 1526 : |
|
|
but it is called Environment), no conversions between different
|
| 1527 : |
|
|
kinds of static environments
|
| 1528 : |
|
|
|
| 1529 : |
|
|
- There is still a notion of a "modmap", but such modmaps are generated
|
| 1530 : |
|
|
on demand at the time when they are needed. This sounds slow, but I
|
| 1531 : |
|
|
sped up the code that generates modmaps enough for this not to lead to
|
| 1532 : |
|
|
a slowdown of the compiler (at least I didn't detect any).
|
| 1533 : |
|
|
|
| 1534 : |
|
|
- To facilitate rapid modmap generation, static environments now
|
| 1535 : |
|
|
contain an (optional) "modtree" structure. Modtree annotations are
|
| 1536 : |
|
|
constructed by the unpickler during unpickling. (This means that
|
| 1537 : |
|
|
the elaborator does not have to worry about modtrees at all.)
|
| 1538 : |
|
|
Modtrees have the advantage that they are compositional in the same
|
| 1539 : |
|
|
way as the environment data structure itself is compositional.
|
| 1540 : |
|
|
As a result, modtrees never hang on to parts of an environment that
|
| 1541 : |
|
|
has already been rendered "stale" by filtering or rebinding.
|
| 1542 : |
|
|
|
| 1543 : |
|
|
- I went through many, many trials and errors before arriving at the
|
| 1544 : |
|
|
current solution. (The initial idea of "linkpaths" did not work.)
|
| 1545 : |
|
|
But the result of all this is that I have touched a lot of files that
|
| 1546 : |
|
|
depend on the "modules" and "types" data structures (most of the
|
| 1547 : |
|
|
elaborator). There were a lot of changes during my "linkpath" trials
|
| 1548 : |
|
|
that could have been reverted to their original state but weren't.
|
| 1549 : |
|
|
Please, don't be too harsh on me for messing with this code a bit more
|
| 1550 : |
|
|
than what was strictly necessary... (I _did_ resist the tempation
|
| 1551 : |
|
|
of doing any "global reformatting" to avoid an untimely death at
|
| 1552 : |
|
|
Dave's hands. :)
|
| 1553 : |
|
|
|
| 1554 : |
|
|
- One positive aspect of the previous point: At least I made sure that
|
| 1555 : |
|
|
all files that I touched now compile without warnings (other than
|
| 1556 : |
|
|
"polyEqual").
|
| 1557 : |
|
|
|
| 1558 : |
|
|
- compiler now tends to run "leaner" (i.e., ties up less memory in
|
| 1559 : |
|
|
redundant modmaps)
|
| 1560 : |
|
|
|
| 1561 : |
|
|
----------------------------------------------------------------------
|
| 1562 : |
leunga |
580 |
Name: Allen Leung
|
| 1563 : |
leunga |
585 |
Date: 2000/03/29 18:00:00
|
| 1564 : |
|
|
Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
|
| 1565 : |
leunga |
586 |
Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
|
| 1566 : |
leunga |
585 |
Description:
|
| 1567 : |
|
|
|
| 1568 : |
|
|
This update contains *MAJOR* changes to the way code is generated from CPS
|
| 1569 : |
|
|
in the module mlriscGen, and in various backend modules.
|
| 1570 : |
|
|
|
| 1571 : |
|
|
CHANGES
|
| 1572 : |
|
|
=======
|
| 1573 : |
|
|
|
| 1574 : |
|
|
1. MLRiscGen: forward propagation fix.
|
| 1575 : |
|
|
|
| 1576 : |
|
|
There was a bug in forward propagation introduced at about the same time
|
| 1577 : |
|
|
as the MLRISC x86 backend, which prohibits coalescing to be
|
| 1578 : |
|
|
performed effectively in loops.
|
| 1579 : |
|
|
|
| 1580 : |
|
|
Effect: speed up of loops in RISC architectures.
|
| 1581 : |
|
|
By itself, this actually slowed down certain benchmarks on the x86.
|
| 1582 : |
|
|
|
| 1583 : |
|
|
2. MLRiscGen: forward propagating addresses from consing.
|
| 1584 : |
|
|
|
| 1585 : |
|
|
I've changed the way consing code is generated. Basically I separated
|
| 1586 : |
|
|
out the initialization part:
|
| 1587 : |
|
|
|
| 1588 : |
|
|
store tag, offset(allocptr)
|
| 1589 : |
|
|
store elem1, offset+4(allocptr)
|
| 1590 : |
|
|
store elem2, offset+8(allocptr)
|
| 1591 : |
|
|
...
|
| 1592 : |
|
|
store elemn, offset+4n(allocptr)
|
| 1593 : |
|
|
|
| 1594 : |
|
|
and the address computation part:
|
| 1595 : |
|
|
|
| 1596 : |
|
|
celladdr <- offset+4+alloctpr
|
| 1597 : |
|
|
|
| 1598 : |
|
|
and move the address computation part
|
| 1599 : |
|
|
|
| 1600 : |
|
|
Effect: register pressure is generally lower as a result. This
|
| 1601 : |
|
|
makes compilation of certain expressions much faster, such as
|
| 1602 : |
|
|
long lists with non-trivial elements.
|
| 1603 : |
|
|
|
| 1604 : |
|
|
[(0,0), (0,0), .... (0,0)]
|
| 1605 : |
|
|
|
| 1606 : |
|
|
3. MLRiscGen: base pointer elimination.
|
| 1607 : |
|
|
|
| 1608 : |
|
|
As part of the linkage mechanism, we generate the sequence:
|
| 1609 : |
|
|
|
| 1610 : |
|
|
L: ... <- start of the code fragment
|
| 1611 : |
|
|
|
| 1612 : |
|
|
L1:
|
| 1613 : |
|
|
base pointer <- linkreg - L1 + L
|
| 1614 : |
|
|
|
| 1615 : |
|
|
The base pointer was then used for computing relocatable addresses
|
| 1616 : |
|
|
in the code fragment. Frequently (such as in lots of continuations)
|
| 1617 : |
|
|
this is not needed. We now eliminate this sequence whenever possible.
|
| 1618 : |
|
|
|
| 1619 : |
|
|
For compile time efficiency, I'm using a very stupid local heuristic.
|
| 1620 : |
|
|
But in general, this should be done as a control flow analysis.
|
| 1621 : |
|
|
|
| 1622 : |
|
|
Effect: Smaller code size. Speed up of most programs.
|
| 1623 : |
|
|
|
| 1624 : |
|
|
4. Hppa back end
|
| 1625 : |
|
|
|
| 1626 : |
|
|
Long jumps in span dependence resolution used to depend on the existence
|
| 1627 : |
|
|
of the base pointer.
|
| 1628 : |
|
|
|
| 1629 : |
|
|
A jump to a long label L was expanded into the following sequence:
|
| 1630 : |
|
|
|
| 1631 : |
|
|
LDIL %hi(L-8192), %r29
|
| 1632 : |
|
|
LDO %lo(L-8192)(%r29), %r29
|
| 1633 : |
|
|
ADD %r29, baseptr, %r29
|
| 1634 : |
|
|
BV,n %r0(%r29)
|
| 1635 : |
|
|
|
| 1636 : |
|
|
In the presence of change (3) above, this will not work. I've changed
|
| 1637 : |
|
|
it so that the following sequence of instructions are generated, which
|
| 1638 : |
|
|
doesn't mention the base pointer at all:
|
| 1639 : |
|
|
|
| 1640 : |
|
|
BL,n L', %r29 /* branch and link, L' + 4 -> %r29 */
|
| 1641 : |
|
|
L': ADDIL L-(L'+4), %r29 /* Compute address of L */
|
| 1642 : |
|
|
BV,n %r0(%r29) /* Jump */
|
| 1643 : |
|
|
|
| 1644 : |
|
|
5. Alpha back end
|
| 1645 : |
|
|
|
| 1646 : |
|
|
New alpha instructions LDB/LDW have been added, as per Fermin's
|
| 1647 : |
|
|
suggestions. This is unrelated to all other changes.
|
| 1648 : |
|
|
|
| 1649 : |
|
|
6. X86 back end
|
| 1650 : |
|
|
|
| 1651 : |
|
|
I've changed andl to testl in the floating point test sequence
|
| 1652 : |
|
|
whenever appropriate. The Intel optimization guide states that
|
| 1653 : |
monnier |
659 |
testl is preferable to andl.
|
| 1654 : |
leunga |
585 |
|
| 1655 : |
|
|
7. RA (x86 only)
|
| 1656 : |
|
|
|
| 1657 : |
|
|
I've improved the spill propagation algorithm, using an approximation
|
| 1658 : |
|
|
of maximal weighted independent sets. This seems to be necessary to
|
| 1659 : |
|
|
alleviate the negative effect in light of the slow down in (1).
|
| 1660 : |
|
|
|
| 1661 : |
|
|
I'll write down the algorithm one of these days.
|
| 1662 : |
|
|
|
| 1663 : |
|
|
8. MLRiscGen: frequencies
|
| 1664 : |
|
|
|
| 1665 : |
|
|
I've added an annotation that states that all call gc blocks have zero
|
| 1666 : |
|
|
execution frequencies. This improves register allocation on the x86.
|
| 1667 : |
|
|
|
| 1668 : |
|
|
BENCHMARKS
|
| 1669 : |
|
|
==========
|
| 1670 : |
|
|
|
| 1671 : |
|
|
I've only perform the comparison on 110.25.
|
| 1672 : |
|
|
|
| 1673 : |
|
|
The platforms are:
|
| 1674 : |
|
|
|
| 1675 : |
|
|
HPPA A four processor HP machine (E9000) with 5G of memory.
|
| 1676 : |
|
|
X86 A 300Hhz Pentium II with 128M of memory, and
|
| 1677 : |
|
|
SPARC An Ultra sparc 2 with 512M of memory.
|
| 1678 : |
|
|
|
| 1679 : |
|
|
I used the following parameters for the SML benchmarks:
|
| 1680 : |
|
|
|
| 1681 : |
|
|
@SMLalloc
|
| 1682 : |
|
|
HPPA 256k
|
| 1683 : |
|
|
SPARC 512k
|
| 1684 : |
|
|
X86 256k
|
| 1685 : |
|
|
|
| 1686 : |
|
|
COMPILATION TIME
|
| 1687 : |
|
|
----------------
|
| 1688 : |
|
|
Here are the numbers comparing the compilation times of the compilers.
|
| 1689 : |
|
|
I've only compared 110.25 compiling the new sources versus
|
| 1690 : |
|
|
a fixpoint version of the new compiler compiling the same.
|
| 1691 : |
|
|
|
| 1692 : |
|
|
110.25 New
|
| 1693 : |
|
|
Total Time in RA Spill+Reload Total Time In RA Spill+Reload
|
| 1694 : |
|
|
HPPA 627s 116s 2684+3584 599s 95s 1003+1879
|
| 1695 : |
|
|
SPARC 892s 173s 2891+3870 708s 116s 1004+1880
|
| 1696 : |
|
|
X86 999s 315s 94006+130691 987s 296s 108877+141957
|
| 1697 : |
|
|
|
| 1698 : |
|
|
110.25 New
|
| 1699 : |
|
|
Code Size Code Size
|
| 1700 : |
|
|
HPPA 8596736 8561421
|
| 1701 : |
|
|
SPARC 8974299 8785143
|
| 1702 : |
|
|
X86 9029180 8716783
|
| 1703 : |
|
|
|
| 1704 : |
|
|
So in summary, things are at least as good as before. Dramatic
|
| 1705 : |
|
|
reduction in compilation is obtained on the Sparc; I can't explain it,
|
| 1706 : |
|
|
but it is reproducible. Perhaps someone should try to reproduce this
|
| 1707 : |
|
|
on their own machines.
|
| 1708 : |
|
|
|
| 1709 : |
|
|
SML BENCHMARKS
|
| 1710 : |
|
|
--------------
|
| 1711 : |
|
|
|
| 1712 : |
|
|
On the average, all benchmarks perform at least as well as before.
|
| 1713 : |
|
|
|
| 1714 : |
|
|
HPPA Compilation Time Spill+Reload Run Time
|
| 1715 : |
|
|
110.25 New 110.25 New 110.25 New
|
| 1716 : |
|
|
|
| 1717 : |
|
|
barnesHut 3.158 3.015 4.75% 1+1 0+0 2.980 2.922 2.00%
|
| 1718 : |
|
|
boyer 6.152 5.708 7.77% 0+0 0+0 0.218 0.213 2.34%
|
| 1719 : |
|
|
count-graphs 1.168 1.120 4.32% 0+0 0+0 22.705 23.073 -1.60%
|
| 1720 : |
|
|
fft 0.877 0.792 10.74% 1+3 1+3 0.602 0.587 2.56%
|
| 1721 : |
|
|
knuthBendix 3.180 2.857 11.32% 0+0 0+0 0.675 0.662 2.02%
|
| 1722 : |
|
|
lexgen 6.190 5.290 17.01% 0+0 0+0 0.913 0.788 15.86%
|
| 1723 : |
|
|
life 0.803 0.703 14.22% 25+25 0+0 0.153 0.140 9.52%
|
| 1724 : |
|
|
logic 2.048 2.007 2.08% 6+6 1+1 4.133 4.008 3.12%
|
| 1725 : |
|
|
mandelbrot 0.077 0.080 -4.17% 0+0 0+0 0.765 0.712 7.49%
|
| 1726 : |
|
|
mlyacc 22.932 20.937 9.53% 154+181 32+57 0.468 0.430 8.91%
|
| 1727 : |
|
|
nucleic 5.183 5.060 2.44% 2+2 0+0 0.125 0.120 4.17%
|
| 1728 : |
|
|
ratio-regions 3.357 3.142 6.84% 0+0 0+0 116.225 113.173 2.70%
|
| 1729 : |
|
|
ray 1.283 1.290 -0.52% 0+0 0+0 2.887 2.855 1.11%
|
| 1730 : |
|
|
simple 6.307 6.032 4.56% 28+30 5+7 3.705 3.658 1.28%
|
| 1731 : |
|
|
tsp 0.888 0.862 3.09% 0+0 0+0 7.040 6.893 2.13%
|
| 1732 : |
|
|
vliw 24.378 23.455 3.94% 106+127 25+45 2.758 2.707 1.91%
|
| 1733 : |
|
|
--------------------------------------------------------------------------
|
| 1734 : |
|
|
Average 6.12% 4.09%
|
| 1735 : |
|
|
|
| 1736 : |
|
|
SPARC Compilation Time Spill+Reload Run Time
|
| 1737 : |
|
|
110.25 New 110.25 New 110.25 New
|
| 1738 : |
|
|
|
| 1739 : |
|
|
barnesHut 3.778 3.592 5.20% 2+2 0+0 3.648 3.453 5.65%
|
| 1740 : |
|
|
boyer 6.632 6.110 8.54% 0+0 0+0 0.258 0.242 6.90%
|
| 1741 : |
|
|
count-graphs 1.435 1.325 8.30% 0+0 0+0 33.672 34.737 -3.07%
|
| 1742 : |
|
|
fft 0.980 0.940 4.26% 3+9 2+6 0.838 0.827 1.41%
|
| 1743 : |
|
|
knuthBendix 3.590 3.138 14.39% 0+0 0+0 0.962 0.967 -0.52%
|
| 1744 : |
|
|
lexgen 6.593 6.072 8.59% 1+1 0+0 1.077 1.078 -0.15%
|
| 1745 : |
|
|
life 0.972 0.868 11.90% 26+26 0+0 0.143 0.140 2.38%
|
| 1746 : |
|
|
logic 2.525 2.387 5.80% 7+7 1+1 5.625 5.158 9.05%
|
| 1747 : |
|
|
mandelbrot 0.090 0.093 -3.57% 0+0 0+0 0.855 0.728 17.39%
|
| 1748 : |
|
|
mlyacc 26.732 23.827 12.19% 162+189 32+57 0.550 0.560 -1.79%
|
| 1749 : |
|
|
nucleic 6.233 6.197 0.59% 3+3 0+0 0.163 0.173 -5.77%
|
| 1750 : |
|
|
ratio-regions 3.780 3.507 7.79% 0+0 0+0 133.993 131.035 2.26%
|
| 1751 : |
|
|
ray 1.595 1.550 2.90% 1+1 0+0 3.440 3.418 0.63%
|
| 1752 : |
|
|
simple 6.972 6.487 7.48% 29+32 5+7 3.523 3.525 -0.05%
|
| 1753 : |
|
|
tsp 1.115 1.063 4.86% 0+0 0+0 7.393 7.265 1.77%
|
| 1754 : |
|
|
vliw 27.765 24.818 11.87% 110+135 25+45 2.265 2.135 6.09%
|
| 1755 : |
|
|
----------------------------------------------------------------------------
|
| 1756 : |
|
|
Average 6.94% 2.64%
|
| 1757 : |
|
|
|
| 1758 : |
|
|
X86 Compilation Time Spill+Reload Run Time
|
| 1759 : |
|
|
110.25 New 110.25 New 110.25 New
|
| 1760 : |
|
|
|
| 1761 : |
|
|
barnesHut 5.530 5.420 2.03% 593+893 597+915 3.532 3.440 2.66%
|
| 1762 : |
|
|
boyer 8.768 7.747 13.19% 493+199 301+289 0.327 0.297 10.11%
|
| 1763 : |
|
|
count-graphs 2.040 2.010 1.49% 298+394 315+457 26.578 28.660 -7.26%
|
| 1764 : |
|
|
fft 1.327 1.302 1.92% 112+209 115+210 1.055 0.962 9.71%
|
| 1765 : |
|
|
knuthBendix 5.218 5.475 -4.69% 451+598 510+650 0.928 0.932 -0.36%
|
| 1766 : |
|
|
lexgen 9.970 9.623 3.60% 1014+841 1157+885 0.947 0.928 1.97%
|
| 1767 : |
|
|
life 1.183 1.183 0.00% 162+182 145+148 0.127 0.103 22.58%
|
| 1768 : |
|
|
logic 3.285 3.512 -6.45% 514+684 591+836 5.682 5.577 1.88%
|
| 1769 : |
|
|
mandelbrot 0.147 0.143 2.33% 38+41 33+54 0.703 0.690 1.93%
|
| 1770 : |
|
|
mlyacc 35.457 32.763 8.22% 3496+4564 3611+4860 0.552 0.550 0.30%
|
| 1771 : |
|
|
nucleic 7.100 6.888 3.07% 239+168 201+158 0.175 0.173 0.96%
|
| 1772 : |
|
|
ratio-regions 6.388 6.843 -6.65% 1182+257 981+300 120.142 120.345 -0.17%
|
| 1773 : |
|
|
ray 2.332 2.338 -0.29% 346+398 402+494 3.593 3.540 1.51%
|
| 1774 : |
|
|
simple 9.912 9.903 0.08% 1475+941 1579+1168 3.057 3.178 -3.83%
|
| 1775 : |
|
|
tsp 1.623 1.532 5.98% 266+200 250+211 8.045 7.878 2.12%
|
| 1776 : |
|
|
vliw 33.947 35.470 -4.29% 2629+2774 2877+3171 2.072 1.890 9.61%
|
| 1777 : |
|
|
----------------------------------------------------------------------------
|
| 1778 : |
|
|
Average 1.22% 3.36%
|
| 1779 : |
|
|
|
| 1780 : |
|
|
----------------------------------------------------------------------
|
| 1781 : |
|
|
Name: Allen Leung
|
| 1782 : |
leunga |
583 |
Date: 2000/03/23 16:25:00
|
| 1783 : |
|
|
Tag: leunga-20000323-fix_x86_alpha
|
| 1784 : |
|
|
Description:
|
| 1785 : |
|
|
|
| 1786 : |
|
|
1. X86 fixes/changes
|
| 1787 : |
|
|
|
| 1788 : |
|
|
a. The old code generated for SETcc was completely wrong.
|
| 1789 : |
|
|
The Intel optimization guide is VERY misleading.
|
| 1790 : |
|
|
|
| 1791 : |
|
|
2. ALPHA fixes/changes
|
| 1792 : |
|
|
|
| 1793 : |
|
|
a. Added the instructions LDBU, LDWU, STB, STW as per Fermin's suggestion.
|
| 1794 : |
|
|
b. Added a new mode byteWordLoadStores to the functor parameter to Alpha()
|
| 1795 : |
|
|
c. Added reassociation code for address computation.
|
| 1796 : |
|
|
|
| 1797 : |
|
|
----------------------------------------------------------------------
|
| 1798 : |
|
|
Name: Allen Leung
|
| 1799 : |
leunga |
580 |
Date: 2000/03/22 01:23:00
|
| 1800 : |
|
|
Tag: leunga-20000322-fix_x86_hppa_ra
|
| 1801 : |
|
|
Description:
|
| 1802 : |
|
|
|
| 1803 : |
|
|
1. X86 fixes/changes
|
| 1804 : |
|
|
|
| 1805 : |
|
|
a. x86Rewrite bug with MUL3 (found by Lal)
|
| 1806 : |
|
|
b. Added the instructions FSTS, FSTL
|
| 1807 : |
|
|
|
| 1808 : |
|
|
2. PA-RISC fixes/changes
|
| 1809 : |
|
|
|
| 1810 : |
|
|
a. B label should not be a delay slot candidate! Why did this work?
|
| 1811 : |
|
|
b. ADDT(32, REG(32, r), LI n) now generates one instruction instead of two,
|
| 1812 : |
|
|
as it should be.
|
| 1813 : |
|
|
c. The assembly syntax for fstds and fstdd was wrong.
|
| 1814 : |
|
|
d. Added the composite instruction COMICLR/LDO, which is the immediate
|
| 1815 : |
|
|
operand variant of COMCLR/LDO.
|
| 1816 : |
|
|
|
| 1817 : |
|
|
3. Generic MLRISC
|
| 1818 : |
|
|
|
| 1819 : |
|
|
a. shuffle.sml rewritten to be slightly more efficient
|
| 1820 : |
|
|
b. DIV bug in mltree-simplify fixed (found by Fermin)
|
| 1821 : |
|
|
|
| 1822 : |
|
|
4. Register Allocator
|
| 1823 : |
|
|
|
| 1824 : |
|
|
a. I now release the interference graph earlier during spilling.
|
| 1825 : |
|
|
May improve memory usage.
|
| 1826 : |
|
|
|
| 1827 : |
|
|
----------------------------------------------------------------------
|
| 1828 : |
blume |
577 |
Name: Matthias Blume
|
| 1829 : |
blume |
578 |
Date: 2000/03/14 14:15:32
|
| 1830 : |
|
|
Tag: blume_main_v110p26p1_2
|
| 1831 : |
|
|
Description:
|
| 1832 : |
|
|
|
| 1833 : |
|
|
1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
|
| 1834 : |
|
|
additional argument called "template" which is an optional string that
|
| 1835 : |
monnier |
659 |
specifies the layout of the tool command line. See the CM manual for
|
| 1836 : |
blume |
578 |
explanation.
|
| 1837 : |
|
|
|
| 1838 : |
monnier |
659 |
2. A special-purpose tool can be "registered" by simply dropping the
|
| 1839 : |
blume |
578 |
corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
|
| 1840 : |
|
|
directory where the .cm file lives that uses this tool. (The
|
| 1841 : |
|
|
behavior/misfeature until now was to look for the tool description
|
| 1842 : |
|
|
files in the current working directory.) As before, tool description
|
| 1843 : |
|
|
files could also be anchored -- in which case they can live anywhere
|
| 1844 : |
|
|
they like. Following the recent e-mail discussion, this change should
|
| 1845 : |
|
|
make it easier to have special-purpose tools that are shipped together
|
| 1846 : |
|
|
with the sources of the program that uses them.
|
| 1847 : |
|
|
|
| 1848 : |
|
|
----------------------------------------------------------------------
|
| 1849 : |
|
|
Name: Matthias Blume
|
| 1850 : |
blume |
577 |
Date: 2000/03/10 07:48:34
|
| 1851 : |
|
|
Tag: blume_main_v110p26p1_1
|
| 1852 : |
|
|
Description:
|
| 1853 : |
|
|
|
| 1854 : |
|
|
I added a re-written version of Dave's fixpt script to src/system.
|
| 1855 : |
|
|
Changes relative to the original version:
|
| 1856 : |
|
|
- sh-ified (not everybody has ksh)
|
| 1857 : |
|
|
- automatically figures out which architecture it runs on
|
| 1858 : |
|
|
- uses ./makeml a bit more cleverly
|
| 1859 : |
|
|
- never invokes ./installml (and, thus, does not clobber your
|
| 1860 : |
|
|
good and working installation of sml in case something goes wrong)
|
| 1861 : |
|
|
- accepts max iteration count using option "-iter <n>"
|
| 1862 : |
|
|
- accepts a "base" name using option "-base <base>"
|
| 1863 : |
|
|
|
| 1864 : |
|
|
It does not build any extraneous heap images but directly rebuilds
|
| 1865 : |
|
|
bin- and boot-hierarchies using makeml's "-rebuild" switch. Finally,
|
| 1866 : |
|
|
it can incorporate existing bin- and boot- hierarchies. For example,
|
| 1867 : |
|
|
suppose the base is set to "sml" (which is the default). Then it
|
| 1868 : |
|
|
successively builds
|
| 1869 : |
|
|
|
| 1870 : |
|
|
sml.bin.<arch>-unix and sml.boot.<arch>-unix
|
| 1871 : |
|
|
then sml1.bin.<arch>-unix and sml1.boot.<arch>-unix
|
| 1872 : |
|
|
then sml2.bin.<arch>-unix and sml2.boot.<arch>-unix
|
| 1873 : |
|
|
...
|
| 1874 : |
|
|
then sml<n>.bin.<arch>-unix and sml<n>.boot.<arch>-unix
|
| 1875 : |
|
|
|
| 1876 : |
|
|
and so on. If any of these already exist, it will just use what's
|
| 1877 : |
|
|
there. In particular, many people will have the initial set of bin
|
| 1878 : |
|
|
and boot files around, so this saves time for at least one full
|
| 1879 : |
|
|
rebuild. Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
|
| 1880 : |
|
|
<k>=1,2,... is normally not a good idea when invoking fixpt. However,
|
| 1881 : |
|
|
they might be the result of an earlier partial run of fixpt (which
|
| 1882 : |
monnier |
659 |
perhaps got accidentally killed). In this case, fixpt will quickly
|
| 1883 : |
blume |
577 |
move through what exists before continuing where it left off earlier,
|
| 1884 : |
|
|
and, thus, saves a lot of time.
|
| 1885 : |
|
|
|
| 1886 : |
|
|
----------------------------------------------------------------------
|
| 1887 : |
leunga |
576 |
Name: Allen Leung
|
| 1888 : |
|
|
Date: 00/03/10 02:20:00
|
| 1889 : |
|
|
Tag: leunga-20000310-fix_x86_asm_ra
|
| 1890 : |
|
|
Description:
|
| 1891 : |
dbm |
570 |
|
| 1892 : |
leunga |
576 |
More assembly output problems involving the indexed addressing mode
|
| 1893 : |
|
|
on the x86 have been found and corrected. Thanks to Fermin Reig for the
|
| 1894 : |
|
|
fix.
|
| 1895 : |
|
|
|
| 1896 : |
|
|
The interface and implementation of the register allocator have been changed
|
| 1897 : |
|
|
slightly to accommodate the possibility to skip the register allocation
|
| 1898 : |
|
|
phases completely and go directly to memory allocation. This is needed
|
| 1899 : |
|
|
for C-- use.
|
| 1900 : |
|
|
|
| 1901 : |
dbm |
570 |
----------------------------------------------------------------------
|
| 1902 : |
blume |
572 |
Name: Matthias Blume
|
| 1903 : |
blume |
575 |
Date: 00/03/09 10:23:53
|
| 1904 : |
|
|
Tag: blume_main_v110p26p1_0
|
| 1905 : |
|
|
Description:
|
| 1906 : |
|
|
|
| 1907 : |
|
|
* Complete re-organization of library names. Many libraries have been
|
| 1908 : |
|
|
consolidated so that they share the same path anchor. For example,
|
| 1909 : |
|
|
all MLRISC-related libraries are anchored at MLRISC, most libraries that
|
| 1910 : |
|
|
are SML/NJ-specific are under "smlnj". Notice that names like
|
| 1911 : |
|
|
host-cmb.cm or host-compiler.cm no longer exist. See system/README
|
| 1912 : |
|
|
for a complete description of the new naming scheme. Quick reference:
|
| 1913 : |
|
|
|
| 1914 : |
|
|
host-cmb.cm -> smlnj/cmb.cm
|
| 1915 : |
|
|
host-compiler.cm -> smlnj/compiler.cm
|
| 1916 : |
|
|
full-cm.cm -> smlnj/cm.cm
|
| 1917 : |
|
|
<arch>-<os>.cm -> smlnj/cmb/<arch>-<os>.cm
|
| 1918 : |
|
|
<arch>-compiler.cm -> smlnj/compiler/<arch>.cm
|
| 1919 : |
|
|
|
| 1920 : |
|
|
* Bug fixes in CM.
|
| 1921 : |
|
|
- exceptions in user code are being passed through (i.e., reach top level)
|
| 1922 : |
|
|
- more bugs in paranoia mode fixed
|
| 1923 : |
|
|
- bug related to checking group owners fixed
|
| 1924 : |
|
|
|
| 1925 : |
|
|
* New install.sh script that automagically fetches archive files:
|
| 1926 : |
|
|
The new file config/srcarchiveurl must contain the URL of the
|
| 1927 : |
|
|
(remote) directory that contains bin files (or other source archives).
|
| 1928 : |
|
|
If install.sh does not find the archive locally, it tries to get
|
| 1929 : |
|
|
it from that remote directory.
|
| 1930 : |
|
|
This should simplify installation further: For machines that have
|
| 1931 : |
|
|
access to the internet, just fetch <version>-config.tgz, unpack it,
|
| 1932 : |
monnier |
659 |
edit config/targets, and go (run config/install.sh). The script will
|
| 1933 : |
blume |
575 |
fetch everything else that it might need all by itself.
|
| 1934 : |
|
|
|
| 1935 : |
|
|
For CVS users, this mechanism is not relevant for source archives, but
|
| 1936 : |
|
|
it is convenient for getting new sets of binfiles.
|
| 1937 : |
|
|
|
| 1938 : |
|
|
Archives should be tar files compressed with either gzip, compress, or
|
| 1939 : |
|
|
bzip2. The script recognizes .tgz, .tar, tar.gz, tz, .tar.Z, and .tar.bz2.
|
| 1940 : |
|
|
|
| 1941 : |
|
|
----------------------------------------------------------------------
|
| 1942 : |
|
|
Name: Matthias Blume
|
| 1943 : |
blume |
572 |
Date: 2000/03/07 04:01:04
|
| 1944 : |
|
|
Tag: blume_main_v110_26_2
|
| 1945 : |
dbm |
570 |
Description:
|
| 1946 : |
blume |
572 |
- size info in BOOTLIST
|
| 1947 : |
|
|
* no fixed upper limits for number of bootfiles or length of
|
| 1948 : |
|
|
bootfile names in runtime
|
| 1949 : |
|
|
* falling back to old behavior if no BOOTLIST size info found
|
| 1950 : |
|
|
- allocation size heuristics in .run-sml
|
| 1951 : |
|
|
* tries to read cache size from /proc/cpuinfo (this is important for
|
| 1952 : |
|
|
small-cache Celeron systems!)
|
| 1953 : |
|
|
- install.sh robustified
|
| 1954 : |
|
|
- CM manual updates
|
| 1955 : |
|
|
- paranoid mode
|
| 1956 : |
|
|
* no more CMB.deliver() (i.e., all done by CMB.make())
|
| 1957 : |
|
|
* can re-use existing sml.boot.* files
|
| 1958 : |
|
|
* init.cmi now treated as library
|
| 1959 : |
|
|
* library stamps for consistency checks
|
| 1960 : |
|
|
- sml.boot.<arch>-<os>/PIDMAP file
|
| 1961 : |
|
|
* This file is read by the CM startup code. This is used to minimize
|
| 1962 : |
|
|
the amount of dynamic state that needs to be stowed away for the
|
| 1963 : |
|
|
purpose of sharing between interactive system and user code.
|
| 1964 : |
|
|
- CM.Anchor.anchor instead of CM.Anchor.{set,cancel}
|
| 1965 : |
|
|
* Upon request by Elsa. Anchors now controlled by get-set-pair
|
| 1966 : |
|
|
like most other CM state variables.
|
| 1967 : |
|
|
- Compiler.CMSA eliminated
|
| 1968 : |
|
|
* No longer supported by CM anyway.
|
| 1969 : |
|
|
- fixed bugs in pickler that kept biting Stefan
|
| 1970 : |
|
|
* past refs to past refs (was caused by the possibility that
|
| 1971 : |
|
|
ad-hoc sharing is more discriminating than hash-cons sharing)
|
| 1972 : |
|
|
* integer overflow on LargeInt.minInt
|
| 1973 : |
|
|
- ml-{lex,yacc} build scripts now use new mechanism
|
| 1974 : |
|
|
for building standalone programs
|
| 1975 : |
|
|
- fixed several gcc -Wall warnings that were caused by missing header
|
| 1976 : |
|
|
files, missing initializations, etc., in runtime (not all warnings
|
| 1977 : |
|
|
eliminated, though)
|
Parent Directory
| 
Revision Log
