Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 1156, Thu Mar 21 22:01:11 2002 UTC	revision 1411, Tue Nov 4 17:50:53 2003 UTC
#	Line 13	Line 13
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: Matthias Blume (blume (at) tti - c (dot) org)
17		Date: 2003/11/04 11:50:00 CST
18		Tag: blume-20031104-move-libraries
19		Description:
20
21		Eliminated the "dont_move_libraries" directive in config/targets.
22		(The mechanism was broken and could not be fixed easily.  Moreover,
23		there does not seem to be any reason not to move all libraries into
24		lib during installation.  I originally implemented this directive as a
25		backward-compatibility feature when I first introduced the new CM.
26		Now that things have been stable for a long time and going back to the
27		old CM is not an option, there is no reason to keep it around.)
28
29		----------------------------------------------------------------------
30		Name: Matthias Blume (blume (at) tti - c (dot) org)
31		Date: 2003/11/03 16:00:00 CST
32		Tag: blume-20031103-installdir
33		Description:
34
35		Made installer honor INSTALLDIR variable again.  (Thanks to Chris
36		Richards for pointing out the problem and providing the solution.)
37
38		----------------------------------------------------------------------
39		Name: Matthias Blume (blume (at) tti - c (dot) org)
40		Date: 2003/10/01 17:05:00 CDT
41		Tag: blume-20031001-lal-mlrisc
42		Description:
43
44		MLRISC bug fix from Lal.
45
46		----------------------------------------------------------------------
47		Name: Matthias Blume (blume (at) tti - c (dot) org)
48		Date: 2003/09/30 16:10:00 CDT
49		Tag: blume-20030930-primio-bat
50		Description:
51
52		1. Added openVector, nullRd, and nullWr to PRIM_IO.
53		2. Improved .bat files (for Win32 port) to make things work under Win95.
54		(thanks to Aaron S. Hawley for this one)
55
56		----------------------------------------------------------------------
57		Name: Matthias Blume (blume (at) tti - c (dot) org)
58		Date: 2003/09/26 16:05:00 CDT
59		Tag: blume-20030926-wrappriv
60		Description:
61
62		Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
63
64		----------------------------------------------------------------------
65		Name: Matthias Blume (blume (at) tti - c (dot) org)
66		Date: 2003/09/26 15:00:00 CDT
67		Tag: blume-20030926-110_43_3
68		Description:
69
70		- additional cleanup
71		- version number bump, NEW BOOTFILES
72
73		----------------------------------------------------------------------
74		Name: Matthias Blume (blume (at) tti - c (dot) org)
75		Date: 2003/09/26 12:00:00 CDT
76		Tag: blume-20030926-ppautoload
77		Description:
78
79		I modified the read-eval-print loop so that the autoloader gets
80		invoked whenever the prettyprinter tries to look up a symbol that
81		is not currently defined in the toplevel environment but which
82		appears in CM's autoload registry.  As a result, we see far fewer of
83		those ?.Foo.Bar.xxx names in the prettyprinter's output.
84
85		In addition to this I tried to clean up some pieces of the Basis
86		implementation (e.g., Socket, Word8Array) in order to prevent other
87		instances of these ?.Foo.Bar.xxx names from being printed.
88
89		The mechanism that picks names for types still needs some work, though.
90		(Right now it seems that if there is a type A.t which is defined to
91		be B.u, but B is unavailable at toplevel, then A.t gets printed as
92		"?.B.u" although the perhaps more sensible solution would be to use
93		"A.t" in this case.  In other words, the prettyprinter should follow
94		a chain of DEFtycs not farther than there are corresponding toplevel
95		names in the current environment.)
96
97		----------------------------------------------------------------------
98		Name: Matthias Blume (blume (at) tti - c (dot) org)
99		Date: 2003/09/24 16:31:00 CDT
100		Tag: blume-20030924-installer
101		Description:
102
103		Another installer tweak:  All the ML code for the installer is now
104		compiled during CMB.make and put into a little library called
105		$smlnj/installer.cm.  The installation then simply invokes
106
107		sml -m $smlnj/installer.cm
108
109		and everything happens automagically.
110
111		Win32:  ML code senses value of environment variable SMLNJ_HOME.
112		Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
113		and BINDIR.
114
115		The new scheme guarantees that the ML code responsible for the installation
116		is in sync with the APIs of the main system.  Also, the installer is
117		somewhat faster because the installer script is precompiled.
118
119		----------------------------------------------------------------------
120		Name: Matthias Blume (blume (at) tti - c (dot) org)
121		Date: 2003/09/24 15:35:00 CDT
122		Tag: blume-20030924-synsock
123		Description:
124
125		Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
126		but excludes all non-blocking operations.  Defined SOCKET (in Basis)
127		and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
128		implementations of non-blocking operations from CML's Socket
129		structure.
130
131		----------------------------------------------------------------------
132		Name: Matthias Blume (blume (at) tti - c (dot) org)
133		Date: 2003/09/24 15:10:05 CDT
134		Tag: blume-20030924-sockets
135		Description:
136
137		1. Fixed SOCKET API and implementation to match Basis spec.
138		This required changing the internal representation of sockets to one
139		that remembers (for each socket file descriptor) whether it is currently
140		blocking or non-blocking.  This state is maintained lazily (i.e., a system
141		call is made only if the state actually needs to change).
142
143		2. OS-specific details of sockets were moved into separate files, thus
144		making it possible to unify the bulk of the socket implementations
145		between Unix and Win32.
146
147		3. CML's socket API changed accordingly.
148		(Note that we need to remove non-blocking functions from this API
149		since they are redundant in the case of CML!)
150
151		4. CML's socket implementation now makes use of non-blocking functions
152		provided by Basis, thus removing all OS-dependent code from this part
153		of CML.
154
155		5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
156
157		----------------------------------------------------------------------
158		Name: Matthias Blume (blume (at) tti - c (dot) org)
159		Date: 2003/09/22 12:10:00 CDT
160		Tag: blume-20030922-110_43_2
161		Description:
162
163		Made a new interim version and bootfiles for developer's bootstrapping
164		convenience.
165
166		110.43.2 -- NEW BOOTFILES
167
168		----------------------------------------------------------------------
169		Name: Matthias Blume (blume (at) tti - c (dot) org)
170		Date: 2003/09/19 15:55:00 CDT
171		Tag: blume-20030919-cmdir
172		Description:
173
174		1. new-install.sh -> install.sh
175		2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
176		3. tweaked installer so that another name instead of .cm can be chosen
177		at install time (by setting the CM_DIR_ARC environment variable
178		during installation); once installation is complete, the name is
179		fixed
180
181		----------------------------------------------------------------------
182		Name: Matthias Blume (blume (at) tti - c (dot) org)
183		Date: 2003/09/18 16:00:00 CDT
184		Tag: blume-20030918-110_43_1
185		Description:
186
187		Made a new interim version and bootfiles for developer's bootstrapping
188		convenience.
189
190		110.43.1 -- NEW BOOTFILES
191
192		----------------------------------------------------------------------
193		Name: Matthias Blume (blume (at) tti - c (dot) org)
194		Date: 2003/09/18 15:20:00 CDT
195		Tag: blume-20030918-misc
196		Description:
197
198		1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
199		this seems to be controversial at the moment) and used those in
200		Posix.ProcEnv.times.
201
202		2. Added Time.{from,to}Nanoseconds to Time.
203
204		3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
205		For example, fromLargeInt never needs to look at more than 3 "big
206		digits" to get its 53 bits of precision.
207
208		----------------------------------------------------------------------
209		Name: Matthias Blume (blume (at) tti - c (dot) org)
210		Date: 2003/09/17 16:30:00 CDT
211		Tag: blume-20030917-real32-slices
212		Description:
213
214		Added an entry to the primitive environment
215		(compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
216		added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
217
218		Removed some of the "magic" constants in real64.sml and replaced them
219		with code that generates these values from their corresponding
220		integer counterparts.
221
222		Made all(?) the slice-related changes to the Basis and made everything
223		compile again...
224
225		----------------------------------------------------------------------
226		Name: Matthias Blume (blume (at) tti - c (dot) org)
227		Date: 2003/09/15 17:45:00 CDT
228		Tag: blume-20030915-rbase
229		Description:
230
231		Fixed bug in Real.fromLargeInt.
232
233		----------------------------------------------------------------------
234		Name: Matthias Blume (blume (at) tti - c (dot) org)
235		Date: 2003/09/13 18:11:00 CDT
236		Tag: blume-20030913-libinstall
237		Description:
238
239		Minor bugfix in config/libinstall (set anchor with path to
240		standalone tool after installing it, otherwise libraries that
241		need ml-lex or ml-yacc won't compile the first time the installer
242		runs).
243
244		----------------------------------------------------------------------
245		Name: Matthias Blume (blume (at) tti - c (dot) org)
246		Date: 2003/09/12 11:45:00 CDT
247		Tag: blume-20030912-various
248		Description:
249
250		- fixed bug in Real.toLargeInt
251		- fixed bug in Posix.ProcEnv.times
252		- changed inputLine functions to return an option
253		- minor installer improvements / bugfixes
254		- changed default @SMLalloc parameter for x86/celeron to 64k
255
256		----------------------------------------------------------------------
257		Name: Matthias Blume (blume (at) tti - c (dot) org)
258		Date: 2003/09/09 22:00:00 CDT
259		Tag: Release_110_43
260		Description:
261
262		New working release 110.43.  New bootfiles.
263
264		----------------------------------------------------------------------
265		Name: Matthias Blume (blume (at) tti - c (dot) org)
266		Date: 2003/09/09 19:20:00 CDT
267		Tag: blume-20030909-installer
268		Description:
269
270		Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
271		Modified config/install.bat to take advantage of it.  Also modified
272		config/install.sh (and called it config/new-install.sh) to take advantage
273		of it on Unix systems.  (The SML code is (supposed to be) platform-
274		independent.)
275
276		The installer can now install everything under Win32
277		as well as under *nix as long as it compiles.
278
279		Other changes:
280
281		- made CML compile again under Win32
282		- made eXene compile under Win32 (by providing a fake structure UnixSock
283		and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
284		- fixed a bug in nowhere: it assumed that type OS.Process.status is the
285		same as type int;  under Win32 it isn't
286		- fixed some slice-related problems in the win32-specific parts of CML
287		- added a functor argument "sameVol" to os-path-fn.sml in the Basis
288		(under Win32, the volume name is case-insensitive, and the
289		OS.Path code compares volume names for equality)
290
291		----------------------------------------------------------------------
292		Name: Matthias Blume (blume (at) tti - c (dot) org)
293		Date: 2003/09/08 11:55:00 CDT
294		Tag: blume-20030908-fullpath
295		Description:
296
297		Made Win32 version of OS.FileSys.fullPath return current directory
298		when given an empty string.  This is what the spec says, and incidentally,
299		CM depends on it.  (CM otherwise goes into an infinite loop in certain
300		cases when presented with the name of a non-existing .cm file.)
301
302		----------------------------------------------------------------------
303		Name: Matthias Blume (blume (at) tti - c (dot) org)
304		Date: 2003/09/04 16:30:00 CDT
305		Tag: blume-20030905-slices-etc
306		Description:
307
308		1. Changed interface to vectors and arrays in Basis to match
309		(draft) Basis spec.
310		2. Added signatures and implementations of slices according to
311		Basis spec.
312		3. Edited source code throughout the system to make it compile again
313		under 1. and 2.  (In some cases code had to be added to have it
314		match the new signatures.)
315		4. MLRISC should be backward-compatible: the copies of the originals
316		of files that needed to change under 3. were retained, the .cm files
317		check the compiler version number and use old versions when
318		appropriate.
319		5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
320		dirstream -> string option (in accordance with Basis spec).
321		6. When generating code that counts lines, ml-lex used function
322		CharVector.foldli, taking advantage of its old interface.
323		This has been replaced with the corresponding code from
324		CharVectorSlice.  (html-lex must be re-lexed!)
325		7. BitArray in smlnj-lib/Util has been extended/modified to match the
326		new MONO_ARRAY signature.  (Do we need BitArraySlice?)
327		8. Removed temporary additions (fromInternal, toInternal) from the
328		(now obsolete) IntInf in smlnj-lib/Util.
329		9. Cleaned up structure Byte.
330		10. Added localOffset, scan, and fromString to Date (according to spec).
331		Cleaned/corrected implementation of Date.
332		(Still need to check for correctness; implement better canonicalizeDate.)
333		11. Added "scan" to signature IEEE_REAL.
334		12. Some improvements to IntInf [in particular: efficiency-hack for
335		mod and rem when second operand is 2 (for parity checks).]
336		13. Changed representation of type Time.time, using a single IntInf.int
337		value counting microseconds.  This considerably simplified the
338		implementation of structure Time.  We now support negative time
339		values; scan and fromString handle signs.
340		14. Functor PrimIO now takes two additional arguments (VectorSlice and
341		ArraySlice).
342
343		----------------------------------------------------------------------
344		Name: Matthias Blume (blume (at) tti - c (dot) org)
345		Date: 2003/08/28 17:00:00 CDT
346		Tag: blume-20030828-intinf
347		Description:
348
349		This is a major update which comes with a version number bump
350		(110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
351		and an implementation of IntInf in the Basis.
352
353		There are a fairly large number of related changes and updates throughout
354		the system:
355
356		Basis:
357		- Implemented IntInf.
358		- Made LargeInt a projection of IntInf (by filtering through INTEGER).
359		- Added some missing Real64 operations, most notably Real.toLargeInt.
360		- Added FixedInt as a synonym for Int32.
361
362		compiler:
363		* Added support for a built-in intinf type.
364		- literals
365		- pattern matching
366		- conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
367		- overloading on literals and operations
368
369		This required adding a primitive type intinf, some additional
370		primops, and implementations for several non-trivial intinf
371		operations in Core.  (The intinf type is completely abstract
372		to the compiler; all operations get delegated back to the Core.)
373
374		* Intinf equality is handled by polyequal.  However, the compiler
375		does not print its usual warning in this case (since polyequal
376		is the right thing to do there).
377
378		* Improved the organization of structure InlineT.
379
380		* A word about conversion primops:
381		If conversions involving intinf do not cancel out during
382		CPS contract, then the compiler must insert calls to Core functions.
383		Since all core access must be resolved already during the FLINT
384		translate phase, it would be too late a the time of CPS contract
385		to add new Core calls.  For this reason, conversion primops
386		for intinf carry two arguments:  1. the numeric argument that
387		they are supposed to convert, and 2. the Core function that
388		can help with this conversion if necessary.  If CPS contract
389		eliminates a primop, then the associated Core function becomes
390		dead and goes away.  Intinf conversion primops that do not get
391		eliminated by CPS contract get rewritten into calls of their
392		core functions by a separate, new phase.
393
394		interactive system:
395		- Control.Print.intinfDepth controls max length of intinf constants
396		being printed. (Analogous to Control.Print.stringDepth.)
397		- Cleanup in printutil and pputil: got rid of unused stuff and
398		duplicates; replaced some of the code with code that makes better
399		use of library functionality.
400
401		CM:
402		Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
403		are no longer silent.
404
405		CKIT:
406		Fixed mismatched uses of Int32 and LargeInt.  I always decided
407		in favor of LargeInt -- which is now the same as IntInf.
408		CKIT-knowledgable people should check whether this is what's
409		intended and otherwise change things back to using Int32 or
410		FixedInt.
411
412		Throughout the code:
413		Started using IntInf.int literals and built-in operations
414		(e.g., comparison with 0) where this seems appropriate.
415
416
417		----------------------------------------------------------------------
418		Name: Dave MacQueen (dbm@cs.uchicago.edu)
419		Date: 2003/08/13 11:36:00 CDT
420		Tag: dbm-20030813-mcz-merge1
421		Description:
422
423		Merging changes from the mcz-branch development branch into trunk.
424		These changes involve replacement of the emulated old prettyprinter
425		interface with direct use of the SML/NJ Lib PP library, and fixing
426		of a couple of bugs (895, 1186) relating to error messages.  A new
427		prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
428		has been added.
429
430		----------------------------------------------------------------------
431		Name: Matthias Blume (blume (at) tti - c (dot) org)
432		Date: 2003/08/11 15:45:00 CDT
433		Tag: blume-20030811-windows
434		Description:
435
436		Version number bumped to 110.42.9.  NEW BOOTFILES!!!
437
438		http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
439
440		This patch restores SML/NJ's ability to run under win32.  There are a
441		number of changes, including fixes for several bugs that had gone
442		unnoticed until now:
443
444		- uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
445		trying to run the win32 version from within cygwin.)
446
447		- There are a number of simple .bat scripts that substitute for their
448		corresponding Unix shell-scripts.  (See below.)
449
450		- The internals of ml-build have been modified slightly.  The main
451		difference is that instead of calling ".link-sml" (or link-sml.bat)
452		using OS.Process.system, the ML process delegates this task back
453		to the script.  Otherwise problems arise in mixed environments such
454		as Cygwin where scripts look and work like Unix scripts, but
455		where OS.Process.system cannot run them.
456
457		- In CM, the srcpath pickler used native pathname syntax -- which
458		is incorrect in the case of cross-compilation.  The new pickle format
459		is independent of platform-specific naming conventions.
460
461		- Path configuration files (such as lib/pathconfig) can now choose
462		between native and standard syntax.  Placing a line of the form
463
464		standard!
465
466		into the file causes all subsequent paths to be interpreted using
467		CM standard pathname syntax (= Unix conventions); a line
468
469		native!
470
471		switches back to native style.  This was needed so that
472		path config files can be written portably, see src/system/pathconfig.
473
474		- Runtime system:
475
476		- win32-filesys.c:  get_file_time and set_file_time now
477		access modification time, not creation time.
478
479		- I/O code made aware of new array representation.
480
481		- Bug fixes in X86.prim.masm.
482
483		- src/system/makeml made aware of win32. (For use under cygwin
484		and other Unix-environments for windows.)
485
486		- In Basis, fixed off-by-one error in win32-io.sml (function vecF)
487		which caused BinIO.inputAll to fail consistently.
488
489		.bat scripts:
490
491		Windows .bat scripts assume that SMLNJ_HOME is defined.
492
493		- sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
494		applications (sml, ml-yacc, ml-lex).
495		- ml-build.bat: analogous to ml-build.
496		- config\install.bat:  Analogous to config/install.sh.  This requires
497		that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
498		(nmake etc. must be on the path, and vcvars32 must have been run.)
499		Moreover, sources for ml-lex and ml-yacc need to exist under src,
500		and the bootfile hierarchy must have been unpacked under
501		sml.boot.x86-win32.
502		The script is very primitive and does a poor job at error checking.
503		It only installs the base system, ml-lex, and ml-yacc.  No other
504		libraries are being installed (i.e., you get only those that
505		are part of the compiler.)
506		- link-sml.bat:  analogous to .link-sml, but not currently used
507
508		Unrelated bug fixes:
509
510		- ml-nlffigen now exports structures ST_* corresponding to incomplete
511		types.
512		- Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
513		otherwise.)
514
515		----------------------------------------------------------------------
516		Name: Dave MacQueen (macqueen@cs.uchicago.edu)
517		Date: 2003/06/17
518		Tag: macqueen-20030617-bug895
519		Description:
520
521		Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
522		Tag will be used for new development branch (mcz-branch) for
523		use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
524
525		----------------------------------------------------------------------
526		Name: Matthias Blume (blume (at) tti - c (dot) org)
527		Date: 2003/05/27 16:55:00 CDT
528		Tag: blume-20030527-polyeq
529		Description:
530
531		Tried to eliminated most cases of polymorphic equality.
532
533		----------------------------------------------------------------------
534		Name: Matthias Blume (blume (at) tti - c (dot) org)
535		Date: 2003/05/21 17:45:00 CDT
536		Tag: blume-20030517-complete
537		Description:
538
539		Two changes:
540
541		1. Added a flag for controlling whether non-exhaustive bindings will
542		be treated as errors (default is false).
543		2. Cleaned up the *entire* source tree so that CMB.make goes through
544		without a single non-exhaustive match- or bind warning.
545
546		----------------------------------------------------------------------
547		Name: Matthias Blume (blume (at) tti - c (dot) org)
548		Date: 2003/05/17 10:20:00 CDT
549		Tag: blume-20030517-absyn
550		Description:
551
552		1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
553
554		This mainly affects the quality of error messages.  However, some
555		of the code is now more straightforward than before. (Treatment of
556		the above four constructs in translate.sml is much simpler than
557		the "macro-expansion" that was going on before.  Plus, the mach-
558		compiler no longer gets invoked just to be able to compile an
559		if-expression.)
560
561		2. The ErrorMsg.Error exception is now caught and absorbed by the
562		interactive loop.
563
564		----------------------------------------------------------------------
565		Name: Allen Leung
566		Date: 2003/05/16 13:05:00 CDT
567		Tag: leunga-20030516-cygwin-runtime
568		Description:
569
570		Ported the runtime system to cygwin, which uses the unix
571		x86-unix bin files.  Missing/buggy features:
572
573		o getnetbyname, getnetbyaddr: these functions seem to be missing in
574		the Cygwin library.
575		o Ctrl-C handling may be flaky.
576		o Windows system calls and Windows I/O are not supported.
577
578		A new set of binfiles is located at:
579
580		http://www.dorsai.org/~leunga/boot.x86-unix.tgz
581
582		This is only needed for bootstrapping the cygwin version of smlnj.
583		Other x86 versions can use the existing binfiles.
584
585		----------------------------------------------------------------------
586		Name: Matthias Blume
587		Date: 2003/04/08 15:42:00 CDT
588		Tag: blume-20030408-listpair
589		Description:
590
591		1. Added a target 'mlrisc' to installer.
592
593		2. Added missing elements to structure ListPair.
594
595		----------------------------------------------------------------------
596		Name: Allen Leung
597		Date: 2003/01/07 10:40:00 EST
598		Tag: leunga-20030107-int-rem
599		Description:
600
601		Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
602		The arguments to the SUBL instruction were swapped.
603
604		----------------------------------------------------------------------
605		Name: Matthias Blume
606		Date: 2002/12/12 16:25:00 EST
607		Tag: blume-20021212-risc-ra
608		Description:
609
610		Fixed a serious bug in the rewrite code for FP spilling/reloading that
611		sent the RA into an infinite loop when floating point registers get
612		spilled.  (Because of this bug, e.g., nucleic stopped compiling between
613		110.37 and 110.38.)
614		There was another set of potential problems related to the handling of
615		MLRISC annotations (but those did not yet cause real problems, apparently).
616
617		----------------------------------------------------------------------
618		Name: Matthias Blume
619		Date: 2002/12/06 22:40:00 EST
620		Tag: blume-20021206-cm-fileid
621		Description:
622
623		Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
624		This fixes the problem of CM getting confused by files that suddenly
625		change their identity (e.g., by getting unlinked and recreated by some
626		text editor such as vi).  There might be a better/cheaper/cleaner way
627		of doing this, but for now this will have to do.
628
629		----------------------------------------------------------------------
630		Name: Matthias Blume
631		Date: 2002/10/28 09:50:00 EST
632		Tag: blume-20021028-typecheck
633		Description:
634
635		Exported structure Typecheck from $smlnj/viscomp/core.cm.
636
637		----------------------------------------------------------------------
638		Name: Matthias Blume
639		Date: 2002/10/17 09:10:00 EDT
640		Tag: Release_110_42
641		Description:
642
643		In good old tradition, there has been a slight hiccup so that we have
644		to patch 110.42 after the fact.   The old release tag has been replaced
645		(see below).
646
647		The change solves a problem with two competing approaches the
648		configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
649		each other's way.
650
651		This change only affects the runtime system code and the installer script.
652		(No new bootfiles.)
653
654		----------------------------------------------------------------------
655		Name: Matthias Blume
656		Date: 2002/10/16 12:00:00 EDT
657		Tag: Release_110_42_removed
658		Description:
659
660		New working release.  New bootfiles.
661
662		----------------------------------------------------------------------
663		Name: Matthias Blume
664		Date: 2002/10/10 13:10:00 EDT
665		Tag: blume-20021010-ppc-divs
666		Description:
667
668		The mltree operator DIVS must be implemented with an overflow check on
669		the PPC because the hardware indicates divide-by-zero using "overflow" as
670		well.
671
672		----------------------------------------------------------------------
673		Name: Matthias Blume
674		Date: 2002/07/23 11:45:00
675		Tag: blume-20020723-smlnj-home
676		Description:
677
678		Sml now senses the SMLNJ_HOME environment variable.  If this is set,
679		then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
680		CM_PATHCONFIG is also set), the path configuration file is assumed
681		to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
682		the entire tree to some other place and everything will "just work".
683
684		(Companion commands such as ml-build and ml-makedepend also sense this
685		variable.)
686
687		----------------------------------------------------------------------
688		Name: Matthias Blume
689		Date: 2002/07/12 21:19:00 EDT
690		Tag: blume-20020712-liveness
691		Description:
692
693		Exported two useful "step" functions from liveness module (MLRISC).
694
695		----------------------------------------------------------------------
696		Name: Matthias Blume
697		Date: 2002/07/05 16:00 EDT
698		Tag: Release_110_41
699		Description:
700
701		New working release.  New bootfiles.
702
703		----------------------------------------------------------------------
704		Name: Matthias Blume
705		Date: 2002/07/05 10:25:00 EDT
706		Tag: blume-20020705-btimp
707		Description:
708
709		Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
710		other clients can set up backtracing support.
711
712		----------------------------------------------------------------------
713		Name: Matthias Blume
714		Date: 2002/06/25 17:23:00 EDT
715		Tag: blume-20020625-fpmax
716		Description:
717
718		Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
719		case.  (The sense of the isNaN test was reversed -- which made min and
720		max always return their first argument.)
721
722		----------------------------------------------------------------------
723		Name: Matthias Blume
724		Date: 2002/06/11
725		Tag: blume-20020611-unixpath
726		Description:
727
728		Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
729
730		----------------------------------------------------------------------
731		Name: Matthias Blume
732		Date: 2002/06/10 16:35:00 EDT
733		Tag: blume-20020610-ieeereal
734		Description:
735
736		I back-ported my implementation of IEEEReal.fromString from the
737		idlbasis-devel branch so that we can test it.
738
739		Another small change is that ppDec tries to give more information
740		than just "<sig>" in the case of functors.  However, this code is
741		broken in some mysterious way if the functor's body's signature
742		has not been declared by ascription but gets inferred from the
743		implementation.  This needs fixing...
744
745		----------------------------------------------------------------------
746		Name: Matthias Blume
747		Date: 2002/05/31
748		Tag: blume-20020531-btrace-mode
749		Description:
750
751		Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
752		the wayside when I switched over to using Controls everywhere.)
753
754		----------------------------------------------------------------------
755		Name: Lal George
756		Date: 2002/05/23 12:21:40 EDT
757		Tag: george-20020523-visual-labels
758		Description:
759
760		Labels are now displayed in the graphical output to make
761		the fall-through and target blocks obvious.
762
763		----------------------------------------------------------------------
764		Name: Matthias Blume
765		Date: 2002/05/22 11:03:00 EDT
766		Tag: blume-20020522-shrink
767		Description:
768
769		John tweaked yesterday's fix for 1131 to handle an out-of-memory
770		situation that comes up when allocating huge arrays.
771
772		----------------------------------------------------------------------
773		Name: Matthias Blume
774		Date: 2002/05/21 16:00:00 EDT
775		Tag: Release_110_40
776		Description:
777
778		New working release (110.40).  New bootfiles.
779
780		[Also: John Reppy fixed GC bug 1131.]
781
782		----------------------------------------------------------------------
783		Name: Matthias Blume
784		Date: 2002/05/21 12:35:00 EDT
785		Tag: blume-20020521-cmdoc
786		Description:
787
788		CM documentation update.
789
790		----------------------------------------------------------------------
791		Name: Matthias Blume
792		Date: 2002/05/21 10:55:00 EDT
793		Tag: blume-20020521-misc
794		Description:
795
796		- John tweaked runtime to be silent on heap export (except when
797		GC messages are on).
798		- I added a few more things (cross-compiling versions of CMB) to
799		config/preloads (as suggestions).
800
801		----------------------------------------------------------------------
802		Name: Matthias Blume
803		Date: 2002/05/20 22:25:00 EDT
804		Tag: blume-20020520-controls
805		Description:
806
807		- Added ControlUtil structure to control-lib.cm.
808		- Use it throughout.
809		- Used Controls facility to define MLRISC controls (as opposed to
810		registering MLRISC control ref cells with Controls after the fact)
811		- Fixed messed-up controls priorities.
812
813		* Removed again all the stuff from config/preloads that one wouldn't
814		be able to preload at the time the initial heap image is built.
815		(Many libraries, e.g., CML, do not exist yet at this time.  The
816		only libraries that can be preloaded via config/preloads are those
817		that come bundled with the bootfiles.)
818
819		----------------------------------------------------------------------
820		Name: Matthias Blume
821		Date: 2002/05/20 10:59:00 EDT
822		Tag: blume-20020520-preloads
823		Description:
824
825		Added a lot of commented-out suggestions for things to be included
826		in config/preloads.
827
828		----------------------------------------------------------------------
829		Name: Allen Leung
830		Date: 2002/05/18 14:20:00 EDT
831		Tag: leunga-20020518-mdl
832		Description:
833
834		o Made the mdl tool stuff compile and run again.
835		o I've disabled all the stuff that depends on RTL specifications; they
836		are all badly broken anyway.
837
838		----------------------------------------------------------------------
839		Name: Matthias Blume
840		Date: 2002/05/17 16:49:00 EDT
841		Tag: blume-20020517-controls
842		Description:
843
844		0. John Reppy made several modifications to the SML/NJ library.
845		In particular, there is a shiny new controls-lib.cm.
846
847		1. Pushed new controls interface through compiler so that everything
848		compiles again.
849
850		2. Added FormatComb and FORMAT_COMB to the CML version of the
851		SML/NJ library (so that CML compiles again).
852
853		3. Modified init scripts because XXX_DEFAULT environment variables
854		are no longer with us.  (Boot-time initialization is now done
855		using the same environment variables that are also used for
856		startup-time initialization of controls.)
857
858		----------------------------------------------------------------------
859		Name: Lal George
860		Date: 2002/05/15 09:20:10 EDT
861		Tag: george-20020515-pseudo-op-decls
862		Description:
863
864		All pseudo-ops emitted before the first segment declaration
865		such as TEXT, DATA, and BSS directives are assumed to be global
866		declarations and are emitted first in the assembly file. This is
867		useful in a number of situations where one has pseudo-ops that are not
868		specific to any segment, and also works around the constraint that one
869		cannot have client pseudo-ops in the TEXT segment.
870
871		Because no segment is associated with these declarations it is
872		an error to allocate any space or objects before the first segment
873		directive and an exception will be raised. However, we cannot make
874		this check for client pseudo-ops.
875
876		These top level declarations are a field in the CFG graph_info.
877		In theory you can continue to add to this field after the CFG has been
878		built -- provided you know what you are doing;-)
879
880		----------------------------------------------------------------------
881		Name: Matthias Blume
882		Date: 2002/05/13 16:40:00 EDT
883		Tag: blume-20020513-pp-etc
884		Description:
885
886		A few minor bugfixes:
887
888		- Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
889		(Bogus printouts for redefined bindings still occur.  Compiler
890		bug should no longer occur now.  We need to redo the prettyprinter
891		from scratch.)
892
893		- CM pathname printer now also adds escape sequences for ( and )
894
895		- commend and docu fixes for ml-nlffi
896
897		----------------------------------------------------------------------
898		Name: Matthias Blume
899		Date: 2002/05/10 16:40:00 EDT
900		Tag: blume-20020510-erg-textio
901		Description:
902
903		Applied the following bugfix provided by Emden Gansner:
904
905		Output is corrupted when outputSubstr is used rather than output.
906
907		The problem occurs when a substring
908
909		ss = (s, dataStart, dataLen)
910
911		where dataStart > 0, fills a stream buffer with avail bytes left.
912		avail bytes of s, starting at index dataStart, are copied into the
913		buffer, the buffer is flushed, and then the remaining dataLen-avail
914		bytes of ss are copied into the beginning of the buffer. Instead of
915		starting this copy at index dataStart+avail in s, the current code
916		starts the copy at index avail.
917
918		Fix:
919		In text-io-fn.sml, change line 695 from
920		val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
921		to
922		val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
923
924		----------------------------------------------------------------------
925		Name: Matthias Blume
926		Date: 2002/04/12 13:55:00 EDT
927		Tag: blume-20020412-assyntax
928		Description:
929
930		1. Grabbed newer assyntax.h from the XFree86 project.
931		2. Fiddled with how to compile X86.prim.asm without warnings.
932		3. (Very) Minor cleanup in CM.
933
934		----------------------------------------------------------------------
935		Name: Matthias Blume
936		Date: 2002/04/01 (no joke!) 17:07:00 EST
937		Tag: blume-20020401-x86div
938		Description:
939
940		Added full support for div/mod/rem/quot on the x86, using the machine
941		instruction's two results (without clumsily recomputing the remainder)
942		directly where appropriate.
943
944		Some more extensive power-of-two support was added to the x86 instruction
945		selector (avoiding expensive divs, mods, and muls where they can be
946		replaced with cheaper shifts and masks).  However, this sort of thing
947		ought to be done earlier, e.g., within the CPS optimizer so that
948		all architectures benefit from it.
949
950		The compiler compiles to a fixed point, but changes might be somewhat
951		fragile nevertheless.  Please, report any strange things that you might
952		see wrt. div/mod/quot/rem...
953
954		----------------------------------------------------------------------
955		Name: Matthias Blume
956		Date: 2002/03/29 17:22:00
957		Tag: blume-20020329-div
958		Description:
959
960		Fixed my broken div/mod logic.  Unfortunately, this means that the
961		inline code for div/mod now has one more comparison than before.
962		Fast paths (quotient > 0 or remainder = 0) are not affected, though.
963		The problem was with quotient = 0, because that alone does not tell
964		us which way the rounding went.  One then has to look at whether
965		remainder and divisor have the same sign...  :(
966
967		Anyway, I replaced the bootfiles with fresh ones...
968
969		----------------------------------------------------------------------
970		Name: Matthias Blume
971		Date: 2002/03/29 14:10:00 EST
972		Tag: blume-20020329-inlprims
973		Description:
974
975		NEW BOOTFILES!!!    Version number bumped to 110.39.3.
976
977		Primops have changed. This means that the bin/boot-file formats have
978		changed as well.
979
980		To make sure that there is no confusion, I made a new version.
981
982
983		CHANGES:
984
985		* removed REMT from mltree (remainder should never overflow).
986
987		* added primops to deal with divisions of all flavors to the frontend
988
989		* handled these primops all the way through so they map to their respective
990		MLRISC support
991
992		* used these primops in the implementation of Int, Int32, Word, Word32
993
994		* removed INLDIV, INLMOD, and INLREM as they are no longer necessary
995
996		* parameterized INLMIN, INLMAX, and INLABS by a numkind
997
998		* translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
999		floating point
1000
1001		* used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
1002		and Real (but Real.abs maps to a separate floating-point-only primop)
1003
1004
1005		TODO items:
1006
1007		* Hacked Alpha32 instruction selection, disabling the selection of REMx
1008		instructions because the machine instruction encoder cannot handle
1009		them.  (Hppa, PPC, and Sparc instruction selection did not handle
1010		REM in the first place, and REM is supported by the x86 machine coder.)
1011
1012		* Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
1013		selection phase.  (The two can be streamlined because the hardware
1014		delivers both quotient and remainder at the same time anyway.)
1015
1016		* Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
1017		(Currently the behavior is inconsistent both across architectures and
1018		wrt. the draft Basis spec.)
1019
1020		* Word8 should eventually be handled natively, too.
1021
1022		* There seems to be one serious bug in mltree-gen.sml.  It appears, though,
1023		as if there currently is no execution path that could trigger it in
1024		SML/NJ.  (The assumptions underlying functions arith and promotable do not
1025		hold for things like multiplication and division.)
1026
1027		----------------------------------------------------------------------
1028		Name: Matthias Blume
1029		Date: 2002/03/27 16:27:00 EST
1030		Tag: blume-20020327-mlrisc-divisions
1031		Description:
1032
1033		Added support for all four division operations (ML's div, mod, quot,
1034		and rem) to MLRISC.  In the course of doing so, I also rationalized
1035		the naming (no more annoying switch-around of DIV and QUOT), by
1036		parameterizing the operation by div_rounding_mode (which can be either
1037		DIV_TO_ZERO or DIV_TO_NEGINF).
1038
1039		The generic MLTreeGen functor takes care of compiling all four
1040		operations down to only round-to-zero div.
1041
1042		Missing pieces:
1043
1044		* Doing something smarter than relying on MLTreeGen on architectures
1045		like, e.g., the x86 where hardware division delivers both quotient and
1046		remainder at the same time.  With this, the implementation of the
1047		round-to-neginf operations could be further streamlined.
1048
1049		* Remove inlining support for div/mod/rem from the frontend and replace it
1050		with primops that get carried through to the backend.  Do this for all
1051		int and word types.
1052
1053		----------------------------------------------------------------------
1054		Name: Matthias Blume
1055		Date: 2002/03/25 17:25:00 EST
1056		Tag: blume-20020325-divmod
1057		Description:
1058
1059		I improved (hopefully without breaking them) the implementation of Int.div,
1060		Int.mod, and Int.rem.   For this, the code in translate.sml now takes
1061		advantage of the following observations:
1062
1063		Let  q = x quot y      r = x rem y
1064		d = x div  y      m = x mod y
1065
1066		where "quot" is the round-to-zero version of integer division that
1067		hardware usually provides.  Then we have:
1068
1069		r = x - q * y        where neither the * nor the - will overflow
1070		d = if q >= 0 orelse x = q * y then q else q - 1
1071		where neither the * nor the - will overflow
1072		m = if q >= 0 orelse r = 0 then r else r + y
1073		where the + will not overflow
1074
1075		This results in substantial simplification of the generated code.
1076		The following table shows the number of CFG nodes and edges generated
1077		for
1078		fun f (x, y) = x OPER y
1079		(* with OPER \in div, mod, quot, rem *)
1080
1081
1082		OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
1083		--------------------------------------------------------
1084		div |         24 |         39 |         12 |         16
1085		mod |         41 |         71 |         12 |         16
1086		quot |          8 |         10 |          8 |         10
1087		rem |         10 |         14 |          8 |         10
1088
1089
1090		----------------------------------------------------------------------
1091		Name: Matthias Blume
1092		Date: 2002/03/25 22:06:00 EST
1093		Tag: blume-20020325-cprotobug
1094		Description:
1095
1096		Fixed a bug in cproto (c prototype decoder).
1097
1098		----------------------------------------------------------------------
1099		Name: Matthias Blume
1100		Date: 2002/03/25 16:00:00 EST
1101		Tag: blume-20020325-raw-primops
1102		Description:
1103
1104		I did some cleanup to Allen's new primop code and
1105		replaced yesterday's bootfiles with new ones.
1106		(But they are stored in the same place.)
1107
1108		----------------------------------------------------------------------
1109		Name: Matthias Blume
1110		Date: 2002/03/24 22:40:00 EST
1111		Tag: blume-20020324-bootfiles
1112		Description:
1113
1114		Made the bootfiles that Allen asked for.
1115
1116		----------------------------------------------------------------------
1117		Name: Allen Leung
1118		Date: 2002/03/23 15:50:00 EST
1119		Tag: leunga-20020323-flint-cps-rcc-primops
1120		Description:
1121
1122		1. Changes to FLINT primops:
1123
1124		(* make a call to a C-function;
1125		* The primop carries C function prototype information and specifies
1126		* which of its (ML-) arguments are floating point. C prototype
1127		* information is for use by the backend, ML information is for
1128		* use by the CPS converter. *)
1129		| RAW_CCALL of { c_proto: CTypes.c_proto,
1130		ml_args: ccall_type list,
1131		ml_res_opt: ccall_type option,
1132		reentrant : bool
1133		} option
1134		(* Allocate uninitialized storage on the heap.
1135		* The record is meant to hold short-lived C objects, i.e., they
1136		* are not ML pointers.  With the tag, the representation is
1137		* the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1138		*)
1139		| RAW_RECORD of {tag:bool,sz:int}
1140		and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1141
1142		2.  These CPS primops are now overloaded:
1143
1144		rawload of {kind:numkind}
1145		rawstore of {kind:numkind}
1146
1147		The one argument form is:
1148
1149		rawload {kind} address
1150
1151		The two argument form is:
1152
1153		rawload {kind} [ml object, byte-offset]
1154
1155		3. RAW_CCALL/RCC now takes two extra arguments:
1156
1157		a. The first is whether the C call is reentrant, i.e., whether
1158		ML state should be saved and restored.
1159		b. The second argument is a string argument specifying the name of
1160		library and the C function.
1161
1162		These things are currently not handled in the code generator, yet.
1163
1164		4. In CProto,
1165
1166		An encoding type of "bool" means "ml object" and is mapped into
1167		C prototype of PTR.  Note that "bool" is different than "string",
1168		even though "string" is also mapped into PTR, because "bool"
1169		is assigned an CPS type of BOGt, while "string" is assigned INT32t.
1170
1171		5. Pickler/unpicker
1172
1173		Changed to handle RAW_RECORD and newest RAW_CCALL
1174
1175		6. MLRiscGen,
1176
1177		1. Changed to handle the new rawload/rawstore/rawrecord operators.
1178		2. Code for handling C Calls has been moved to a new module CPSCCalls,
1179		in the file CodeGen/cpscompile/cps-c-calls.sml
1180
1181		7. Added the conditional move operator
1182
1183		condmove of branch
1184
1185		to cps.  Generation of this is still buggy so it is currently
1186		disabled.
1187
1188		----------------------------------------------------------------------
1189		Name: Lal George
1190		Date: 2002/03/22 14:18:25 EST
1191		Tag: george-20020322-cps-branch-prob
1192		Description:
1193
1194		Implemented the Ball-Larus branch prediction-heuristics, and
1195		incorporated graphical viewers for control flow graphs.
1196
1197		Ball-Larus Heuristics:
1198		---------------------
1199		See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
1200
1201		By design it uses the Dempster-Shafer theory for combining
1202		probabilities.  For example, in the function:
1203
1204		fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
1205
1206		the ball-larus heuristics predicts that the n=0 is unlikely
1207		(OH-heuristic), and the 'then' branch is unlikely because of the
1208		RH-heuristic -- giving the 'then' branch an even lower combined
1209		probability using the Dempster-Shafer theory.
1210
1211		Finally, John Reppy's loop analysis in MLRISC, further lowers the
1212		probability of the 'then' branch because of the loop in the else
1213		branch.
1214
1215
1216		Graphical Viewing:
1217		------------------
1218		I merely plugged in Allen's graphical viewers into the compiler. The
1219		additional code is not much. At the top level, saying:
1220
1221		Control.MLRISC.getFlag "cfg-graphical-view" := true;
1222
1223		will display the graphical view of the control flow graph just before
1224		back-patching.  daVinci must be in your path for this to work. If
1225		daVinci is not available, then the default viewer can be changed
1226		using:
1227
1228		Control.MLRISC.getString "viewer"
1229
1230		which can be set to "dot" or "vcg" for the corresponding viewers. Of
1231		course, these viewers must be in your path.
1232
1233		The above will display the compilation unit at the level of clusters,
1234		many of which are small, boring, and un-interesting. Also setting:
1235
1236		Control.MLRISC.getInt "cfg-graphical-view_size"
1237
1238		will display clusters that are larger than the value set by the above.
1239
1240
1241		----------------------------------------------------------------------
1242		Name: Matthias Blume
1243		Date: 2002/03/21 22:20:00 EST
1244		Tag: blume-20020321-kmp-bugfix
1245		Description:
1246
1247		Changed the interface to the KMP routine in PreString and fixed
1248		a minor bug in one place where it was used.
1249
1250		----------------------------------------------------------------------
1251		Name: Allen Leung
1252		Date: 2002/03/21 20:30:00 EST
1253		Tag: leunga-20020321-cfg
1254		Description:
1255
1256		Fixed a potential problem in cfg edge splitting.
1257
1258		----------------------------------------------------------------------
1259	Name: Allen Leung	Name: Allen Leung
1260	Date: 2002/03/21 17:15:00 EST	Date: 2002/03/21 17:15:00 EST
1261	Tag: leunga-20020321-x86-fp-cfg	Tag: leunga-20020321-x86-fp-cfg

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.1156   changed lines   Added in v.1411

---

root@smlnj-gforge.cs.uchicago.edu	ViewVC Help
Powered by ViewVC 1.0.0