Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 972, Sun Oct 28 00:37:02 2001 UTC	revision 1066, Thu Feb 14 16:50:02 2002 UTC
#	Line 8	Line 8
8	The form of an entry should be:	The form of an entry should be:
9
10	Name:	Name:
11	Date:	Date: yyyy/mm/dd
12	Tag: <post-commit CVS tag>	Tag: <post-commit CVS tag>
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2002/02/14 11:50:00 EST
18		Tag: blume-20020214-quote
19		Description:
20
21		Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
22		because of the new and improved tag generation scheme.  (Thanks to Allen
23		Leung for pointing it out.)
24
25		----------------------------------------------------------------------
26		Name: Lal George
27		Date: 2002/02/14 09:55:27 EST 2002
28		Tag: george-20020214-isabelle-bug
29		Description:
30
31		Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
32		of Isabelle on the x86.
33
34		From Allen:
35		-----------
36		I've found the problem:
37
38		in ra-core.sml, I use the counter "blocked" to keep track of the
39		true number of elements in the freeze queue.  When the counter goes
40		to zero, I skip examining the queue.  But I've messed up the
41		bookkeeping in combine():
42
43		else ();
44		case !ucol of
45		PSEUDO => (if !cntv > 0 then
46		(if !cntu > 0 then blocked := !blocked - 1 else ();
47		^^^^^^^^^^^^^^^^^^^^^^^
48		moveu := mergeMoveList(!movev, !moveu)
49		)
50		else ();
51
52		combine() is called to coalesce two nodes u and v.
53		I think I was thinking that if the move counts of u and v are both
54		greater than zero then after they are coalesced then one node is
55		removed from the freeze queue.  Apparently I was thinking that
56		both u and v are of low degree, but that's clearly not necessarily true.
57
58
59		02/12/2002:
60		Here's the patch.  HOL now compiles.
61
62		I don't know how this impact on performance (compile
63		time or runtime).  This bug caused the RA (especially on the x86)
64		to go thru the potential spill phase when there are still nodes on the
65		freeze queue.
66
67
68
69
70		----------------------------------------------------------------------
71		Name: Matthias Blume
72		Date: 2002/02/13 22:40:00 EST
73		Tag: blume-20020213-fptr-rtti
74		Description:
75
76		Fixed a bug in ml-nlffigen that was introduced with one of the previous
77		updates.
78
79		----------------------------------------------------------------------
80		Name: Matthias Blume
81		Date: 2002/02/13 16:41:00 EST
82		Tag: blume-20020213-cmlpq
83		Description:
84
85		Added new priority queue export symbols (which have just been added to
86		smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
87		not compile and the installer would choke.)
88
89		----------------------------------------------------------------------
90		Name: Matthias Blume
91		Date: 2002/02/13 16:15:00 EST
92		Tag: blume-20020213-various
93		Description:
94
95		1. More tweaks to ml-nlffigen:
96
97		- better internal datastructures (resulting in slight speedup)
98		- "-match" option requires exact match
99		- "localized" gensym counters (untagged structs/unions nested within
100		other structs/unions or within typedefs get a fresh counter; their
101		tag will be prefixed by a concatenation of their parents' tags)
102		- bug fixes (related to calculation of transitive closure of types
103		to be included in the output)
104
105		2. Minor Basis updates:
106
107		- added implementations for List.collate and Option.app
108
109		----------------------------------------------------------------------
110		Name: Matthias Blume
111		Date: 2002/02/11 15:55:00 EST
112		Tag: blume-20020211-gensym
113		Description:
114
115		Added a "-gensym" option to command line of ml-nlffigen.  This can be
116		used to specify a "stem" -- a string that is inserted in all "gensym'd"
117		names (ML structure names that correspond to unnamed C structs, unions,
118		and enums), so that separate runs of ml-nlffigen do not clash.
119
120		----------------------------------------------------------------------
121		Name: Matthias Blume
122		Date: 2002/02/11 12:05:00 EST
123		Tag: blume-20020211-gensml
124		Description:
125
126		A quick fix for a problem with GenSML (in the pgraph-util library):
127		Make generation of toplevel "local" optional.  (Strictly speaking,
128		signature definitions within "local" are not legal SML.)
129
130		Other than that: updates to INSTALL and cm/TODO.
131
132		----------------------------------------------------------------------
133		Name: Matthias Blume
134		Date: 2002/02/08 15:00:00 EST
135		Tag: blume-20020208-uniquepid
136		Description:
137
138		0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
139
140		1. The installer (config/install.sh) has gotten smarter:
141
142		- Configuration options are a bit easier to specify now
143		(in config/targets).
144		- Bug in recognizing .tar.bz2 files fixed.
145		- Installer automatically resolves dependencies between
146		configuration options (e.g., if you ask for eXene, you will
147		also get cml -- regardless whether you asked for it or not).
148		- Installer can run in "quieter mode" by setting the environment
149		variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
150		"completely silent", though.
151		- Build HashCons library as part of smlnj-lib.
152
153		2. A new scheme for assigning persistent identifiers to compilation
154		units (and, by extension, to types etc.) has been put into place.
155		This fixes a long-standing bug where types and even dynamic values
156		can get internally confused, thereby compromising type safety
157		(abstraction) and dynamic correctness.  See
158
159		http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
160
161		for an example of how things could go wrong until now.
162
163		The downside of the new scheme is that pids are not quite as
164		persistent as they used to be: CM will generate a fresh pid
165		for every compilation unit that it thinks it sees for the first
166		time.  That means that if you compile starting from a clean, fresh
167		source tree at two different times, you end up with different
168		binaries.
169
170		Cutoff recompilation, however, has not been compromised because
171		CM keeps pid information in special caches between runs.
172
173		----------------------------------------------------------------------
174		Name: Lal George
175		Date: 2002/02/07 15:34:13 EST 2002
176		Tag: <none>
177		Description:
178
179		Compilers that generate assembly code may produce  global labels
180		whose value is resolved at link time. The various peephole optimization
181		modules did not take this in account.
182
183		TODO. The Labels.addrOf function should really return an option
184		type so that clients are forced to deal with this issue, rather
185		than an exception being raised.
186
187		----------------------------------------------------------------------
188		Name: Lal George
189		Date: 2002/02/06 13:55:02 EST
190		Tag: george-20020206-ra-breakup
191		Description:
192
193		1. A bug fix from Allen.
194
195		A typo causes extra fstp %st(0)'s to be generated at compensation
196		edges, which might cause stack underflow traps at runtime.  This
197		occurs in fft where there are extraneous fstps right before the 'into'
198		trap instruction (in this case they are harmless since none of the
199		integers overflow.)
200
201		2. Pulled out various utility modules that were embedded in the modules
202		of the register allocator. I need these modules for other purposes, but
203		they are not complete enough to put into a library (just yet).
204		----------------------------------------------------------------------
205		Name: Matthias Blume
206		Date: 2002/01/31 16:05:00 EST
207		Tag: blume-20020131-sparc-ccalls
208		Description:
209
210		1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
211		of extra stack space by mistake.  Fixed.
212
213		2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
214
215		----------------------------------------------------------------------
216		Name: Allen Leung
217		Date: 2002/01/30
218		Tag: leunga-20020130-nowhere-bug-fix
219		Description:
220
221		MLRISC bug fixes:
222		1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
223		2. MachineInt.fromString was negating its value.
224
225		----------------------------------------------------------------------
226		Name: Matthias Blume
227		Date: 2002/01/29
228		Tag: blume-20020129-INSTALL
229		Description:
230
231		- Added somewhat detailed installation instructions (file INSTALL).
232		- Fixed curl-detection bug in config/install.sh.
233		- It is now possible to select the URL getter using the URLGETTER
234		environment variable:
235
236		not set / "unknown"      --> automatic detection (script tries wget,
237		curl, and lynx)
238		"wget" / "curl" / "lynx" --> use the specified program (script "knows"
239		how to properly invoke them)
240		other                    --> use $URLGETTER directly, it must take
241		precisely two command-line arguments
242		(source URL and destination file name)
243
244		----------------------------------------------------------------------
245		Name: Matthias Blume
246		Date: 2002/01/28
247		Tag: blume-20020128-sparc-ccalls
248		Description:
249
250		- Fixed problem with calculation of "used" registers in sparc-c-calls.
251		- Make use of the allocParam argument in sparc-c-calls.
252
253		----------------------------------------------------------------------
254		Name: Matthias Blume
255		Date: 2002/01/28
256		Tag: blume-20020128-allocParam
257		Description:
258
259		John Reppy:  Changes c-calls API to accept client-callback for
260		allocating extra stack space.
261		me: Corresponding changes to mlriscGen (using a dummy argument that
262		does not change the current behavior).
263
264		----------------------------------------------------------------------
265		Name: Matthias Blume
266		Date: 2002/01/28 12:00:00
267		Tag: Release_110_38
268		Description:
269
270		This time for real!!!
271
272		----------------------------------------------------------------------
273		Name: Matthias Blume
274		Date: 2002/01/28 10:56:00 EST
275		Tag: blume-20020128-retraction
276		Description:
277
278		0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
279		with blume-Release_110_38-retracted.)
280
281		1. Fixed a problem with incorrect rounding modes in real64.sml.
282		(Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
283
284		2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
285		and enums fixed.  The naming of corresponding ML identifiers should
286		now be consistent again.
287
288		----------------------------------------------------------------------
289		Name: Allen Leung
290		Date: 2002/01/27
291		Tag: leunga-20020127-nowhere
292		Description:
293
294		Added a target called nowhere in the configuration scripts.
295		Enabling this will build the MLRISC 'nowhere' tool (for translating
296		programs with where-clauses into legal SML code) during installation.
297
298		----------------------------------------------------------------------
299		Name: Matthias Blume
300		Date: 2002/01/25 21:27:00 EST
301		Tag: blume-Release_110_38-retracted
302		Description:
303
304		Call it a (working) release!  Version is 110.38. Bootfiles are ready.
305
306		README will be added later.
307
308		!!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
309		(Release_110_38) removed.  Reason: Last-minute bug fixes.
310
311		----------------------------------------------------------------------
312		Name: Matthias Blume
313		Date: 2002/01/25
314		Tag: blume-20020125-ffi
315		Description:
316
317		A large number of tweaks and improvements to ml-nlffi-lib and
318		ml-nlffigen:
319
320		- ML represenation types have been streamlined
321		- getter and setter functions work with concrete values, not abstract
322		ones where possible
323		- ml-nlffigen command line more flexible (see README file there)
324		- some bugs have been fixed (hopefully)
325
326		----------------------------------------------------------------------
327		Name: Lal George
328		Date: 2002/01/24
329		Tag: george-20020124-risc-ra-interface
330		Description:
331
332		There is a dramatic simplification in the interface to the
333		register allocator for RISC architectures as a result of making
334		parallel copy instructions explicit.
335
336		----------------------------------------------------------------------
337		Name: Matthias Blume
338		Date: 2002/01/22
339		Tag: blume-20020122-x86-ccalls
340		Description:
341
342		Bug fix for c-calls on x86 (having to do with how char- and
343		short-arguments are being handled).
344
345		----------------------------------------------------------------------
346		Name: Matthias Blume
347		Date: 2002/01/21
348		Tag: blume-20020121-ff
349		Description:
350
351		Another day of fiddling with the FFI...
352
353		1. Bug fix/workaround:  CKIT does not complain about negative array
354		dimensions, so ml-nlffigen has to guard itself against this possibility.
355		(Otherwise a negative dimension would send it into an infinite loop.)
356
357		2. Some of the abstract types (light objects, light pointers, most "base"
358		types) in structure C are now eqtypes.
359
360		3. Added constructors and test functions for NULL function pointers.
361
362		----------------------------------------------------------------------
363		Name: Matthias Blume
364		Date: 2002/01/18
365		Tag: blume-20020118-ready-for-new-release
366		Description:
367
368		Made config/srcarchiveurl point to a new place.  (Will provide boot
369		files shortly.)
370
371		Maybe we christen this to be 110.38?
372
373		----------------------------------------------------------------------
374		Name: Matthias Blume
375		Date: 2002/01/18
376		Tag: blume-20020118-more-ffifiddle
377		Description:
378
379		Today's FFI fiddling:
380
381		- Provided a structure CGetSet with "convenient" versions of C.Get.* and
382		C.Set.* that use concrete (MLRep.*) arguments and results instead
383		of abstract ones.
384
385		- Provided word-style bit operations etc. for "int" representation
386		types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
387		and Long.
388
389		----------------------------------------------------------------------
390		Name: Matthias Blume
391		Date: 2002/01/18
392		Tag: blume-20020118-use-x86-fp
393		Description:
394
395		Now that x86-fast-fp seems to be working, I turned it back on again
396		by default.  (Seems to work fine now, even with the FFI.)
397
398		Other than that, I added some documentation about the FFI to
399		src/ml-nlffigen/README and updated the FFI test examples in
400		src/ml-nlffi-lib/Tests/*.
401
402		----------------------------------------------------------------------
403		Name: Allen Leung
404		Date: 2002/01/17
405		Tag: leunga-20020117-x86-fast-fp-call
406		Description:
407
408		1. Fixed a problem with handling return fp values when x86's fast fp
409		mode is turned on.
410
411		2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
412		of %f32.
413
414		3. Added a constructor INT32lit to the ast of MLRISC tools.
415
416		----------------------------------------------------------------------
417		Name: Matthias Blume
418		Date: 2002/01/16
419		Tag: blume-20020116-ffifiddle
420		Description:
421
422		More fiddling with the FFI interface:
423
424		- Make constness 'c instead of rw wherever possible.  This eliminates
425		the need for certain explicit coercions.  (However, due to ML's
426		value polymorphism, there will still be many cases where explicit
427		coercions are necessary.  Phantom types are not the whole answer
428		to modeling a subtyping relationship in ML.)
429
430		- ro/rw coersions for pointers added.  (Avoids the detour through */&.)
431
432		- "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
433		clumsy workaround for varargs problem.)
434
435		----------------------------------------------------------------------
436		Name: Lal George
437		Date: 2002/01/15
438		Tag: <none>
439		Description:
440
441		1. Since COPY instructions are no longer native to the architecture,
442		a generic functor can be used to implement the expandCopies function.
443
444		2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
445		TEXT segment.
446
447		----------------------------------------------------------------------
448		Name: Matthias Blume
449		Date: 2002/01/15
450		Tag: blume-20020115-ffiupdates
451		Description:
452
453		1. Fix for bug resulting in single-precision float values being returned
454		incorrectly from FFI calls.
455
456		2. Small modifications to C FFI API:
457
458		- memory-allocation routines return straight objects (no options)
459		and raise an exception in out-of-memory situations
460		- unsafe extensions to cast between function pointers and pointers
461		from/to ints
462		- added structure C_Debug as an alternative to structure C where
463		pointer-dereferencing (|*| and |*!) always check for null-pointers
464		- added open_lib' to DynLinkage;  open_lib' works like open_lib
465		but also takes a (possibly empty) list of existing library handles
466		that the current library depends on
467
468		----------------------------------------------------------------------
469		Name: Matthias Blume
470		Date: 2002/01/10
471		Tag: blume-20020110-newffigen
472		Description:
473
474		1. Updates to portable graph code.
475
476		2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
477		more scalable now so that even huge interfaces such as the one
478		for GTK compile in finite time and space. :-)
479		See src/ml-nlffigen/README for details on what's new.
480
481		----------------------------------------------------------------------
482		Name: Lal George
483		Date: 2001/01/09 14:31:35 EST 2002
484		Tag: george-20011206-rm-native-copy
485		Description:
486
487		Removed the native COPY and FCOPY instructions
488		from all the architectures and replaced it with the
489		explicit COPY instruction from the previous commit.
490
491		It is now possible to simplify many of the optimizations
492		modules that manipulate copies. This has not been
493		done in this change.
494
495		----------------------------------------------------------------------
496		Name: Lal George
497		Date: 2001/12/06 16:50:13 EST 2001
498		Tag: george-20011206-mlrisc-instruction
499		Description:
500
501		Changed the representation of instructions from being fully abstract
502		to being partially concrete. That is to say:
503
504		from
505		type instruction
506
507		to
508		type instr                              (* machine instruction *)
509
510		datatype instruction =
511		LIVE of {regs: C.cellset, spilled: C.cellset}
512		| KILL of {regs: C.cellset, spilled: C.cellset}
513		| COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
514		| ANNOTATION of {i: instruction, a: Annotations.annotation}
515		| INSTR of instr
516
517		This makes the handling of certain special instructions that appear on
518		all architectures easier and uniform.
519
520		LIVE and KILL say that a list of registers are live or killed at the
521		program point where they appear. No spill code is generated when an
522		element of the 'regs' field is spilled, but the register is moved to
523		the 'spilled' (which is present, more for debugging than anything else).
524
525		LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
526		We used to generate:
527
528		DEFFREG f1
529		f1 := f2 + f3
530		trapb
531
532		but now generate:
533
534		f1 := f2 + f3
535		trapb
536		LIVE {regs=[f1,f2,f3], spilled=[]}
537
538		Furthermore, the DEFFREG (hack) required that all floating point instruction
539		use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
540		defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
541		in a cleaner alpha implementation. (Hopefully, intel will not get rid of
542		this architecture).
543
544		COPYXXX is intended to replace the parallel COPY and FCOPY  available on
545		all the architectures. This will result in further simplification of the
546		register allocator that must be aware of them for coalescing purposes, and
547		will also simplify certain aspects of the machine description that provides
548		callbacks related to parallel copies.
549
550		ANNOTATION should be obvious, and now INSTR represents the honest to God
551		machine instruction set!
552
553		The <arch>/instructions/<arch>Instr.sml files define certain utility
554		functions for making porting easier -- essentially converting upper case
555		to lower case. All machine instructions (of type instr) are in upper case,
556		and the lower case form generates an MLRISC instruction. For example on
557		the alpha we have:
558
559		datatype instr =
560		LDA of {r:cell, b:cell, d:operand}
561		| ...
562
563		val lda : {r:cell, b:cell, d:operand} -> instruction
564		...
565
566		where lda is just (INSTR o LDA), etc.
567
568		----------------------------------------------------------------------
569		Name: Matthias Blume
570		Date: 2001/11/22 21:40:00 EST
571		Tag: Release_110_37
572		Description:
573
574		Release 110.37.  This time for real.
575
576		----------------------------------------------------------------------
577		Name: Matthias Blume
578		Date: 2001/11/21 16:35:00 EST
579		Tag: blume-20011121-foot-in-mouth
580		Description:
581
582		Removed the "Release_110_37" tag because of a serious bug.
583		This will be re-tagged once the bug is fixed.
584
585		----------------------------------------------------------------------
586		Name: Matthias Blume
587		Date: 2001/11/21 16:14:00 EST
588		Tag: blume-20011121-forgottenfile
589		Description:
590
591		Forgot to add a file.  (Just a .tex-file -- part of
592		the CM manual source.)
593
594		----------------------------------------------------------------------
595		Name: Matthias Blume
596		Date: 2001/11/21 16:10:00 EST
597		Tag: blume-20011121-invalid_110_37
598		Description:
599
600		Note: I removed the original tag "Release_110_37" from this commit
601		because we found a serious bug in all non-x86 backends.
602		- Matthias
603
604		1. Modifications to the SML/NJ code generator and to the runtime system
605		so that code object name strings are directly inserted into code
606		objects at code generation time.  The only business the runtime system
607		has with this is now to read the name strings on occasions.
608		(The encoding of the name string has also changed somewhat.)
609
610		2. CM now implements a simple "set calculus" for specifying export lists.
611		In particular, it is now possible to refer to the export lists of
612		other libraries/groups/sources and form unions as well as differences.
613		See the latest CM manual for details.
614
615		3. An separate notion of "proxy" libraries has again be eliminated from
616		CM's model.  (Proxy libraries are now simply a special case of using
617		the export list calculus.)
618
619		4. Some of the existing libraries now take advantage of the new set
620		calculus.
621		(Notice that not all libraries have been converted because some
622		of the existing .cm-files are supposed to be backward compatible
623		with 110.0.x.)
624
625		5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
626		"General.exnMessage"!  The former relies on a certain hook to be
627		initialized, and that often does not happen in the stand-alone case.)
628
629		----------------------------------------------------------------------
630		Name: Lal George
631		Date: 2001/11/21  13:56:18 EST
632		Tag: george-2001121-pseudo-ops
633		Description:
634
635		Implemented a complete redesign of MLRISC pseudo-ops. Now there
636		ought to never be any question of incompatabilities with
637		pseudo-op syntax expected by host assemblers.
638
639		For now, only modules supporting GAS syntax are implemented
640		but more should follow, such as MASM, and vendor assembler
641		syntax, e.g. IBM as, Sun as, etc.
642
643		----------------------------------------------------------------------
644		Name: Matthias Blume
645		Date: 2001/11/14 11:52:00 EST
646		Tag: blume-20011114-srcname
647		Description:
648
649		1. Routed the name of the current source file to mlriscgen where it
650		should be directly emitted into the code object.  (This last part
651		is yet to be done.)
652
653		2. Some cleanup of the pgraph code to make it match the proposal that
654		I put out the other day.  (The proposal notwithstanding, things are
655		still in flux here.)
656
657		----------------------------------------------------------------------
658		Name: Lal George
659		Date: 2001/11/14 09:44:04 EST
660		Tag:
661		Description:
662
663		Fix for a backpatching bug reported by Allen.
664
665		Because the boundary between short and long span-dependent
666		instructions is +/- 128, there are an astounding number of
667		span-dependent instructions whose size is over estimated.
668
669		Allen came up with the idea of letting the size of span
670		dependent instructions be non-monotonic, for a maxIter
671		number of times, after which the size must be monotonically
672		increasing.
673
674		This table shows the number of span-dependent instructions
675		whose size was over-estimated as a function of maxIter, for the
676		file Parse/parse/ml.grm.sml:
677
678		maxIter            # of instructions:
679		10                      687
680		20                      438
681		30                      198
682		40                        0
683
684		In compiling the compiler, there is no significant difference in
685		compilation speed between maxIter=10 and maxIter=40. Actually,
686		my measurements showed that maxIter=40 was a tad faster than
687		maxIter=10! Also 96% of the  files in the compiler reach a fix
688		point within 13 iterations, so fixing maxIter at 40, while high,
689		is okay.
690
691		----------------------------------------------------------------------
692		Name: Matthias Blume
693		Date: 2001/10/31 15:25:00 EST
694		Tag: blume-20011031-pgraph
695		Description:
696
697		CKIT:
698		* Changed the "Function" constructor of type Ast.ctype to carry optional
699		argument identifiers.
700		* Changed the return type of TypeUtil.getFunction accordingly.
701		* Type equality ignores the argument names.
702		* TypeUtil.composite tries to preserve argument names but gives up quickly
703		if there is a mismatch.
704
705		installation script:
706		* attempts to use "curl" if available (unless "wget" is available as well)
707
708		CM:
709		* has an experimental implementation of "portable graphs" which I will
710		soon propose as an implementation-independent library format
711		* there are also new libraries $/pgraph.cm and $/pgraph-util.cm
712
713		NLFFI-LIB:
714		* some cleanup (all cosmetic)
715
716		NLFFIGEN:
717		* temporarily disabled the mechanism that suppresses ML output for
718		C definitions whose identifiers start with an underscore character
719		* generate val bindings for enum constants
720		* user can request that only one style (light or heavy) is being used;
721		default is to use both (command-line arguments: -heavy and -light)
722		* fixed bug in handling of function types involving incomplete pointers
723		* generate ML entry points that take record arguments (i.e., using
724		named arguments) for C functions that have a prototype with named
725		arguments
726		(see changes to CKIT)
727
728		----------------------------------------------------------------------
729	Name: Allen Leung	Name: Allen Leung
730	Date: 2001/10/27 20:34:00 EDT	Date: 2001/10/27 20:34:00 EDT
731	Tag: leunga-20011027-x86-fast-fp-call	Tag: leunga-20011027-x86-fast-fp-call
#	Line 109	Line 822
822
823	----------------------------------------------------------------------	----------------------------------------------------------------------
824	Name: Matthias Blume	Name: Matthias Blume
	>>>>>>> 1.169
825	Date: 2001/09/18 15:35:00 EDT	Date: 2001/09/18 15:35:00 EDT
826	Tag: blume-20010918-readme11036	Tag: blume-20010918-readme11036
827	Description:	Description:

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.972   changed lines   Added in v.1066

---

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