Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 756, Thu Dec 14 16:01:33 2000 UTC	revision 854, Wed Jun 27 19:11:38 2001 UTC
#	Line 11	Line 11
11	Date:	Date:
12	Tag: <post-commit CVS tag>	Tag: <post-commit CVS tag>
13	Description:	Description:
14
15		----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2001/06/27 15:10:00 EDT
18		Tag: blume-20010627-concur
19		Description:
20
21		Fixed a small bug in CM's handling of parallel compilation.
22		(You could observe the bug by Control-C-interrupting an ordinary
23		CMB.make or CM.stabilize and then attaching some compile servers.
24		The result was that all of a sudden the previously interrupted
25		compilation would continue on its own.  This was because of
26		an over-optimization: CM did not bother to clean out certain queues
27		when no servers were attached "anyway", resulting in the contents
28		of these queues to grab control when new servers did get attached.)
29
30		There is also another minor update to the CM manual.
31
32		----------------------------------------------------------------------
33		Name: Matthias Blume
34		Date: 2001/06/26 16:15:00 EDT
35		Tag: blume-20010626-cmdoc
36		Description:
37
38		Minor typo fixed in CM manual (syntax diagram for libraries).
39
40		----------------------------------------------------------------------
41		Name: Matthias Blume
42		Date: 2001/06/25 22:55:00 EDT
43		Tag: blume-20010625-x86pc
44		Description:
45
46		Fixed a nasty bug in the X86 assembly code that caused signal
47		handlers to fail (crash) randomly.
48
49		----------------------------------------------------------------------
50		Name: Matthias Blume
51		Date: 2001/06/25 12:05:00 EDT
52		Tag: blume-20010625-nlffigen
53		Description:
54
55		This update fixes a number of minor bugs in ml-nlffigen as reported by
56		Nick Carter <nbc@andrew.cmu.edu>.
57
58		1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
59		2. Default names for generated files are now derived from the name of
60		the C file *without its directory*.  In particular, this causes generated
61		files to be placed locally even if the C file is in some system directory.
62		3. Default names for generated signatures and structures are also derived
63		from the C file name without its directory.  This avoids silly things
64		like "structure GL/GL".
65		(Other silly names are still possible because ml-nlffigen does not do
66		a thorough check of whether generated names are legal ML identifiers.
67		When in doubt, use command line arguments to force particular names.)
68
69		----------------------------------------------------------------------
70		Name: Matthias Blume
71		Date: 2001/06/21 12:25:00 EDT
72		Tag: blume-20010621-eXene
73		Description:
74
75		eXene now compiles and (sort of) works again.
76
77		The library name (for version > 110.33) is $/eXene.cm.
78
79		I also added an new example in src/eXene/examples/nbody.  See the
80		README file there for details.
81
82		----------------------------------------------------------------------
83		Name: Matthias Blume
84		Date: 2001/06/20 16:40:00 EDT
85		Tag: blume-20010620-cml
86		Description:
87
88		CML now compiles and works again.
89
90		Libraries (for version > 110.33):
91
92		$cml/cml.cm            Main CML library.
93		$cml/basis.cm          CML's version of $/basis.cm.
94		$cml/cml-internal.cm   Internal helper library.
95		$cml/core-cml.cm       Internal helper library.
96		$cml-lib/trace-cml.cm  Tracing facility.
97		$cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
98
99		The installer (config/install.sh) has been taught how to properly
100		install this stuff.
101
102		----------------------------------------------------------------------
103		Name: Matthias Blume
104		Date: 2001/06/19 17:55:00 EDT
105		Tag: blume-20010619-instantiate
106		Description:
107
108		This un-breaks the fix for bug 1432.
109		(The bug was originally fixed in 110.9 but I broke it again some
110		time after that.)
111
112		----------------------------------------------------------------------
113		Name: Matthias Blume
114		Date: 2001/06/19 17:25:00 EDT
115		Tag: blume-20010619-signals
116		Description:
117
118		This should (hopefully) fix the long-standing signal handling bug.
119		(The runtime system was constructing a continuation record with an
120		incorrect descriptor which would cause the GC to drop data on the floor...)
121
122		----------------------------------------------------------------------
123		Name: Matthias Blume
124		Date: 2001/06/15 15:05:00 EDT
125		Tag: blume-20010615-moresparc
126		Description:
127
128		Here is a short late-hour update related to Sparc c-calls:
129
130		-- made handling of double-word arguments a bit smarter
131
132		-- instruction selection phase tries to collapse certain clumsily
133		constructed ML-Trees; typical example:
134
135		ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
136
137		This currently has no further impact on SML/NJ since mlriscGen does
138		not seem to generate such patterns in the first place, and c-calls
139		(which did generate them in the beginning) has meanwhile been fixed
140		so as to avoid them as well.
141
142		----------------------------------------------------------------------
143		Name: Matthias Blume
144		Date: 2001/06/15 15:05:00 EDT
145		Tag: blume-20010615-sparc
146		Description:
147
148		The purpose of this update is to provide an implementation of NLFFI
149		on Sparc machines.
150
151		Here are the changes in detail:
152
153		* src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
154		the Sparc implementation of the c-calls API.
155		* The Sparc backend of SML/NJ has been modified to uniformely use %fp
156		for accessing the ML frame.  Thus, we have a real frame pointer and
157		can freely modify %sp without need for an omit-frame-ptr phase.
158		The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
159		this case.
160		* ml-nlffigen has been taught to produce code for different architectures
161		and calling conventions.
162		* In a way similar to what was done in the x86 case, the Sparc
163		backend uses its own specific extension to mltree.  (For example,
164		it needs to be able to generate UNIMP instructions which are part
165		of the calling convention.)
166		* ml-nlffi-lib was reorganized to make it more modular (in particular,
167		to make it easier to plug in new machine- and os-dependent parts).
168
169		There are some other fairly unrelated bug fixes and cleanups as well:
170
171		* I further hacked the .cm files for MLRISC tools (like MDLGen) so
172		that they properly share their libraries with existing SML/NJ libraries.
173		* I fixed a minor cosmetic bug in CM, supressing certain spurious
174		follow-up error messages.
175		* Updates to CM/CMB documentation.
176
177		TODO items:
178
179		* MLRISC should use a different register as its asmTemp on the Sparc.
180		(The current %o2 is a really bad choice because it is part of the
181		calling conventions, so things might interfere in unexpected ways.)
182
183		----------------------------------------------------------------------
184		Name: Matthias Blume
185		Date: 2001/06/07
186		Tag: blume-20010607-calls
187		Description:
188
189		A number of internal changes related to C calls and calling conventions:
190
191		1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
192		number of bytes popped implicitly (by the callee).  In most cases
193		this field is 0 but on x86/win32 it is some non-zero value.  This
194		is information provided for the benefit of the "omit-frameptr" pass.
195		2. The CALL instruction on the x86 carries a similar "pops" field.
196		The instruction selection phase copies its value from the ML-Tree
197		CALL statement.
198		3. On all other architectures, the instruction selection phase checks
199		whether "pops=0" and complains if not.
200		4. The c-calls implementation for x86 now accepts two calling conventions:
201		"ccall" and "stdcall".  When "ccall" is selected, the caller cleans
202		up after the call and pops is set to 0.  For "stdcall", the caller
203		does nothing, leaving the cleanup to the callee; pops is set to
204		the number of bytes that were pushed onto the stack.
205		5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
206		distinguish between "ccall" and "stdcall".
207		6. The UNIMP instruction has been added to the supported Sparc instruction
208		set. (This is needed for implementing the official C calling convention
209		on this architecture.)
210		7. I fixed some of the .cm files under src/MLRISC/Tools to make them
211		work with the latest CM.
212
213		----------------------------------------------------------------------
214		Name: Matthias Blume
215		Date: 2001/06/05 15:10:00 EDT
216		Tag: blume-20010605-cm-index
217		Description:
218
219		0. The "lambdasplit" parameter for class "sml" in CM has been documented.
220
221		1. CM can now generate "index files".  These are human-readable files
222		that list on a per-.cm-file basis each toplevel symbol defined or
223		imported.  The location of the index file for
224		<p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
225		To enable index-file generation, set CM.Control.generate_index to true
226		or export an environment-symbol: export CM_GENERATE_INDEX=true.
227
228		The CM manual has been updated accordingly.
229
230		2. I made some slight modifications to the c-calls API in MLRISC.
231
232		a) There is now a callback to support saving/restoring of
233		dedicated but caller-save registers around the actual call
234		instruction.
235		b) One can optionally specify a comment-annotation for the
236		call instruction.
237
238		3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
239		(For example, the comment annotation shows the C prototype of
240		the function being called.)
241
242		----------------------------------------------------------------------
243		Name: Matthias Blume
244		Date: 2001/06/01 13:30:00 EDT
245		Tag: blume-20010601-nlffi-cleanup
246		Description:
247
248		This is mostly a cleanup of MLFFI stuff:
249
250		- some signature files have been put into a more exposed place
251		- the ugly 'f type parameter is gone (simplifies types tremendously!)
252		- ml-nlffigen changed accordingly
253		- tutorial updated
254
255		Other changes:
256
257		- author's affiliation in CM manual(s) updated
258		- some more recognized keywords added to Allen's sml.sty
259
260		----------------------------------------------------------------------
261		Name: Matthias Blume
262		Date: 2001/05/25 15:30:00 EDT
263		Tag: blume-20010525-iptr
264		Description:
265
266		- put the official 110.33-README (as it appears on the ftp server) under
267		CVS
268		- fixed a small bug related to incomplete pointer types in
269		ml-nlffigen
270		- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
271		(it does not need the 'f type parameter)
272
273		----------------------------------------------------------------------
274		Name: Matthias Blume
275		Date: 2001/05/23 14:30:00 EDT
276		Tag: Release_110_33
277		Description:
278
279		New version number (110.33).  New bootfiles.
280
281		----------------------------------------------------------------------
282		Name: Matthias Blume
283		Date: 2001/05/22 18:06:00 EDT
284		Tag: blume-20010522-targets
285		Description:
286
287		Made install.sh use file config/targets.customized if it exists, falling
288		back to config/targets if it doesn't.  This way one can have a customized
289		version of the targets file without touching the "real thing", thus
290		eliminating the constant fear of accidentally checking something bogus
291		back into the CVS repository...  (File config/targets.customized must
292		not be added to the repository!)
293
294		----------------------------------------------------------------------
295		Name: Matthias Blume
296		Date: 2001/05/22 16:30:00 EDT
297		Tag: blume-20010522-minitut
298		Description:
299
300		1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
301		struct returns.
302		2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
303		incomplete, preliminary documentation for NLFFI.
304
305		----------------------------------------------------------------------
306		Name: Matthias Blume
307		Date: 2001/05/14 11:30:00 EDT
308		Tag: blume-20010514-script
309		Description:
310
311		Some bugs in install script fixed.
312
313		In addition to that I also made a slight change to the NLFFI API:
314		Functors generated by ml-nlffigen now take the dynamic library as a
315		straight functor argument, not as a suspended one.  (The original
316		functor code used to force the suspension right away anyway, so there
317		was nothing gained by this complication of the interface.)
318
319		----------------------------------------------------------------------
320		Name: Matthias Blume
321		Date: 2001/05/11 14:35:00 EDT
322		Tag: blume-20010511-ml-nlffi
323		Description:
324
325		I finally took the plunge and added my new FFI code to the main
326		repository.  For x86-linux it is now ready for prime-time.
327
328		There are two new subdirectories of "src":
329
330		- ml-nlffi-lib:
331		The utility library for programs using the FFI interface.
332		Here is the implementation of $/c.cm and its associated low-level
333		partners $/c-int.cm and $/memory.cm.
334		- ml-nlffigen:
335		A stand-alone program for generating ML glue code from C source
336		code.
337
338		Building ml-nlffigen requires $/ckit-lib.cm.
339
340		The config/install.sh script has been updates to do the Right Thing
341		(hopefully).
342
343		Notice that the source tree for the C-Kit will not be put under "src"
344		but directly under the installation root directory.  (This is the
345		structure that currently exists on the CVS server when you check out
346		module "sml".)  Fortunately, config/install.sh knows about this oddity.
347
348		Bugs: No documentation yet.
349
350		----------------------------------------------------------------------
351		Name: Matthias Blume
352		Date: 2001/05/09 16:35:00 EDT
353		Tag: blume-20010509-cpscontract
354		Description:
355
356		Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
357		wrapper/unwrapper elimination did not decrement usage counts and some
358		dead variables got overlooked by the dead-up logic.)
359
360		----------------------------------------------------------------------
361		Name: Lal George
362		Date: 2001/05/08  17:26:09 EDT
363		Tag: george-20010508-omit-frameptr
364		Description:
365
366		Changes to implement the omit-frame-pointer optimization to support
367		raw C calls. For now, there is only support on the Intel x86, but
368		other architectures will follow as more experience is gained with this.
369
370
371		----------------------------------------------------------------------
372		Name: Matthias Blume
373		Date: 2001/05/07 14:40:00 EDT
374		Tag: blume-20010507-proxies
375		Description:
376
377		I made into "proxy libraries" all libraries that qualify for such a
378		change.  (A qualifying library is a library that has another library or
379		groups as its sole member and repeats that member's export list
380		verbatim.  A proxy library avoids this repetition by omitting its export
381		list, effectively inheriting the list that its (only) member exports.
382		See the CM manual for more explanation.)
383		The main effect is that explicit export lists for these libraries
384		do not have to be kepts in sync, making maintenance a bit easier.
385
386		I also added copyright notices to many .cm-files.
387
388		Last but not least, I made a new set of bootfiles.
389
390		----------------------------------------------------------------------
391		Name: Matthias Blume
392		Date: 2001/05/04 17:00:00 EDT
393		Tag: blume-20010504-cm-lsplit
394		Description:
395
396		0. John merged pending changes to $/smlnj-lib.cm
397
398		1. Allen's previous change accidentally backed out of one of Lal's
399		earlier changes.  I undid this mistake (re-introducing Lal's change).
400
401		2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
402		within the compiler where applicable.  There is some code simplification
403		because of that.
404
405		3. The "split" phase (in FLINT) is now part of the default list of phases.
406		Compiler.Control.LambdaSplitting.* can be used to globally control the
407		lambda-splitting (cross-module-inlining) engine.  In addition to that,
408		it can now also be controlled on a per-source basis: CM has been taught
409		a new tool parameter applicable to ML source files.
410
411		- To turn lambda-splitting off completely:
412		local open Compiler.Control.LambdaSplitting in
413		val _ = set Off
414		end
415		- To make "no lambda-splitting" the global default (but allow per-source
416		overriding); this is the initial setting:
417		local open Compiler.Control.LambdaSplitting in
418		val _ = set (Default NONE)
419		end
420		- To make "lambda-splitting with aggressiveness a" the global default
421		(and allow per-source overriding):
422		local open Compiler.Control.LambdaSplitting in
423		val _ = set (Default (SOME a))
424		end
425
426		- To turn lambda-splitting off for a given ML souce file (say: a.sml)
427		write (in the respective .cm-file):
428		a.sml (lambdasplitting:off)
429		- To turn lambda-splitting for a.sml on with minimal aggressiveness:
430		a.sml (lambdasplitting:on)
431		- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
432		<a> is a decimal non-negative integer):
433		a.sml (lambdasplitting:<a>)
434		- To turn lambda-splitting for a.sml on with maximal aggressiveness:
435		a.sml (lambdasplitting:infinity)
436		- To use the global default for a.sml:
437		a.sml (lambdasplitting:default)
438		or simply
439		a.sml
440
441		----------------------------------------------------------------------
442		Name: Allen Leung
443		Date: 2001/05/04 01:57:00 EDT
444		Tag: leunga-20010504-sync
445		Description:
446
447		MLRISC features.
448
449		1. Fix to CMPXCHG instructions.
450		2. Changed RA interface to allow annotations in callbacks.
451		3. Added a new method to the stream interface to allow annotations updates.
452
453		----------------------------------------------------------------------
454		Name: Matthias Blume
455		Date: 2001/05/01 11:45:00 EDT
456		Tag: blume-20010501-pcedittmp
457		Description:
458
459		Changed install.sh to use the current working directory instead of
460		/usr/tmp for a temporary file (pcedittmp).  The previous choice
461		of /usr/tmp caused trouble with MacOS X because of file premission
462		problems.
463
464		----------------------------------------------------------------------
465		Name: Matthias Blume
466		Date: 2001/04/20 11:10:00 EDT
467		Tag: blume-20010420-inMLflag
468		Description:
469
470		- added vp_limitPtrMask to vproc-state.h
471		(for use by the raw-C-calls mechanism to implement proper interrupt
472		handling)
473		- made the ML compiler aware of various data-structure offsets so it
474		can generate code for accessing the vp_inML flag and vp_limitPtrMask
475		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
476		raw C-calls
477
478		----------------------------------------------------------------------
479		Name: Lal George
480		Date: 2001/04/20 09:15:28 EDT
481		Tag: george-20010420-macosX
482		Description:
483
484		- Changes to port to Mac OS X; Darwin.
485
486		- In the process I found that sqrt was broken on the PPC, because the
487		fsqrt instruction is not implemented.
488
489		----------------------------------------------------------------------
490		Name: Matthias Blume
491		Date: 2001/04/18 12:45:00 EDT
492		Tag: blume-20010418-ccalls
493		Description:
494
495		- fixed two off-by-4 errors in the x86-specific c-calls implementation
496		(this bug prevented structure arguments containing pointers from being
497		passed correctly)
498		- changed the raw-C-call code in mlriscGen.sml in such a way that
499		structure arguments are represented as a pointer to the beginning
500		of the structure (instead of having a series of synthesized arguments,
501		one for each structure member)
502
503		- made makeml script's verbosity level configurable via environment
504		variable (MAKEML_VERBOSITY)
505
506		- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
507		in rawmem-x86.sml; we are now using the real thing
508
509		----------------------------------------------------------------------
510		Name: Matthias Blume
511		Date: 2001/03/22 16:25:00 EST
512		Tag: blume-20010322-bootfiles
513		Description:
514
515		Created a new set of bootfiles (for your automatic installation convenience).
516
517		----------------------------------------------------------------------
518		Name: Matthias Blume
519		Date: 2001/03/22 15:10:00 EST
520		Tag: blume-20010322-rawmem-parcm
521		Description:
522
523		1. All "raw memory access" primitives for the new FFI are implemented now
524		(at least on the x86).
525		2. Some further cleanup of CM's parallel make mechanism.
526
527		----------------------------------------------------------------------
528		Name: Matthias Blume
529		Date: 2001/03/19 17:53:00 EST
530		Tag: blume-20010319-parallel
531		Description:
532
533		Parallel make (using compile servers) now works again.
534
535		To this end, CM.stabilize and CMB.make have been modified to work in
536		two passes when compile servers are attached:
537		1. Compile everything, do not perform stabilization; this pass
538		uses compile servers
539		2. Stabilize everything; this pass does not use compile servers
540		If there are no compile servers, the two passes are combined into one
541		(as before).  Splitting the passes increases the inherent parallelism
542		in the dependency graph because the entire graph including all
543		libraries is available at the same time.  This, in turn, improves
544		server utilization.  The downside is that the master process will
545		have to do some extra work after compilation is done (because for
546		technical reasons it must re-read all the binfiles during stabilization).
547
548		----------------------------------------------------------------------
549		Name: Matthias Blume
550		Date: 2001/03/16 12:22:00 EST
551		Tag: blume-20010316-bootfiles
552		Description:
553
554		Created a new set of bootfiles (for your automatic installation convenience).
555
556		----------------------------------------------------------------------
557		Name: Matthias Blume
558		Date: 2001/03/16 11:00:00 EST
559		Tag: blume-20010316-MLTREE-fixup
560		Description:
561
562		This is a minor fixup for an (untagged) earlier commit by Allen.
563		(A file was missing).
564
565		----------------------------------------------------------------------
566		Name: Allen Leung
567		Date: Mon Mar  5 18:54:57 EST 2001
568		Tag: leunga-20010305-cut-support
569
570		1. New support for alternative control-flow in MLTREE.
571		Currently we support
572
573		FLOW_TO(CALL ...., [k1,...,kn])
574
575		This is needed for 'cuts to' in C-- and try/handle-like constructs
576		in Moby
577
578		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
579
580		2. Register Allocator
581
582		Changes in interface [from Fermin, John]
583
584		3. Alpha 8-bit SLL support [Fermin]
585
586		4. All architectures
587
588		A new module (ClusterExpandCopies) for expanding parallel copies.
589
590		----------------------------------------------------------------------
591		Name: Allen Leung
592		Date: 2001/02/27 23:07:00 EST
593		Tag: leunga-20010227-minor-stuff
594
595		1. Alpha bug fix for CMOVNE
596		2. Handle mltree COND(..,FCMP ...,...)
597		3. Bug fix in simplifier
598
599		----------------------------------------------------------------------
600		Name: Matthias Blume
601		Date: 2001/01/30 17:50:00 EST
602		Tag: blume-20010130-sync
603		Description:
604
605		This is just a minor update to sync my devel branch with the main brach.
606		The only visible change is the addition of some README files.
607
608		----------------------------------------------------------------------
609		Name: Matthias Blume
610		Date: 2001/01/12 23:30:00 JST
611		Tag: blume-20010112-bootfiles
612		Description:
613
614		Made a new set of bootfiles that goes with the current state of the
615		repository.
616
617		----------------------------------------------------------------------
618		Name: Matthias Blume
619		Date: 2001/01/12 21:20:00 JST
620		Tag: blume-20010112-sync
621		Description:
622
623		I am just flushing out some minor changes that had accumulated in
624		my private branch in order to sync with the main tree.  (This is
625		mainly because I had CVS trouble when trying to merge _into_ my
626		private branch.)
627
628		Most people should be completely unaffected by this.
629
630		----------------------------------------------------------------------
631		Name: Allen Leung
632		Date: Thu Jan 11 21:03:00 EST 2001
633		Tag: leunga-20010111-labexp=mltree
634		Description:
635
636		1.  Removed the type LabelExp and replace it by MLTree.
637		2.  Rewritten mltree-simplify with the pattern matcher tool.
638		3.  There were some bugs in alpha code generator which would break
639		64-bit code generation.
640		4.  Redo the tools to generate code with the
641		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
642		are now generated by perl scripts.
643
644		----------------------------------------------------------------------
645		Name: Matthias Blume
646		Date: 2001/01/10 21:55:00 JST
647		Tag: blume-20010110-rcc
648		Description:
649
650		The RCC stuff now seems to work (but only on the x86).
651		This required hacking of the c-calls interface (and -implementation) in
652		MLRISC.
653
654		Normal compiler users should be unaffected.
655
656		----------------------------------------------------------------------
657		Name: Matthias Blume
658		Date: 2001/01/09 01:20:00 JST
659		Tag: blume-20010109-rcc
660		Description:
661
662		This is a fairly big patch, flushing out a large number of pending
663		changes that I made to my development copy over the last couple of days.
664
665		Of practical relevance at this moment is a workaround for a pickling
666		bug that Allen ran into the other day.  The cause of the bug itself is
667		still unknown and it might be hard to fix it properly, but the
668		workaround has some merits of its own (namely somewhat reducing pickling
669		overhead for certain libraries).  Therefore, I think this solution should
670		be satisfactory at this time.
671
672		The rest of the changes (i.e., the vast majority) has to do with my
673		ongoing efforts of providing direct support for C function calls from
674		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
675		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
676		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
677		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
678		handle RCC.
679
680		The last part (the changes to mlriscGen.sml) are still known to be
681		wrong on the x86 and not implemented on all other architectures.  But
682		the infrastructure is in place. I had to change a few functor
683		signatures in the backend to be able to route the CCalls interface
684		from MLRISC there, and I had to specialize the mltree type (on the
685		x86) to include the necessary extensions. (The extensions themselves
686		were already there and redy to go in MLRISC/x86).
687
688		Everything should be very happy as soon as someone helps me with
689		mlriscGen.sml...
690
691		In any case, nothing of this should matter to anyone as long as the
692		new primop is not being used (which is going to be the case unless you
693		find it where I hid it :). The rest of the compiler is completely
694		unaffected.
695
696		----------------------------------------------------------------------
697		Name: Matthias Blume
698		Date: 2001/01/05 00:30:00 JST
699		Tag: blume-20010105-primops
700		Description:
701
702		Added some experimental support for work that I am doing right now.
703		These changes mostly concern added primops, but there is also a new
704		experimental C library in the runtime system (but currently not enabled
705		anywhere except on Linux/X86).
706
707		In the course of adding primops (and playing with them), I discovered that
708		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
709		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
710		To recover, I made type information in INL_PRIM mandatory and changed
711		prim.sml as well as built-in.sml accordingly.  The InLine structure now
712		has complete, correct type information (i.e., no bottom types).
713
714		Since all these changes mean that we need new binfiles, I also bumped the
715		version number to 110.32.1.
716
717		----------------------------------------------------------------------
718		Name: Matthias Blume
719		Date: 2000/12/30 22:10:00 JST
720		Tag: blume-20001230-various
721		Description:
722
723		Added proxy libraries for MLRISC and let MLRISC libraries refer
724		to each other using path anchors.  (See CM manual for explanation.)
725
726		Updated CM documentation.
727
728		Fixed some bugs in CM.
729
730		Implemented "proxy" libraries (= syntactic sugar for CM).
731
732		Added "-quiet" option to makeml and changed runtime system accordingly.
733
734		Added cleanup handler for exportML to reset timers and compiler stats.
735
736		----------------------------------------------------------------------
737		Name: Lal George
738		Date: 2000/12/22 22:22:58 EST 2000
739		Tag: Release_110_32
740		Description:
741
742		Infinite precision used throughout MLRISC.
743		see MLRISC/mltree/machine-int.sig
744
745		----------------------------------------------------------------------
746		Name: Matthias Blume
747		Date: 2000/12/22 23:16:00 JST
748		Tag: blume-20001222-warn
749		Description:
750
751		Corrected wording and formatting of some CM warning message which I
752		broke in my previous patch.
753
754		----------------------------------------------------------------------
755		Name: Matthias Blume
756		Date: 2000/12/22 21:20:00 JST
757		Tag: blume-20001222-anchorenv
758		Description:
759
760		Fixed CM's handling of anchor environments in connection with CMB.make.
761
762		----------------------------------------------------------------------
763		Name: Matthias Blume
764		Date: 2000/12/22 13:15:00 JST
765		Tag: blume-20001222-cleanup
766		Description:
767
768		Removed src/cm/ffi which does not (and did not) belong here.
769
770		----------------------------------------------------------------------
771		Name: Matthias Blume
772		Date: 2000/12/21 23:55:00 JST
773		Tag: blume-20001221-exn
774		Description:
775
776		Probably most important: CM no longer silently swallows all exceptions
777		in the compiler.
778		Plus: some other minor CM changes.  For example, CM now reports some
779		sizes for generated binfiles (code, data, envpickle, lambdapickle).
780
781	----------------------------------------------------------------------	----------------------------------------------------------------------
782	Name: Matthias Blume	Name: Matthias Blume
783	Date: 2000/12/15 00:01:05 JST	Date: 2000/12/15 00:01:05 JST

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.756   changed lines   Added in v.854

---

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