Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 844, Wed Jun 20 20:39:15 2001 UTC	revision 939, Tue Sep 25 19:25:48 2001 UTC
#	Line 14	Line 14
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16	Name: Matthias Blume	Name: Matthias Blume
17		Date: 2001/09/25 15:25:00 EDT
18		Tag: blume-20010925-exninfo
19		Description:
20
21		I provided a non-hook implementation of exnName (at the toplevel) and
22		made the "dummy" implementation of exnMessage (at the toplevel) more
23		useful: if nothing gets "hooked in", then at least you are going to
24		see the exception name and a message indicating why you don't see more.
25
26		[For the time being, programs that need exnMessage and want to use
27		ml-build should either use General.exnMessage (strongly recommended) or
28		refer to structure General at some other point so that CM sees a
29		static dependency.]
30
31		[Similar remarks go for "print" and "use":  If you want to use their
32		functionality in stand-alone programs generated by ml-build, then use
33		TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
34
35		----------------------------------------------------------------------
36		Name: Matthias Blume
37		Date: 2001/09/20 17:28:00 EDT
38		Tag: blume-20010920-slowfp
39		Description:
40
41		Allen says that x86-fast-fp is not safe yet, so I turned it off again...
42
43		----------------------------------------------------------------------
44		Name: Matthias Blume
45		Date: 2001/09/20 17:20:00 EDT
46		Tag: blume-20010920-canonicalpaths
47		Description:
48
49		0. Updated the BOOT file (something that I forgot to do earlier).
50
51		1. Small internal change to CM so that it avoids "/../" in filenames
52		as much as possible (but only where it is safe).
53
54		2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
55		that arguments that contain delimiters are passed through correctly.
56		This change also means that all "special" arguments of the form
57		@SMLxxx... must come first.
58
59		3. Changed install script to put relative anchor names for tool commands
60		into pathconfig.
61
62		----------------------------------------------------------------------
63		Name: Matthias Blume
64		Date: 2001/09/18 15:35:00 EDT
65		Tag: blume-20010918-readme11036
66		Description:
67
68		Added README files.
69
70		----------------------------------------------------------------------
71		Name: Matthias Blume
72		Date: 2001/09/18 11:45:00 EDT
73		Tag: Release_110_36 (retag)
74		Description:
75
76		Fixed mistake in config/preloads. Retagged as 110.36.
77
78		----------------------------------------------------------------------
79		Name: Matthias Blume
80		Date: 2001/09/18 09:40:00 EDT
81		Tag: Release_110_36_orig (tag changed)
82		Description:
83
84		New version (110.36).  New bootfiles.
85
86		----------------------------------------------------------------------
87		Name: Matthias Blume
88		Date: 2001/09/14 16:15:00 EDT
89		Tag: blume-20010914-x86fastfp
90		Description:
91
92		John committed some changes that Allen made, in particular a (hopefully)
93		correctly working version of the x86-fp module.
94
95		I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
96		flag to "true".  Everything seems to compile to a fixpoint ok, and
97		"mandelbrot" speeds up by about 15%.
98
99		----------------------------------------------------------------------
100		Name: Matthias Blume
101		Date: 2001/09/13 11:20:00 EDT
102		Tag: blume-20010913-minimal
103		Description:
104
105		1. Stefan Monnier's patch to fix a miscompilation problem that
106		was brought to light by John Reppy's work on Moby.
107
108		2. Implemented a minimal "structure Compiler" that contains just
109		"version" and "architecture".  The minimal version will be
110		available when the full version is not.  This is for backward-
111		compatibility with code that wants to test Compiler.version.
112
113		----------------------------------------------------------------------
114		Name: Matthias Blume
115		Date: 2001/08/28 14:03:00 EDT
116		Tag: blume-20010828-ml-lex
117		Description:
118
119		Fix for bug 1581, received from Neophytos Michael.
120
121		----------------------------------------------------------------------
122		Name: Matthias Blume
123		Date: 2001/08/27 11:20:00 EDT
124		Tag: blume-20010827-readme11035
125		Description:
126
127		Fleshed out the README file for 110.35.
128
129		----------------------------------------------------------------------
130		Name: Matthias Blume
131		Date: 2001/08/24 17:10:00 EDT
132		Tag: Release_110_35
133		Description:
134
135		New version number (110.35).  New bootfiles.
136
137		----------------------------------------------------------------------
138		Name: Lal George
139		Date: 2001/08/24 13:47:18 EDT 2001
140		Tag: george-20010824-MLRISC-graphs
141		Description:
142
143		removed clusters from MLRISC completely and replaced with graphs.
144
145		----------------------------------------------------------------------
146		Name: Matthias Blume
147		Date: 2001/08/23 17:50:00 EDT
148		Tag: blume-20010823-toplevel
149		Description:
150
151		- some reorganization of the code that implements various kinds of
152		environments in the compiler (static, dynamic, symbolic, combined)
153		- re-implemented the EnvRef module so that evalStream works properly
154		(if the stream contains references to "use", "CM.make", etc.)
155		- cleaned up evalloop.sml and interact.sml (but they need more cleaning)
156
157		----------------------------------------------------------------------
158		Name: Matthias Blume
159		Date: 2001/08/20 15:50 EDT
160		Tag: blume20010820-slipup
161		Description:
162
163		I forgot to commit a few files.  Here they are...
164
165		----------------------------------------------------------------------
166		Name: Matthias Blume
167		Date: 2001/08/20 15:35:00 EDT
168		Tag: blume-20010820-debugprof
169		Description:
170
171		!!!! NEW BOOTFILES !!!!
172
173		This is another round of reorganizing the compiler sources.  This
174		time the main goal was to factor out all the "instrumentation"
175		passes (for profiling and backtracing) into their own library.
176		The difficulty was to do it in such a way that it does not depend
177		on elaborate.cm but only on elabdata.cm.
178
179		Therefore there have been further changes to both elaborate.cm and
180		elabdata.cm -- more "generic" things have been moved from the former
181		to the latter.  As a result, I was forced to split the assignment
182		of numbers indicating "primtyc"s into two portions: SML-generic and
183		SML/NJ-specific.  Since it would have been awkward to maintain,
184		I bit the bullet and actually _changed_ the mapping between these
185		numbers and primtycs.  The bottom line of this is that you need
186		a new set of bin- and bootfiles.
187
188		I have built new bootfiles for all architectures, so doing a fresh
189		checkout and config/install.sh should be all you need.
190
191		The newly created library's name is
192
193		$smlnj/viscomp/debugprof.cm
194
195		and its sources live under
196
197		src/compiler/DebugProf
198
199		----------------------------------------------------------------------
200		Name: Matthias Blume
201		Date: 2001/08/15 17:15:00 EDT
202		Tag: blume-20010815-compreorg
203		Description:
204
205		This is a first cut at reorganizing the CM libraries that make up the
206		core of the compiler.  The idea is to separate out pieces that could
207		be used independently by tools, e.g., the parser, the typechecker, etc.
208
209		The current status is a step in this direction, but it is not quite
210		satisfactory yet.  Expect more changes in the future.
211
212		Here is the current (new) organization...
213
214		What used to be $smlnj/viscomp/core.cm is now divided into
215		six CM libraries:
216
217		$smlnj/viscomp/basics.cm
218		/parser.cm
219		/elabdata.cm
220		/elaborate.cm
221		/execute.cm
222		/core.cm
223
224		The CM files for these libraries live under src/system/smlnj/viscomp.
225		All these libraries are proxy libraries that contain precisely
226		one CM library component.  Here are the locations of the components
227		(all within the src/compiler tree):
228
229		Basics/basics.cm
230		Parse/parser.cm
231		ElabData/elabdata.cm
232		Elaborator/elaborate.cm
233		Execution/execute.cm
234		core.cm
235
236		[This organization is the same that has been used already
237		for a while for the architecture-specific parts of the visible
238		compiler and for the old version of core.cm.]
239
240		As you will notice, many source files have been moved from their
241		respective original locations to a new home in one of the above
242		subtrees.
243
244		The division of labor between the new libraries is the following:
245
246		basics.cm:
247		- Simple, basic definitions that pertain to many (or all) of
248		the other libraries.
249		parser.cm:
250		- The SML parser, producing output of type Ast.dec.
251		- The type family for Ast is also defined and exported here.
252		elabdata.cm:
253		- The datatypes that describe input and output of the elaborator.
254		This includes types, absyn, and static environments.
255		elaborator.cm:
256		- The SML/NJ type checker and elaborator.
257		This maps an Ast.dec (with a given static environment) to
258		an Absyn.dec (with a new static environment).
259		- This libraries implements certain modules that used to be
260		structures as functors (to remove dependencies on FLINT).
261		execute.cm:
262		- Everything having to do with executing binary code objects.
263		- Dynamic environments.
264		core.cm:
265		- SML/NJ-specific instantiations of the elaborator and MLRISC.
266		- Top-level modules.
267		- FLINT (this should eventually become its own library)
268
269		Notes:
270
271		I am not 100% happy with the way I separated the elaborator (and its
272		data structures) from FLINT.  Two instances of the same problem:
273
274		1. Data structures contain certain fields that carry FLINT-specific
275		information.  I hacked around this using exn and the property list
276		module from smlnj-lib.  But the fact that there are middle-end
277		specific fields around at all is a bit annoying.
278
279		2. The elaborator calculates certain FLINT-related information.  I tried
280		to make this as abstract as I could using functorization, but, again,
281		the fact that the elaborator has to perform calculations on behalf
282		of the middle-end at all is not nice.
283
284		3. Having to used exn and property lists is unfortunate because it
285		weakens type checking.  The other alternative (parameterizing
286		nearly *everything*) is not appealing, though.
287
288		I removed the "rebinding =" warning hack because due to the new organization
289		it was awkward to maintain it.  As a result, the compiler now issues some of
290		these warnings when compiling init.cmi during bootstrap compilation. On
291		the plus side, you also get a warning when you do, for example:
292		val op = = Int32.+
293		which was not the case up to now.
294
295		I placed "assign" and "deref" into the _Core structure so that the
296		code that deals with the "lazy" keyword can find them there.  This
297		removes the need for having access to the primitive environment
298		during elaboration.
299
300		----------------------------------------------------------------------
301		Name: Matthias Blume
302		Date: 2001/08/13
303		Tag: blume-20010813-closures
304		Description:
305
306		This fix was sent to us by Zhong Shao.  It is supposed to improve the
307		performance of certain loops by avoiding needless closure allocation.
308
309		----------------------------------------------------------------------
310		Name: Lal George
311		Date: 2001/07/31 10:03:23 EDT 2001
312		Tag: george-20010731-x86-fmalloc
313		Description: Fixed bug in x86 calls
314
315		There was a bug where call instructions would mysteriously
316		vanish. The call instruction had to be one that returned
317		a floating point value.
318
319		----------------------------------------------------------------------
320		Name: Lal George
321		Date: 2001/07/19 16:36:29 EDT 2001
322		Tag: george-20010719-simple-cells
323		Description:
324
325		I have dramatically simplified the interface for CELLS in MLRISC.
326
327		In summary, the cells interface is broken up into three parts:
328
329		1. CellsBasis : CELLS_BASIS
330
331		CellsBasis is a top level structure and common for all
332		architectures.  it contains the definitions of basic datatypes
333		and utility  functions over these types.
334
335		2. functor Cells() : CELLS
336
337		Cells generates an interface for CELLS that incorporates the
338		specific resources on the target architecture, such as the
339		presence of special register classes, their number and size,
340		and various useful substructures.
341
342		3. <ARCH>CELLS
343
344		e.g. SparcCells: SPARCCELLS
345
346		<ARCH>CELLS usually contains additional bindings for special
347		registers  on the architecture, such as:
348
349		val r0 : cell           (* register zero *)
350		val y : cell            (* Y register *)
351		val psr : cell          (* processor status register *)
352		...
353
354		The structure returned by applying the Cells functor is opened
355		in this interface.
356
357		The main implication of all this is that the datatypes for cells is
358		split between CellsBasis and CELLS -- a fairly simple change for user
359		code.
360
361		In the old scheme the CELLS interface had a definitional binding of
362		the form:
363
364		signature CELLS = sig
365
366		structure CellsBasis = CellsBasis
367
368		...
369
370		end
371
372		With all the sharing constraints that goes on in MLRISC, this old
373		design  quickly leads to errors such as:
374
375		"structure definition spec inside of sharing ... "
376
377
378		and appears to require an unacceptable amount of sharing and where
379		constraint hackery.
380
381		I think this error message (the interaction of definitional specs and
382		sharing) requires more explanation on our web page.
383
384		----------------------------------------------------------------------
385		Name: Matthias Blume
386		Date: 2001/07/19 15:00:00 EDT
387		Tag: blume-20010719-libreorg
388		Description:
389
390		This update puts together a fairly extensive but straightforward change
391		to the way the libraries that implement the interactive system are
392		organized:
393
394		The biggest change is the elimination of structure Compiler.  As a
395		replacement for this structure, there is now a CM library
396		(known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
397		that exports all the substructures of the original structure Compiler
398		directly.  So instead of saying Compiler.Foo.bar one now simply
399		says Foo.bar.  (The CM libraries actually export a collection of
400		structures that is richer than the collection of substructures of
401		structure Compiler.)
402
403		To make the transition smooth, there is a separate library called
404		$smlnj/compiler/compiler.cm which puts together and exports the
405		original structure Compiler (or at least something very close to it).
406
407		There are five members of the original structure Compiler
408		that are not exported directly but which instead became members
409		of a new structure Backend (described by signature BACKEND).  These are:
410		structure Profile (: PROFILE), structure Compile (: COMPILE), structure
411		Interact (: INTERACT), structure Machine (: MACHINE), and val
412		architecture (: string).
413
414		Structure Compiler.Version has become structure CompilerVersion.
415
416		Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
417		by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
418		or x86, respectively.
419		Each of these exports the same frontend structures that
420		$smlnj/compiler.cm exports.  But they do not have a structure Backend
421		and instead export some structure <Arch>Backend where <Arch> is Alpha32,
422		Hppa, PPC, Sparc, or X86, respectively.
423
424		Library $smlnj/compiler/all.cm exports the union of the exports of
425		$smlnj/compiler/<arch>.cm
426
427		There are no structures <Arch>Compiler anymore, use
428		$smlnj/compiler/<arch>.cm instead.
429
430		Library host-compiler-0.cm is gone.  Instead, the internal library
431		that instantiates CM is now called cm0.cm.  Selection of the host
432		compiler (backend) is no longer done here but. (Responsibility for it
433		now lies with $smlnj/compiler/current.cm.  This seems to be more
434		logical.)
435
436		Many individual files have been moved or renamed.  Some files have
437		been split into multiple files, and some "dead" files have been deleted.
438
439		Aside from these changes to library organization, there are also changes
440		to the way the code itself is organized:
441
442		Structure Binfile has been re-implemented in such a way that it no
443		longer needs any knowledge of the compiler.  It exclusively deals
444		with the details of binfile layout.  It no longer invokes the
445		compiler (for the purpose of creating new prospective binfile
446		content), and it no longer has any knowledge of how to interpret
447		pickles.
448
449		Structure Compile (: COMPILE) has been stripped down to the bare
450		essentials of compilation.  It no longer deals with linking/execution.
451		The interface has been cleaned up considerably.
452
453		Utility routines for dealing with linking and execution have been
454		moved into their own substructures.
455
456		(The ultimate goal of these changes is to provide a light-weight
457		binfile loader/linker (at least for, e.g., stable libraries) that
458		does not require CM or the compiler to be present.)
459
460		CM documentation has been updated to reflect the changes to library
461		organization.
462
463		----------------------------------------------------------------------
464		Name: Matthias Blume
465		Date: 2001/07/10 17:30:00 EDT
466		Tag: Release_110_34
467		Description:
468
469		Minor tweak to 110.34 (re-tagged):
470
471		- README.html file added to CVS repository
472		- runtime compiles properly under FreeBSD 3.X and 4.X
473
474		----------------------------------------------------------------------
475		Name: Matthias Blume
476		Date: 2001/07/10 17:30:00 EDT
477		Tag: Release_110_34
478		Description:
479
480		New version number (110.34). New bootfiles.
481
482		----------------------------------------------------------------------
483		Name: Matthias Blume
484		Date: 2001/07/09 16:00:00 EDT
485		Tag: blume-20010709-more-varargs
486		Description:
487
488		I changed the handling of varargs in ml-nlffigen again:
489		The ellipsis ... will now simply be ignored (with an accompanying warning).
490
491		The immediate effect is that you can actually call a varargs function
492		from ML -- but you can't actually supply any arguments beyond the ones
493		specified explicitly.  (For example, you can call printf with its format
494		string, but you cannot pass additional arguments.)
495
496		This behavior is only marginally more useful than the one before, but
497		it has the advantage that a function or, more importantly, a function
498		type never gets dropped on the floor, thus avoiding follow-up problems with
499		other types that refer to the offending one.
500
501		----------------------------------------------------------------------
502		Name: Matthias Blume
503		Date: 2001/07/09 11:25:00 EDT
504		Tag: blume-20010709-varargs
505		Description:
506
507		1. ckit-lib.cm now exports structure Error
508		2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
509		with a warning accompanied by a source location.  Moreover, it
510		merely skips the offending function or type and proceeds with the
511		rest of its work.u  As a result, one can safely feed C code containing
512		"..." to ml-nlffigen.
513		3. There are some internal improvements to CM, providing slightly
514		more general string substitutions in the tools subsystem.
515
516		----------------------------------------------------------------------
517		Name: Matthias Blume
518		Date: 2001/06/27 15:10:00 EDT
519		Tag: blume-20010627-concur
520		Description:
521
522		Fixed a small bug in CM's handling of parallel compilation.
523		(You could observe the bug by Control-C-interrupting an ordinary
524		CMB.make or CM.stabilize and then attaching some compile servers.
525		The result was that all of a sudden the previously interrupted
526		compilation would continue on its own.  This was because of
527		an over-optimization: CM did not bother to clean out certain queues
528		when no servers were attached "anyway", resulting in the contents
529		of these queues to grab control when new servers did get attached.)
530
531		There is also another minor update to the CM manual.
532
533		----------------------------------------------------------------------
534		Name: Matthias Blume
535		Date: 2001/06/26 16:15:00 EDT
536		Tag: blume-20010626-cmdoc
537		Description:
538
539		Minor typo fixed in CM manual (syntax diagram for libraries).
540
541		----------------------------------------------------------------------
542		Name: Matthias Blume
543		Date: 2001/06/25 22:55:00 EDT
544		Tag: blume-20010625-x86pc
545		Description:
546
547		Fixed a nasty bug in the X86 assembly code that caused signal
548		handlers to fail (crash) randomly.
549
550		----------------------------------------------------------------------
551		Name: Matthias Blume
552		Date: 2001/06/25 12:05:00 EDT
553		Tag: blume-20010625-nlffigen
554		Description:
555
556		This update fixes a number of minor bugs in ml-nlffigen as reported by
557		Nick Carter <nbc@andrew.cmu.edu>.
558
559		1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
560		2. Default names for generated files are now derived from the name of
561		the C file *without its directory*.  In particular, this causes generated
562		files to be placed locally even if the C file is in some system directory.
563		3. Default names for generated signatures and structures are also derived
564		from the C file name without its directory.  This avoids silly things
565		like "structure GL/GL".
566		(Other silly names are still possible because ml-nlffigen does not do
567		a thorough check of whether generated names are legal ML identifiers.
568		When in doubt, use command line arguments to force particular names.)
569
570		----------------------------------------------------------------------
571		Name: Matthias Blume
572		Date: 2001/06/21 12:25:00 EDT
573		Tag: blume-20010621-eXene
574		Description:
575
576		eXene now compiles and (sort of) works again.
577
578		The library name (for version > 110.33) is $/eXene.cm.
579
580		I also added an new example in src/eXene/examples/nbody.  See the
581		README file there for details.
582
583		----------------------------------------------------------------------
584		Name: Matthias Blume
585	Date: 2001/06/20 16:40:00 EDT	Date: 2001/06/20 16:40:00 EDT
586	Tag: blume-20010620-cml	Tag: blume-20010620-cml
587	Description:	Description:
#	Line 2469	Line 3037
3037	elaborator). There were a lot of changes during my "linkpath" trials	elaborator). There were a lot of changes during my "linkpath" trials
3038	that could have been reverted to their original state but weren't.	that could have been reverted to their original state but weren't.
3039	Please, don't be too harsh on me for messing with this code a bit more	Please, don't be too harsh on me for messing with this code a bit more
3040	than what was strictly necessary...  (I _did_ resist the tempation	than what was strictly necessary...  (I _did_ resist the temptation
3041	of doing any "global reformatting" to avoid an untimely death at	of doing any "global reformatting" to avoid an untimely death at
3042	Dave's hands. :)	Dave's hands. :)
3043

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.844   changed lines   Added in v.939

---

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