Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 670, Sun Jun 18 13:10:57 2000 UTC	revision 715, Thu Nov 2 07:51:04 2000 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: 2000/11/02 17:00:00 JST
18		Tag: blume-20001102-condcomp
19		Description:
20
21		- Small tweaks to pickler -- new BOOTFILES!
22		- Version bumped to 110.29.2.
23		- Added conditional compilation facility to init.cmi (see comment there).
24		----------------------------------------------------------------------
25		Name: Allen Leung
26		Date: 2000/10/23 19:31:00
27		Tag: leunga-20001023-demo-ra
28
29		1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
30		2. Test programs for the graph library updated
31		3. Some new MLRISC demo programs added
32
33		----------------------------------------------------------------------
34		Name: Matthias Blume
35		Date: 2000/08/31 22:15:00 JST
36		Tag: blume-20001017-errmsg
37		Description:
38
39		More error message grief: Where there used to be no messages, there
40		now were some that had bogus error regions.  Fixed.
41
42		----------------------------------------------------------------------
43		Name: Matthias Blume
44		Date: 2000/08/31 17:30:00 JST
45		Tag: blume-20001017-v110p29p1
46		Description:
47
48		I made a version 110.29.1 with new bootfiles.
49
50		Changes:  Modified pickler/unpickler for faster and leaner unpickling.
51		CM documentation changes and a small bugfix in CM's error reporting.
52
53		----------------------------------------------------------------------
54		Name: Lal George
55		Date: 2000/09/27 14:42:35 EDT
56		Tag: george-20000927-nodestatus
57		Description:
58
59		Changed the type of the nodestatus, so that:
60
61		SPILLED(~1)             is now SPILLED
62		SPILLED(m) where m>=0   is now MEMREG(m)
63		SPILLED(s) where s<~1   is now SPILL_LOC(~s)
64
65		----------------------------------------------------------------------
66		Name: Matthias Blume
67		Date: 2000/09/07 14:45:00 JST
68		Tag: blume-20000907-cmerrmsg
69		Description:
70
71		Small tweak to CM to avoid getting ML syntax error messages twice.
72
73		----------------------------------------------------------------------
74		Name: Matthias Blume
75		Date: 2000/08/31 18:00:00 JST
76		Tag: blume-20000831-cvsbootfiles
77		Description:
78
79		New URL for boot files (because the 110.29 files on the BL server do
80		now work correctly with my updated install scripts for yacc and lex).
81
82		----------------------------------------------------------------------
83		Name: Matthias Blume
84		Date: 2000/08/08 12:33:00 JST
85		Tag: blume-20000808-manual
86		Description:
87
88		Tiny update to CM manual.
89
90		----------------------------------------------------------------------
91		Name: Allen Leung
92		Date: 2000/08/7 19:31:00
93		Tag: leunga-20000807-a-whole-bunch-of-stuff
94
95		Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
96		and Mobdy.
97
98		1.  x86
99
100		a.  Fixes to peephole module by John and Dan.
101		b.  Assembly fix to SETcc by Allen.
102		c.  Fix to c-call by John.
103		d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
104		e.  Instruction selection optimization to SETcc as suggested by John.
105
106		For example,
107
108		MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
109
110		should generate:
111
112		MOVL a, x
113		SUBL b, x
114		SHRL 31, x
115
116		2.  IR stuff
117
118		A bunch of new DJ-graph related algorithms added.  These
119		speed up SSA construction.
120
121		3.  SSA + Scheduling
122
123		Added code for SSA and scheduling to the repository
124
125		----------------------------------------------------------------------
126		Name: Lal George
127		Date: 2000/07/27 11:53:14 EDT
128
129		Tag: lal-20000727-linux-ppc
130		Description:
131
132		Made changes to support Linux PPC.
133		p.s. I have confirmation that the 110.29 boot files work fine.
134
135		----------------------------------------------------------------------
136		Name: Matthias Blume
137		Date: 2000/07/27 17:40:00 JST
138		Tag: blume-20000727-scripts
139		Description:
140
141		!!!! WARNING !!!!
142		You must recompile the runtime system!
143		!!!! WARNING !!!!
144
145		This is basically another round of script-enhancements:
146
147		1. sml, ml-build, and ml-makedepend accept options -D and -U to define
148		and undefine CM preprocessor symbols.
149
150		2. ml-build avoids generating a new heap image if it finds that the
151		existing one is still ok.  (The condition is that no ML file had to
152		be recompiled and all ML files are found to be older that the heap
153		file.)
154
155		To make this work smoothly, I also hacked the runtime system as
156		well as SMLofNJ.SysInfo to get access to the heap image suffix
157		(.sparc-solaris, ...) that is currently being used.
158
159		Moreover, the signature of CM.mk_standalone has changed.  See the
160		CM manual.
161
162		3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
163		CM manual for details.)
164
165		4. More CM manual updates:
166		- all of the above has been documented.
167		- there is now a section describing the (CM-related) command line
168		arguments that are accepted by the "sml" command
169
170		----------------------------------------------------------------------
171		Name: Matthias Blume
172		Date: 2000/07/25 16:20:00 JST
173		Tag: blume-20000725-makedepend
174		Description:
175
176		Added a script called ml-makedepend.  This can be used in makefiles
177		for Unix' make in a way very similar to the "makedepend" command for
178		C.
179
180		The script internally uses function CM.sources.
181
182		Synopsis:
183
184		ml-makedepend [-f makefile] cmfile targetname
185
186		The default for the makefile is "makefile" (or "Makefile" should
187		"makefile" not exist).
188
189		ml-makedepend adds a cmfile/targetname-specific section to this
190		makefile (after removing the previous version of this section).  The
191		section contains a single dependency specification with targetname on
192		the LHS (targetname is an arbitrary name), and a list of files derived
193		from the cmfile on the RHS.  Some of the files on the RHS are
194		ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
195		"make" variables $(ARCH) and $(OPSYS) in place of the corresponding
196		path names.  The makefile writer is responsible for making sure that
197		these variables have correct at the time "make" is invoked.
198
199		----------------------------------------------------------------------
200		Name: Matthias Blume
201		Date: 2000/07/22 23:30:00 JST
202		Tag: blume-20000722-urlupdate
203		Description:
204
205		Changed BOOT and config/srcarchiveurl to point to BL server:
206
207		ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
208
209		----------------------------------------------------------------------
210		Name: Matthias Blume
211		Date: 2000/07/18 18:00:00 JST
212		Tag: blume-20000718-Version_110_29
213		Description:
214
215		1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
216
217		2. Updated config/version to 110.29
218
219		3. Updated config/srcarchiveurl
220
221		3. New boot files!
222		ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
223
224		----------------------------------------------------------------------
225		Name: Matthias Blume
226		Date: 2000/07/11 13:58:00 JST
227		Tag: blume-20000711-doctypo
228		Description:
229
230		Fixed a few typos in CM manual.
231
232		----------------------------------------------------------------------
233		Name: Allen Leung
234		Date: 2000/06/15 00:38:00
235		Tag: leunga-20000704-sparc-x86
236
237		1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
238		2. fix to x86 RET bug [found by Dan Grossman]
239		3. sparc assembly bug fix for ticc instructions [found by Fermin]
240
241		Affects c-- and moby only
242
243		----------------------------------------------------------------------
244		Name: Matthias Blume
245		Date: 2000/07/04 15:26:00
246		Tag: blume-20000704-trigger
247		Description:
248
249		1. Improvements to CM manual.
250		2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
251		of getting a back-trace.  The function, when called, raises an
252		internal exception which explicitly carries the full back-trace history,
253		so it is unaffected by any intervening handle-raise pairs ("trivial"
254		or not).  The interactive loop will print that history once it arrives
255		at top level.
256		Short of having all exceptions implicitly carry the full history, the
257		recommended way of using this facility is:
258		- compile your program with instrumentation "on"
259		- run it, when it raises an exception, look at the history
260		- if the history is "cut off" because of some handler, go and modify
261		your program so that it explicitly calls BTrace.trigger
262		- recompile (still instrumented), and rerun; look at the full history
263
264		----------------------------------------------------------------------
265		Name: Matthias Blume
266		Date: 2000/07/03 15:36:00 JST
267		Tag: blume-20000702-manual
268		Description:
269
270		Small corrections and updates to CM manual.
271
272		----------------------------------------------------------------------
273		Name: Matthias Blume
274		Date: 2000/06/29 16:04:00 JST
275		Tag: blume-20000629-yacctool
276		Description:
277
278		Changes:
279
280		1. Class "mlyacc" now takes separate arguments to pass options to
281		generated .sml- and .sig-files independently.
282		2. Corresponding CM manual updates.
283		3. BTrace module now also reports call sites.  (However, for loop clusters
284		it only shows from where the cluster was entered.)  There are associated
285		modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
286
287		----------------------------------------------------------------------
288		Name: Matthias Blume
289		Date: 2000/06/27 16:51:00 JST
290		Tag: blume-20000627-noweb
291		Description:
292
293		Changes:
294
295		1. Implemented "subdir" and "witness" options for noweb tool.
296		This caused some slight internal changes in CM's tool implementation.
297		2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
298		some remaining issues from earlier path anchor changes.
299		3. Updated CM manual accordingly.
300
301		4. Changed implementation of back-tracing so that I now consider it
302		ready for prime-time.
303
304		In particular, you don't have to explicitly trigger the back-trace
305		anymore.  Instead, if you are running BTrace-instrumented code and
306		there is an uncaught exception (regardless of whether or not it was
307		raised in instrumented code), the top-level evalloop will print
308		the back-trace.
309
310		Features:
311
312		- Instrumented and uninstrumented code work together seemlessly.
313		(Of course, uninstrumented code is never mentioned in actual
314		back-traces.)
315
316		- Asymptotic time- and space-complexity of instrumented code is
317		equal to that of uninstrumented code.  (This means that
318		tail-recursion is preserved by the instrumentation phase.)
319
320		- Modules whose code has been instrumented in different sessions
321		work together without problem.
322
323		- There is no penalty whatsoever on uninstrumented code.
324
325		- There is no penalty on "raise" expressions, even in
326		instrumented code.
327
328		A potential bug (or perhaps it is a feature, too):
329
330		A back-trace reaches no further than the outermost instrumented
331		non-trivial "raise".  Here, a "trivial" raise is one that is the
332		sole RHS of a "handle" rule.  Thus, back-traces reach trough
333
334		<exp> handle e => raise e
335
336		and even
337
338		<exp> handle Foo => raise Bar
339
340		and, of course, through
341
342		<exp> handle Foo => ...
343
344		if the exception was not Foo.
345
346		Back-traces always reach right through any un-instrumented code
347		including any of its "handle" expressions, trivial or not.
348
349		To try this out, do the following:
350
351		- Erase all existing binfiles for your program.
352		(You may keep binfiles for those modules where you think you
353		definitely don't need back-tracing.)
354		- Turn on back-trace instrumentation:
355		SMLofNJ.Internals.BTrace.mode (SOME true);
356		- Recompile your program.  (I.e., run "CM.make" or "use".)
357		- You may now turn instrumentation off again (if you want):
358		SMLofNJ.Internals.BTrace.mode (SOME false);
359		- Run your program as usual.  If it raises an exception that
360		reaches the interactive toplevel, then a back-trace will
361		automatically be printed.  After that, the toplevel loop
362		will print the exception history as usual.
363
364		----------------------------------------------------------------------
365		Name: Matthias Blume
366		Date: 2000/06/26 09:56:46 JST
367		Tag: blume-20000626-setup
368		Description:
369
370		CM: - setup-parameter to "sml" added; this can be used to run arbitrary
371		ML code before and after compiling a file (e.g., to set compiler
372		flags)
373
374		Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
375		- associated changes to btrace.sml (BTrace instrumentation pass)
376		- cleaner implementation of btimp.sml (BTrace tracing and report
377		module)
378
379		CM manual: * new path encoding documented
380		* description of setup-parameter to "sml" added
381
382		The biggest user-visible change to back-tracing is that it is no
383		longer necessary to compile all traced modules within the same
384		session.  (This was a real limitation.)
385
386		----------------------------------------------------------------------
387		Name: Matthias Blume
388		Date: 2000/06/24 12:40:00 JST
389		Tag: blume-20000624-startup
390		Description:
391
392		Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
393		bit too often -- to put it mildly. :)
394
395		----------------------------------------------------------------------
396		Name: Matthias Blume
397		Date: 2000/06/23 18:20:00 JST
398		Tag: blume-20000623-btrace
399		Description:
400
401		This updates adds a backtrace facility to aid programmers in debugging
402		their programs.  This involves the following changes:
403
404		1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
405		keeping track of the current call stack.  When programs are compiled
406		in a special mode, the compiler will insert calls to these hooks
407		into the user program.
408		"Hook" means that it is possible for different implementations of
409		back-tracing to register themselves (at different times).
410
411		2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
412		as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
413		It can be turned on using this call:
414		SMLofNJ.Internals.BTrace.mode (SOME true);
415		Turning it off again:
416		SMLofNJ.Internals.BTrace.mode (SOME false);
417		Querying the current status:
418		SMLofNJ.Internals.BTrace.mode NONE;
419		Annotated programs are about twice as big as normal ones, and they
420		run a factor of 2 to 4 slower with a dummy back-trace plugin (one
421		where all hooks do nothing).  The slowdown with a plugin that is
422		actually useful (such as the one supplied by default) is even greater,
423		but in the case of the default plugin it is still only an constant
424		factor (amortized).
425
426		3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
427		with a sub-structure BTrace for controlling back-tracing.  In particular,
428		the above-mentioned function "mode" controls whether the annotation
429		phase is invoked by the compiler.  Another important function is
430		"trigger": when called it aborts the current execution and causes
431		the top-level loop to print a full back-trace.
432
433		4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
434		for back-tracing.  It keeps track of the dynamic call stack and in
435		addition to that it keeps a partial history at each "level" of that
436		stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
437		d and b (at separate times, dynamically), then the report will show:
438
439		GOTO   d
440		/c
441		GOTO  \b
442		CALL   a
443
444		This shows that there was an initial non-tail call of a, then a
445		tail-call to b or c, looping behavior in a cluster of functions that
446		consist of b and c, and then a goto from that cluster (i.e., either from
447		b or from c) to d.
448
449		Note that (depending on the user program) the amount of information
450		that the back-trace module has to keep track of at each level is bounded
451		by a constant.  Thus, the whole implementation has the same asymptotical
452		complexity as the original program (both in space and in time).
453
454		5. compiler/TopLevel/interact/evalloop.sml has been modified to
455		handle the special exception SMLofNJ.Internals.BTrace.BTrace
456		which is raised by the "trigger" function mentioned above.
457
458		Notes on usage:
459
460		- Annotated code works well together with unannotated code:
461		Unannotated calls simply do not show up at all in the backtrace.
462
463		- It is not a good idea to let modules that were annotated during
464		different sessions run at the same time.  This is because the compiler
465		chooses small integers to identify individual functions, and there
466		will be clashes if different modules were compiled in separate sessions.
467		(Nothing will crash, and you will even be told about the clashes, but
468		back-trace information will in general not be useful.)
469
470		- Back-tracing can be confused by callcc and capture.
471
472		- The only way of getting a back-trace right now is to explicitly
473		invoke the "trigger" function from your user program.  Eventually, we
474		should make every exception carry back-trace information (if
475		available).  But since this creates more overhead at "raise"-time
476		(similar to the current exnHistory overhead), I have not yet
477		implemented this.  (The implementation will be rather easy.)  With
478		exceptions carrying back-trace information, this facility will be even
479		more useful because users don't need to modify their programs...
480
481		- While it is possible to compile the compiler with back-trace
482		annotations turned on (I did it to get some confidence in
483		correctness), you must make absolutely sure that core.sml and
484		btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
485		be compiled with annotation because there is no core access yet, but
486		if you compile btimp.sml with annotation, then the system will go into
487		an infinite recursion and crash.)
488		Since CM currently does not know about BTrace, the only way to turn
489		annotations on and off for different modules of the compiler is to
490		interrupt CMB.make, change the settings, and re-invoke it.  Of course,
491		this is awkward and clumsy.
492
493		Sample sessions:
494
495		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
496		- SMLofNJ.Internals.BTrace.mode (SOME true);
497		[autoloading]
498		[autoloading done]
499		val it = false : bool
500		- structure X = struct
501		-     fun main n = let
502		-         fun a (x, 0) = d x
503		-           | a (x, n) = b (x, n - 1)
504		-         and b (x, n) = c (x, n)
505		-         and c (x, n) = a (x, n)
506		-         and d x = e (x, 3)
507		-         and e (x, 0) = f x
508		-           | e (x, n) = e (x, n - 1)
509		-         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
510		-           | f n = n * g (n - 1)
511		-         and g n = a (n, 3)
512		-     in
513		-         f n
514		-     end
515		- end;
516		structure X : sig val main : int -> int end
517		- X.main 3;
518		*** BACK-TRACE ***
519		GOTO   stdIn:4.2-13.20: X.main[2].f
520		GOTO-( stdIn:4.2-13.20: X.main[2].e
521		GOTO   stdIn:4.2-13.20: X.main[2].d
522		/ stdIn:4.2-13.20: X.main[2].a
523		| stdIn:4.2-13.20: X.main[2].b
524		GOTO-\ stdIn:4.2-13.20: X.main[2].c
525		CALL   stdIn:4.2-13.20: X.main[2].g
526		GOTO   stdIn:4.2-13.20: X.main[2].f
527		GOTO-( stdIn:4.2-13.20: X.main[2].e
528		GOTO   stdIn:4.2-13.20: X.main[2].d
529		/ stdIn:4.2-13.20: X.main[2].a
530		| stdIn:4.2-13.20: X.main[2].b
531		GOTO-\ stdIn:4.2-13.20: X.main[2].c
532		CALL   stdIn:4.2-13.20: X.main[2].g
533		GOTO   stdIn:4.2-13.20: X.main[2].f
534		GOTO-( stdIn:4.2-13.20: X.main[2].e
535		GOTO   stdIn:4.2-13.20: X.main[2].d
536		/ stdIn:4.2-13.20: X.main[2].a
537		| stdIn:4.2-13.20: X.main[2].b
538		GOTO-\ stdIn:4.2-13.20: X.main[2].c
539		CALL   stdIn:4.2-13.20: X.main[2].g
540		GOTO   stdIn:4.2-13.20: X.main[2].f
541		CALL   stdIn:2.15-17.4: X.main[2]
542		-
543
544		(Note that because of a FLINt bug the above code currently does not
545		compile without BTrace turned on.)
546
547		Here is another example, using my modified Tiger compiler:
548
549		Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
550		- SMLofNJ.Internals.BTrace.mode (SOME true);
551		[autoloading]
552		[autoloading done]
553		val it = false : bool
554		- CM.make "sources.cm";
555		[autoloading]
556		...
557		[autoloading done]
558		[scanning sources.cm]
559		[parsing (sources.cm):parse.sml]
560		[creating directory CM/SKEL ...]
561		[parsing (sources.cm):tiger.lex.sml]
562		...
563		[wrote CM/sparc-unix/semant.sml]
564		[compiling (sources.cm):main.sml]
565		[wrote CM/sparc-unix/main.sml]
566		[New bindings added.]
567		val it = true : bool
568		- Main.compile ("../testcases/merge.tig", "foo.out");
569		*** BACK-TRACE ***
570		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
571		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
572		CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
573		GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
574		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
575		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
576		CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
577		/ lib/semant.sml:411.65-543.8: SemantFun[2].transDec
578		CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
579		CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
580		CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
581		CALL   main.sml:1.18-118.4: Main.compile[2]
582		-
583
584		----------------------------------------------------------------------
585		Name: Matthias Blumen
586		Date: 2000/06/21 18:00:00 JST
587		Tag: blume-20000621-manual
588		Description:
589
590		CM manual update: Path environments documented.
591
592		----------------------------------------------------------------------
593		Name: Matthias Blume
594		Date: 2000/06/19 13:40:00
595		Tag: blume-20000619-manual
596		Description:
597
598		CM manual and system/README update.  This only covers the fact that
599		there are no more implicit anchors.  (Path environments and the "bind"
600		option to "cm" have yet to be documented.)
601
602		----------------------------------------------------------------------
603		Name: Matthias Blume
604		Date: 2000/06/19 11:05:00 JST
605		Tag: blume-20000619-chdir-bugfix
606		Description:
607
608		Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
609
610	----------------------------------------------------------------------	----------------------------------------------------------------------
611	Name: Matthias Blume	Name: Matthias Blume
612	Date: 2000/06/18 22:00:10 JST	Date: 2000/06/18 22:00:10 JST

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.670   changed lines   Added in v.715

---

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