Home My Page Projects Code Snippets Project Openings SML/NJ
Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

SCM Repository

[smlnj] Diff of /sml/trunk/HISTORY
ViewVC logotype

Diff of /sml/trunk/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 586, Thu Mar 30 05:08:07 2000 UTC revision 731, Fri Nov 10 22:57:45 2000 UTC
# Line 14  Line 14 
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16  Name: Allen Leung  Name: Allen Leung
17    Date: 2000/11/10 18:00:00
18    Tag: leunga-20001110-new-x86-fp
19    
20    A new x86 floating point code generator has been added.
21    By default this is turned off.  To turn this on, do:
22    
23        CM.autoload "$smlnj/compiler.cm";
24        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
25    
26    Changes:
27    
28    1.  Changed FTAN to FPTAN so that the assembly output is correct.
29    2.  Changed the extension callback for FTANGENT to generate:
30    
31              fptan
32              fstp  %st(0)
33        instead of
34              fptan
35              fstpl ftempmem
36    
37    3.  Numerous assembly fixes for x86.
38    
39    5.  Cleaned up the machine code output module x86/x86MC.sml and added
40        support for a whole bunch of instructions and addressing modes:
41    
42          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
43          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
44          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
45          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
46          fxch %st(n)
47          fld %st(n)
48          fst %st(n)
49          fst mem
50          fstp %st(n)
51          fucom %st(n)
52          fucomp %st(n)
53    
54        All these are now generated when the fast fp mode is turned on.
55    
56    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
57    
58    ----------------------------------------------------------------------
59    Name: Matthias Blume
60    Date: 2000/11/09 11:20:00 JST
61    Tag: blume-20001109-scc
62    Description:
63    
64    Eliminated some code duplication:
65    
66    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
67       (Otherwise the functor is useless.)
68    2. Used GraphSCCFn where SCCUtilFun was used previously.
69    3. Got rid of SCCUtilFun (in comp-lib.cm).
70    
71    ----------------------------------------------------------------------
72    Name: Lal George
73    Date: 2000/11/06 09:02:21 EST 2000
74    Tag: Release_110_30
75    Description:
76    
77    - Version 110.30
78    ----------------------------------------------------------------------
79    Name: Matthias Blume
80    Date: 2000/11/04 14:45:00
81    Tag: blume-20001104-mlbuild
82    Description:
83    
84    - Made ml-build faster on startup.
85    - Documentation fixes.
86    
87    ----------------------------------------------------------------------
88    Name: Matthias Blume
89    Date: 2000/11/02 17:00:00 JST
90    Tag: blume-20001102-condcomp
91    Description:
92    
93    - Small tweaks to pickler -- new BOOTFILES!
94    - Version bumped to 110.29.2.
95    - Added conditional compilation facility to init.cmi (see comment there).
96    ----------------------------------------------------------------------
97    Name: Allen Leung
98    Date: 2000/10/23 19:31:00
99    Tag: leunga-20001023-demo-ra
100    
101    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
102    2. Test programs for the graph library updated
103    3. Some new MLRISC demo programs added
104    
105    ----------------------------------------------------------------------
106    Name: Matthias Blume
107    Date: 2000/08/31 22:15:00 JST
108    Tag: blume-20001017-errmsg
109    Description:
110    
111    More error message grief: Where there used to be no messages, there
112    now were some that had bogus error regions.  Fixed.
113    
114    ----------------------------------------------------------------------
115    Name: Matthias Blume
116    Date: 2000/08/31 17:30:00 JST
117    Tag: blume-20001017-v110p29p1
118    Description:
119    
120    I made a version 110.29.1 with new bootfiles.
121    
122    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
123              CM documentation changes and a small bugfix in CM's error reporting.
124    
125    ----------------------------------------------------------------------
126    Name: Lal George
127    Date: 2000/09/27 14:42:35 EDT
128    Tag: george-20000927-nodestatus
129    Description:
130    
131    Changed the type of the nodestatus, so that:
132    
133            SPILLED(~1)             is now SPILLED
134            SPILLED(m) where m>=0   is now MEMREG(m)
135            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
136    
137    ----------------------------------------------------------------------
138    Name: Matthias Blume
139    Date: 2000/09/07 14:45:00 JST
140    Tag: blume-20000907-cmerrmsg
141    Description:
142    
143    Small tweak to CM to avoid getting ML syntax error messages twice.
144    
145    ----------------------------------------------------------------------
146    Name: Matthias Blume
147    Date: 2000/08/31 18:00:00 JST
148    Tag: blume-20000831-cvsbootfiles
149    Description:
150    
151    New URL for boot files (because the 110.29 files on the BL server do
152    now work correctly with my updated install scripts for yacc and lex).
153    
154    ----------------------------------------------------------------------
155    Name: Matthias Blume
156    Date: 2000/08/08 12:33:00 JST
157    Tag: blume-20000808-manual
158    Description:
159    
160    Tiny update to CM manual.
161    
162    ----------------------------------------------------------------------
163    Name: Allen Leung
164    Date: 2000/08/7 19:31:00
165    Tag: leunga-20000807-a-whole-bunch-of-stuff
166    
167      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
168    and Mobdy.
169    
170    1.  x86
171    
172       a.  Fixes to peephole module by John and Dan.
173       b.  Assembly fix to SETcc by Allen.
174       c.  Fix to c-call by John.
175       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
176       e.  Instruction selection optimization to SETcc as suggested by John.
177    
178           For example,
179    
180            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
181    
182           should generate:
183    
184            MOVL a, x
185            SUBL b, x
186            SHRL 31, x
187    
188    2.  IR stuff
189    
190         A bunch of new DJ-graph related algorithms added.  These
191         speed up SSA construction.
192    
193    3.  SSA + Scheduling
194    
195         Added code for SSA and scheduling to the repository
196    
197    ----------------------------------------------------------------------
198    Name: Lal George
199    Date: 2000/07/27 11:53:14 EDT
200    
201    Tag: lal-20000727-linux-ppc
202    Description:
203    
204     Made changes to support Linux PPC.
205     p.s. I have confirmation that the 110.29 boot files work fine.
206    
207    ----------------------------------------------------------------------
208    Name: Matthias Blume
209    Date: 2000/07/27 17:40:00 JST
210    Tag: blume-20000727-scripts
211    Description:
212    
213    !!!! WARNING !!!!
214    You must recompile the runtime system!
215    !!!! WARNING !!!!
216    
217    This is basically another round of script-enhancements:
218    
219    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
220       and undefine CM preprocessor symbols.
221    
222    2. ml-build avoids generating a new heap image if it finds that the
223       existing one is still ok.  (The condition is that no ML file had to
224       be recompiled and all ML files are found to be older that the heap
225       file.)
226    
227       To make this work smoothly, I also hacked the runtime system as
228       well as SMLofNJ.SysInfo to get access to the heap image suffix
229       (.sparc-solaris, ...) that is currently being used.
230    
231       Moreover, the signature of CM.mk_standalone has changed.  See the
232       CM manual.
233    
234    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
235       CM manual for details.)
236    
237    4. More CM manual updates:
238        - all of the above has been documented.
239        - there is now a section describing the (CM-related) command line
240          arguments that are accepted by the "sml" command
241    
242    ----------------------------------------------------------------------
243    Name: Matthias Blume
244    Date: 2000/07/25 16:20:00 JST
245    Tag: blume-20000725-makedepend
246    Description:
247    
248    Added a script called ml-makedepend.  This can be used in makefiles
249    for Unix' make in a way very similar to the "makedepend" command for
250    C.
251    
252    The script internally uses function CM.sources.
253    
254    Synopsis:
255    
256        ml-makedepend [-f makefile] cmfile targetname
257    
258    The default for the makefile is "makefile" (or "Makefile" should
259    "makefile" not exist).
260    
261    ml-makedepend adds a cmfile/targetname-specific section to this
262    makefile (after removing the previous version of this section).  The
263    section contains a single dependency specification with targetname on
264    the LHS (targetname is an arbitrary name), and a list of files derived
265    from the cmfile on the RHS.  Some of the files on the RHS are
266    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
267    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
268    path names.  The makefile writer is responsible for making sure that
269    these variables have correct at the time "make" is invoked.
270    
271    ----------------------------------------------------------------------
272    Name: Matthias Blume
273    Date: 2000/07/22 23:30:00 JST
274    Tag: blume-20000722-urlupdate
275    Description:
276    
277    Changed BOOT and config/srcarchiveurl to point to BL server:
278    
279        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
280    
281    ----------------------------------------------------------------------
282    Name: Matthias Blume
283    Date: 2000/07/18 18:00:00 JST
284    Tag: blume-20000718-Version_110_29
285    Description:
286    
287    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
288    
289    2. Updated config/version to 110.29
290    
291    3. Updated config/srcarchiveurl
292    
293    3. New boot files!
294       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
295    
296    ----------------------------------------------------------------------
297    Name: Matthias Blume
298    Date: 2000/07/11 13:58:00 JST
299    Tag: blume-20000711-doctypo
300    Description:
301    
302    Fixed a few typos in CM manual.
303    
304    ----------------------------------------------------------------------
305    Name: Allen Leung
306    Date: 2000/06/15 00:38:00
307    Tag: leunga-20000704-sparc-x86
308    
309    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
310    2. fix to x86 RET bug [found by Dan Grossman]
311    3. sparc assembly bug fix for ticc instructions [found by Fermin]
312    
313       Affects c-- and moby only
314    
315    ----------------------------------------------------------------------
316    Name: Matthias Blume
317    Date: 2000/07/04 15:26:00
318    Tag: blume-20000704-trigger
319    Description:
320    
321    1. Improvements to CM manual.
322    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
323       of getting a back-trace.  The function, when called, raises an
324       internal exception which explicitly carries the full back-trace history,
325       so it is unaffected by any intervening handle-raise pairs ("trivial"
326       or not).  The interactive loop will print that history once it arrives
327       at top level.
328       Short of having all exceptions implicitly carry the full history, the
329       recommended way of using this facility is:
330         - compile your program with instrumentation "on"
331         - run it, when it raises an exception, look at the history
332         - if the history is "cut off" because of some handler, go and modify
333           your program so that it explicitly calls BTrace.trigger
334         - recompile (still instrumented), and rerun; look at the full history
335    
336    ----------------------------------------------------------------------
337    Name: Matthias Blume
338    Date: 2000/07/03 15:36:00 JST
339    Tag: blume-20000702-manual
340    Description:
341    
342    Small corrections and updates to CM manual.
343    
344    ----------------------------------------------------------------------
345    Name: Matthias Blume
346    Date: 2000/06/29 16:04:00 JST
347    Tag: blume-20000629-yacctool
348    Description:
349    
350    Changes:
351    
352    1. Class "mlyacc" now takes separate arguments to pass options to
353       generated .sml- and .sig-files independently.
354    2. Corresponding CM manual updates.
355    3. BTrace module now also reports call sites.  (However, for loop clusters
356       it only shows from where the cluster was entered.)  There are associated
357       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
358    
359    ----------------------------------------------------------------------
360    Name: Matthias Blume
361    Date: 2000/06/27 16:51:00 JST
362    Tag: blume-20000627-noweb
363    Description:
364    
365    Changes:
366    
367     1. Implemented "subdir" and "witness" options for noweb tool.
368        This caused some slight internal changes in CM's tool implementation.
369     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
370        some remaining issues from earlier path anchor changes.
371     3. Updated CM manual accordingly.
372    
373     4. Changed implementation of back-tracing so that I now consider it
374        ready for prime-time.
375    
376        In particular, you don't have to explicitly trigger the back-trace
377        anymore.  Instead, if you are running BTrace-instrumented code and
378        there is an uncaught exception (regardless of whether or not it was
379        raised in instrumented code), the top-level evalloop will print
380        the back-trace.
381    
382        Features:
383    
384          - Instrumented and uninstrumented code work together seemlessly.
385            (Of course, uninstrumented code is never mentioned in actual
386             back-traces.)
387    
388          - Asymptotic time- and space-complexity of instrumented code is
389            equal to that of uninstrumented code.  (This means that
390            tail-recursion is preserved by the instrumentation phase.)
391    
392          - Modules whose code has been instrumented in different sessions
393            work together without problem.
394    
395          - There is no penalty whatsoever on uninstrumented code.
396    
397          - There is no penalty on "raise" expressions, even in
398            instrumented code.
399    
400        A potential bug (or perhaps it is a feature, too):
401    
402          A back-trace reaches no further than the outermost instrumented
403          non-trivial "raise".  Here, a "trivial" raise is one that is the
404          sole RHS of a "handle" rule.  Thus, back-traces reach trough
405    
406               <exp> handle e => raise e
407    
408          and even
409    
410               <exp> handle Foo => raise Bar
411    
412          and, of course, through
413    
414               <exp> handle Foo => ...
415    
416         if the exception was not Foo.
417    
418         Back-traces always reach right through any un-instrumented code
419         including any of its "handle" expressions, trivial or not.
420    
421       To try this out, do the following:
422    
423         - Erase all existing binfiles for your program.
424           (You may keep binfiles for those modules where you think you
425            definitely don't need back-tracing.)
426         - Turn on back-trace instrumentation:
427              SMLofNJ.Internals.BTrace.mode (SOME true);
428         - Recompile your program.  (I.e., run "CM.make" or "use".)
429         - You may now turn instrumentation off again (if you want):
430              SMLofNJ.Internals.BTrace.mode (SOME false);
431         - Run your program as usual.  If it raises an exception that
432           reaches the interactive toplevel, then a back-trace will
433           automatically be printed.  After that, the toplevel loop
434           will print the exception history as usual.
435    
436    ----------------------------------------------------------------------
437    Name: Matthias Blume
438    Date: 2000/06/26 09:56:46 JST
439    Tag: blume-20000626-setup
440    Description:
441    
442    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
443          ML code before and after compiling a file (e.g., to set compiler
444          flags)
445    
446    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
447              - associated changes to btrace.sml (BTrace instrumentation pass)
448              - cleaner implementation of btimp.sml (BTrace tracing and report
449                module)
450    
451    CM manual: * new path encoding documented
452               * description of setup-parameter to "sml" added
453    
454    The biggest user-visible change to back-tracing is that it is no
455    longer necessary to compile all traced modules within the same
456    session.  (This was a real limitation.)
457    
458    ----------------------------------------------------------------------
459    Name: Matthias Blume
460    Date: 2000/06/24 12:40:00 JST
461    Tag: blume-20000624-startup
462    Description:
463    
464    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
465    bit too often -- to put it mildly. :)
466    
467    ----------------------------------------------------------------------
468    Name: Matthias Blume
469    Date: 2000/06/23 18:20:00 JST
470    Tag: blume-20000623-btrace
471    Description:
472    
473    This updates adds a backtrace facility to aid programmers in debugging
474    their programs.  This involves the following changes:
475    
476    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
477       keeping track of the current call stack.  When programs are compiled
478       in a special mode, the compiler will insert calls to these hooks
479       into the user program.
480       "Hook" means that it is possible for different implementations of
481       back-tracing to register themselves (at different times).
482    
483    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
484       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
485       It can be turned on using this call:
486         SMLofNJ.Internals.BTrace.mode (SOME true);
487       Turning it off again:
488         SMLofNJ.Internals.BTrace.mode (SOME false);
489       Querying the current status:
490         SMLofNJ.Internals.BTrace.mode NONE;
491       Annotated programs are about twice as big as normal ones, and they
492       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
493       where all hooks do nothing).  The slowdown with a plugin that is
494       actually useful (such as the one supplied by default) is even greater,
495       but in the case of the default plugin it is still only an constant
496       factor (amortized).
497    
498    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
499       with a sub-structure BTrace for controlling back-tracing.  In particular,
500       the above-mentioned function "mode" controls whether the annotation
501       phase is invoked by the compiler.  Another important function is
502       "trigger": when called it aborts the current execution and causes
503       the top-level loop to print a full back-trace.
504    
505    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
506       for back-tracing.  It keeps track of the dynamic call stack and in
507       addition to that it keeps a partial history at each "level" of that
508       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
509       d and b (at separate times, dynamically), then the report will show:
510    
511       GOTO   d
512             /c
513       GOTO  \b
514       CALL   a
515    
516       This shows that there was an initial non-tail call of a, then a
517       tail-call to b or c, looping behavior in a cluster of functions that
518       consist of b and c, and then a goto from that cluster (i.e., either from
519       b or from c) to d.
520    
521       Note that (depending on the user program) the amount of information
522       that the back-trace module has to keep track of at each level is bounded
523       by a constant.  Thus, the whole implementation has the same asymptotical
524       complexity as the original program (both in space and in time).
525    
526    5. compiler/TopLevel/interact/evalloop.sml has been modified to
527       handle the special exception SMLofNJ.Internals.BTrace.BTrace
528       which is raised by the "trigger" function mentioned above.
529    
530    Notes on usage:
531    
532    - Annotated code works well together with unannotated code:
533    Unannotated calls simply do not show up at all in the backtrace.
534    
535    - It is not a good idea to let modules that were annotated during
536    different sessions run at the same time.  This is because the compiler
537    chooses small integers to identify individual functions, and there
538    will be clashes if different modules were compiled in separate sessions.
539    (Nothing will crash, and you will even be told about the clashes, but
540    back-trace information will in general not be useful.)
541    
542    - Back-tracing can be confused by callcc and capture.
543    
544    - The only way of getting a back-trace right now is to explicitly
545    invoke the "trigger" function from your user program.  Eventually, we
546    should make every exception carry back-trace information (if
547    available).  But since this creates more overhead at "raise"-time
548    (similar to the current exnHistory overhead), I have not yet
549    implemented this.  (The implementation will be rather easy.)  With
550    exceptions carrying back-trace information, this facility will be even
551    more useful because users don't need to modify their programs...
552    
553    - While it is possible to compile the compiler with back-trace
554    annotations turned on (I did it to get some confidence in
555    correctness), you must make absolutely sure that core.sml and
556    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
557    be compiled with annotation because there is no core access yet, but
558    if you compile btimp.sml with annotation, then the system will go into
559    an infinite recursion and crash.)
560    Since CM currently does not know about BTrace, the only way to turn
561    annotations on and off for different modules of the compiler is to
562    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
563    this is awkward and clumsy.
564    
565    Sample sessions:
566    
567    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
568    - SMLofNJ.Internals.BTrace.mode (SOME true);
569    [autoloading]
570    [autoloading done]
571    val it = false : bool
572    - structure X = struct
573    -     fun main n = let
574    -         fun a (x, 0) = d x
575    -           | a (x, n) = b (x, n - 1)
576    -         and b (x, n) = c (x, n)
577    -         and c (x, n) = a (x, n)
578    -         and d x = e (x, 3)
579    -         and e (x, 0) = f x
580    -           | e (x, n) = e (x, n - 1)
581    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
582    -           | f n = n * g (n - 1)
583    -         and g n = a (n, 3)
584    -     in
585    -         f n
586    -     end
587    - end;
588    structure X : sig val main : int -> int end
589    - X.main 3;
590    *** BACK-TRACE ***
591    GOTO   stdIn:4.2-13.20: X.main[2].f
592    GOTO-( stdIn:4.2-13.20: X.main[2].e
593    GOTO   stdIn:4.2-13.20: X.main[2].d
594         / stdIn:4.2-13.20: X.main[2].a
595         | stdIn:4.2-13.20: X.main[2].b
596    GOTO-\ stdIn:4.2-13.20: X.main[2].c
597    CALL   stdIn:4.2-13.20: X.main[2].g
598    GOTO   stdIn:4.2-13.20: X.main[2].f
599    GOTO-( stdIn:4.2-13.20: X.main[2].e
600    GOTO   stdIn:4.2-13.20: X.main[2].d
601         / stdIn:4.2-13.20: X.main[2].a
602         | stdIn:4.2-13.20: X.main[2].b
603    GOTO-\ stdIn:4.2-13.20: X.main[2].c
604    CALL   stdIn:4.2-13.20: X.main[2].g
605    GOTO   stdIn:4.2-13.20: X.main[2].f
606    GOTO-( stdIn:4.2-13.20: X.main[2].e
607    GOTO   stdIn:4.2-13.20: X.main[2].d
608         / stdIn:4.2-13.20: X.main[2].a
609         | stdIn:4.2-13.20: X.main[2].b
610    GOTO-\ stdIn:4.2-13.20: X.main[2].c
611    CALL   stdIn:4.2-13.20: X.main[2].g
612    GOTO   stdIn:4.2-13.20: X.main[2].f
613    CALL   stdIn:2.15-17.4: X.main[2]
614    -
615    
616    (Note that because of a FLINt bug the above code currently does not
617    compile without BTrace turned on.)
618    
619    Here is another example, using my modified Tiger compiler:
620    
621    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
622    - SMLofNJ.Internals.BTrace.mode (SOME true);
623    [autoloading]
624    [autoloading done]
625    val it = false : bool
626    - CM.make "sources.cm";
627    [autoloading]
628    ...
629    [autoloading done]
630    [scanning sources.cm]
631    [parsing (sources.cm):parse.sml]
632    [creating directory CM/SKEL ...]
633    [parsing (sources.cm):tiger.lex.sml]
634    ...
635    [wrote CM/sparc-unix/semant.sml]
636    [compiling (sources.cm):main.sml]
637    [wrote CM/sparc-unix/main.sml]
638    [New bindings added.]
639    val it = true : bool
640    - Main.compile ("../testcases/merge.tig", "foo.out");
641    *** BACK-TRACE ***
642    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
643    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
644    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
645    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
646    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
647    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
648    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
649         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
650    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
651    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
652    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
653    CALL   main.sml:1.18-118.4: Main.compile[2]
654    -
655    
656    ----------------------------------------------------------------------
657    Name: Matthias Blumen
658    Date: 2000/06/21 18:00:00 JST
659    Tag: blume-20000621-manual
660    Description:
661    
662    CM manual update: Path environments documented.
663    
664    ----------------------------------------------------------------------
665    Name: Matthias Blume
666    Date: 2000/06/19 13:40:00
667    Tag: blume-20000619-manual
668    Description:
669    
670    CM manual and system/README update.  This only covers the fact that
671    there are no more implicit anchors.  (Path environments and the "bind"
672    option to "cm" have yet to be documented.)
673    
674    ----------------------------------------------------------------------
675    Name: Matthias Blume
676    Date: 2000/06/19 11:05:00 JST
677    Tag: blume-20000619-chdir-bugfix
678    Description:
679    
680    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
681    
682    ----------------------------------------------------------------------
683    Name: Matthias Blume
684    Date: 2000/06/18 22:00:10 JST
685    Tag: blume-20000618-implicit-anchors-really-gone
686    Description:
687    
688    I updates the previous HISTORY entry where I forgot to mention that
689    implicit anchors are no longer with us.
690    
691    The current update also gets rid of the (now useless) controller
692    CM.Control.implicit_anchors.
693    
694    ----------------------------------------------------------------------
695    Name: Matthias Blume
696    Date: 2000/06/16 17:30:00 JST
697    Tag: blume-20000616-anchorenv
698    Description:
699    
700    This patch implements the long anticipated (just kidding :) "anchor
701    environment" mechanism.  In the course of doing this, I also
702    re-implemented CM's internal "SrcPath" module from scratch.  The new
703    one should be more robust in certain boundary cases.  In any case, it
704    is a lot cleaner than its predecessor (IMHO).
705    
706    This time, although there is yet another boot file format change, I
707    kept the unpickler backward-compatible.  As a result, no new bootfiles
708    are necessary and bootstrapping is straightforward.  (You cannot read
709    new bootfiles into an old system, but the other way around is no
710    problem.)
711    
712    Visible changes:
713    
714    ** 0. Implicit path anchors (without the leading $-symbol) are no
715    longer recognized at all. This means that such path names are not
716    illegal either.  For example, the name basis.cm simply refers to a
717    local file called "basis.cm" (i.e, the name is an ordinary path
718    relative to .cm-files directory).  Or, to put it differently, only
719    names that start with $ are anchored paths.
720    
721    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
722    vanished.
723    
724    John (Reppy) had critizised this as soon as I originally proposed and
725    implemented it, but at that time I did not really deeply believe
726    him. :) Now I came full-circle because I need the $<singlearc> syntax
727    in another place where it cannot be seen as an abbreviation for
728    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
729    seems to mean (i.e., it "expands" into the corresponding anchor
730    value).
731    
732    However, when paths are used as members in CM description files, it
733    continues to be true that there must be at least another arc after the
734    anchor.  This is now enforced separately during semantic analysis
735    (i.e., from a lexical/syntactical point of view, the notation is ok.)
736    
737    ** 2. The "cm" class now accepts an option "bind".  The option's value
738    is a sub-option list of precisely two items -- one labeled "anchor"
739    and the other one labeled "value".  As you might expect, "anchor" is
740    used to specify an anchor name to be bound, and "value" specifies what
741    the anchor is being bound to.
742    
743    The value must be a directory name and can be given in either standard
744    syntax (including the possibility that it is itself an anchored path)
745    or native syntax.
746    
747    Examples:
748    
749       foo.cm (bind:(anchor:bar value:$mystuff/bar))
750       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
751    
752    and so on.
753    
754    The meaning of this is that the .cm-file will be processed with an
755    augmented anchor environment where the given anchor(s) is/are bound to
756    the given values(s).
757    
758    The rationale for having this feature is this: Suppose you are trying
759    to use two different (already stable) libraries a.cm and b.cm (that
760    you perhaps didn't write yourself).  Further, suppose each of these
761    two libraries internally uses its own auxiliary library $aux/lib.cm.
762    Normally you would now have a problem because the anchor "lib" can not
763    be bound to more than one value globally.  Therefore, the project that
764    uses both a.cm and b.cm must locally redirect the anchor to some other
765    place:
766    
767       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
768       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
769    
770    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
771    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
772    
773    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
774    you when you do so at the time of CM.stabilize).  Therefore, you can
775    also use an anchored path as the value:
776    
777      a.cm (bind:(anchor:lib value:$a-lib))
778      b.cm (bind:(anchor:lib value:$b-lib))
779    
780    Now you can globally configure (using the usual CM.Anchor.anchor or
781    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
782    itself is always locally bound, setting it globally is no longer
783    meaningful or necessary (but it does not hurt either).  In fact, "lib"
784    can still be used as a global anchor for separate purposes.  As a
785    matter of fact, one can locally define "lib" in terms of a global
786    "lib":
787    
788      a.cm (bind:(anchor:lib value:$lib/a))
789      b.cm (bind:(anchor:lib value:$lib/b))
790    
791    ** 3: The encoding of path names has changed.  This affects the way
792    path names are shown in CM's progress report and also the internal
793    protocol encoding used for parallel make.
794    
795    The encoding now uses one or more ':'-separated segments.  Each
796    segments corresponds to a file that has been specified relative to the
797    file given by its preceding segment.  The first segment is either
798    relative to the CWD, absolute, or anchored.  Each segment itself is
799    basically a Unix pathname; all segments but the first are relative.
800    
801    Example:
802    
803       $foo/bar/baz.cm:a/b/c.sml
804    
805    This path denotes the file bar/a/b/c.sml relative to the directory
806    denoted by anchor "foo".  Notice that the encoding also includes
807    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
808    paths are resolved relative to the .cm-files directory, so baz.cm must
809    be ignored to get the "real" pathname.
810    
811    To make this fact more obvious, CM puts the names of such "virtual
812    arcs" into parentheses when they appear in progress reports. (No
813    parentheses will appear in the internal protocol encoding.)  Thus,
814    what you really see is:
815    
816      $foo/bar/(baz.cm):a/b/c.sml
817    
818    I find this notation to be much more informative than before.
819    
820    Another new feature of the encoding is that special characters
821    including parentheses, colons, (back)slashes, and white space are
822    written as \ddd (where ddd is the decimal encoding of the character).
823    
824    *** The CM manual still needs to be updated.
825    
826    ----------------------------------------------------------------------
827    Name: Allen Leung
828    Date: 2000/06/15 00:38:00
829    Tag: leunga-20000615-x86-peephole
830    
831    x86 Peephole fix by Fermin.  Affects c-- and moby only.
832    
833    ----------------------------------------------------------------------
834    Name: Matthias Blume
835    Date: 2000/06/12 11:40:00
836    Tag: blume-20000612-parmakefix
837    Description:
838    
839    More cleanup after changing the file naming scheme: This time I
840    repaired the parallel make mechanism for CMB.make which I broke earlier.
841    
842    ----------------------------------------------------------------------
843    Name: Allen Leung
844    Date: 2000/06/09 01:25:00
845    Tag: leunga-20000609-various
846    
847    None of these things should affect normal SML/NJ operations
848    
849    1. Peephole improvements provided by Fermin (c--)
850    2. New annotation DEFUSE for adding extra dependence (moby)
851    3. New X86 LOCK instructions (moby)
852    4. New machine description language for reservation tables (scheduling)
853    5. Fixes to various optimization/analysis modules (branch chaining, dominator
854       trees etc.)
855    6. I've changed the CM files so that they can work with versions
856       110.0.6, 110.25 and 110.28
857    
858    ----------------------------------------------------------------------
859    Name: Matthias Blume
860    Date: 2000/06/09 12:40:00
861    Tag: blume-20000609-log
862    Description:
863    
864    - Removed all(?) remaining RCS Log entries from sources.
865    
866    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
867      anchored paths).
868    
869    ----------------------------------------------------------------------
870    Name: Matthias Blume
871    Date: 2000/06/07 17:00:00 JST
872    Tag: blume-20000607-no-implicit-anchors
873    Description:
874    
875    1. This update changes the default setting for
876    CM.Control.implicit_anchors from true to false.  This means that
877    implicit anchors are no longer permitted by default.  I also tried to
878    make sure that nothing else still relies on implicit anchors.
879    (This is the next step on the schedule towards a CM that does not even
880    have the notion of implicit anchors anymore.)
881    
882    2. More CM manual updates.
883    
884    3. I managed to track down and fix the pickling bug I mentioned last
885    time.  Because of the previously existing workaround, this entails no
886    immediate practical changes.
887    
888    ----------------------------------------------------------------------
889    Name: Matthias Blume
890    Date: 2000/06/06 11:15:00 JST
891    Tag: blume-20000606-lazierpickle
892    Description:
893    
894    !!!! NEW BOOT FILES !!!!
895    
896    * The main purpose of this update is to make library pickles lazier in
897    order to reduce the initial space penalty for autoloading a library.
898    As a result, it is now possible to have $smlnj/compiler.cm
899    pre-registered.  This should take care of the many complaints or
900    inquiries about missing structure Compiler.  This required changes to
901    CM's internal data structures and small tweaks to some algorithms.
902    
903    As a neat additional effect, it is no longer necessary (for the sake
904    of lean heap image files) to distinguish between a "minimal" CM and a
905    "full" CM.  Now, there is only one CM (i.e., the "full" version:
906    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
907    interactive top level. ($smlnj/cm/minimal.cm is gone.)
908    
909    To make the life of compiler-hackers easier, "makeml" now also
910    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
911    words, after you bootstrap a new sml for the first time, you will not
912    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
913    around you will still have to do it, though.)
914    
915    * A second change consists of major updates to the CM manual.  There
916    are now several appendices with summary information and also a full
917    specification of the CM description file syntax.
918    
919    * In directory src/system I added the script "allcross".  This script
920    invokes sml and cross-compiles the compiler for all supported
921    architectures.  (Useful when providing a new set of boot files.)
922    
923    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
924    added a small tweak to pickle-util.sml to work around this problem,
925    but it is not a proper fix yet.  I will investigate further.  (The
926    effect of the bug was an inflation of library pickle size.)
927    
928    * Version number increased to 110.28.1 (to avoid compatibility problems).
929    
930    ----------------------------------------------------------------------
931    Name: Allen Leung
932    Date: 2000/05/25 17:28 EDT
933    Tag: leunga-20000525-ra
934    Description:
935    
936      Fixed a bug in freezing phase of the register allocator.
937    
938    ----------------------------------------------------------------------
939    Name: Allen Leung
940    Date: 2000/05/15 22:53 EDT
941    Tag: leunga-20000515-alpha-x86-ra
942    Description:
943    
944      1. Alpha
945    
946          Slight cleanup.  Removed the instruction SGNXL
947    
948      2. X86
949    
950          Added the following instructions to the instruction set:
951    
952            ROLx, RORx,
953            BTx, BTSx, BTLx, BTRx,
954            XCHGx, and variants with the LOCK prefix
955    
956      3. Register Allocation
957    
958          The module ra-rewrite-with-renaming has been improved.
959    
960      These have no effect on SML/NJ.
961    
962    ----------------------------------------------------------------------
963    Name: Matthias Blume
964    Date: 2000/05/15 16:20:00 JST
965    Tag: blume-20000515-lightrebuild
966    Description:
967    
968    1. I added an alternative to "-rebuild" to "makeml".  The difference is
969       that prior to calling CMB.make' the CM-variable "LIGHT" will be
970       defined.  In effect, the command will not build any cross-compiler
971       backends and therefore finish more quickly.
972    
973       The "fixpt" script also takes a "-light" switch to be able to use
974       this new facility while compiling for a fixpoint.
975    
976    2. I replaced all mentions of anchored paths in group owner specifications
977       with simple relative paths (usually starting with "..").
978       The rationale is that a library's internal workings should not be
979       compromised by the lack of some anchor.  (An anchor is necessary
980       for someone who wants to refer to the library by an anchored path,
981       but it should not be necessary to build the same library in the first
982       place.)
983    
984    3. I changed the way CM's tool mechanism determines the shell command
985       string used for things like ml-yacc etc. so that it does not break
986       when CM.Control.implicit_anchors is turned off.
987    
988    ----------------------------------------------------------------------
989    Name: Matthias Blume
990    Date: 2000/05/12 18:20:00 JST
991    Tag: blume-20000512-ml-build
992    Description:
993    
994    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
995    getting installed properly (by config/install.sh).
996    
997    ----------------------------------------------------------------------
998    Name: Matthias Blume
999    Date: 2000/05/12 17:30:00 JST
1000    Tag: blume-20000512-anchors
1001    Description:
1002    
1003    !!! NEW BOOT FILES !!!
1004    
1005    This change is in preparation of fading out support for "implicitly
1006    anchored path names".  I went through all sources and used the
1007    explicit (and relatively new) $-notation.  See system/README and the
1008    CM manual for more info on this.
1009    
1010    I also modified the anchoring scheme for some things such as "smlnj",
1011    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1012    anchors are more expressive: anchor name and first arc do not have to
1013    coincide.  This entails the following user-visible change:
1014    
1015    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1016    particular, when you fire up sml with a command-line argument, say,
1017    e.g.:
1018    
1019       sml '$smlnj/cmb.cm'
1020    
1021    At the ML toplevel prompt:
1022    
1023       CM.autoload "$smlnj/cmb.cm";
1024    
1025    There is also a new controller in CM.Control that can be used to turn
1026    off all remaining support for implicit anchors by saying:
1027    
1028        CM.autoload "$smlnj/
1029        #set CM.Control.implicit_anchors false;
1030    
1031    This causes CM to reject implicitly anchored paths.  This is (for the
1032    time being) less permissive than the "final" version where there will
1033    be no more such implicit anchors and relative paths will be just that:
1034    relative.
1035    
1036    The next step (version after next version?) will be to make the
1037    default for CM.Control.implicit_anchors false.  After the dust has
1038    settled, I can then produce the "final" version of this...
1039    
1040    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1041    
1042    ----------------------------------------------------------------------
1043    Name: Matthias Blume
1044    Date: 2000/05/11 16:30:00 JST
1045    Tag: blume-20000511-sources
1046    Description:
1047    
1048    The main change is that I added function CM.sources as a generalized
1049    version of the earlier CM.makedepend.  This entails the following
1050    additional changes:
1051    
1052      - CM.makedepend has been dropped.
1053    
1054      - CM manual has been updated.
1055    
1056      - TOOLS signature and API have been changed.
1057    
1058    ----------------------------------------------------------------------
1059    Name: Allen Leung
1060    Date: 2000/05/10 21:17 EDT
1061    Tag: leunga-20000510-moby-c--ssa
1062    Description:
1063    
1064      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1065    None of these affect SML/NJ.
1066    
1067    1. Register Allocation
1068    
1069        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1070           This module tries to remove local (i.e. basic block level) redundancies
1071           during spilling.
1072    
1073        b. A new framework for performing region based register allocation.
1074           Not yet entirely functional.
1075    
1076    2. X86
1077    
1078       a. DefUse for POP was missing the stack pointer [found by Lal]
1079       b. Reload for CALL was incorrect in X86Spill [found by John]
1080       c. Various fixes in X86Spill so that it can be used correctly for
1081          the new spilling module.
1082    
1083    3. SSA/IR
1084    
1085       a. New module ir/dj-dataflow.sml implements elimination based
1086          data flow analysis.
1087    
1088    4. MLRiscGen
1089    
1090       a. Fix for gc type annotation
1091    
1092    5. MDGen
1093    
1094       Various fixes for machine description -> ml code translation.  For ssa
1095       only.
1096    
1097    ----------------------------------------------------------------------
1098    Name: Allen Leung
1099    Date: 2000/05/08 22:17 EDT
1100    Tag: leunga-20000508-labexp
1101    Description:
1102    
1103      Fermin has found a few assembly problems with constant expressions
1104      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1105      which choke on dumb assemblers.  This is his fix.
1106    
1107    ----------------------------------------------------------------------
1108    Name: Dave MacQueen
1109    Date: 2000/04/09 14:00 EDT
1110    Tag: dbm-20000502-Version_110_28
1111    Description:
1112    
1113    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1114    
1115    2. Updated config/version to 110.28
1116    
1117    3. Updated config/srcarchiveurl
1118    
1119    3. New boot files!
1120       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1121    
1122    ----------------------------------------------------------------------
1123    Name: Matthias Blume
1124    Date: 2000/05/01 19:05:00 JST
1125    Tag: blume-20000501-noweb
1126    Description:
1127    
1128    A new noweb tool has been added.  The existing system is entirely
1129    unaffected by this, but some CM users have asked for renewed noweb
1130    support.  Everything is documented in the CM manual.
1131    
1132    New (plugin) libraries:
1133    
1134       noweb-tool.cm
1135       nw-ext.cm
1136    
1137    ----------------------------------------------------------------------
1138    Name: Dave MacQueen
1139    Date: 2000/04/30 12:40PM EDT
1140    Tag: dbm-20000430-bug_fixes
1141    Description:
1142    
1143    1. Fix for bug 1498
1144       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1145       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1146         added toRealArray function
1147       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1148         added check for tag Obj.RealArray to array printing case in ppObj
1149    
1150    2. Fix for bug 1510
1151       smlnj/src/compiler/Semant/types/typesutil.sml
1152         fixed definition of dummyargs (used by equalTycon) so that
1153         dummy args are distinct types
1154    
1155    ----------------------------------------------------------------------
1156    Name: Matthias Blume
1157    Date: 2000/04/30 01:00:00 JST
1158    Tag: blume-20000430-versions
1159    Description:
1160    
1161    1. CM version numbering added.  This is an implementation of Lal's
1162       proposal for adding version numbers and version checking to .cm
1163       files.  Lal said that his proposal was just that -- a proposal.
1164       For the time being I went ahead and implemented it so that people
1165       can comment on it.  Everything is completely backward-compatible
1166       (except for the stable library format, i.e., new bootfiles!).
1167    
1168       As usual, see the CM manual for details.
1169    
1170    2. An alternative syntax for anchored paths has been implemented.
1171       Dave has recently voiced the same concerns that I had when I did
1172       this, so there should be some support.  My take is that eventually
1173       I will let support for the current syntax (where anchors are
1174       "implicit") fade out in favor of the new, explicit syntax.
1175       In order to be backward-compatible, both old and new syntax are
1176       currently supported.
1177    
1178       Again, see the CM manual for details.
1179    
1180    3. Parallel make is trying to be slightly smarter:  When the master
1181       process finds a "bottleneck", i.e., when there is only one
1182       compilation unit that can be compiled and everybody else is
1183       waiting on it, then it will simply compile it directly instead
1184       of clumsily telling one of the slaves to do it.
1185    
1186    4. Support for "unsharing" added.  This is necessary in order to be
1187       able to have two different versions of the same library running
1188       at the same time (e.g., for trying out a new MLRISC while still
1189       having the old MLRISC linked into the current compiler, etc.)
1190       See the CM manual.
1191    
1192    5. Simple "makedepend" functionality added for generating Makefile
1193       dependency information.  (This is rather crude at the moment.
1194       Expect some changes here in the future.)
1195    
1196    6. ".fun" added as a recognized suffix for ML files. Also documented
1197       explicitly in the manual that the fallback behavior (unknown suffix
1198       -> ML file) is not an official feature!
1199    
1200    7. Small changes to the pickler for stable libraries.
1201    
1202    8. Several internal changes to CM (for cleanup/improvement).
1203    
1204    
1205    !!!! NEW BINFILES !!!!
1206    
1207    ----------------------------------------------------------------------
1208    Name: Matthias Blume
1209    Date: 2000/04/28 17:30:00 JST
1210    Tag: blume-20000428-pathconfig
1211    Description:
1212    
1213    1. I changed config/install.sh to remove duplicate entries from the
1214       lib/pathconfig file at the end.  Moreover, the final version of
1215       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1216       in src/system/installml.
1217    
1218    2. The config/install.sh script now consistently uses relative
1219       pathnames in lib/pathconfig whenever the anchor is in the lib
1220       directory.  (So far this was true for the libraries that come
1221       pre-compiled and bundled as part of the bootfiles but not for
1222       libraries that are compiled by the script itself.)
1223    
1224    ----------------------------------------------------------------------
1225    Name: Matthias Blume
1226    Date: 2000/04/26 13:10:00 JST
1227    Tag: blume-20000426-fun_suffix
1228    Description:
1229    
1230    Added ".fun" as a recognized file name suffix (for ML code).
1231    
1232    ----------------------------------------------------------------------
1233    Name: Allen Leung
1234    Date: 2000/04/25 17:00:00 EST
1235    Tag: leunga-20000425-alpha-ra
1236    Description:
1237    
1238    1. Alpha
1239    
1240        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1241    in C--.
1242    
1243    2. RA
1244    
1245       Added a flag "ra-dump-size" to print out the size of the flowgraph
1246       and the interference graph.
1247    
1248    ----------------------------------------------------------------------
1249    Name: Dave MacQueen
1250    Date: 2000/04/25/
1251    Tag: dbm-20000425-mlyacc_doc_examples
1252    Description:
1253      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1254      Updated all three examples in src/ml-yacc/examples to run
1255      under 110.* using CM.make.
1256    
1257    ----------------------------------------------------------------------
1258    Name: Allen Leung
1259    Date: 2000/04/20 23:04:00 EST
1260    Tag: leunga-20000420-ssa-c---stuff
1261    Description:
1262    
1263      This update synchronizes my repository with Yale's.  Most of these
1264    changes, however, do not affect SML/NJ at all (the RA is an exception).
1265    
1266    1. Register Allocator
1267    
1268       a. An improvement in the interference graph construction:
1269          Given a copy
1270    
1271                s <- t
1272    
1273          no interference edge between s and t is added for this definition of s.
1274    
1275       b. I've added two new spill heuristic modules that Fermin and I developed
1276          (in the new library RA.cm). These are unused in SML/NJ but maybe
1277          useful for others (Moby?)
1278    
1279    2. X86
1280    
1281       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1282    
1283    3. Alpha
1284    
1285       a. Added the BSR instruction and code generation that goes with it [C--]
1286       b. Other fixes too numerous to recount provided by Fermin [C--]
1287    
1288    4. Regmaps
1289    
1290       a. The regmaps are not initialized with the identity physical bindings
1291          at creation time.  This is unneeded.
1292    
1293    5. MLRISC Optimizations
1294    
1295       a. The DJ-Graph module can now compute the iterated dominance frontiers
1296          intersects with liveness incrementally in linear time! Woohoo!
1297          This is now used in my new SSA construction algorithm.
1298    
1299       b. THe branch reorganization module is now smarter about linear chains of
1300          basic blocks.
1301    
1302    
1303    ----------------------------------------------------------------------
1304    Name: Matthias Blume
1305    Date: 2000/04/12 13:52:00 JST
1306    Tag: blume_main_v110p27_1
1307    Description:
1308    
1309    Changed install.sh script to handle archive files without version number
1310    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1311    name of the boot file archive.
1312    
1313    ----------------------------------------------------------------------
1314    Name: Dave MacQueen
1315    Date: 2000/04/09 14:00 EDT
1316    Tag: dbm-20000410-Version_110_27
1317    Description:
1318    
1319    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1320    
1321    2. Updated src/config/version to 110.27
1322    
1323    3. New boot files!
1324    
1325    ----------------------------------------------------------------------
1326    Name: Allen Leung
1327    Date: 2000/04/09 19:09:00 EST
1328    Tag: leunga-20000409-misc
1329    Description:
1330    
1331    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1332    
1333    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1334    
1335    ----------------------------------------------------------------------
1336    Name: Stefan
1337    Date: 2000/04/07 10:00:00 EDT
1338    Tag: monnier-20000406-branch-handling
1339    Description:
1340    
1341    Improved handling of branches (mostly those generated from
1342    polymorphic equality), removed switchoff and changed the
1343    default optimization settings (more cpsopt and less flintopt).
1344    
1345    ----------------------------------------------------------------------
1346    Name: Allen Leung
1347    Date: 2000/04/06 01:30:00 EST
1348    Tag: leunga-20000406-peephole-x86-SSA-2
1349    Description:
1350    
1351       Forgot a few files.
1352    
1353    ----------------------------------------------------------------------
1354    Name: Allen Leung
1355    Date: 2000/04/06 00:36:00 EST
1356    Tag: leunga-20000406-peephole-x86-SSA
1357    Description:
1358    
1359    1.  New Peephole code
1360    
1361    2.  Minor improvement to X86 instruction selection
1362    
1363    3.  Various fixes to SSA and machine description -> code translator
1364    
1365    ----------------------------------------------------------------------
1366    Name: Matthias Blume
1367    Date: 2000/04/05 12:30:00 JST
1368    Tag: blume_main_v110p26p2_3
1369    Description:
1370    
1371    This update just merges three minor cosmetic updates to CM's sources
1372    to get ready for the 110.27 code freeze on Friday.  No functionality
1373    has changed.
1374    
1375    ----------------------------------------------------------------------
1376    Name: Allen Leung
1377    Date: 2000/04/04 19:39:00 EST
1378    Tag: leunga-20000404-x86-asm
1379    Description:
1380    
1381    1.  Fixed a problem in X86 assembly.
1382    
1383        Things like
1384    
1385           jmp %eax
1386           jmp (%eax)
1387    
1388        should be output as
1389    
1390           jmp *%eax
1391           jmp *(%eax)
1392    
1393    2.  Assembly output
1394    
1395          Added a new flag
1396    
1397              "asm-indent-copies" (default to false)
1398    
1399          When this flag is on, parallel copies will be indented an extra level.
1400    
1401    ----------------------------------------------------------------------
1402    Name: Allen Leung
1403    Date: 2000/04/04 03:18:00 EST
1404    Tag: leunga-20000404-C--Moby
1405    Description:
1406    
1407        All of these fixes are related to C--, Moby, and my own optimization
1408        stuff; so they shouldn't affect SML/NJ.
1409    
1410    1.  X86
1411    
1412        Various fixes related floating point, and extensions.
1413    
1414    2.  Alpha
1415    
1416        Some extra patterns related to loads with signed/zero extension
1417        provided by Fermin.
1418    
1419    3.  Assembly
1420    
1421        When generating assembly, resolve the value of client defined constants,
1422        instead of generating symbolic values.  This is controlled by the
1423        new flag "asm-resolve-constants", which is default to true.
1424    
1425    4.  Machine Descriptions
1426    
1427        a. The precedence parser was slightly broken when parsing infixr symbols.
1428        b. The type generalizing code had the bound variables reversed, resulting
1429           in a problem during arity raising.
1430        c. Various fixes in machine descriptions.
1431    
1432    ----------------------------------------------------------------------
1433    Name: Matthias Blume
1434    Date: 2000/04/03 16:05:00 JST
1435    Tag: blume_main_v110p26p2_2
1436    Description:
1437    
1438    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1439    now done via the ordinary static environment that is context to each
1440    compilation unit.
1441    
1442    To this end, I arranged that instead of "structure Core" as "structure
1443    _Core" is bound in the pervasive environment.  Core access is done via
1444    _Core (which can never be accidentally rebound because _Core is not a
1445    legal surface-syntax symbol).
1446    
1447    The current solution is much cleaner because the core environment is
1448    now simply part of the pervasive environment which is part of every
1449    compilation unit's context anyway.  In particular, this eliminates all
1450    special-case handling that was necessary until now in order to deal
1451    with dynamic and symbolic parts of the core environment.
1452    
1453    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1454    compilation manager's bootstrap compiler (actually: in the "init group"
1455    handling).  See the comments in src/system/smlnj/init/init.cmi for
1456    more details.
1457    
1458    I also tried to track down all mentions of "Core" (as string argument
1459    to Symbol.strSymbol) in the compiler and replaced them with a
1460    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1461    name appears in one place only.
1462    
1463    Binfile and bootfile format have not changed, but the switchover from
1464    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1465    bootfiles anyway.
1466    
1467    ----------------------------------------------------------------------
1468    Name: Allen Leung
1469    Date: 2000/04/02 21:17:00 EST
1470    Tag: leunga-20000402-mltree
1471    Description:
1472    
1473       1. Renamed the constructor CALL in MLTREE by popular demand.
1474       2. Added a bunch of files from my repository.  These are currently
1475          used by other non-SMLNJ backends.
1476    
1477    ----------------------------------------------------------------------
1478    Name: Allen Leung
1479    Date: 2000/03/31 21:15:00 EST
1480    Tag: leunga-20000331-aliasing
1481    Description:
1482    
1483    This update contains a rewritten (and hopefully more correct) module
1484    for extracting aliasing information from CPS.
1485    
1486       To turn on this feature:
1487    
1488            Compiler.Control.CG.memDisambiguate := true
1489    
1490       To pretty print the region information with assembly
1491    
1492           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1493    
1494       To control how many levels of aliasing information are printed, use:
1495    
1496           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1497    
1498       The default of n is 3.
1499    
1500    ----------------------------------------------------------------------
1501    Name: David MacQueen
1502    Date: 2000/03/31 11:15:00 EST
1503    Tag: dbm-20000331-runtime_fix
1504    Description:
1505    
1506    This update contains:
1507    
1508    1. runtime/c-lib/c-libraries.c
1509       includes added in revision 1.2 caused compilation errors on hppa-hpux
1510    
1511    2. fix for bug 1556
1512       system/Basis/Implementation/NJ/internal-signals.sml
1513    
1514    ----------------------------------------------------------------------
1515    Name: Matthias Blume
1516    Date: 2000/03/31 18:00:00 JST
1517    Tag: blume_main_v110p26p2_1
1518    Description:
1519    
1520    This update contains:
1521    
1522    1. A small change to CM's handling of stable libraries:
1523       CM now maintains one "global" modmap that is used for all stable
1524       libraries.  The use of such a global modmap maximizes sharing and
1525       minimizes the need for re-traversing parts of environments during
1526       modmap construction.  (However, this has minor impact since modmap
1527       construction seems to account for just one percent or less of total
1528       compile time.)
1529    
1530    2. I added a "genmap" phase to the statistics.  This is where I got the
1531       "one percent" number (see above).
1532    
1533    3. CM's new tool parameter mechanism just became _even_ better. :)
1534       - The parser understands named parameters and recursive options.
1535       - The "make" and "shell" tools use these new features.
1536         (This makes it a lot easier to cascade these tools.)
1537       - There is a small syntax change: named parameters use a
1538    
1539           <name> : ( <option> ... )            or
1540           <name> : <string>
1541    
1542         syntax.  Previously, named parameters were implemented in an
1543         ad-hoc fashion by each tool individually (by parsing strings)
1544         and had the form
1545    
1546           <name>=<string>
1547    
1548       See the CM manual for a full description of these issues.
1549    
1550    ----------------------------------------------------------------------
1551    Name: Matthias Blume
1552    Date: 2000/03/30 18:00:00 JST
1553    Tag: blume_main_v110p26p2_0
1554    Description:
1555    
1556    !!!!! WARNING !!!!!!
1557    !!  New binfiles  !!
1558    !!!!!!!!!!!!!!!!!!!!
1559    
1560    This update contains:
1561    
1562    1. Moderate changes to CM:
1563    
1564       - Changes to CM's tools mechanism.  In particular, it is now possible
1565       to have tools that accept additional "command line" parameters
1566       (specified in the .cm file at each instance where the tool's class is
1567       used).
1568    
1569       This was done to accommodate the new "make" and "shell" tools which
1570       facilitate fairly seamless hookup to portions of code managed using
1571       Makefiles or Shell scripts.
1572    
1573       There are no classes "shared" or "private" anymore.  Instead, the
1574       sharing annotation is now a parameter to the "sml" class.
1575    
1576       There is a bit of generic machinery for implementing one's own
1577       tools that accept command-line parameters.  However, I am not yet fully
1578       satisfied with that part, so expect changes here in the future.
1579    
1580       All existing tools are described in the CM manual.
1581    
1582       - Slightly better error handling.  (CM now suppresses many followup
1583       error messages that tended to be more annoying than helpful.)
1584    
1585    2. Major changes to the compiler's static environment data structures.
1586    
1587       - no CMStaticEnv anymore.
1588            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
1589              but it is called Environment), no conversions between different
1590              kinds of static environments
1591    
1592       - There is still a notion of a "modmap", but such modmaps are generated
1593         on demand at the time when they are needed.  This sounds slow, but I
1594         sped up the code that generates modmaps enough for this not to lead to
1595         a slowdown of the compiler (at least I didn't detect any).
1596    
1597       - To facilitate rapid modmap generation, static environments now
1598         contain an (optional) "modtree" structure.  Modtree annotations are
1599         constructed by the unpickler during unpickling.  (This means that
1600         the elaborator does not have to worry about modtrees at all.)
1601         Modtrees have the advantage that they are compositional in the same
1602         way as the environment data structure itself is compositional.
1603         As a result, modtrees never hang on to parts of an environment that
1604         has already been rendered "stale" by filtering or rebinding.
1605    
1606       - I went through many, many trials and errors before arriving at the
1607         current solution.  (The initial idea of "linkpaths" did not work.)
1608         But the result of all this is that I have touched a lot of files that
1609         depend on the "modules" and "types" data structures (most of the
1610         elaborator). There were a lot of changes during my "linkpath" trials
1611         that could have been reverted to their original state but weren't.
1612         Please, don't be too harsh on me for messing with this code a bit more
1613         than what was strictly necessary...  (I _did_ resist the tempation
1614         of doing any "global reformatting" to avoid an untimely death at
1615         Dave's hands. :)
1616    
1617       - One positive aspect of the previous point:  At least I made sure that
1618         all files that I touched now compile without warnings (other than
1619         "polyEqual").
1620    
1621       - compiler now tends to run "leaner" (i.e., ties up less memory in
1622         redundant modmaps)
1623    
1624    ----------------------------------------------------------------------
1625    Name: Allen Leung
1626  Date: 2000/03/29 18:00:00  Date: 2000/03/29 18:00:00
1627  Tag: leunga-20000327-mlriscGen_hppa_alpha_x86  Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
1628  Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz  Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
# Line 104  Line 1713 
1713    
1714       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1715       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1716       testl is perferable to andl.       testl is preferable to andl.
1717    
1718  7. RA (x86 only)  7. RA (x86 only)
1719    
# Line 286  Line 1895 
1895    
1896  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1897  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
1898  specifiel the layout of the tool command line.  See the CM manual for  specifies the layout of the tool command line.  See the CM manual for
1899  explanation.  explanation.
1900    
1901  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
1902  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1903  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
1904  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 333  Line 1942 
1942  rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for  rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
1943  <k>=1,2,... is normally not a good idea when invoking fixpt.  However,  <k>=1,2,... is normally not a good idea when invoking fixpt.  However,
1944  they might be the result of an earlier partial run of fixpt (which  they might be the result of an earlier partial run of fixpt (which
1945  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
1946  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
1947  and, thus, saves a lot of time.  and, thus, saves a lot of time.
1948    
# Line 383  Line 1992 
1992    it from that remote directory.    it from that remote directory.
1993    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
1994    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
1995    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
1996    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
1997    
1998    For CVS users, this mechanism is not relevant for source archives, but    For CVS users, this mechanism is not relevant for source archives, but

Legend:
Removed from v.586  
changed lines
  Added in v.731

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