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 570, Wed Mar 8 17:30:13 2000 UTC revision 742, Thu Nov 30 14:09:32 2000 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16  Name:  Name: Matthias Blume
17  Date:  Date: 2000/11/30 23:12:00 JST
18  Tag:  Tag: blume-20001130-filereorg
19  Description:  Description:
20    
21    Some manual updates and some file reorganizations in CM.
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume
25    Date: 2000/11/24 17:45:00 JST
26    Tag: blume-20001124-link
27    Description:
28    
29    Drastically improved link traversal code for the case that the dynamic
30    value was already loaded at bootstrap time.  As a result, CM and CMB
31    now both load blazingly fast -- even on a very slow machine.  Also,
32    memory consumption has been further reduced by this.
33    
34    Warning: The format of the PIDMAP file has changed.  THerefore, to
35    bootstrap you have to do this:
36    
37    1. Run CMB.make
38    2. Make a symbolic link for the boot directory:
39         ln -s sml.boot.ARCH-OS xxx
40    3. "Rebuild" the boot directory:
41         ./makeml -boot xxx -rebuild sml ; rm xxx
42    4. Boot normally:
43          ./makeml
44    
45    ----------------------------------------------------------------------
46    Name: Matthias Blume
47    Date: 2000/11/21 21:20:00 JST
48    Tag: blume-20001121-tools
49    Description:
50    
51    Continued hacking on autoloading problem -- with success this time.
52    Also changed tool-plugin mechanism.  See new CM manual.
53    
54    ----------------------------------------------------------------------
55    Name: Matthias Blume
56    Date: 2000/11/19 14:30:00 JST
57    Tag:  blume-20001119-autoload
58    Description:
59    
60    Some hacking to make autoloading faster.  Success for CMB, no success
61    so far for CM.  There is a reduced structure CM' that autoloads faster.
62    (This is a temporary, non-documented hack to be eliminated again when
63    the general problem is solved.)
64    
65    ----------------------------------------------------------------------
66    Name: Matthias Blume
67    Date: 2000/11/17 14:10:00 JST
68    Tag: blume-20001117-pickle-lib
69    Description:
70    
71    1. Eliminated comp-lib.cm
72    2. Made pickle-lib.cm
73    3. Eliminated all uses of intset.sml (from comp-lib.cm)
74    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
75       equivalent constructs from smlnj-lib.cm (INtHashTable).
76    5. Point 4. also goes for those uses of intmap.* in MLRISC.
77       Duplicated intmap modules thrown out.
78    6. Hunted down all duplicated SCC code and replaced it with
79       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
80    7. Rewrote Feedback module.
81    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
82       should be thrown out and equivalent modules from smlnj-lib.cm
83       should be used (IntRedBlackSet, IntListSet, ...).
84    
85    Confirmed that compiler compiles to fixpoint.
86    
87    ----------------------------------------------------------------------
88    Name: Allen Leung
89    Date: 2000/11/10 18:00:00
90    Tag: leunga-20001110-new-x86-fp
91    
92    A new x86 floating point code generator has been added.
93    By default this is turned off.  To turn this on, do:
94    
95        CM.autoload "$smlnj/compiler.cm";
96        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
97    
98    Changes:
99    
100    1.  Changed FTAN to FPTAN so that the assembly output is correct.
101    2.  Changed the extension callback for FTANGENT to generate:
102    
103              fptan
104              fstp  %st(0)
105        instead of
106              fptan
107              fstpl ftempmem
108    
109    3.  Numerous assembly fixes for x86.
110    
111    5.  Cleaned up the machine code output module x86/x86MC.sml and added
112        support for a whole bunch of instructions and addressing modes:
113    
114          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
115          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
116          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
117          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
118          fxch %st(n)
119          fld %st(n)
120          fst %st(n)
121          fst mem
122          fstp %st(n)
123          fucom %st(n)
124          fucomp %st(n)
125    
126        All these are now generated when the fast fp mode is turned on.
127    
128    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
129    
130    ----------------------------------------------------------------------
131    Name: Matthias Blume
132    Date: 2000/11/09 11:20:00 JST
133    Tag: blume-20001109-scc
134    Description:
135    
136    Eliminated some code duplication:
137    
138    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
139       (Otherwise the functor is useless.)
140    2. Used GraphSCCFn where SCCUtilFun was used previously.
141    3. Got rid of SCCUtilFun (in comp-lib.cm).
142    
143    ----------------------------------------------------------------------
144    Name: Lal George
145    Date: 2000/11/06 09:02:21 EST 2000
146    Tag: Release_110_30
147    Description:
148    
149    - Version 110.30
150    ----------------------------------------------------------------------
151    Name: Matthias Blume
152    Date: 2000/11/04 14:45:00
153    Tag: blume-20001104-mlbuild
154    Description:
155    
156    - Made ml-build faster on startup.
157    - Documentation fixes.
158    
159    ----------------------------------------------------------------------
160    Name: Matthias Blume
161    Date: 2000/11/02 17:00:00 JST
162    Tag: blume-20001102-condcomp
163    Description:
164    
165    - Small tweaks to pickler -- new BOOTFILES!
166    - Version bumped to 110.29.2.
167    - Added conditional compilation facility to init.cmi (see comment there).
168    ----------------------------------------------------------------------
169    Name: Allen Leung
170    Date: 2000/10/23 19:31:00
171    Tag: leunga-20001023-demo-ra
172    
173    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
174    2. Test programs for the graph library updated
175    3. Some new MLRISC demo programs added
176    
177    ----------------------------------------------------------------------
178    Name: Matthias Blume
179    Date: 2000/08/31 22:15:00 JST
180    Tag: blume-20001017-errmsg
181    Description:
182    
183    More error message grief: Where there used to be no messages, there
184    now were some that had bogus error regions.  Fixed.
185    
186    ----------------------------------------------------------------------
187    Name: Matthias Blume
188    Date: 2000/08/31 17:30:00 JST
189    Tag: blume-20001017-v110p29p1
190    Description:
191    
192    I made a version 110.29.1 with new bootfiles.
193    
194    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
195              CM documentation changes and a small bugfix in CM's error reporting.
196    
197    ----------------------------------------------------------------------
198    Name: Lal George
199    Date: 2000/09/27 14:42:35 EDT
200    Tag: george-20000927-nodestatus
201    Description:
202    
203    Changed the type of the nodestatus, so that:
204    
205            SPILLED(~1)             is now SPILLED
206            SPILLED(m) where m>=0   is now MEMREG(m)
207            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
208    
209    ----------------------------------------------------------------------
210    Name: Matthias Blume
211    Date: 2000/09/07 14:45:00 JST
212    Tag: blume-20000907-cmerrmsg
213    Description:
214    
215    Small tweak to CM to avoid getting ML syntax error messages twice.
216    
217    ----------------------------------------------------------------------
218    Name: Matthias Blume
219    Date: 2000/08/31 18:00:00 JST
220    Tag: blume-20000831-cvsbootfiles
221    Description:
222    
223    New URL for boot files (because the 110.29 files on the BL server do
224    now work correctly with my updated install scripts for yacc and lex).
225    
226    ----------------------------------------------------------------------
227    Name: Matthias Blume
228    Date: 2000/08/08 12:33:00 JST
229    Tag: blume-20000808-manual
230    Description:
231    
232    Tiny update to CM manual.
233    
234    ----------------------------------------------------------------------
235    Name: Allen Leung
236    Date: 2000/08/7 19:31:00
237    Tag: leunga-20000807-a-whole-bunch-of-stuff
238    
239      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
240    and Mobdy.
241    
242    1.  x86
243    
244       a.  Fixes to peephole module by John and Dan.
245       b.  Assembly fix to SETcc by Allen.
246       c.  Fix to c-call by John.
247       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
248       e.  Instruction selection optimization to SETcc as suggested by John.
249    
250           For example,
251    
252            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
253    
254           should generate:
255    
256            MOVL a, x
257            SUBL b, x
258            SHRL 31, x
259    
260    2.  IR stuff
261    
262         A bunch of new DJ-graph related algorithms added.  These
263         speed up SSA construction.
264    
265    3.  SSA + Scheduling
266    
267         Added code for SSA and scheduling to the repository
268    
269    ----------------------------------------------------------------------
270    Name: Lal George
271    Date: 2000/07/27 11:53:14 EDT
272    
273    Tag: lal-20000727-linux-ppc
274    Description:
275    
276     Made changes to support Linux PPC.
277     p.s. I have confirmation that the 110.29 boot files work fine.
278    
279    ----------------------------------------------------------------------
280    Name: Matthias Blume
281    Date: 2000/07/27 17:40:00 JST
282    Tag: blume-20000727-scripts
283    Description:
284    
285    !!!! WARNING !!!!
286    You must recompile the runtime system!
287    !!!! WARNING !!!!
288    
289    This is basically another round of script-enhancements:
290    
291    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
292       and undefine CM preprocessor symbols.
293    
294    2. ml-build avoids generating a new heap image if it finds that the
295       existing one is still ok.  (The condition is that no ML file had to
296       be recompiled and all ML files are found to be older that the heap
297       file.)
298    
299       To make this work smoothly, I also hacked the runtime system as
300       well as SMLofNJ.SysInfo to get access to the heap image suffix
301       (.sparc-solaris, ...) that is currently being used.
302    
303       Moreover, the signature of CM.mk_standalone has changed.  See the
304       CM manual.
305    
306    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
307       CM manual for details.)
308    
309    4. More CM manual updates:
310        - all of the above has been documented.
311        - there is now a section describing the (CM-related) command line
312          arguments that are accepted by the "sml" command
313    
314    ----------------------------------------------------------------------
315    Name: Matthias Blume
316    Date: 2000/07/25 16:20:00 JST
317    Tag: blume-20000725-makedepend
318    Description:
319    
320    Added a script called ml-makedepend.  This can be used in makefiles
321    for Unix' make in a way very similar to the "makedepend" command for
322    C.
323    
324    The script internally uses function CM.sources.
325    
326    Synopsis:
327    
328        ml-makedepend [-f makefile] cmfile targetname
329    
330    The default for the makefile is "makefile" (or "Makefile" should
331    "makefile" not exist).
332    
333    ml-makedepend adds a cmfile/targetname-specific section to this
334    makefile (after removing the previous version of this section).  The
335    section contains a single dependency specification with targetname on
336    the LHS (targetname is an arbitrary name), and a list of files derived
337    from the cmfile on the RHS.  Some of the files on the RHS are
338    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
339    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
340    path names.  The makefile writer is responsible for making sure that
341    these variables have correct at the time "make" is invoked.
342    
343    ----------------------------------------------------------------------
344    Name: Matthias Blume
345    Date: 2000/07/22 23:30:00 JST
346    Tag: blume-20000722-urlupdate
347    Description:
348    
349    Changed BOOT and config/srcarchiveurl to point to BL server:
350    
351        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
352    
353    ----------------------------------------------------------------------
354    Name: Matthias Blume
355    Date: 2000/07/18 18:00:00 JST
356    Tag: blume-20000718-Version_110_29
357    Description:
358    
359    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
360    
361    2. Updated config/version to 110.29
362    
363    3. Updated config/srcarchiveurl
364    
365    3. New boot files!
366       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
367    
368    ----------------------------------------------------------------------
369    Name: Matthias Blume
370    Date: 2000/07/11 13:58:00 JST
371    Tag: blume-20000711-doctypo
372    Description:
373    
374    Fixed a few typos in CM manual.
375    
376    ----------------------------------------------------------------------
377    Name: Allen Leung
378    Date: 2000/06/15 00:38:00
379    Tag: leunga-20000704-sparc-x86
380    
381    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
382    2. fix to x86 RET bug [found by Dan Grossman]
383    3. sparc assembly bug fix for ticc instructions [found by Fermin]
384    
385       Affects c-- and moby only
386    
387    ----------------------------------------------------------------------
388    Name: Matthias Blume
389    Date: 2000/07/04 15:26:00
390    Tag: blume-20000704-trigger
391    Description:
392    
393    1. Improvements to CM manual.
394    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
395       of getting a back-trace.  The function, when called, raises an
396       internal exception which explicitly carries the full back-trace history,
397       so it is unaffected by any intervening handle-raise pairs ("trivial"
398       or not).  The interactive loop will print that history once it arrives
399       at top level.
400       Short of having all exceptions implicitly carry the full history, the
401       recommended way of using this facility is:
402         - compile your program with instrumentation "on"
403         - run it, when it raises an exception, look at the history
404         - if the history is "cut off" because of some handler, go and modify
405           your program so that it explicitly calls BTrace.trigger
406         - recompile (still instrumented), and rerun; look at the full history
407    
408    ----------------------------------------------------------------------
409    Name: Matthias Blume
410    Date: 2000/07/03 15:36:00 JST
411    Tag: blume-20000702-manual
412    Description:
413    
414    Small corrections and updates to CM manual.
415    
416    ----------------------------------------------------------------------
417    Name: Matthias Blume
418    Date: 2000/06/29 16:04:00 JST
419    Tag: blume-20000629-yacctool
420    Description:
421    
422    Changes:
423    
424    1. Class "mlyacc" now takes separate arguments to pass options to
425       generated .sml- and .sig-files independently.
426    2. Corresponding CM manual updates.
427    3. BTrace module now also reports call sites.  (However, for loop clusters
428       it only shows from where the cluster was entered.)  There are associated
429       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
430    
431    ----------------------------------------------------------------------
432    Name: Matthias Blume
433    Date: 2000/06/27 16:51:00 JST
434    Tag: blume-20000627-noweb
435    Description:
436    
437    Changes:
438    
439     1. Implemented "subdir" and "witness" options for noweb tool.
440        This caused some slight internal changes in CM's tool implementation.
441     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
442        some remaining issues from earlier path anchor changes.
443     3. Updated CM manual accordingly.
444    
445     4. Changed implementation of back-tracing so that I now consider it
446        ready for prime-time.
447    
448        In particular, you don't have to explicitly trigger the back-trace
449        anymore.  Instead, if you are running BTrace-instrumented code and
450        there is an uncaught exception (regardless of whether or not it was
451        raised in instrumented code), the top-level evalloop will print
452        the back-trace.
453    
454        Features:
455    
456          - Instrumented and uninstrumented code work together seemlessly.
457            (Of course, uninstrumented code is never mentioned in actual
458             back-traces.)
459    
460          - Asymptotic time- and space-complexity of instrumented code is
461            equal to that of uninstrumented code.  (This means that
462            tail-recursion is preserved by the instrumentation phase.)
463    
464          - Modules whose code has been instrumented in different sessions
465            work together without problem.
466    
467          - There is no penalty whatsoever on uninstrumented code.
468    
469          - There is no penalty on "raise" expressions, even in
470            instrumented code.
471    
472        A potential bug (or perhaps it is a feature, too):
473    
474          A back-trace reaches no further than the outermost instrumented
475          non-trivial "raise".  Here, a "trivial" raise is one that is the
476          sole RHS of a "handle" rule.  Thus, back-traces reach trough
477    
478               <exp> handle e => raise e
479    
480          and even
481    
482               <exp> handle Foo => raise Bar
483    
484          and, of course, through
485    
486               <exp> handle Foo => ...
487    
488         if the exception was not Foo.
489    
490         Back-traces always reach right through any un-instrumented code
491         including any of its "handle" expressions, trivial or not.
492    
493       To try this out, do the following:
494    
495         - Erase all existing binfiles for your program.
496           (You may keep binfiles for those modules where you think you
497            definitely don't need back-tracing.)
498         - Turn on back-trace instrumentation:
499              SMLofNJ.Internals.BTrace.mode (SOME true);
500         - Recompile your program.  (I.e., run "CM.make" or "use".)
501         - You may now turn instrumentation off again (if you want):
502              SMLofNJ.Internals.BTrace.mode (SOME false);
503         - Run your program as usual.  If it raises an exception that
504           reaches the interactive toplevel, then a back-trace will
505           automatically be printed.  After that, the toplevel loop
506           will print the exception history as usual.
507    
508    ----------------------------------------------------------------------
509    Name: Matthias Blume
510    Date: 2000/06/26 09:56:46 JST
511    Tag: blume-20000626-setup
512    Description:
513    
514    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
515          ML code before and after compiling a file (e.g., to set compiler
516          flags)
517    
518    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
519              - associated changes to btrace.sml (BTrace instrumentation pass)
520              - cleaner implementation of btimp.sml (BTrace tracing and report
521                module)
522    
523    CM manual: * new path encoding documented
524               * description of setup-parameter to "sml" added
525    
526    The biggest user-visible change to back-tracing is that it is no
527    longer necessary to compile all traced modules within the same
528    session.  (This was a real limitation.)
529    
530    ----------------------------------------------------------------------
531    Name: Matthias Blume
532    Date: 2000/06/24 12:40:00 JST
533    Tag: blume-20000624-startup
534    Description:
535    
536    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
537    bit too often -- to put it mildly. :)
538    
539    ----------------------------------------------------------------------
540    Name: Matthias Blume
541    Date: 2000/06/23 18:20:00 JST
542    Tag: blume-20000623-btrace
543    Description:
544    
545    This updates adds a backtrace facility to aid programmers in debugging
546    their programs.  This involves the following changes:
547    
548    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
549       keeping track of the current call stack.  When programs are compiled
550       in a special mode, the compiler will insert calls to these hooks
551       into the user program.
552       "Hook" means that it is possible for different implementations of
553       back-tracing to register themselves (at different times).
554    
555    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
556       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
557       It can be turned on using this call:
558         SMLofNJ.Internals.BTrace.mode (SOME true);
559       Turning it off again:
560         SMLofNJ.Internals.BTrace.mode (SOME false);
561       Querying the current status:
562         SMLofNJ.Internals.BTrace.mode NONE;
563       Annotated programs are about twice as big as normal ones, and they
564       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
565       where all hooks do nothing).  The slowdown with a plugin that is
566       actually useful (such as the one supplied by default) is even greater,
567       but in the case of the default plugin it is still only an constant
568       factor (amortized).
569    
570    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
571       with a sub-structure BTrace for controlling back-tracing.  In particular,
572       the above-mentioned function "mode" controls whether the annotation
573       phase is invoked by the compiler.  Another important function is
574       "trigger": when called it aborts the current execution and causes
575       the top-level loop to print a full back-trace.
576    
577    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
578       for back-tracing.  It keeps track of the dynamic call stack and in
579       addition to that it keeps a partial history at each "level" of that
580       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
581       d and b (at separate times, dynamically), then the report will show:
582    
583       GOTO   d
584             /c
585       GOTO  \b
586       CALL   a
587    
588       This shows that there was an initial non-tail call of a, then a
589       tail-call to b or c, looping behavior in a cluster of functions that
590       consist of b and c, and then a goto from that cluster (i.e., either from
591       b or from c) to d.
592    
593       Note that (depending on the user program) the amount of information
594       that the back-trace module has to keep track of at each level is bounded
595       by a constant.  Thus, the whole implementation has the same asymptotical
596       complexity as the original program (both in space and in time).
597    
598    5. compiler/TopLevel/interact/evalloop.sml has been modified to
599       handle the special exception SMLofNJ.Internals.BTrace.BTrace
600       which is raised by the "trigger" function mentioned above.
601    
602    Notes on usage:
603    
604    - Annotated code works well together with unannotated code:
605    Unannotated calls simply do not show up at all in the backtrace.
606    
607    - It is not a good idea to let modules that were annotated during
608    different sessions run at the same time.  This is because the compiler
609    chooses small integers to identify individual functions, and there
610    will be clashes if different modules were compiled in separate sessions.
611    (Nothing will crash, and you will even be told about the clashes, but
612    back-trace information will in general not be useful.)
613    
614    - Back-tracing can be confused by callcc and capture.
615    
616    - The only way of getting a back-trace right now is to explicitly
617    invoke the "trigger" function from your user program.  Eventually, we
618    should make every exception carry back-trace information (if
619    available).  But since this creates more overhead at "raise"-time
620    (similar to the current exnHistory overhead), I have not yet
621    implemented this.  (The implementation will be rather easy.)  With
622    exceptions carrying back-trace information, this facility will be even
623    more useful because users don't need to modify their programs...
624    
625    - While it is possible to compile the compiler with back-trace
626    annotations turned on (I did it to get some confidence in
627    correctness), you must make absolutely sure that core.sml and
628    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
629    be compiled with annotation because there is no core access yet, but
630    if you compile btimp.sml with annotation, then the system will go into
631    an infinite recursion and crash.)
632    Since CM currently does not know about BTrace, the only way to turn
633    annotations on and off for different modules of the compiler is to
634    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
635    this is awkward and clumsy.
636    
637    Sample sessions:
638    
639    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
640    - SMLofNJ.Internals.BTrace.mode (SOME true);
641    [autoloading]
642    [autoloading done]
643    val it = false : bool
644    - structure X = struct
645    -     fun main n = let
646    -         fun a (x, 0) = d x
647    -           | a (x, n) = b (x, n - 1)
648    -         and b (x, n) = c (x, n)
649    -         and c (x, n) = a (x, n)
650    -         and d x = e (x, 3)
651    -         and e (x, 0) = f x
652    -           | e (x, n) = e (x, n - 1)
653    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
654    -           | f n = n * g (n - 1)
655    -         and g n = a (n, 3)
656    -     in
657    -         f n
658    -     end
659    - end;
660    structure X : sig val main : int -> int end
661    - X.main 3;
662    *** BACK-TRACE ***
663    GOTO   stdIn:4.2-13.20: X.main[2].f
664    GOTO-( stdIn:4.2-13.20: X.main[2].e
665    GOTO   stdIn:4.2-13.20: X.main[2].d
666         / stdIn:4.2-13.20: X.main[2].a
667         | stdIn:4.2-13.20: X.main[2].b
668    GOTO-\ stdIn:4.2-13.20: X.main[2].c
669    CALL   stdIn:4.2-13.20: X.main[2].g
670    GOTO   stdIn:4.2-13.20: X.main[2].f
671    GOTO-( stdIn:4.2-13.20: X.main[2].e
672    GOTO   stdIn:4.2-13.20: X.main[2].d
673         / stdIn:4.2-13.20: X.main[2].a
674         | stdIn:4.2-13.20: X.main[2].b
675    GOTO-\ stdIn:4.2-13.20: X.main[2].c
676    CALL   stdIn:4.2-13.20: X.main[2].g
677    GOTO   stdIn:4.2-13.20: X.main[2].f
678    GOTO-( stdIn:4.2-13.20: X.main[2].e
679    GOTO   stdIn:4.2-13.20: X.main[2].d
680         / stdIn:4.2-13.20: X.main[2].a
681         | stdIn:4.2-13.20: X.main[2].b
682    GOTO-\ stdIn:4.2-13.20: X.main[2].c
683    CALL   stdIn:4.2-13.20: X.main[2].g
684    GOTO   stdIn:4.2-13.20: X.main[2].f
685    CALL   stdIn:2.15-17.4: X.main[2]
686    -
687    
688    (Note that because of a FLINt bug the above code currently does not
689    compile without BTrace turned on.)
690    
691    Here is another example, using my modified Tiger compiler:
692    
693    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
694    - SMLofNJ.Internals.BTrace.mode (SOME true);
695    [autoloading]
696    [autoloading done]
697    val it = false : bool
698    - CM.make "sources.cm";
699    [autoloading]
700    ...
701    [autoloading done]
702    [scanning sources.cm]
703    [parsing (sources.cm):parse.sml]
704    [creating directory CM/SKEL ...]
705    [parsing (sources.cm):tiger.lex.sml]
706    ...
707    [wrote CM/sparc-unix/semant.sml]
708    [compiling (sources.cm):main.sml]
709    [wrote CM/sparc-unix/main.sml]
710    [New bindings added.]
711    val it = true : bool
712    - Main.compile ("../testcases/merge.tig", "foo.out");
713    *** BACK-TRACE ***
714    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
715    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
716    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
717    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
718    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
719    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
720    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
721         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
722    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
723    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
724    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
725    CALL   main.sml:1.18-118.4: Main.compile[2]
726    -
727    
728    ----------------------------------------------------------------------
729    Name: Matthias Blumen
730    Date: 2000/06/21 18:00:00 JST
731    Tag: blume-20000621-manual
732    Description:
733    
734    CM manual update: Path environments documented.
735    
736    ----------------------------------------------------------------------
737    Name: Matthias Blume
738    Date: 2000/06/19 13:40:00
739    Tag: blume-20000619-manual
740    Description:
741    
742    CM manual and system/README update.  This only covers the fact that
743    there are no more implicit anchors.  (Path environments and the "bind"
744    option to "cm" have yet to be documented.)
745    
746    ----------------------------------------------------------------------
747    Name: Matthias Blume
748    Date: 2000/06/19 11:05:00 JST
749    Tag: blume-20000619-chdir-bugfix
750    Description:
751    
752    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
753    
754    ----------------------------------------------------------------------
755    Name: Matthias Blume
756    Date: 2000/06/18 22:00:10 JST
757    Tag: blume-20000618-implicit-anchors-really-gone
758    Description:
759    
760    I updates the previous HISTORY entry where I forgot to mention that
761    implicit anchors are no longer with us.
762    
763    The current update also gets rid of the (now useless) controller
764    CM.Control.implicit_anchors.
765    
766    ----------------------------------------------------------------------
767    Name: Matthias Blume
768    Date: 2000/06/16 17:30:00 JST
769    Tag: blume-20000616-anchorenv
770    Description:
771    
772    This patch implements the long anticipated (just kidding :) "anchor
773    environment" mechanism.  In the course of doing this, I also
774    re-implemented CM's internal "SrcPath" module from scratch.  The new
775    one should be more robust in certain boundary cases.  In any case, it
776    is a lot cleaner than its predecessor (IMHO).
777    
778    This time, although there is yet another boot file format change, I
779    kept the unpickler backward-compatible.  As a result, no new bootfiles
780    are necessary and bootstrapping is straightforward.  (You cannot read
781    new bootfiles into an old system, but the other way around is no
782    problem.)
783    
784    Visible changes:
785    
786    ** 0. Implicit path anchors (without the leading $-symbol) are no
787    longer recognized at all. This means that such path names are not
788    illegal either.  For example, the name basis.cm simply refers to a
789    local file called "basis.cm" (i.e, the name is an ordinary path
790    relative to .cm-files directory).  Or, to put it differently, only
791    names that start with $ are anchored paths.
792    
793    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
794    vanished.
795    
796    John (Reppy) had critizised this as soon as I originally proposed and
797    implemented it, but at that time I did not really deeply believe
798    him. :) Now I came full-circle because I need the $<singlearc> syntax
799    in another place where it cannot be seen as an abbreviation for
800    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
801    seems to mean (i.e., it "expands" into the corresponding anchor
802    value).
803    
804    However, when paths are used as members in CM description files, it
805    continues to be true that there must be at least another arc after the
806    anchor.  This is now enforced separately during semantic analysis
807    (i.e., from a lexical/syntactical point of view, the notation is ok.)
808    
809    ** 2. The "cm" class now accepts an option "bind".  The option's value
810    is a sub-option list of precisely two items -- one labeled "anchor"
811    and the other one labeled "value".  As you might expect, "anchor" is
812    used to specify an anchor name to be bound, and "value" specifies what
813    the anchor is being bound to.
814    
815    The value must be a directory name and can be given in either standard
816    syntax (including the possibility that it is itself an anchored path)
817    or native syntax.
818    
819    Examples:
820    
821       foo.cm (bind:(anchor:bar value:$mystuff/bar))
822       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
823    
824    and so on.
825    
826    The meaning of this is that the .cm-file will be processed with an
827    augmented anchor environment where the given anchor(s) is/are bound to
828    the given values(s).
829    
830    The rationale for having this feature is this: Suppose you are trying
831    to use two different (already stable) libraries a.cm and b.cm (that
832    you perhaps didn't write yourself).  Further, suppose each of these
833    two libraries internally uses its own auxiliary library $aux/lib.cm.
834    Normally you would now have a problem because the anchor "lib" can not
835    be bound to more than one value globally.  Therefore, the project that
836    uses both a.cm and b.cm must locally redirect the anchor to some other
837    place:
838    
839       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
840       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
841    
842    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
843    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
844    
845    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
846    you when you do so at the time of CM.stabilize).  Therefore, you can
847    also use an anchored path as the value:
848    
849      a.cm (bind:(anchor:lib value:$a-lib))
850      b.cm (bind:(anchor:lib value:$b-lib))
851    
852    Now you can globally configure (using the usual CM.Anchor.anchor or
853    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
854    itself is always locally bound, setting it globally is no longer
855    meaningful or necessary (but it does not hurt either).  In fact, "lib"
856    can still be used as a global anchor for separate purposes.  As a
857    matter of fact, one can locally define "lib" in terms of a global
858    "lib":
859    
860      a.cm (bind:(anchor:lib value:$lib/a))
861      b.cm (bind:(anchor:lib value:$lib/b))
862    
863    ** 3: The encoding of path names has changed.  This affects the way
864    path names are shown in CM's progress report and also the internal
865    protocol encoding used for parallel make.
866    
867    The encoding now uses one or more ':'-separated segments.  Each
868    segments corresponds to a file that has been specified relative to the
869    file given by its preceding segment.  The first segment is either
870    relative to the CWD, absolute, or anchored.  Each segment itself is
871    basically a Unix pathname; all segments but the first are relative.
872    
873    Example:
874    
875       $foo/bar/baz.cm:a/b/c.sml
876    
877    This path denotes the file bar/a/b/c.sml relative to the directory
878    denoted by anchor "foo".  Notice that the encoding also includes
879    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
880    paths are resolved relative to the .cm-files directory, so baz.cm must
881    be ignored to get the "real" pathname.
882    
883    To make this fact more obvious, CM puts the names of such "virtual
884    arcs" into parentheses when they appear in progress reports. (No
885    parentheses will appear in the internal protocol encoding.)  Thus,
886    what you really see is:
887    
888      $foo/bar/(baz.cm):a/b/c.sml
889    
890    I find this notation to be much more informative than before.
891    
892    Another new feature of the encoding is that special characters
893    including parentheses, colons, (back)slashes, and white space are
894    written as \ddd (where ddd is the decimal encoding of the character).
895    
896    *** The CM manual still needs to be updated.
897    
898    ----------------------------------------------------------------------
899    Name: Allen Leung
900    Date: 2000/06/15 00:38:00
901    Tag: leunga-20000615-x86-peephole
902    
903    x86 Peephole fix by Fermin.  Affects c-- and moby only.
904    
905    ----------------------------------------------------------------------
906    Name: Matthias Blume
907    Date: 2000/06/12 11:40:00
908    Tag: blume-20000612-parmakefix
909    Description:
910    
911    More cleanup after changing the file naming scheme: This time I
912    repaired the parallel make mechanism for CMB.make which I broke earlier.
913    
914    ----------------------------------------------------------------------
915    Name: Allen Leung
916    Date: 2000/06/09 01:25:00
917    Tag: leunga-20000609-various
918    
919    None of these things should affect normal SML/NJ operations
920    
921    1. Peephole improvements provided by Fermin (c--)
922    2. New annotation DEFUSE for adding extra dependence (moby)
923    3. New X86 LOCK instructions (moby)
924    4. New machine description language for reservation tables (scheduling)
925    5. Fixes to various optimization/analysis modules (branch chaining, dominator
926       trees etc.)
927    6. I've changed the CM files so that they can work with versions
928       110.0.6, 110.25 and 110.28
929    
930    ----------------------------------------------------------------------
931    Name: Matthias Blume
932    Date: 2000/06/09 12:40:00
933    Tag: blume-20000609-log
934    Description:
935    
936    - Removed all(?) remaining RCS Log entries from sources.
937    
938    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
939      anchored paths).
940    
941    ----------------------------------------------------------------------
942    Name: Matthias Blume
943    Date: 2000/06/07 17:00:00 JST
944    Tag: blume-20000607-no-implicit-anchors
945    Description:
946    
947    1. This update changes the default setting for
948    CM.Control.implicit_anchors from true to false.  This means that
949    implicit anchors are no longer permitted by default.  I also tried to
950    make sure that nothing else still relies on implicit anchors.
951    (This is the next step on the schedule towards a CM that does not even
952    have the notion of implicit anchors anymore.)
953    
954    2. More CM manual updates.
955    
956    3. I managed to track down and fix the pickling bug I mentioned last
957    time.  Because of the previously existing workaround, this entails no
958    immediate practical changes.
959    
960    ----------------------------------------------------------------------
961    Name: Matthias Blume
962    Date: 2000/06/06 11:15:00 JST
963    Tag: blume-20000606-lazierpickle
964    Description:
965    
966    !!!! NEW BOOT FILES !!!!
967    
968    * The main purpose of this update is to make library pickles lazier in
969    order to reduce the initial space penalty for autoloading a library.
970    As a result, it is now possible to have $smlnj/compiler.cm
971    pre-registered.  This should take care of the many complaints or
972    inquiries about missing structure Compiler.  This required changes to
973    CM's internal data structures and small tweaks to some algorithms.
974    
975    As a neat additional effect, it is no longer necessary (for the sake
976    of lean heap image files) to distinguish between a "minimal" CM and a
977    "full" CM.  Now, there is only one CM (i.e., the "full" version:
978    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
979    interactive top level. ($smlnj/cm/minimal.cm is gone.)
980    
981    To make the life of compiler-hackers easier, "makeml" now also
982    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
983    words, after you bootstrap a new sml for the first time, you will not
984    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
985    around you will still have to do it, though.)
986    
987    * A second change consists of major updates to the CM manual.  There
988    are now several appendices with summary information and also a full
989    specification of the CM description file syntax.
990    
991    * In directory src/system I added the script "allcross".  This script
992    invokes sml and cross-compiles the compiler for all supported
993    architectures.  (Useful when providing a new set of boot files.)
994    
995    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
996    added a small tweak to pickle-util.sml to work around this problem,
997    but it is not a proper fix yet.  I will investigate further.  (The
998    effect of the bug was an inflation of library pickle size.)
999    
1000    * Version number increased to 110.28.1 (to avoid compatibility problems).
1001    
1002    ----------------------------------------------------------------------
1003    Name: Allen Leung
1004    Date: 2000/05/25 17:28 EDT
1005    Tag: leunga-20000525-ra
1006    Description:
1007    
1008      Fixed a bug in freezing phase of the register allocator.
1009    
1010    ----------------------------------------------------------------------
1011    Name: Allen Leung
1012    Date: 2000/05/15 22:53 EDT
1013    Tag: leunga-20000515-alpha-x86-ra
1014    Description:
1015    
1016      1. Alpha
1017    
1018          Slight cleanup.  Removed the instruction SGNXL
1019    
1020      2. X86
1021    
1022          Added the following instructions to the instruction set:
1023    
1024            ROLx, RORx,
1025            BTx, BTSx, BTLx, BTRx,
1026            XCHGx, and variants with the LOCK prefix
1027    
1028      3. Register Allocation
1029    
1030          The module ra-rewrite-with-renaming has been improved.
1031    
1032      These have no effect on SML/NJ.
1033    
1034    ----------------------------------------------------------------------
1035    Name: Matthias Blume
1036    Date: 2000/05/15 16:20:00 JST
1037    Tag: blume-20000515-lightrebuild
1038    Description:
1039    
1040    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1041       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1042       defined.  In effect, the command will not build any cross-compiler
1043       backends and therefore finish more quickly.
1044    
1045       The "fixpt" script also takes a "-light" switch to be able to use
1046       this new facility while compiling for a fixpoint.
1047    
1048    2. I replaced all mentions of anchored paths in group owner specifications
1049       with simple relative paths (usually starting with "..").
1050       The rationale is that a library's internal workings should not be
1051       compromised by the lack of some anchor.  (An anchor is necessary
1052       for someone who wants to refer to the library by an anchored path,
1053       but it should not be necessary to build the same library in the first
1054       place.)
1055    
1056    3. I changed the way CM's tool mechanism determines the shell command
1057       string used for things like ml-yacc etc. so that it does not break
1058       when CM.Control.implicit_anchors is turned off.
1059    
1060    ----------------------------------------------------------------------
1061    Name: Matthias Blume
1062    Date: 2000/05/12 18:20:00 JST
1063    Tag: blume-20000512-ml-build
1064    Description:
1065    
1066    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1067    getting installed properly (by config/install.sh).
1068    
1069    ----------------------------------------------------------------------
1070    Name: Matthias Blume
1071    Date: 2000/05/12 17:30:00 JST
1072    Tag: blume-20000512-anchors
1073    Description:
1074    
1075    !!! NEW BOOT FILES !!!
1076    
1077    This change is in preparation of fading out support for "implicitly
1078    anchored path names".  I went through all sources and used the
1079    explicit (and relatively new) $-notation.  See system/README and the
1080    CM manual for more info on this.
1081    
1082    I also modified the anchoring scheme for some things such as "smlnj",
1083    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1084    anchors are more expressive: anchor name and first arc do not have to
1085    coincide.  This entails the following user-visible change:
1086    
1087    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1088    particular, when you fire up sml with a command-line argument, say,
1089    e.g.:
1090    
1091       sml '$smlnj/cmb.cm'
1092    
1093    At the ML toplevel prompt:
1094    
1095       CM.autoload "$smlnj/cmb.cm";
1096    
1097    There is also a new controller in CM.Control that can be used to turn
1098    off all remaining support for implicit anchors by saying:
1099    
1100        CM.autoload "$smlnj/
1101        #set CM.Control.implicit_anchors false;
1102    
1103    This causes CM to reject implicitly anchored paths.  This is (for the
1104    time being) less permissive than the "final" version where there will
1105    be no more such implicit anchors and relative paths will be just that:
1106    relative.
1107    
1108    The next step (version after next version?) will be to make the
1109    default for CM.Control.implicit_anchors false.  After the dust has
1110    settled, I can then produce the "final" version of this...
1111    
1112    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1113    
1114    ----------------------------------------------------------------------
1115    Name: Matthias Blume
1116    Date: 2000/05/11 16:30:00 JST
1117    Tag: blume-20000511-sources
1118    Description:
1119    
1120    The main change is that I added function CM.sources as a generalized
1121    version of the earlier CM.makedepend.  This entails the following
1122    additional changes:
1123    
1124      - CM.makedepend has been dropped.
1125    
1126      - CM manual has been updated.
1127    
1128      - TOOLS signature and API have been changed.
1129    
1130    ----------------------------------------------------------------------
1131    Name: Allen Leung
1132    Date: 2000/05/10 21:17 EDT
1133    Tag: leunga-20000510-moby-c--ssa
1134    Description:
1135    
1136      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1137    None of these affect SML/NJ.
1138    
1139    1. Register Allocation
1140    
1141        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1142           This module tries to remove local (i.e. basic block level) redundancies
1143           during spilling.
1144    
1145        b. A new framework for performing region based register allocation.
1146           Not yet entirely functional.
1147    
1148    2. X86
1149    
1150       a. DefUse for POP was missing the stack pointer [found by Lal]
1151       b. Reload for CALL was incorrect in X86Spill [found by John]
1152       c. Various fixes in X86Spill so that it can be used correctly for
1153          the new spilling module.
1154    
1155    3. SSA/IR
1156    
1157       a. New module ir/dj-dataflow.sml implements elimination based
1158          data flow analysis.
1159    
1160    4. MLRiscGen
1161    
1162       a. Fix for gc type annotation
1163    
1164    5. MDGen
1165    
1166       Various fixes for machine description -> ml code translation.  For ssa
1167       only.
1168    
1169    ----------------------------------------------------------------------
1170    Name: Allen Leung
1171    Date: 2000/05/08 22:17 EDT
1172    Tag: leunga-20000508-labexp
1173    Description:
1174    
1175      Fermin has found a few assembly problems with constant expressions
1176      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1177      which choke on dumb assemblers.  This is his fix.
1178    
1179    ----------------------------------------------------------------------
1180    Name: Dave MacQueen
1181    Date: 2000/04/09 14:00 EDT
1182    Tag: dbm-20000502-Version_110_28
1183    Description:
1184    
1185    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1186    
1187    2. Updated config/version to 110.28
1188    
1189    3. Updated config/srcarchiveurl
1190    
1191    3. New boot files!
1192       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1193    
1194    ----------------------------------------------------------------------
1195    Name: Matthias Blume
1196    Date: 2000/05/01 19:05:00 JST
1197    Tag: blume-20000501-noweb
1198    Description:
1199    
1200    A new noweb tool has been added.  The existing system is entirely
1201    unaffected by this, but some CM users have asked for renewed noweb
1202    support.  Everything is documented in the CM manual.
1203    
1204    New (plugin) libraries:
1205    
1206       noweb-tool.cm
1207       nw-ext.cm
1208    
1209    ----------------------------------------------------------------------
1210    Name: Dave MacQueen
1211    Date: 2000/04/30 12:40PM EDT
1212    Tag: dbm-20000430-bug_fixes
1213    Description:
1214    
1215    1. Fix for bug 1498
1216       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1217       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1218         added toRealArray function
1219       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1220         added check for tag Obj.RealArray to array printing case in ppObj
1221    
1222    2. Fix for bug 1510
1223       smlnj/src/compiler/Semant/types/typesutil.sml
1224         fixed definition of dummyargs (used by equalTycon) so that
1225         dummy args are distinct types
1226    
1227    ----------------------------------------------------------------------
1228    Name: Matthias Blume
1229    Date: 2000/04/30 01:00:00 JST
1230    Tag: blume-20000430-versions
1231    Description:
1232    
1233    1. CM version numbering added.  This is an implementation of Lal's
1234       proposal for adding version numbers and version checking to .cm
1235       files.  Lal said that his proposal was just that -- a proposal.
1236       For the time being I went ahead and implemented it so that people
1237       can comment on it.  Everything is completely backward-compatible
1238       (except for the stable library format, i.e., new bootfiles!).
1239    
1240       As usual, see the CM manual for details.
1241    
1242    2. An alternative syntax for anchored paths has been implemented.
1243       Dave has recently voiced the same concerns that I had when I did
1244       this, so there should be some support.  My take is that eventually
1245       I will let support for the current syntax (where anchors are
1246       "implicit") fade out in favor of the new, explicit syntax.
1247       In order to be backward-compatible, both old and new syntax are
1248       currently supported.
1249    
1250       Again, see the CM manual for details.
1251    
1252    3. Parallel make is trying to be slightly smarter:  When the master
1253       process finds a "bottleneck", i.e., when there is only one
1254       compilation unit that can be compiled and everybody else is
1255       waiting on it, then it will simply compile it directly instead
1256       of clumsily telling one of the slaves to do it.
1257    
1258    4. Support for "unsharing" added.  This is necessary in order to be
1259       able to have two different versions of the same library running
1260       at the same time (e.g., for trying out a new MLRISC while still
1261       having the old MLRISC linked into the current compiler, etc.)
1262       See the CM manual.
1263    
1264    5. Simple "makedepend" functionality added for generating Makefile
1265       dependency information.  (This is rather crude at the moment.
1266       Expect some changes here in the future.)
1267    
1268    6. ".fun" added as a recognized suffix for ML files. Also documented
1269       explicitly in the manual that the fallback behavior (unknown suffix
1270       -> ML file) is not an official feature!
1271    
1272    7. Small changes to the pickler for stable libraries.
1273    
1274    8. Several internal changes to CM (for cleanup/improvement).
1275    
1276    
1277    !!!! NEW BINFILES !!!!
1278    
1279    ----------------------------------------------------------------------
1280    Name: Matthias Blume
1281    Date: 2000/04/28 17:30:00 JST
1282    Tag: blume-20000428-pathconfig
1283    Description:
1284    
1285    1. I changed config/install.sh to remove duplicate entries from the
1286       lib/pathconfig file at the end.  Moreover, the final version of
1287       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1288       in src/system/installml.
1289    
1290    2. The config/install.sh script now consistently uses relative
1291       pathnames in lib/pathconfig whenever the anchor is in the lib
1292       directory.  (So far this was true for the libraries that come
1293       pre-compiled and bundled as part of the bootfiles but not for
1294       libraries that are compiled by the script itself.)
1295    
1296    ----------------------------------------------------------------------
1297    Name: Matthias Blume
1298    Date: 2000/04/26 13:10:00 JST
1299    Tag: blume-20000426-fun_suffix
1300    Description:
1301    
1302    Added ".fun" as a recognized file name suffix (for ML code).
1303    
1304    ----------------------------------------------------------------------
1305    Name: Allen Leung
1306    Date: 2000/04/25 17:00:00 EST
1307    Tag: leunga-20000425-alpha-ra
1308    Description:
1309    
1310    1. Alpha
1311    
1312        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1313    in C--.
1314    
1315    2. RA
1316    
1317       Added a flag "ra-dump-size" to print out the size of the flowgraph
1318       and the interference graph.
1319    
1320    ----------------------------------------------------------------------
1321    Name: Dave MacQueen
1322    Date: 2000/04/25/
1323    Tag: dbm-20000425-mlyacc_doc_examples
1324    Description:
1325      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1326      Updated all three examples in src/ml-yacc/examples to run
1327      under 110.* using CM.make.
1328    
1329    ----------------------------------------------------------------------
1330    Name: Allen Leung
1331    Date: 2000/04/20 23:04:00 EST
1332    Tag: leunga-20000420-ssa-c---stuff
1333    Description:
1334    
1335      This update synchronizes my repository with Yale's.  Most of these
1336    changes, however, do not affect SML/NJ at all (the RA is an exception).
1337    
1338    1. Register Allocator
1339    
1340       a. An improvement in the interference graph construction:
1341          Given a copy
1342    
1343                s <- t
1344    
1345          no interference edge between s and t is added for this definition of s.
1346    
1347       b. I've added two new spill heuristic modules that Fermin and I developed
1348          (in the new library RA.cm). These are unused in SML/NJ but maybe
1349          useful for others (Moby?)
1350    
1351    2. X86
1352    
1353       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1354    
1355    3. Alpha
1356    
1357       a. Added the BSR instruction and code generation that goes with it [C--]
1358       b. Other fixes too numerous to recount provided by Fermin [C--]
1359    
1360    4. Regmaps
1361    
1362       a. The regmaps are not initialized with the identity physical bindings
1363          at creation time.  This is unneeded.
1364    
1365    5. MLRISC Optimizations
1366    
1367       a. The DJ-Graph module can now compute the iterated dominance frontiers
1368          intersects with liveness incrementally in linear time! Woohoo!
1369          This is now used in my new SSA construction algorithm.
1370    
1371       b. THe branch reorganization module is now smarter about linear chains of
1372          basic blocks.
1373    
1374    
1375    ----------------------------------------------------------------------
1376    Name: Matthias Blume
1377    Date: 2000/04/12 13:52:00 JST
1378    Tag: blume_main_v110p27_1
1379    Description:
1380    
1381    Changed install.sh script to handle archive files without version number
1382    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1383    name of the boot file archive.
1384    
1385    ----------------------------------------------------------------------
1386    Name: Dave MacQueen
1387    Date: 2000/04/09 14:00 EDT
1388    Tag: dbm-20000410-Version_110_27
1389    Description:
1390    
1391    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1392    
1393    2. Updated src/config/version to 110.27
1394    
1395    3. New boot files!
1396    
1397    ----------------------------------------------------------------------
1398    Name: Allen Leung
1399    Date: 2000/04/09 19:09:00 EST
1400    Tag: leunga-20000409-misc
1401    Description:
1402    
1403    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1404    
1405    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1406    
1407    ----------------------------------------------------------------------
1408    Name: Stefan
1409    Date: 2000/04/07 10:00:00 EDT
1410    Tag: monnier-20000406-branch-handling
1411    Description:
1412    
1413    Improved handling of branches (mostly those generated from
1414    polymorphic equality), removed switchoff and changed the
1415    default optimization settings (more cpsopt and less flintopt).
1416    
1417    ----------------------------------------------------------------------
1418    Name: Allen Leung
1419    Date: 2000/04/06 01:30:00 EST
1420    Tag: leunga-20000406-peephole-x86-SSA-2
1421    Description:
1422    
1423       Forgot a few files.
1424    
1425    ----------------------------------------------------------------------
1426    Name: Allen Leung
1427    Date: 2000/04/06 00:36:00 EST
1428    Tag: leunga-20000406-peephole-x86-SSA
1429    Description:
1430    
1431    1.  New Peephole code
1432    
1433    2.  Minor improvement to X86 instruction selection
1434    
1435    3.  Various fixes to SSA and machine description -> code translator
1436    
1437    ----------------------------------------------------------------------
1438    Name: Matthias Blume
1439    Date: 2000/04/05 12:30:00 JST
1440    Tag: blume_main_v110p26p2_3
1441    Description:
1442    
1443    This update just merges three minor cosmetic updates to CM's sources
1444    to get ready for the 110.27 code freeze on Friday.  No functionality
1445    has changed.
1446    
1447    ----------------------------------------------------------------------
1448    Name: Allen Leung
1449    Date: 2000/04/04 19:39:00 EST
1450    Tag: leunga-20000404-x86-asm
1451    Description:
1452    
1453    1.  Fixed a problem in X86 assembly.
1454    
1455        Things like
1456    
1457           jmp %eax
1458           jmp (%eax)
1459    
1460        should be output as
1461    
1462           jmp *%eax
1463           jmp *(%eax)
1464    
1465    2.  Assembly output
1466    
1467          Added a new flag
1468    
1469              "asm-indent-copies" (default to false)
1470    
1471          When this flag is on, parallel copies will be indented an extra level.
1472    
1473    ----------------------------------------------------------------------
1474    Name: Allen Leung
1475    Date: 2000/04/04 03:18:00 EST
1476    Tag: leunga-20000404-C--Moby
1477    Description:
1478    
1479        All of these fixes are related to C--, Moby, and my own optimization
1480        stuff; so they shouldn't affect SML/NJ.
1481    
1482    1.  X86
1483    
1484        Various fixes related floating point, and extensions.
1485    
1486    2.  Alpha
1487    
1488        Some extra patterns related to loads with signed/zero extension
1489        provided by Fermin.
1490    
1491    3.  Assembly
1492    
1493        When generating assembly, resolve the value of client defined constants,
1494        instead of generating symbolic values.  This is controlled by the
1495        new flag "asm-resolve-constants", which is default to true.
1496    
1497    4.  Machine Descriptions
1498    
1499        a. The precedence parser was slightly broken when parsing infixr symbols.
1500        b. The type generalizing code had the bound variables reversed, resulting
1501           in a problem during arity raising.
1502        c. Various fixes in machine descriptions.
1503    
1504    ----------------------------------------------------------------------
1505    Name: Matthias Blume
1506    Date: 2000/04/03 16:05:00 JST
1507    Tag: blume_main_v110p26p2_2
1508    Description:
1509    
1510    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1511    now done via the ordinary static environment that is context to each
1512    compilation unit.
1513    
1514    To this end, I arranged that instead of "structure Core" as "structure
1515    _Core" is bound in the pervasive environment.  Core access is done via
1516    _Core (which can never be accidentally rebound because _Core is not a
1517    legal surface-syntax symbol).
1518    
1519    The current solution is much cleaner because the core environment is
1520    now simply part of the pervasive environment which is part of every
1521    compilation unit's context anyway.  In particular, this eliminates all
1522    special-case handling that was necessary until now in order to deal
1523    with dynamic and symbolic parts of the core environment.
1524    
1525    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1526    compilation manager's bootstrap compiler (actually: in the "init group"
1527    handling).  See the comments in src/system/smlnj/init/init.cmi for
1528    more details.
1529    
1530    I also tried to track down all mentions of "Core" (as string argument
1531    to Symbol.strSymbol) in the compiler and replaced them with a
1532    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1533    name appears in one place only.
1534    
1535    Binfile and bootfile format have not changed, but the switchover from
1536    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1537    bootfiles anyway.
1538    
1539    ----------------------------------------------------------------------
1540    Name: Allen Leung
1541    Date: 2000/04/02 21:17:00 EST
1542    Tag: leunga-20000402-mltree
1543    Description:
1544    
1545       1. Renamed the constructor CALL in MLTREE by popular demand.
1546       2. Added a bunch of files from my repository.  These are currently
1547          used by other non-SMLNJ backends.
1548    
1549    ----------------------------------------------------------------------
1550    Name: Allen Leung
1551    Date: 2000/03/31 21:15:00 EST
1552    Tag: leunga-20000331-aliasing
1553    Description:
1554    
1555    This update contains a rewritten (and hopefully more correct) module
1556    for extracting aliasing information from CPS.
1557    
1558       To turn on this feature:
1559    
1560            Compiler.Control.CG.memDisambiguate := true
1561    
1562       To pretty print the region information with assembly
1563    
1564           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1565    
1566       To control how many levels of aliasing information are printed, use:
1567    
1568           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1569    
1570       The default of n is 3.
1571    
1572    ----------------------------------------------------------------------
1573    Name: David MacQueen
1574    Date: 2000/03/31 11:15:00 EST
1575    Tag: dbm-20000331-runtime_fix
1576    Description:
1577    
1578    This update contains:
1579    
1580    1. runtime/c-lib/c-libraries.c
1581       includes added in revision 1.2 caused compilation errors on hppa-hpux
1582    
1583    2. fix for bug 1556
1584       system/Basis/Implementation/NJ/internal-signals.sml
1585    
1586    ----------------------------------------------------------------------
1587    Name: Matthias Blume
1588    Date: 2000/03/31 18:00:00 JST
1589    Tag: blume_main_v110p26p2_1
1590    Description:
1591    
1592    This update contains:
1593    
1594    1. A small change to CM's handling of stable libraries:
1595       CM now maintains one "global" modmap that is used for all stable
1596       libraries.  The use of such a global modmap maximizes sharing and
1597       minimizes the need for re-traversing parts of environments during
1598       modmap construction.  (However, this has minor impact since modmap
1599       construction seems to account for just one percent or less of total
1600       compile time.)
1601    
1602    2. I added a "genmap" phase to the statistics.  This is where I got the
1603       "one percent" number (see above).
1604    
1605    3. CM's new tool parameter mechanism just became _even_ better. :)
1606       - The parser understands named parameters and recursive options.
1607       - The "make" and "shell" tools use these new features.
1608         (This makes it a lot easier to cascade these tools.)
1609       - There is a small syntax change: named parameters use a
1610    
1611           <name> : ( <option> ... )            or
1612           <name> : <string>
1613    
1614         syntax.  Previously, named parameters were implemented in an
1615         ad-hoc fashion by each tool individually (by parsing strings)
1616         and had the form
1617    
1618           <name>=<string>
1619    
1620       See the CM manual for a full description of these issues.
1621    
1622    ----------------------------------------------------------------------
1623    Name: Matthias Blume
1624    Date: 2000/03/30 18:00:00 JST
1625    Tag: blume_main_v110p26p2_0
1626    Description:
1627    
1628    !!!!! WARNING !!!!!!
1629    !!  New binfiles  !!
1630    !!!!!!!!!!!!!!!!!!!!
1631    
1632    This update contains:
1633    
1634    1. Moderate changes to CM:
1635    
1636       - Changes to CM's tools mechanism.  In particular, it is now possible
1637       to have tools that accept additional "command line" parameters
1638       (specified in the .cm file at each instance where the tool's class is
1639       used).
1640    
1641       This was done to accommodate the new "make" and "shell" tools which
1642       facilitate fairly seamless hookup to portions of code managed using
1643       Makefiles or Shell scripts.
1644    
1645       There are no classes "shared" or "private" anymore.  Instead, the
1646       sharing annotation is now a parameter to the "sml" class.
1647    
1648       There is a bit of generic machinery for implementing one's own
1649       tools that accept command-line parameters.  However, I am not yet fully
1650       satisfied with that part, so expect changes here in the future.
1651    
1652       All existing tools are described in the CM manual.
1653    
1654       - Slightly better error handling.  (CM now suppresses many followup
1655       error messages that tended to be more annoying than helpful.)
1656    
1657    2. Major changes to the compiler's static environment data structures.
1658    
1659       - no CMStaticEnv anymore.
1660            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
1661              but it is called Environment), no conversions between different
1662              kinds of static environments
1663    
1664       - There is still a notion of a "modmap", but such modmaps are generated
1665         on demand at the time when they are needed.  This sounds slow, but I
1666         sped up the code that generates modmaps enough for this not to lead to
1667         a slowdown of the compiler (at least I didn't detect any).
1668    
1669       - To facilitate rapid modmap generation, static environments now
1670         contain an (optional) "modtree" structure.  Modtree annotations are
1671         constructed by the unpickler during unpickling.  (This means that
1672         the elaborator does not have to worry about modtrees at all.)
1673         Modtrees have the advantage that they are compositional in the same
1674         way as the environment data structure itself is compositional.
1675         As a result, modtrees never hang on to parts of an environment that
1676         has already been rendered "stale" by filtering or rebinding.
1677    
1678       - I went through many, many trials and errors before arriving at the
1679         current solution.  (The initial idea of "linkpaths" did not work.)
1680         But the result of all this is that I have touched a lot of files that
1681         depend on the "modules" and "types" data structures (most of the
1682         elaborator). There were a lot of changes during my "linkpath" trials
1683         that could have been reverted to their original state but weren't.
1684         Please, don't be too harsh on me for messing with this code a bit more
1685         than what was strictly necessary...  (I _did_ resist the tempation
1686         of doing any "global reformatting" to avoid an untimely death at
1687         Dave's hands. :)
1688    
1689       - One positive aspect of the previous point:  At least I made sure that
1690         all files that I touched now compile without warnings (other than
1691         "polyEqual").
1692    
1693       - compiler now tends to run "leaner" (i.e., ties up less memory in
1694         redundant modmaps)
1695    
1696    ----------------------------------------------------------------------
1697    Name: Allen Leung
1698    Date: 2000/03/29 18:00:00
1699    Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
1700    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
1701    Description:
1702    
1703       This update contains *MAJOR* changes to the way code is generated from CPS
1704    in the module mlriscGen, and in various backend modules.
1705    
1706    CHANGES
1707    =======
1708    
1709    1. MLRiscGen: forward propagation fix.
1710    
1711       There was a bug in forward propagation introduced at about the same time
1712       as the MLRISC x86 backend, which prohibits coalescing to be
1713       performed effectively in loops.
1714    
1715       Effect: speed up of loops in RISC architectures.
1716               By itself, this actually slowed down certain benchmarks on the x86.
1717    
1718    2. MLRiscGen:  forward propagating addresses from consing.
1719    
1720       I've changed the way consing code is generated.  Basically I separated
1721       out the initialization part:
1722    
1723            store tag,   offset(allocptr)
1724            store elem1, offset+4(allocptr)
1725            store elem2, offset+8(allocptr)
1726            ...
1727            store elemn, offset+4n(allocptr)
1728    
1729       and the address computation part:
1730    
1731            celladdr <- offset+4+alloctpr
1732    
1733       and move the address computation part
1734    
1735       Effect:  register pressure is generally lower as a result.  This
1736                makes compilation of certain expressions much faster, such as
1737                long lists with non-trivial elements.
1738    
1739                 [(0,0), (0,0), .... (0,0)]
1740    
1741    3. MLRiscGen: base pointer elimination.
1742    
1743        As part of the linkage mechanism, we generate the sequence:
1744    
1745         L:  ...  <- start of the code fragment
1746    
1747         L1:
1748             base pointer <- linkreg - L1 + L
1749    
1750         The base pointer was then used for computing relocatable addresses
1751       in the code fragment.  Frequently (such as in lots of continuations)
1752       this is not needed.  We now eliminate this sequence whenever possible.
1753    
1754         For compile time efficiency, I'm using a very stupid local heuristic.
1755       But in general, this should be done as a control flow analysis.
1756    
1757       Effect:  Smaller code size.  Speed up of most programs.
1758    
1759    4. Hppa back end
1760    
1761         Long jumps in span dependence resolution used to depend on the existence
1762      of the base pointer.
1763    
1764         A jump to a long label L was expanded into the following sequence:
1765    
1766          LDIL %hi(L-8192), %r29
1767          LDO  %lo(L-8192)(%r29), %r29
1768          ADD  %r29, baseptr, %r29
1769          BV,n %r0(%r29)
1770    
1771         In the presence of change (3) above, this will not work.  I've changed
1772       it so that the following sequence of instructions are generated, which
1773       doesn't mention the base pointer at all:
1774    
1775             BL,n  L', %r29           /* branch and link, L' + 4 -> %r29 */
1776        L':  ADDIL L-(L'+4), %r29     /* Compute address of L */
1777             BV,n  %r0(%r29)          /* Jump */
1778    
1779    5. Alpha back end
1780    
1781          New alpha instructions LDB/LDW have been added, as per Fermin's
1782       suggestions.   This is unrelated to all other changes.
1783    
1784    6. X86 back end
1785    
1786         I've changed andl to testl in the floating point test sequence
1787         whenever appropriate.  The Intel optimization guide states that
1788         testl is preferable to andl.
1789    
1790    7. RA (x86 only)
1791    
1792         I've improved the spill propagation algorithm, using an approximation
1793       of maximal weighted independent sets.   This seems to be necessary to
1794       alleviate the negative effect in light of the slow down in (1).
1795    
1796         I'll write down the algorithm one of these days.
1797    
1798    8. MLRiscGen: frequencies
1799    
1800         I've added an annotation that states that all call gc blocks have zero
1801       execution frequencies.  This improves register allocation on the x86.
1802    
1803    BENCHMARKS
1804    ==========
1805    
1806       I've only perform the comparison on 110.25.
1807    
1808       The platforms are:
1809    
1810        HPPA  A four processor HP machine (E9000) with 5G of memory.
1811        X86   A 300Hhz Pentium II with 128M of memory, and
1812        SPARC An Ultra sparc 2 with 512M of memory.
1813    
1814       I used the following parameters for the SML benchmarks:
1815    
1816                 @SMLalloc
1817         HPPA    256k
1818         SPARC   512k
1819         X86     256k
1820    
1821    COMPILATION TIME
1822    ----------------
1823       Here are the numbers comparing the compilation times of the compilers.
1824       I've only compared 110.25 compiling the new sources versus
1825       a fixpoint version of the new compiler compiling the same.
1826    
1827                     110.25                                  New
1828               Total  Time in RA  Spill+Reload   Total  Time In RA Spill+Reload
1829         HPPA   627s    116s        2684+3584     599s    95s       1003+1879
1830         SPARC  892s    173s        2891+3870     708s    116s      1004+1880
1831         X86    999s    315s       94006+130691   987s    296s    108877+141957
1832    
1833                   110.25         New
1834                Code Size      Code Size
1835         HPPA   8596736         8561421
1836         SPARC  8974299         8785143
1837         X86    9029180         8716783
1838    
1839       So in summary, things are at least as good as before.   Dramatic
1840       reduction in compilation is obtained on the Sparc; I can't explain it,
1841       but it is reproducible.  Perhaps someone should try to reproduce this
1842       on their own machines.
1843    
1844    SML BENCHMARKS
1845    --------------
1846    
1847        On the average, all benchmarks perform at least as well as before.
1848    
1849          HPPA         Compilation Time     Spill+Reload      Run Time
1850                     110.25  New            110.25    New   110.25  New
1851    
1852          barnesHut  3.158  3.015  4.75%    1+1       0+0   2.980  2.922   2.00%
1853              boyer  6.152  5.708  7.77%    0+0       0+0   0.218  0.213   2.34%
1854       count-graphs  1.168  1.120  4.32%    0+0       0+0  22.705 23.073  -1.60%
1855                fft  0.877  0.792 10.74%    1+3       1+3   0.602  0.587   2.56%
1856        knuthBendix  3.180  2.857 11.32%    0+0       0+0   0.675  0.662   2.02%
1857             lexgen  6.190  5.290 17.01%    0+0       0+0   0.913  0.788  15.86%
1858               life  0.803  0.703 14.22%   25+25      0+0   0.153  0.140   9.52%
1859              logic  2.048  2.007  2.08%    6+6       1+1   4.133  4.008   3.12%
1860         mandelbrot  0.077  0.080 -4.17%    0+0       0+0   0.765  0.712   7.49%
1861             mlyacc 22.932 20.937  9.53%  154+181    32+57  0.468  0.430   8.91%
1862            nucleic  5.183  5.060  2.44%    2+2       0+0   0.125  0.120   4.17%
1863      ratio-regions  3.357  3.142  6.84%    0+0       0+0  116.225 113.173 2.70%
1864                ray  1.283  1.290 -0.52%    0+0       0+0   2.887  2.855   1.11%
1865             simple  6.307  6.032  4.56%   28+30      5+7   3.705  3.658   1.28%
1866                tsp  0.888  0.862  3.09%    0+0       0+0   7.040  6.893   2.13%
1867               vliw 24.378 23.455  3.94%  106+127    25+45  2.758  2.707   1.91%
1868      --------------------------------------------------------------------------
1869       Average                     6.12%                                   4.09%
1870    
1871          SPARC        Compilation Time     Spill+Reload      Run Time
1872                     110.25  New            110.25    New   110.25  New
1873    
1874          barnesHut  3.778  3.592  5.20%    2+2       0+0   3.648  3.453    5.65%
1875              boyer  6.632  6.110  8.54%    0+0       0+0   0.258  0.242    6.90%
1876       count-graphs  1.435  1.325  8.30%    0+0       0+0  33.672 34.737   -3.07%
1877                fft  0.980  0.940  4.26%    3+9       2+6   0.838  0.827    1.41%
1878        knuthBendix  3.590  3.138 14.39%    0+0       0+0   0.962  0.967   -0.52%
1879             lexgen  6.593  6.072  8.59%    1+1       0+0   1.077  1.078   -0.15%
1880               life  0.972  0.868 11.90%   26+26      0+0   0.143  0.140    2.38%
1881              logic  2.525  2.387  5.80%    7+7       1+1   5.625  5.158    9.05%
1882         mandelbrot  0.090  0.093 -3.57%    0+0       0+0   0.855  0.728   17.39%
1883             mlyacc 26.732 23.827 12.19%  162+189    32+57  0.550  0.560   -1.79%
1884            nucleic  6.233  6.197  0.59%    3+3       0+0   0.163  0.173   -5.77%
1885      ratio-regions  3.780  3.507  7.79%    0+0       0+0 133.993 131.035   2.26%
1886                ray  1.595  1.550  2.90%    1+1       0+0   3.440  3.418    0.63%
1887             simple  6.972  6.487  7.48%   29+32      5+7   3.523  3.525   -0.05%
1888                tsp  1.115  1.063  4.86%    0+0       0+0   7.393  7.265    1.77%
1889               vliw 27.765 24.818 11.87%  110+135    25+45  2.265  2.135    6.09%
1890      ----------------------------------------------------------------------------
1891       Average                     6.94%                                    2.64%
1892    
1893          X86          Compilation Time     Spill+Reload      Run Time
1894                     110.25  New            110.25    New   110.25  New
1895    
1896          barnesHut  5.530  5.420  2.03%  593+893   597+915   3.532  3.440   2.66%
1897              boyer  8.768  7.747 13.19%  493+199   301+289   0.327  0.297  10.11%
1898       count-graphs  2.040  2.010  1.49%  298+394   315+457  26.578 28.660  -7.26%
1899                fft  1.327  1.302  1.92%  112+209   115+210   1.055  0.962   9.71%
1900        knuthBendix  5.218  5.475 -4.69%  451+598   510+650   0.928  0.932  -0.36%
1901             lexgen  9.970  9.623  3.60% 1014+841  1157+885   0.947  0.928   1.97%
1902               life  1.183  1.183  0.00%  162+182   145+148   0.127  0.103  22.58%
1903              logic  3.285  3.512 -6.45%  514+684   591+836   5.682  5.577   1.88%
1904         mandelbrot  0.147  0.143  2.33%   38+41     33+54    0.703  0.690   1.93%
1905             mlyacc 35.457 32.763  8.22% 3496+4564 3611+4860  0.552  0.550   0.30%
1906            nucleic  7.100  6.888  3.07%  239+168   201+158   0.175  0.173   0.96%
1907      ratio-regions  6.388  6.843 -6.65% 1182+257   981+300  120.142 120.345 -0.17%
1908                ray  2.332  2.338 -0.29%  346+398   402+494   3.593  3.540   1.51%
1909             simple  9.912  9.903  0.08% 1475+941  1579+1168  3.057  3.178  -3.83%
1910                tsp  1.623  1.532  5.98%  266+200   250+211   8.045  7.878   2.12%
1911               vliw 33.947 35.470 -4.29% 2629+2774 2877+3171  2.072  1.890   9.61%
1912      ----------------------------------------------------------------------------
1913       Average                     1.22%                                     3.36%
1914    
1915    ----------------------------------------------------------------------
1916    Name: Allen Leung
1917    Date: 2000/03/23 16:25:00
1918    Tag: leunga-20000323-fix_x86_alpha
1919    Description:
1920    
1921    1. X86 fixes/changes
1922    
1923       a.  The old code generated for SETcc was completely wrong.
1924           The Intel optimization guide is VERY misleading.
1925    
1926    2. ALPHA fixes/changes
1927    
1928       a.  Added the instructions LDBU, LDWU, STB, STW as per Fermin's suggestion.
1929       b.  Added a new mode byteWordLoadStores to the functor parameter to Alpha()
1930       c.  Added reassociation code for address computation.
1931    
1932    ----------------------------------------------------------------------
1933    Name: Allen Leung
1934    Date: 2000/03/22 01:23:00
1935    Tag: leunga-20000322-fix_x86_hppa_ra
1936    Description:
1937    
1938    1. X86 fixes/changes
1939    
1940       a.  x86Rewrite bug with MUL3 (found by Lal)
1941       b.  Added the instructions FSTS, FSTL
1942    
1943    2. PA-RISC fixes/changes
1944    
1945       a.  B label should not be a delay slot candidate!  Why did this work?
1946       b.  ADDT(32, REG(32, r), LI n) now generates one instruction instead of two,
1947           as it should be.
1948       c.  The assembly syntax for fstds and fstdd was wrong.
1949       d.  Added the composite instruction COMICLR/LDO, which is the immediate
1950           operand variant of COMCLR/LDO.
1951    
1952    3. Generic MLRISC
1953    
1954       a.  shuffle.sml rewritten to be slightly more efficient
1955       b.  DIV bug in mltree-simplify fixed (found by Fermin)
1956    
1957    4. Register Allocator
1958    
1959       a.  I now release the interference graph earlier during spilling.
1960           May improve memory usage.
1961    
1962    ----------------------------------------------------------------------
1963    Name: Matthias Blume
1964    Date: 2000/03/14 14:15:32
1965    Tag: blume_main_v110p26p1_2
1966    Description:
1967    
1968    1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1969    additional argument called "template" which is an optional string that
1970    specifies the layout of the tool command line.  See the CM manual for
1971    explanation.
1972    
1973    2. A special-purpose tool can be "registered" by simply dropping the
1974    corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1975    directory where the .cm file lives that uses this tool.  (The
1976    behavior/misfeature until now was to look for the tool description
1977    files in the current working directory.)  As before, tool description
1978    files could also be anchored -- in which case they can live anywhere
1979    they like.  Following the recent e-mail discussion, this change should
1980    make it easier to have special-purpose tools that are shipped together
1981    with the sources of the program that uses them.
1982    
1983    ----------------------------------------------------------------------
1984    Name: Matthias Blume
1985    Date: 2000/03/10 07:48:34
1986    Tag: blume_main_v110p26p1_1
1987    Description:
1988    
1989    I added a re-written version of Dave's fixpt script to src/system.
1990    Changes relative to the original version:
1991      - sh-ified (not everybody has ksh)
1992      - automatically figures out which architecture it runs on
1993      - uses ./makeml a bit more cleverly
1994      - never invokes ./installml (and, thus, does not clobber your
1995        good and working installation of sml in case something goes wrong)
1996      - accepts max iteration count using option "-iter <n>"
1997      - accepts a "base" name using option "-base <base>"
1998    
1999    It does not build any extraneous heap images but directly rebuilds
2000    bin- and boot-hierarchies using makeml's "-rebuild" switch. Finally,
2001    it can incorporate existing bin- and boot- hierarchies.  For example,
2002    suppose the base is set to "sml" (which is the default).  Then it
2003    successively builds
2004    
2005            sml.bin.<arch>-unix and sml.boot.<arch>-unix
2006    then    sml1.bin.<arch>-unix and sml1.boot.<arch>-unix
2007    then    sml2.bin.<arch>-unix and sml2.boot.<arch>-unix
2008    ...
2009    then    sml<n>.bin.<arch>-unix and sml<n>.boot.<arch>-unix
2010    
2011    and so on.  If any of these already exist, it will just use what's
2012    there.  In particular, many people will have the initial set of bin
2013    and boot files around, so this saves time for at least one full
2014    rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
2015    <k>=1,2,... is normally not a good idea when invoking fixpt.  However,
2016    they might be the result of an earlier partial run of fixpt (which
2017    perhaps got accidentally killed).  In this case, fixpt will quickly
2018    move through what exists before continuing where it left off earlier,
2019    and, thus, saves a lot of time.
2020    
2021    ----------------------------------------------------------------------
2022    Name: Allen Leung
2023    Date: 00/03/10 02:20:00
2024    Tag: leunga-20000310-fix_x86_asm_ra
2025    Description:
2026    
2027    More assembly output problems involving the indexed addressing mode
2028    on the x86 have been found and corrected. Thanks to Fermin Reig for the
2029    fix.
2030    
2031    The interface and implementation of the register allocator have been changed
2032    slightly to accommodate the possibility to skip the register allocation
2033    phases completely and go directly to memory allocation.  This is needed
2034    for C-- use.
2035    
2036    ----------------------------------------------------------------------
2037    Name: Matthias Blume
2038    Date: 00/03/09 10:23:53
2039    Tag: blume_main_v110p26p1_0
2040    Description:
2041    
2042    * Complete re-organization of library names.  Many libraries have been
2043    consolidated so that they share the same path anchor.  For example,
2044    all MLRISC-related libraries are anchored at MLRISC, most libraries that
2045    are SML/NJ-specific are under "smlnj".  Notice that names like
2046    host-cmb.cm or host-compiler.cm no longer exist.  See system/README
2047    for a complete description of the new naming scheme.  Quick reference:
2048    
2049       host-cmb.cm        -> smlnj/cmb.cm
2050       host-compiler.cm   -> smlnj/compiler.cm
2051       full-cm.cm         -> smlnj/cm.cm
2052       <arch>-<os>.cm     -> smlnj/cmb/<arch>-<os>.cm
2053       <arch>-compiler.cm -> smlnj/compiler/<arch>.cm
2054    
2055    * Bug fixes in CM.
2056        - exceptions in user code are being passed through (i.e., reach top level)
2057        - more bugs in paranoia mode fixed
2058        - bug related to checking group owners fixed
2059    
2060    * New install.sh script that automagically fetches archive files:
2061      The new file config/srcarchiveurl must contain the URL of the
2062      (remote) directory that contains bin files (or other source archives).
2063      If install.sh does not find the archive locally, it tries to get
2064      it from that remote directory.
2065      This should simplify installation further:  For machines that have
2066      access to the internet, just fetch <version>-config.tgz, unpack it,
2067      edit config/targets, and go (run config/install.sh).  The script will
2068      fetch everything else that it might need all by itself.
2069    
2070      For CVS users, this mechanism is not relevant for source archives, but
2071      it is convenient for getting new sets of binfiles.
2072    
2073      Archives should be tar files compressed with either gzip, compress, or
2074      bzip2.  The script recognizes .tgz, .tar, tar.gz, tz, .tar.Z, and .tar.bz2.
2075    
2076    ----------------------------------------------------------------------
2077    Name: Matthias Blume
2078    Date: 2000/03/07 04:01:04
2079    Tag: blume_main_v110_26_2
2080    Description:
2081    - size info in BOOTLIST
2082         * no fixed upper limits for number of bootfiles or length of
2083           bootfile names in runtime
2084         * falling back to old behavior if no BOOTLIST size info found
2085    - allocation size heuristics in .run-sml
2086         * tries to read cache size from /proc/cpuinfo (this is important for
2087            small-cache Celeron systems!)
2088    - install.sh robustified
2089    - CM manual updates
2090    - paranoid mode
2091         * no more CMB.deliver() (i.e., all done by CMB.make())
2092         * can re-use existing sml.boot.* files
2093         * init.cmi now treated as library
2094         * library stamps for consistency checks
2095    - sml.boot.<arch>-<os>/PIDMAP file
2096         * This file is read by the CM startup code.  This is used to minimize
2097           the amount of dynamic state that needs to be stowed away for the
2098           purpose of sharing between interactive system and user code.
2099    - CM.Anchor.anchor instead of CM.Anchor.{set,cancel}
2100         * Upon request by Elsa.  Anchors now controlled by get-set-pair
2101           like most other CM state variables.
2102    - Compiler.CMSA eliminated
2103         * No longer supported by CM anyway.
2104    - fixed bugs in pickler that kept biting Stefan
2105         * past refs to past refs (was caused by the possibility that
2106           ad-hoc sharing is more discriminating than hash-cons sharing)
2107         * integer overflow on LargeInt.minInt
2108    - ml-{lex,yacc} build scripts now use new mechanism
2109      for building standalone programs
2110    - fixed several gcc -Wall warnings that were caused by missing header
2111      files, missing initializations, etc., in runtime (not all warnings
2112      eliminated, though)

Legend:
Removed from v.570  
changed lines
  Added in v.742

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