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 595, Wed Apr 5 03:30:17 2000 UTC revision 760, Fri Dec 22 14:15:24 2000 UTC
# Line 13  Line 13 
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15  Name: Matthias Blume  Name: Matthias Blume
16    Date: 2000/12/22 21:20:00 JST
17    Tag: blume-20001222-anchorenv
18    Description:
19    ----------------------------------------------------------------------
20    Name: Matthias Blume
21    Date: 2000/12/22 23:16:00 JST
22    Tag: blume-20001222-warn
23    Description:
24    
25    Corrected wording and formatting of some CM warning message which I
26    broke in my previous patch.
27    
28    ----------------------------------------------------------------------
29    Name: Matthias Blume
30    Date: 2000/12/22 13:15:00 JST
31    Tag: blume-20001222-cleanup
32    Description:
33    
34    Removed src/cm/ffi which does not (and did not) belong here.
35    
36    ----------------------------------------------------------------------
37    Name: Matthias Blume
38    Date: 2000/12/21 23:55:00 JST
39    Tag: blume-20001221-exn
40    Description:
41    
42    Probably most important: CM no longer silently swallows all exceptions
43    in the compiler.
44    Plus: some other minor CM changes.  For example, CM now reports some
45    sizes for generated binfiles (code, data, envpickle, lambdapickle).
46    
47    ----------------------------------------------------------------------
48    Name: Matthias Blume
49    Date: 2000/12/15 00:01:05 JST
50    Tag: blume-20001215-dirtool
51    Description:
52    
53    - "dir" tool added.
54    - improvements and cleanup to Tools structure
55    - documentation updates
56    
57    ----------------------------------------------------------------------
58    Name: Allen Leung
59    Date: Thu Dec 14 03:45:24 EST 2000
60    Description:
61    Tag:  leunga-20001214-int-inf
62    Description:
63    
64       In IntInf, added these standard functions, which are missing from our
65    implementation:
66    
67        andb : int * int -> int
68        xorb : int * int -> int
69        orb  : int * int -> int
70        notb : int -> int
71         <<   : int * word -> int
72        ~>>  : int * word -> int
73    
74       Not tested, I hope they are correct.
75    
76    ----------------------------------------------------------------------
77    Name: Allen Leung
78    Date: Fri Dec  8 19:23:26 EST 2000
79    Description:
80    Tag:  leunga-20001208-nowhere
81    Description:
82    
83      Slight improvements to the 'nowhere' tool to handle OR-patterns,
84    to generate better error messages etc.  Plus a brief manual.
85    
86    ----------------------------------------------------------------------
87    Name: Lal George
88    Date: 2000/12/08 09:54:02 EST 2000
89    Tag: Release_110_31
90    Description:
91    
92    - Version 110.31
93    ----------------------------------------------------------------------
94    Name: Allen Leung
95    Date: Thu Dec  7 22:01:04 EST 2000
96    Tag:  leunga-20001207-cell-monster-hack
97    Description:
98    
99    Major MLRISC internal changes.  Affect all clients.
100    Summary:
101    
102    1.  Type CELLS.cell = int is now replaced by a datatype.
103        As a result, the old regmap is now gone.  Almost all interfaces
104        in MLRISC change as a consequence.
105    
106    2.  A new brand version of machine description tool (v3.0) that generates
107        modules expecting the new interface.  The old version is removed.
108    
109    3.  The RA interface has been further abstracted into two new functors.
110        RISC_RA and X86RA.  These functors have much simpler interfaces.
111        [See also directory MLRISC/demo.]
112    
113    4.  Some other new source->source code generation tools are available:
114    
115        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
116        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
117           I use this tool to generate the peephole optimizers---with the new
118           cell type changes, peephole rules are becoming difficult to write
119           without conditional pattern matching.
120    
121    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
122        cover the entire MLRISC source tree so many things broke.
123    
124    6.  CM files have been moved to the subdirectory MLRISC/cm.
125        They are moved because there are a lot of them and they clutter up the
126        root dir.
127    
128    7.  More detailed documentation to come...
129    
130        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
131        a makeml -rebuild first.  This is because of other other
132        changes that Matthias has made (see below).
133    
134    
135    ----------------------------------------------------------------------
136    Name: Matthias Blume
137    Date: 2000/11/30 23:12:00 JST
138    Tag: blume-20001130-filereorg
139    Description:
140    
141    Some manual updates and some file reorganizations in CM.
142    
143    ----------------------------------------------------------------------
144    Name: Matthias Blume
145    Date: 2000/11/24 17:45:00 JST
146    Tag: blume-20001124-link
147    Description:
148    
149    Drastically improved link traversal code for the case that the dynamic
150    value was already loaded at bootstrap time.  As a result, CM and CMB
151    now both load blazingly fast -- even on a very slow machine.  Also,
152    memory consumption has been further reduced by this.
153    
154    Warning: The format of the PIDMAP file has changed.  THerefore, to
155    bootstrap you have to do this:
156    
157    1. Run CMB.make
158    2. Make a symbolic link for the boot directory:
159         ln -s sml.boot.ARCH-OS xxx
160    3. "Rebuild" the boot directory:
161         ./makeml -boot xxx -rebuild sml ; rm xxx
162    4. Boot normally:
163          ./makeml
164    
165    ----------------------------------------------------------------------
166    Name: Matthias Blume
167    Date: 2000/11/21 21:20:00 JST
168    Tag: blume-20001121-tools
169    Description:
170    
171    Continued hacking on autoloading problem -- with success this time.
172    Also changed tool-plugin mechanism.  See new CM manual.
173    
174    ----------------------------------------------------------------------
175    Name: Matthias Blume
176    Date: 2000/11/19 14:30:00 JST
177    Tag:  blume-20001119-autoload
178    Description:
179    
180    Some hacking to make autoloading faster.  Success for CMB, no success
181    so far for CM.  There is a reduced structure CM' that autoloads faster.
182    (This is a temporary, non-documented hack to be eliminated again when
183    the general problem is solved.)
184    
185    ----------------------------------------------------------------------
186    Name: Matthias Blume
187    Date: 2000/11/17 14:10:00 JST
188    Tag: blume-20001117-pickle-lib
189    Description:
190    
191    1. Eliminated comp-lib.cm
192    2. Made pickle-lib.cm
193    3. Eliminated all uses of intset.sml (from comp-lib.cm)
194    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
195       equivalent constructs from smlnj-lib.cm (INtHashTable).
196    5. Point 4. also goes for those uses of intmap.* in MLRISC.
197       Duplicated intmap modules thrown out.
198    6. Hunted down all duplicated SCC code and replaced it with
199       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
200    7. Rewrote Feedback module.
201    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
202       should be thrown out and equivalent modules from smlnj-lib.cm
203       should be used (IntRedBlackSet, IntListSet, ...).
204    
205    Confirmed that compiler compiles to fixpoint.
206    
207    ----------------------------------------------------------------------
208    Name: Allen Leung
209    Date: 2000/11/10 18:00:00
210    Tag: leunga-20001110-new-x86-fp
211    
212    A new x86 floating point code generator has been added.
213    By default this is turned off.  To turn this on, do:
214    
215        CM.autoload "$smlnj/compiler.cm";
216        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
217    
218    Changes:
219    
220    1.  Changed FTAN to FPTAN so that the assembly output is correct.
221    2.  Changed the extension callback for FTANGENT to generate:
222    
223              fptan
224              fstp  %st(0)
225        instead of
226              fptan
227              fstpl ftempmem
228    
229    3.  Numerous assembly fixes for x86.
230    
231    5.  Cleaned up the machine code output module x86/x86MC.sml and added
232        support for a whole bunch of instructions and addressing modes:
233    
234          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
235          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
236          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
237          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
238          fxch %st(n)
239          fld %st(n)
240          fst %st(n)
241          fst mem
242          fstp %st(n)
243          fucom %st(n)
244          fucomp %st(n)
245    
246        All these are now generated when the fast fp mode is turned on.
247    
248    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
249    
250    ----------------------------------------------------------------------
251    Name: Matthias Blume
252    Date: 2000/11/09 11:20:00 JST
253    Tag: blume-20001109-scc
254    Description:
255    
256    Eliminated some code duplication:
257    
258    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
259       (Otherwise the functor is useless.)
260    2. Used GraphSCCFn where SCCUtilFun was used previously.
261    3. Got rid of SCCUtilFun (in comp-lib.cm).
262    
263    ----------------------------------------------------------------------
264    Name: Lal George
265    Date: 2000/11/06 09:02:21 EST 2000
266    Tag: Release_110_30
267    Description:
268    
269    - Version 110.30
270    ----------------------------------------------------------------------
271    Name: Matthias Blume
272    Date: 2000/11/04 14:45:00
273    Tag: blume-20001104-mlbuild
274    Description:
275    
276    - Made ml-build faster on startup.
277    - Documentation fixes.
278    
279    ----------------------------------------------------------------------
280    Name: Matthias Blume
281    Date: 2000/11/02 17:00:00 JST
282    Tag: blume-20001102-condcomp
283    Description:
284    
285    - Small tweaks to pickler -- new BOOTFILES!
286    - Version bumped to 110.29.2.
287    - Added conditional compilation facility to init.cmi (see comment there).
288    ----------------------------------------------------------------------
289    Name: Allen Leung
290    Date: 2000/10/23 19:31:00
291    Tag: leunga-20001023-demo-ra
292    
293    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
294    2. Test programs for the graph library updated
295    3. Some new MLRISC demo programs added
296    
297    ----------------------------------------------------------------------
298    Name: Matthias Blume
299    Date: 2000/08/31 22:15:00 JST
300    Tag: blume-20001017-errmsg
301    Description:
302    
303    More error message grief: Where there used to be no messages, there
304    now were some that had bogus error regions.  Fixed.
305    
306    ----------------------------------------------------------------------
307    Name: Matthias Blume
308    Date: 2000/08/31 17:30:00 JST
309    Tag: blume-20001017-v110p29p1
310    Description:
311    
312    I made a version 110.29.1 with new bootfiles.
313    
314    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
315              CM documentation changes and a small bugfix in CM's error reporting.
316    
317    ----------------------------------------------------------------------
318    Name: Lal George
319    Date: 2000/09/27 14:42:35 EDT
320    Tag: george-20000927-nodestatus
321    Description:
322    
323    Changed the type of the nodestatus, so that:
324    
325            SPILLED(~1)             is now SPILLED
326            SPILLED(m) where m>=0   is now MEMREG(m)
327            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
328    
329    ----------------------------------------------------------------------
330    Name: Matthias Blume
331    Date: 2000/09/07 14:45:00 JST
332    Tag: blume-20000907-cmerrmsg
333    Description:
334    
335    Small tweak to CM to avoid getting ML syntax error messages twice.
336    
337    ----------------------------------------------------------------------
338    Name: Matthias Blume
339    Date: 2000/08/31 18:00:00 JST
340    Tag: blume-20000831-cvsbootfiles
341    Description:
342    
343    New URL for boot files (because the 110.29 files on the BL server do
344    now work correctly with my updated install scripts for yacc and lex).
345    
346    ----------------------------------------------------------------------
347    Name: Matthias Blume
348    Date: 2000/08/08 12:33:00 JST
349    Tag: blume-20000808-manual
350    Description:
351    
352    Tiny update to CM manual.
353    
354    ----------------------------------------------------------------------
355    Name: Allen Leung
356    Date: 2000/08/7 19:31:00
357    Tag: leunga-20000807-a-whole-bunch-of-stuff
358    
359      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
360    and Mobdy.
361    
362    1.  x86
363    
364       a.  Fixes to peephole module by John and Dan.
365       b.  Assembly fix to SETcc by Allen.
366       c.  Fix to c-call by John.
367       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
368       e.  Instruction selection optimization to SETcc as suggested by John.
369    
370           For example,
371    
372            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
373    
374           should generate:
375    
376            MOVL a, x
377            SUBL b, x
378            SHRL 31, x
379    
380    2.  IR stuff
381    
382         A bunch of new DJ-graph related algorithms added.  These
383         speed up SSA construction.
384    
385    3.  SSA + Scheduling
386    
387         Added code for SSA and scheduling to the repository
388    
389    ----------------------------------------------------------------------
390    Name: Lal George
391    Date: 2000/07/27 11:53:14 EDT
392    
393    Tag: lal-20000727-linux-ppc
394    Description:
395    
396     Made changes to support Linux PPC.
397     p.s. I have confirmation that the 110.29 boot files work fine.
398    
399    ----------------------------------------------------------------------
400    Name: Matthias Blume
401    Date: 2000/07/27 17:40:00 JST
402    Tag: blume-20000727-scripts
403    Description:
404    
405    !!!! WARNING !!!!
406    You must recompile the runtime system!
407    !!!! WARNING !!!!
408    
409    This is basically another round of script-enhancements:
410    
411    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
412       and undefine CM preprocessor symbols.
413    
414    2. ml-build avoids generating a new heap image if it finds that the
415       existing one is still ok.  (The condition is that no ML file had to
416       be recompiled and all ML files are found to be older that the heap
417       file.)
418    
419       To make this work smoothly, I also hacked the runtime system as
420       well as SMLofNJ.SysInfo to get access to the heap image suffix
421       (.sparc-solaris, ...) that is currently being used.
422    
423       Moreover, the signature of CM.mk_standalone has changed.  See the
424       CM manual.
425    
426    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
427       CM manual for details.)
428    
429    4. More CM manual updates:
430        - all of the above has been documented.
431        - there is now a section describing the (CM-related) command line
432          arguments that are accepted by the "sml" command
433    
434    ----------------------------------------------------------------------
435    Name: Matthias Blume
436    Date: 2000/07/25 16:20:00 JST
437    Tag: blume-20000725-makedepend
438    Description:
439    
440    Added a script called ml-makedepend.  This can be used in makefiles
441    for Unix' make in a way very similar to the "makedepend" command for
442    C.
443    
444    The script internally uses function CM.sources.
445    
446    Synopsis:
447    
448        ml-makedepend [-f makefile] cmfile targetname
449    
450    The default for the makefile is "makefile" (or "Makefile" should
451    "makefile" not exist).
452    
453    ml-makedepend adds a cmfile/targetname-specific section to this
454    makefile (after removing the previous version of this section).  The
455    section contains a single dependency specification with targetname on
456    the LHS (targetname is an arbitrary name), and a list of files derived
457    from the cmfile on the RHS.  Some of the files on the RHS are
458    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
459    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
460    path names.  The makefile writer is responsible for making sure that
461    these variables have correct at the time "make" is invoked.
462    
463    ----------------------------------------------------------------------
464    Name: Matthias Blume
465    Date: 2000/07/22 23:30:00 JST
466    Tag: blume-20000722-urlupdate
467    Description:
468    
469    Changed BOOT and config/srcarchiveurl to point to BL server:
470    
471        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
472    
473    ----------------------------------------------------------------------
474    Name: Matthias Blume
475    Date: 2000/07/18 18:00:00 JST
476    Tag: blume-20000718-Version_110_29
477    Description:
478    
479    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
480    
481    2. Updated config/version to 110.29
482    
483    3. Updated config/srcarchiveurl
484    
485    3. New boot files!
486       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
487    
488    ----------------------------------------------------------------------
489    Name: Matthias Blume
490    Date: 2000/07/11 13:58:00 JST
491    Tag: blume-20000711-doctypo
492    Description:
493    
494    Fixed a few typos in CM manual.
495    
496    ----------------------------------------------------------------------
497    Name: Allen Leung
498    Date: 2000/06/15 00:38:00
499    Tag: leunga-20000704-sparc-x86
500    
501    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
502    2. fix to x86 RET bug [found by Dan Grossman]
503    3. sparc assembly bug fix for ticc instructions [found by Fermin]
504    
505       Affects c-- and moby only
506    
507    ----------------------------------------------------------------------
508    Name: Matthias Blume
509    Date: 2000/07/04 15:26:00
510    Tag: blume-20000704-trigger
511    Description:
512    
513    1. Improvements to CM manual.
514    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
515       of getting a back-trace.  The function, when called, raises an
516       internal exception which explicitly carries the full back-trace history,
517       so it is unaffected by any intervening handle-raise pairs ("trivial"
518       or not).  The interactive loop will print that history once it arrives
519       at top level.
520       Short of having all exceptions implicitly carry the full history, the
521       recommended way of using this facility is:
522         - compile your program with instrumentation "on"
523         - run it, when it raises an exception, look at the history
524         - if the history is "cut off" because of some handler, go and modify
525           your program so that it explicitly calls BTrace.trigger
526         - recompile (still instrumented), and rerun; look at the full history
527    
528    ----------------------------------------------------------------------
529    Name: Matthias Blume
530    Date: 2000/07/03 15:36:00 JST
531    Tag: blume-20000702-manual
532    Description:
533    
534    Small corrections and updates to CM manual.
535    
536    ----------------------------------------------------------------------
537    Name: Matthias Blume
538    Date: 2000/06/29 16:04:00 JST
539    Tag: blume-20000629-yacctool
540    Description:
541    
542    Changes:
543    
544    1. Class "mlyacc" now takes separate arguments to pass options to
545       generated .sml- and .sig-files independently.
546    2. Corresponding CM manual updates.
547    3. BTrace module now also reports call sites.  (However, for loop clusters
548       it only shows from where the cluster was entered.)  There are associated
549       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
550    
551    ----------------------------------------------------------------------
552    Name: Matthias Blume
553    Date: 2000/06/27 16:51:00 JST
554    Tag: blume-20000627-noweb
555    Description:
556    
557    Changes:
558    
559     1. Implemented "subdir" and "witness" options for noweb tool.
560        This caused some slight internal changes in CM's tool implementation.
561     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
562        some remaining issues from earlier path anchor changes.
563     3. Updated CM manual accordingly.
564    
565     4. Changed implementation of back-tracing so that I now consider it
566        ready for prime-time.
567    
568        In particular, you don't have to explicitly trigger the back-trace
569        anymore.  Instead, if you are running BTrace-instrumented code and
570        there is an uncaught exception (regardless of whether or not it was
571        raised in instrumented code), the top-level evalloop will print
572        the back-trace.
573    
574        Features:
575    
576          - Instrumented and uninstrumented code work together seemlessly.
577            (Of course, uninstrumented code is never mentioned in actual
578             back-traces.)
579    
580          - Asymptotic time- and space-complexity of instrumented code is
581            equal to that of uninstrumented code.  (This means that
582            tail-recursion is preserved by the instrumentation phase.)
583    
584          - Modules whose code has been instrumented in different sessions
585            work together without problem.
586    
587          - There is no penalty whatsoever on uninstrumented code.
588    
589          - There is no penalty on "raise" expressions, even in
590            instrumented code.
591    
592        A potential bug (or perhaps it is a feature, too):
593    
594          A back-trace reaches no further than the outermost instrumented
595          non-trivial "raise".  Here, a "trivial" raise is one that is the
596          sole RHS of a "handle" rule.  Thus, back-traces reach trough
597    
598               <exp> handle e => raise e
599    
600          and even
601    
602               <exp> handle Foo => raise Bar
603    
604          and, of course, through
605    
606               <exp> handle Foo => ...
607    
608         if the exception was not Foo.
609    
610         Back-traces always reach right through any un-instrumented code
611         including any of its "handle" expressions, trivial or not.
612    
613       To try this out, do the following:
614    
615         - Erase all existing binfiles for your program.
616           (You may keep binfiles for those modules where you think you
617            definitely don't need back-tracing.)
618         - Turn on back-trace instrumentation:
619              SMLofNJ.Internals.BTrace.mode (SOME true);
620         - Recompile your program.  (I.e., run "CM.make" or "use".)
621         - You may now turn instrumentation off again (if you want):
622              SMLofNJ.Internals.BTrace.mode (SOME false);
623         - Run your program as usual.  If it raises an exception that
624           reaches the interactive toplevel, then a back-trace will
625           automatically be printed.  After that, the toplevel loop
626           will print the exception history as usual.
627    
628    ----------------------------------------------------------------------
629    Name: Matthias Blume
630    Date: 2000/06/26 09:56:46 JST
631    Tag: blume-20000626-setup
632    Description:
633    
634    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
635          ML code before and after compiling a file (e.g., to set compiler
636          flags)
637    
638    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
639              - associated changes to btrace.sml (BTrace instrumentation pass)
640              - cleaner implementation of btimp.sml (BTrace tracing and report
641                module)
642    
643    CM manual: * new path encoding documented
644               * description of setup-parameter to "sml" added
645    
646    The biggest user-visible change to back-tracing is that it is no
647    longer necessary to compile all traced modules within the same
648    session.  (This was a real limitation.)
649    
650    ----------------------------------------------------------------------
651    Name: Matthias Blume
652    Date: 2000/06/24 12:40:00 JST
653    Tag: blume-20000624-startup
654    Description:
655    
656    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
657    bit too often -- to put it mildly. :)
658    
659    ----------------------------------------------------------------------
660    Name: Matthias Blume
661    Date: 2000/06/23 18:20:00 JST
662    Tag: blume-20000623-btrace
663    Description:
664    
665    This updates adds a backtrace facility to aid programmers in debugging
666    their programs.  This involves the following changes:
667    
668    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
669       keeping track of the current call stack.  When programs are compiled
670       in a special mode, the compiler will insert calls to these hooks
671       into the user program.
672       "Hook" means that it is possible for different implementations of
673       back-tracing to register themselves (at different times).
674    
675    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
676       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
677       It can be turned on using this call:
678         SMLofNJ.Internals.BTrace.mode (SOME true);
679       Turning it off again:
680         SMLofNJ.Internals.BTrace.mode (SOME false);
681       Querying the current status:
682         SMLofNJ.Internals.BTrace.mode NONE;
683       Annotated programs are about twice as big as normal ones, and they
684       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
685       where all hooks do nothing).  The slowdown with a plugin that is
686       actually useful (such as the one supplied by default) is even greater,
687       but in the case of the default plugin it is still only an constant
688       factor (amortized).
689    
690    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
691       with a sub-structure BTrace for controlling back-tracing.  In particular,
692       the above-mentioned function "mode" controls whether the annotation
693       phase is invoked by the compiler.  Another important function is
694       "trigger": when called it aborts the current execution and causes
695       the top-level loop to print a full back-trace.
696    
697    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
698       for back-tracing.  It keeps track of the dynamic call stack and in
699       addition to that it keeps a partial history at each "level" of that
700       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
701       d and b (at separate times, dynamically), then the report will show:
702    
703       GOTO   d
704             /c
705       GOTO  \b
706       CALL   a
707    
708       This shows that there was an initial non-tail call of a, then a
709       tail-call to b or c, looping behavior in a cluster of functions that
710       consist of b and c, and then a goto from that cluster (i.e., either from
711       b or from c) to d.
712    
713       Note that (depending on the user program) the amount of information
714       that the back-trace module has to keep track of at each level is bounded
715       by a constant.  Thus, the whole implementation has the same asymptotical
716       complexity as the original program (both in space and in time).
717    
718    5. compiler/TopLevel/interact/evalloop.sml has been modified to
719       handle the special exception SMLofNJ.Internals.BTrace.BTrace
720       which is raised by the "trigger" function mentioned above.
721    
722    Notes on usage:
723    
724    - Annotated code works well together with unannotated code:
725    Unannotated calls simply do not show up at all in the backtrace.
726    
727    - It is not a good idea to let modules that were annotated during
728    different sessions run at the same time.  This is because the compiler
729    chooses small integers to identify individual functions, and there
730    will be clashes if different modules were compiled in separate sessions.
731    (Nothing will crash, and you will even be told about the clashes, but
732    back-trace information will in general not be useful.)
733    
734    - Back-tracing can be confused by callcc and capture.
735    
736    - The only way of getting a back-trace right now is to explicitly
737    invoke the "trigger" function from your user program.  Eventually, we
738    should make every exception carry back-trace information (if
739    available).  But since this creates more overhead at "raise"-time
740    (similar to the current exnHistory overhead), I have not yet
741    implemented this.  (The implementation will be rather easy.)  With
742    exceptions carrying back-trace information, this facility will be even
743    more useful because users don't need to modify their programs...
744    
745    - While it is possible to compile the compiler with back-trace
746    annotations turned on (I did it to get some confidence in
747    correctness), you must make absolutely sure that core.sml and
748    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
749    be compiled with annotation because there is no core access yet, but
750    if you compile btimp.sml with annotation, then the system will go into
751    an infinite recursion and crash.)
752    Since CM currently does not know about BTrace, the only way to turn
753    annotations on and off for different modules of the compiler is to
754    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
755    this is awkward and clumsy.
756    
757    Sample sessions:
758    
759    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
760    - SMLofNJ.Internals.BTrace.mode (SOME true);
761    [autoloading]
762    [autoloading done]
763    val it = false : bool
764    - structure X = struct
765    -     fun main n = let
766    -         fun a (x, 0) = d x
767    -           | a (x, n) = b (x, n - 1)
768    -         and b (x, n) = c (x, n)
769    -         and c (x, n) = a (x, n)
770    -         and d x = e (x, 3)
771    -         and e (x, 0) = f x
772    -           | e (x, n) = e (x, n - 1)
773    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
774    -           | f n = n * g (n - 1)
775    -         and g n = a (n, 3)
776    -     in
777    -         f n
778    -     end
779    - end;
780    structure X : sig val main : int -> int end
781    - X.main 3;
782    *** BACK-TRACE ***
783    GOTO   stdIn:4.2-13.20: X.main[2].f
784    GOTO-( stdIn:4.2-13.20: X.main[2].e
785    GOTO   stdIn:4.2-13.20: X.main[2].d
786         / stdIn:4.2-13.20: X.main[2].a
787         | stdIn:4.2-13.20: X.main[2].b
788    GOTO-\ stdIn:4.2-13.20: X.main[2].c
789    CALL   stdIn:4.2-13.20: X.main[2].g
790    GOTO   stdIn:4.2-13.20: X.main[2].f
791    GOTO-( stdIn:4.2-13.20: X.main[2].e
792    GOTO   stdIn:4.2-13.20: X.main[2].d
793         / stdIn:4.2-13.20: X.main[2].a
794         | stdIn:4.2-13.20: X.main[2].b
795    GOTO-\ stdIn:4.2-13.20: X.main[2].c
796    CALL   stdIn:4.2-13.20: X.main[2].g
797    GOTO   stdIn:4.2-13.20: X.main[2].f
798    GOTO-( stdIn:4.2-13.20: X.main[2].e
799    GOTO   stdIn:4.2-13.20: X.main[2].d
800         / stdIn:4.2-13.20: X.main[2].a
801         | stdIn:4.2-13.20: X.main[2].b
802    GOTO-\ stdIn:4.2-13.20: X.main[2].c
803    CALL   stdIn:4.2-13.20: X.main[2].g
804    GOTO   stdIn:4.2-13.20: X.main[2].f
805    CALL   stdIn:2.15-17.4: X.main[2]
806    -
807    
808    (Note that because of a FLINt bug the above code currently does not
809    compile without BTrace turned on.)
810    
811    Here is another example, using my modified Tiger compiler:
812    
813    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
814    - SMLofNJ.Internals.BTrace.mode (SOME true);
815    [autoloading]
816    [autoloading done]
817    val it = false : bool
818    - CM.make "sources.cm";
819    [autoloading]
820    ...
821    [autoloading done]
822    [scanning sources.cm]
823    [parsing (sources.cm):parse.sml]
824    [creating directory CM/SKEL ...]
825    [parsing (sources.cm):tiger.lex.sml]
826    ...
827    [wrote CM/sparc-unix/semant.sml]
828    [compiling (sources.cm):main.sml]
829    [wrote CM/sparc-unix/main.sml]
830    [New bindings added.]
831    val it = true : bool
832    - Main.compile ("../testcases/merge.tig", "foo.out");
833    *** BACK-TRACE ***
834    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
835    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
836    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
837    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
838    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
839    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
840    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
841         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
842    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
843    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
844    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
845    CALL   main.sml:1.18-118.4: Main.compile[2]
846    -
847    
848    ----------------------------------------------------------------------
849    Name: Matthias Blumen
850    Date: 2000/06/21 18:00:00 JST
851    Tag: blume-20000621-manual
852    Description:
853    
854    CM manual update: Path environments documented.
855    
856    ----------------------------------------------------------------------
857    Name: Matthias Blume
858    Date: 2000/06/19 13:40:00
859    Tag: blume-20000619-manual
860    Description:
861    
862    CM manual and system/README update.  This only covers the fact that
863    there are no more implicit anchors.  (Path environments and the "bind"
864    option to "cm" have yet to be documented.)
865    
866    ----------------------------------------------------------------------
867    Name: Matthias Blume
868    Date: 2000/06/19 11:05:00 JST
869    Tag: blume-20000619-chdir-bugfix
870    Description:
871    
872    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
873    
874    ----------------------------------------------------------------------
875    Name: Matthias Blume
876    Date: 2000/06/18 22:00:10 JST
877    Tag: blume-20000618-implicit-anchors-really-gone
878    Description:
879    
880    I updates the previous HISTORY entry where I forgot to mention that
881    implicit anchors are no longer with us.
882    
883    The current update also gets rid of the (now useless) controller
884    CM.Control.implicit_anchors.
885    
886    ----------------------------------------------------------------------
887    Name: Matthias Blume
888    Date: 2000/06/16 17:30:00 JST
889    Tag: blume-20000616-anchorenv
890    Description:
891    
892    This patch implements the long anticipated (just kidding :) "anchor
893    environment" mechanism.  In the course of doing this, I also
894    re-implemented CM's internal "SrcPath" module from scratch.  The new
895    one should be more robust in certain boundary cases.  In any case, it
896    is a lot cleaner than its predecessor (IMHO).
897    
898    This time, although there is yet another boot file format change, I
899    kept the unpickler backward-compatible.  As a result, no new bootfiles
900    are necessary and bootstrapping is straightforward.  (You cannot read
901    new bootfiles into an old system, but the other way around is no
902    problem.)
903    
904    Visible changes:
905    
906    ** 0. Implicit path anchors (without the leading $-symbol) are no
907    longer recognized at all. This means that such path names are not
908    illegal either.  For example, the name basis.cm simply refers to a
909    local file called "basis.cm" (i.e, the name is an ordinary path
910    relative to .cm-files directory).  Or, to put it differently, only
911    names that start with $ are anchored paths.
912    
913    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
914    vanished.
915    
916    John (Reppy) had critizised this as soon as I originally proposed and
917    implemented it, but at that time I did not really deeply believe
918    him. :) Now I came full-circle because I need the $<singlearc> syntax
919    in another place where it cannot be seen as an abbreviation for
920    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
921    seems to mean (i.e., it "expands" into the corresponding anchor
922    value).
923    
924    However, when paths are used as members in CM description files, it
925    continues to be true that there must be at least another arc after the
926    anchor.  This is now enforced separately during semantic analysis
927    (i.e., from a lexical/syntactical point of view, the notation is ok.)
928    
929    ** 2. The "cm" class now accepts an option "bind".  The option's value
930    is a sub-option list of precisely two items -- one labeled "anchor"
931    and the other one labeled "value".  As you might expect, "anchor" is
932    used to specify an anchor name to be bound, and "value" specifies what
933    the anchor is being bound to.
934    
935    The value must be a directory name and can be given in either standard
936    syntax (including the possibility that it is itself an anchored path)
937    or native syntax.
938    
939    Examples:
940    
941       foo.cm (bind:(anchor:bar value:$mystuff/bar))
942       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
943    
944    and so on.
945    
946    The meaning of this is that the .cm-file will be processed with an
947    augmented anchor environment where the given anchor(s) is/are bound to
948    the given values(s).
949    
950    The rationale for having this feature is this: Suppose you are trying
951    to use two different (already stable) libraries a.cm and b.cm (that
952    you perhaps didn't write yourself).  Further, suppose each of these
953    two libraries internally uses its own auxiliary library $aux/lib.cm.
954    Normally you would now have a problem because the anchor "lib" can not
955    be bound to more than one value globally.  Therefore, the project that
956    uses both a.cm and b.cm must locally redirect the anchor to some other
957    place:
958    
959       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
960       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
961    
962    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
963    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
964    
965    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
966    you when you do so at the time of CM.stabilize).  Therefore, you can
967    also use an anchored path as the value:
968    
969      a.cm (bind:(anchor:lib value:$a-lib))
970      b.cm (bind:(anchor:lib value:$b-lib))
971    
972    Now you can globally configure (using the usual CM.Anchor.anchor or
973    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
974    itself is always locally bound, setting it globally is no longer
975    meaningful or necessary (but it does not hurt either).  In fact, "lib"
976    can still be used as a global anchor for separate purposes.  As a
977    matter of fact, one can locally define "lib" in terms of a global
978    "lib":
979    
980      a.cm (bind:(anchor:lib value:$lib/a))
981      b.cm (bind:(anchor:lib value:$lib/b))
982    
983    ** 3: The encoding of path names has changed.  This affects the way
984    path names are shown in CM's progress report and also the internal
985    protocol encoding used for parallel make.
986    
987    The encoding now uses one or more ':'-separated segments.  Each
988    segments corresponds to a file that has been specified relative to the
989    file given by its preceding segment.  The first segment is either
990    relative to the CWD, absolute, or anchored.  Each segment itself is
991    basically a Unix pathname; all segments but the first are relative.
992    
993    Example:
994    
995       $foo/bar/baz.cm:a/b/c.sml
996    
997    This path denotes the file bar/a/b/c.sml relative to the directory
998    denoted by anchor "foo".  Notice that the encoding also includes
999    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1000    paths are resolved relative to the .cm-files directory, so baz.cm must
1001    be ignored to get the "real" pathname.
1002    
1003    To make this fact more obvious, CM puts the names of such "virtual
1004    arcs" into parentheses when they appear in progress reports. (No
1005    parentheses will appear in the internal protocol encoding.)  Thus,
1006    what you really see is:
1007    
1008      $foo/bar/(baz.cm):a/b/c.sml
1009    
1010    I find this notation to be much more informative than before.
1011    
1012    Another new feature of the encoding is that special characters
1013    including parentheses, colons, (back)slashes, and white space are
1014    written as \ddd (where ddd is the decimal encoding of the character).
1015    
1016    *** The CM manual still needs to be updated.
1017    
1018    ----------------------------------------------------------------------
1019    Name: Allen Leung
1020    Date: 2000/06/15 00:38:00
1021    Tag: leunga-20000615-x86-peephole
1022    
1023    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1024    
1025    ----------------------------------------------------------------------
1026    Name: Matthias Blume
1027    Date: 2000/06/12 11:40:00
1028    Tag: blume-20000612-parmakefix
1029    Description:
1030    
1031    More cleanup after changing the file naming scheme: This time I
1032    repaired the parallel make mechanism for CMB.make which I broke earlier.
1033    
1034    ----------------------------------------------------------------------
1035    Name: Allen Leung
1036    Date: 2000/06/09 01:25:00
1037    Tag: leunga-20000609-various
1038    
1039    None of these things should affect normal SML/NJ operations
1040    
1041    1. Peephole improvements provided by Fermin (c--)
1042    2. New annotation DEFUSE for adding extra dependence (moby)
1043    3. New X86 LOCK instructions (moby)
1044    4. New machine description language for reservation tables (scheduling)
1045    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1046       trees etc.)
1047    6. I've changed the CM files so that they can work with versions
1048       110.0.6, 110.25 and 110.28
1049    
1050    ----------------------------------------------------------------------
1051    Name: Matthias Blume
1052    Date: 2000/06/09 12:40:00
1053    Tag: blume-20000609-log
1054    Description:
1055    
1056    - Removed all(?) remaining RCS Log entries from sources.
1057    
1058    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1059      anchored paths).
1060    
1061    ----------------------------------------------------------------------
1062    Name: Matthias Blume
1063    Date: 2000/06/07 17:00:00 JST
1064    Tag: blume-20000607-no-implicit-anchors
1065    Description:
1066    
1067    1. This update changes the default setting for
1068    CM.Control.implicit_anchors from true to false.  This means that
1069    implicit anchors are no longer permitted by default.  I also tried to
1070    make sure that nothing else still relies on implicit anchors.
1071    (This is the next step on the schedule towards a CM that does not even
1072    have the notion of implicit anchors anymore.)
1073    
1074    2. More CM manual updates.
1075    
1076    3. I managed to track down and fix the pickling bug I mentioned last
1077    time.  Because of the previously existing workaround, this entails no
1078    immediate practical changes.
1079    
1080    ----------------------------------------------------------------------
1081    Name: Matthias Blume
1082    Date: 2000/06/06 11:15:00 JST
1083    Tag: blume-20000606-lazierpickle
1084    Description:
1085    
1086    !!!! NEW BOOT FILES !!!!
1087    
1088    * The main purpose of this update is to make library pickles lazier in
1089    order to reduce the initial space penalty for autoloading a library.
1090    As a result, it is now possible to have $smlnj/compiler.cm
1091    pre-registered.  This should take care of the many complaints or
1092    inquiries about missing structure Compiler.  This required changes to
1093    CM's internal data structures and small tweaks to some algorithms.
1094    
1095    As a neat additional effect, it is no longer necessary (for the sake
1096    of lean heap image files) to distinguish between a "minimal" CM and a
1097    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1098    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1099    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1100    
1101    To make the life of compiler-hackers easier, "makeml" now also
1102    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1103    words, after you bootstrap a new sml for the first time, you will not
1104    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1105    around you will still have to do it, though.)
1106    
1107    * A second change consists of major updates to the CM manual.  There
1108    are now several appendices with summary information and also a full
1109    specification of the CM description file syntax.
1110    
1111    * In directory src/system I added the script "allcross".  This script
1112    invokes sml and cross-compiles the compiler for all supported
1113    architectures.  (Useful when providing a new set of boot files.)
1114    
1115    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1116    added a small tweak to pickle-util.sml to work around this problem,
1117    but it is not a proper fix yet.  I will investigate further.  (The
1118    effect of the bug was an inflation of library pickle size.)
1119    
1120    * Version number increased to 110.28.1 (to avoid compatibility problems).
1121    
1122    ----------------------------------------------------------------------
1123    Name: Allen Leung
1124    Date: 2000/05/25 17:28 EDT
1125    Tag: leunga-20000525-ra
1126    Description:
1127    
1128      Fixed a bug in freezing phase of the register allocator.
1129    
1130    ----------------------------------------------------------------------
1131    Name: Allen Leung
1132    Date: 2000/05/15 22:53 EDT
1133    Tag: leunga-20000515-alpha-x86-ra
1134    Description:
1135    
1136      1. Alpha
1137    
1138          Slight cleanup.  Removed the instruction SGNXL
1139    
1140      2. X86
1141    
1142          Added the following instructions to the instruction set:
1143    
1144            ROLx, RORx,
1145            BTx, BTSx, BTLx, BTRx,
1146            XCHGx, and variants with the LOCK prefix
1147    
1148      3. Register Allocation
1149    
1150          The module ra-rewrite-with-renaming has been improved.
1151    
1152      These have no effect on SML/NJ.
1153    
1154    ----------------------------------------------------------------------
1155    Name: Matthias Blume
1156    Date: 2000/05/15 16:20:00 JST
1157    Tag: blume-20000515-lightrebuild
1158    Description:
1159    
1160    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1161       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1162       defined.  In effect, the command will not build any cross-compiler
1163       backends and therefore finish more quickly.
1164    
1165       The "fixpt" script also takes a "-light" switch to be able to use
1166       this new facility while compiling for a fixpoint.
1167    
1168    2. I replaced all mentions of anchored paths in group owner specifications
1169       with simple relative paths (usually starting with "..").
1170       The rationale is that a library's internal workings should not be
1171       compromised by the lack of some anchor.  (An anchor is necessary
1172       for someone who wants to refer to the library by an anchored path,
1173       but it should not be necessary to build the same library in the first
1174       place.)
1175    
1176    3. I changed the way CM's tool mechanism determines the shell command
1177       string used for things like ml-yacc etc. so that it does not break
1178       when CM.Control.implicit_anchors is turned off.
1179    
1180    ----------------------------------------------------------------------
1181    Name: Matthias Blume
1182    Date: 2000/05/12 18:20:00 JST
1183    Tag: blume-20000512-ml-build
1184    Description:
1185    
1186    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1187    getting installed properly (by config/install.sh).
1188    
1189    ----------------------------------------------------------------------
1190    Name: Matthias Blume
1191    Date: 2000/05/12 17:30:00 JST
1192    Tag: blume-20000512-anchors
1193    Description:
1194    
1195    !!! NEW BOOT FILES !!!
1196    
1197    This change is in preparation of fading out support for "implicitly
1198    anchored path names".  I went through all sources and used the
1199    explicit (and relatively new) $-notation.  See system/README and the
1200    CM manual for more info on this.
1201    
1202    I also modified the anchoring scheme for some things such as "smlnj",
1203    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1204    anchors are more expressive: anchor name and first arc do not have to
1205    coincide.  This entails the following user-visible change:
1206    
1207    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1208    particular, when you fire up sml with a command-line argument, say,
1209    e.g.:
1210    
1211       sml '$smlnj/cmb.cm'
1212    
1213    At the ML toplevel prompt:
1214    
1215       CM.autoload "$smlnj/cmb.cm";
1216    
1217    There is also a new controller in CM.Control that can be used to turn
1218    off all remaining support for implicit anchors by saying:
1219    
1220        CM.autoload "$smlnj/
1221        #set CM.Control.implicit_anchors false;
1222    
1223    This causes CM to reject implicitly anchored paths.  This is (for the
1224    time being) less permissive than the "final" version where there will
1225    be no more such implicit anchors and relative paths will be just that:
1226    relative.
1227    
1228    The next step (version after next version?) will be to make the
1229    default for CM.Control.implicit_anchors false.  After the dust has
1230    settled, I can then produce the "final" version of this...
1231    
1232    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1233    
1234    ----------------------------------------------------------------------
1235    Name: Matthias Blume
1236    Date: 2000/05/11 16:30:00 JST
1237    Tag: blume-20000511-sources
1238    Description:
1239    
1240    The main change is that I added function CM.sources as a generalized
1241    version of the earlier CM.makedepend.  This entails the following
1242    additional changes:
1243    
1244      - CM.makedepend has been dropped.
1245    
1246      - CM manual has been updated.
1247    
1248      - TOOLS signature and API have been changed.
1249    
1250    ----------------------------------------------------------------------
1251    Name: Allen Leung
1252    Date: 2000/05/10 21:17 EDT
1253    Tag: leunga-20000510-moby-c--ssa
1254    Description:
1255    
1256      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1257    None of these affect SML/NJ.
1258    
1259    1. Register Allocation
1260    
1261        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1262           This module tries to remove local (i.e. basic block level) redundancies
1263           during spilling.
1264    
1265        b. A new framework for performing region based register allocation.
1266           Not yet entirely functional.
1267    
1268    2. X86
1269    
1270       a. DefUse for POP was missing the stack pointer [found by Lal]
1271       b. Reload for CALL was incorrect in X86Spill [found by John]
1272       c. Various fixes in X86Spill so that it can be used correctly for
1273          the new spilling module.
1274    
1275    3. SSA/IR
1276    
1277       a. New module ir/dj-dataflow.sml implements elimination based
1278          data flow analysis.
1279    
1280    4. MLRiscGen
1281    
1282       a. Fix for gc type annotation
1283    
1284    5. MDGen
1285    
1286       Various fixes for machine description -> ml code translation.  For ssa
1287       only.
1288    
1289    ----------------------------------------------------------------------
1290    Name: Allen Leung
1291    Date: 2000/05/08 22:17 EDT
1292    Tag: leunga-20000508-labexp
1293    Description:
1294    
1295      Fermin has found a few assembly problems with constant expressions
1296      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1297      which choke on dumb assemblers.  This is his fix.
1298    
1299    ----------------------------------------------------------------------
1300    Name: Dave MacQueen
1301    Date: 2000/04/09 14:00 EDT
1302    Tag: dbm-20000502-Version_110_28
1303    Description:
1304    
1305    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1306    
1307    2. Updated config/version to 110.28
1308    
1309    3. Updated config/srcarchiveurl
1310    
1311    3. New boot files!
1312       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1313    
1314    ----------------------------------------------------------------------
1315    Name: Matthias Blume
1316    Date: 2000/05/01 19:05:00 JST
1317    Tag: blume-20000501-noweb
1318    Description:
1319    
1320    A new noweb tool has been added.  The existing system is entirely
1321    unaffected by this, but some CM users have asked for renewed noweb
1322    support.  Everything is documented in the CM manual.
1323    
1324    New (plugin) libraries:
1325    
1326       noweb-tool.cm
1327       nw-ext.cm
1328    
1329    ----------------------------------------------------------------------
1330    Name: Dave MacQueen
1331    Date: 2000/04/30 12:40PM EDT
1332    Tag: dbm-20000430-bug_fixes
1333    Description:
1334    
1335    1. Fix for bug 1498
1336       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1337       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1338         added toRealArray function
1339       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1340         added check for tag Obj.RealArray to array printing case in ppObj
1341    
1342    2. Fix for bug 1510
1343       smlnj/src/compiler/Semant/types/typesutil.sml
1344         fixed definition of dummyargs (used by equalTycon) so that
1345         dummy args are distinct types
1346    
1347    ----------------------------------------------------------------------
1348    Name: Matthias Blume
1349    Date: 2000/04/30 01:00:00 JST
1350    Tag: blume-20000430-versions
1351    Description:
1352    
1353    1. CM version numbering added.  This is an implementation of Lal's
1354       proposal for adding version numbers and version checking to .cm
1355       files.  Lal said that his proposal was just that -- a proposal.
1356       For the time being I went ahead and implemented it so that people
1357       can comment on it.  Everything is completely backward-compatible
1358       (except for the stable library format, i.e., new bootfiles!).
1359    
1360       As usual, see the CM manual for details.
1361    
1362    2. An alternative syntax for anchored paths has been implemented.
1363       Dave has recently voiced the same concerns that I had when I did
1364       this, so there should be some support.  My take is that eventually
1365       I will let support for the current syntax (where anchors are
1366       "implicit") fade out in favor of the new, explicit syntax.
1367       In order to be backward-compatible, both old and new syntax are
1368       currently supported.
1369    
1370       Again, see the CM manual for details.
1371    
1372    3. Parallel make is trying to be slightly smarter:  When the master
1373       process finds a "bottleneck", i.e., when there is only one
1374       compilation unit that can be compiled and everybody else is
1375       waiting on it, then it will simply compile it directly instead
1376       of clumsily telling one of the slaves to do it.
1377    
1378    4. Support for "unsharing" added.  This is necessary in order to be
1379       able to have two different versions of the same library running
1380       at the same time (e.g., for trying out a new MLRISC while still
1381       having the old MLRISC linked into the current compiler, etc.)
1382       See the CM manual.
1383    
1384    5. Simple "makedepend" functionality added for generating Makefile
1385       dependency information.  (This is rather crude at the moment.
1386       Expect some changes here in the future.)
1387    
1388    6. ".fun" added as a recognized suffix for ML files. Also documented
1389       explicitly in the manual that the fallback behavior (unknown suffix
1390       -> ML file) is not an official feature!
1391    
1392    7. Small changes to the pickler for stable libraries.
1393    
1394    8. Several internal changes to CM (for cleanup/improvement).
1395    
1396    
1397    !!!! NEW BINFILES !!!!
1398    
1399    ----------------------------------------------------------------------
1400    Name: Matthias Blume
1401    Date: 2000/04/28 17:30:00 JST
1402    Tag: blume-20000428-pathconfig
1403    Description:
1404    
1405    1. I changed config/install.sh to remove duplicate entries from the
1406       lib/pathconfig file at the end.  Moreover, the final version of
1407       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1408       in src/system/installml.
1409    
1410    2. The config/install.sh script now consistently uses relative
1411       pathnames in lib/pathconfig whenever the anchor is in the lib
1412       directory.  (So far this was true for the libraries that come
1413       pre-compiled and bundled as part of the bootfiles but not for
1414       libraries that are compiled by the script itself.)
1415    
1416    ----------------------------------------------------------------------
1417    Name: Matthias Blume
1418    Date: 2000/04/26 13:10:00 JST
1419    Tag: blume-20000426-fun_suffix
1420    Description:
1421    
1422    Added ".fun" as a recognized file name suffix (for ML code).
1423    
1424    ----------------------------------------------------------------------
1425    Name: Allen Leung
1426    Date: 2000/04/25 17:00:00 EST
1427    Tag: leunga-20000425-alpha-ra
1428    Description:
1429    
1430    1. Alpha
1431    
1432        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1433    in C--.
1434    
1435    2. RA
1436    
1437       Added a flag "ra-dump-size" to print out the size of the flowgraph
1438       and the interference graph.
1439    
1440    ----------------------------------------------------------------------
1441    Name: Dave MacQueen
1442    Date: 2000/04/25/
1443    Tag: dbm-20000425-mlyacc_doc_examples
1444    Description:
1445      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1446      Updated all three examples in src/ml-yacc/examples to run
1447      under 110.* using CM.make.
1448    
1449    ----------------------------------------------------------------------
1450    Name: Allen Leung
1451    Date: 2000/04/20 23:04:00 EST
1452    Tag: leunga-20000420-ssa-c---stuff
1453    Description:
1454    
1455      This update synchronizes my repository with Yale's.  Most of these
1456    changes, however, do not affect SML/NJ at all (the RA is an exception).
1457    
1458    1. Register Allocator
1459    
1460       a. An improvement in the interference graph construction:
1461          Given a copy
1462    
1463                s <- t
1464    
1465          no interference edge between s and t is added for this definition of s.
1466    
1467       b. I've added two new spill heuristic modules that Fermin and I developed
1468          (in the new library RA.cm). These are unused in SML/NJ but maybe
1469          useful for others (Moby?)
1470    
1471    2. X86
1472    
1473       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1474    
1475    3. Alpha
1476    
1477       a. Added the BSR instruction and code generation that goes with it [C--]
1478       b. Other fixes too numerous to recount provided by Fermin [C--]
1479    
1480    4. Regmaps
1481    
1482       a. The regmaps are not initialized with the identity physical bindings
1483          at creation time.  This is unneeded.
1484    
1485    5. MLRISC Optimizations
1486    
1487       a. The DJ-Graph module can now compute the iterated dominance frontiers
1488          intersects with liveness incrementally in linear time! Woohoo!
1489          This is now used in my new SSA construction algorithm.
1490    
1491       b. THe branch reorganization module is now smarter about linear chains of
1492          basic blocks.
1493    
1494    
1495    ----------------------------------------------------------------------
1496    Name: Matthias Blume
1497    Date: 2000/04/12 13:52:00 JST
1498    Tag: blume_main_v110p27_1
1499    Description:
1500    
1501    Changed install.sh script to handle archive files without version number
1502    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1503    name of the boot file archive.
1504    
1505    ----------------------------------------------------------------------
1506    Name: Dave MacQueen
1507    Date: 2000/04/09 14:00 EDT
1508    Tag: dbm-20000410-Version_110_27
1509    Description:
1510    
1511    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1512    
1513    2. Updated src/config/version to 110.27
1514    
1515    3. New boot files!
1516    
1517    ----------------------------------------------------------------------
1518    Name: Allen Leung
1519    Date: 2000/04/09 19:09:00 EST
1520    Tag: leunga-20000409-misc
1521    Description:
1522    
1523    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1524    
1525    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1526    
1527    ----------------------------------------------------------------------
1528    Name: Stefan
1529    Date: 2000/04/07 10:00:00 EDT
1530    Tag: monnier-20000406-branch-handling
1531    Description:
1532    
1533    Improved handling of branches (mostly those generated from
1534    polymorphic equality), removed switchoff and changed the
1535    default optimization settings (more cpsopt and less flintopt).
1536    
1537    ----------------------------------------------------------------------
1538    Name: Allen Leung
1539    Date: 2000/04/06 01:30:00 EST
1540    Tag: leunga-20000406-peephole-x86-SSA-2
1541    Description:
1542    
1543       Forgot a few files.
1544    
1545    ----------------------------------------------------------------------
1546    Name: Allen Leung
1547    Date: 2000/04/06 00:36:00 EST
1548    Tag: leunga-20000406-peephole-x86-SSA
1549    Description:
1550    
1551    1.  New Peephole code
1552    
1553    2.  Minor improvement to X86 instruction selection
1554    
1555    3.  Various fixes to SSA and machine description -> code translator
1556    
1557    ----------------------------------------------------------------------
1558    Name: Matthias Blume
1559  Date: 2000/04/05 12:30:00 JST  Date: 2000/04/05 12:30:00 JST
1560  Tag: blume_main_v110p26p2_3  Tag: blume_main_v110p26p2_3
1561  Description:  Description:
# Line 67  Line 1610 
1610    
1611  3.  Assembly  3.  Assembly
1612    
1613      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
1614      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
1615      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
1616    
# Line 90  Line 1633 
1633    
1634  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
1635  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
1636  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
1637  legal surface-syntax symbol).  legal surface-syntax symbol).
1638    
1639  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 100  Line 1643 
1643  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
1644    
1645  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1646  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
1647  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
1648  more details.  more details.
1649    
# Line 215  Line 1758 
1758     (specified in the .cm file at each instance where the tool's class is     (specified in the .cm file at each instance where the tool's class is
1759     used).     used).
1760    
1761     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1762     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1763     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1764    
1765     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 228  Line 1771 
1771    
1772     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1773    
1774     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1775     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1776    
1777  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 362  Line 1905 
1905    
1906       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1907       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1908       testl is perferable to andl.       testl is preferable to andl.
1909    
1910  7. RA (x86 only)  7. RA (x86 only)
1911    
# Line 544  Line 2087 
2087    
2088  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2089  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2090  specifiel the layout of the tool command line.  See the CM manual for  specifies the layout of the tool command line.  See the CM manual for
2091  explanation.  explanation.
2092    
2093  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2094  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2095  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2096  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 591  Line 2134 
2134  rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for  rebuild.  Having sets of the form <base><k>.{bin,boot}.<arch>-unix for
2135  <k>=1,2,... is normally not a good idea when invoking fixpt.  However,  <k>=1,2,... is normally not a good idea when invoking fixpt.  However,
2136  they might be the result of an earlier partial run of fixpt (which  they might be the result of an earlier partial run of fixpt (which
2137  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2138  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2139  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2140    
# Line 641  Line 2184 
2184    it from that remote directory.    it from that remote directory.
2185    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2186    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2187    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2188    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2189    
2190    For CVS users, this mechanism is not relevant for source archives, but    For CVS users, this mechanism is not relevant for source archives, but

Legend:
Removed from v.595  
changed lines
  Added in v.760

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