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 585, Wed Mar 29 23:55:35 2000 UTC revision 737, Fri Nov 24 08:43:56 2000 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume
17    Date: 2000/11/24 17:45:00 JST
18    Tag: blume-20001124-link
19    Description:
20    
21    Drastically improved link traversal code for the case that the dynamic
22    value was already loaded at bootstrap time.  As a result, CM and CMB
23    now both load blazingly fast -- even on a very slow machine.  Also,
24    memory consumption has been further reduced by this.
25    
26    Warning: The format of the PIDMAP file has changed.  THerefore, to
27    bootstrap you have to do this:
28    
29    1. Run CMB.make
30    2. Make a symbolic link for the boot directory:
31         ln -s sml.boot.ARCH-OS xxx
32    3. "Rebuild" the boot directory:
33         ./makeml -boot xxx -rebuild sml ; rm xxx
34    4. Boot normally:
35          ./makeml
36    
37    ----------------------------------------------------------------------
38    Name: Matthias Blume
39    Date: 2000/11/21 21:20:00 JST
40    Tag: blume-20001121-tools
41    Description:
42    
43    Continued hacking on autoloading problem -- with success this time.
44    Also changed tool-plugin mechanism.  See new CM manual.
45    
46    ----------------------------------------------------------------------
47    Name: Matthias Blume
48    Date: 2000/11/19 14:30:00 JST
49    Tag:  blume-20001119-autoload
50    Description:
51    
52    Some hacking to make autoloading faster.  Success for CMB, no success
53    so far for CM.  There is a reduced structure CM' that autoloads faster.
54    (This is a temporary, non-documented hack to be eliminated again when
55    the general problem is solved.)
56    
57    ----------------------------------------------------------------------
58    Name: Matthias Blume
59    Date: 2000/11/17 14:10:00 JST
60    Tag: blume-20001117-pickle-lib
61    Description:
62    
63    1. Eliminated comp-lib.cm
64    2. Made pickle-lib.cm
65    3. Eliminated all uses of intset.sml (from comp-lib.cm)
66    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
67       equivalent constructs from smlnj-lib.cm (INtHashTable).
68    5. Point 4. also goes for those uses of intmap.* in MLRISC.
69       Duplicated intmap modules thrown out.
70    6. Hunted down all duplicated SCC code and replaced it with
71       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
72    7. Rewrote Feedback module.
73    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
74       should be thrown out and equivalent modules from smlnj-lib.cm
75       should be used (IntRedBlackSet, IntListSet, ...).
76    
77    Confirmed that compiler compiles to fixpoint.
78    
79    ----------------------------------------------------------------------
80    Name: Allen Leung
81    Date: 2000/11/10 18:00:00
82    Tag: leunga-20001110-new-x86-fp
83    
84    A new x86 floating point code generator has been added.
85    By default this is turned off.  To turn this on, do:
86    
87        CM.autoload "$smlnj/compiler.cm";
88        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
89    
90    Changes:
91    
92    1.  Changed FTAN to FPTAN so that the assembly output is correct.
93    2.  Changed the extension callback for FTANGENT to generate:
94    
95              fptan
96              fstp  %st(0)
97        instead of
98              fptan
99              fstpl ftempmem
100    
101    3.  Numerous assembly fixes for x86.
102    
103    5.  Cleaned up the machine code output module x86/x86MC.sml and added
104        support for a whole bunch of instructions and addressing modes:
105    
106          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
107          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
108          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
109          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
110          fxch %st(n)
111          fld %st(n)
112          fst %st(n)
113          fst mem
114          fstp %st(n)
115          fucom %st(n)
116          fucomp %st(n)
117    
118        All these are now generated when the fast fp mode is turned on.
119    
120    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
121    
122    ----------------------------------------------------------------------
123    Name: Matthias Blume
124    Date: 2000/11/09 11:20:00 JST
125    Tag: blume-20001109-scc
126    Description:
127    
128    Eliminated some code duplication:
129    
130    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
131       (Otherwise the functor is useless.)
132    2. Used GraphSCCFn where SCCUtilFun was used previously.
133    3. Got rid of SCCUtilFun (in comp-lib.cm).
134    
135    ----------------------------------------------------------------------
136    Name: Lal George
137    Date: 2000/11/06 09:02:21 EST 2000
138    Tag: Release_110_30
139    Description:
140    
141    - Version 110.30
142    ----------------------------------------------------------------------
143    Name: Matthias Blume
144    Date: 2000/11/04 14:45:00
145    Tag: blume-20001104-mlbuild
146    Description:
147    
148    - Made ml-build faster on startup.
149    - Documentation fixes.
150    
151    ----------------------------------------------------------------------
152    Name: Matthias Blume
153    Date: 2000/11/02 17:00:00 JST
154    Tag: blume-20001102-condcomp
155    Description:
156    
157    - Small tweaks to pickler -- new BOOTFILES!
158    - Version bumped to 110.29.2.
159    - Added conditional compilation facility to init.cmi (see comment there).
160    ----------------------------------------------------------------------
161    Name: Allen Leung
162    Date: 2000/10/23 19:31:00
163    Tag: leunga-20001023-demo-ra
164    
165    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
166    2. Test programs for the graph library updated
167    3. Some new MLRISC demo programs added
168    
169    ----------------------------------------------------------------------
170    Name: Matthias Blume
171    Date: 2000/08/31 22:15:00 JST
172    Tag: blume-20001017-errmsg
173    Description:
174    
175    More error message grief: Where there used to be no messages, there
176    now were some that had bogus error regions.  Fixed.
177    
178    ----------------------------------------------------------------------
179    Name: Matthias Blume
180    Date: 2000/08/31 17:30:00 JST
181    Tag: blume-20001017-v110p29p1
182    Description:
183    
184    I made a version 110.29.1 with new bootfiles.
185    
186    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
187              CM documentation changes and a small bugfix in CM's error reporting.
188    
189    ----------------------------------------------------------------------
190    Name: Lal George
191    Date: 2000/09/27 14:42:35 EDT
192    Tag: george-20000927-nodestatus
193    Description:
194    
195    Changed the type of the nodestatus, so that:
196    
197            SPILLED(~1)             is now SPILLED
198            SPILLED(m) where m>=0   is now MEMREG(m)
199            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
200    
201    ----------------------------------------------------------------------
202    Name: Matthias Blume
203    Date: 2000/09/07 14:45:00 JST
204    Tag: blume-20000907-cmerrmsg
205    Description:
206    
207    Small tweak to CM to avoid getting ML syntax error messages twice.
208    
209    ----------------------------------------------------------------------
210    Name: Matthias Blume
211    Date: 2000/08/31 18:00:00 JST
212    Tag: blume-20000831-cvsbootfiles
213    Description:
214    
215    New URL for boot files (because the 110.29 files on the BL server do
216    now work correctly with my updated install scripts for yacc and lex).
217    
218    ----------------------------------------------------------------------
219    Name: Matthias Blume
220    Date: 2000/08/08 12:33:00 JST
221    Tag: blume-20000808-manual
222    Description:
223    
224    Tiny update to CM manual.
225    
226    ----------------------------------------------------------------------
227    Name: Allen Leung
228    Date: 2000/08/7 19:31:00
229    Tag: leunga-20000807-a-whole-bunch-of-stuff
230    
231      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
232    and Mobdy.
233    
234    1.  x86
235    
236       a.  Fixes to peephole module by John and Dan.
237       b.  Assembly fix to SETcc by Allen.
238       c.  Fix to c-call by John.
239       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
240       e.  Instruction selection optimization to SETcc as suggested by John.
241    
242           For example,
243    
244            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
245    
246           should generate:
247    
248            MOVL a, x
249            SUBL b, x
250            SHRL 31, x
251    
252    2.  IR stuff
253    
254         A bunch of new DJ-graph related algorithms added.  These
255         speed up SSA construction.
256    
257    3.  SSA + Scheduling
258    
259         Added code for SSA and scheduling to the repository
260    
261    ----------------------------------------------------------------------
262    Name: Lal George
263    Date: 2000/07/27 11:53:14 EDT
264    
265    Tag: lal-20000727-linux-ppc
266    Description:
267    
268     Made changes to support Linux PPC.
269     p.s. I have confirmation that the 110.29 boot files work fine.
270    
271    ----------------------------------------------------------------------
272    Name: Matthias Blume
273    Date: 2000/07/27 17:40:00 JST
274    Tag: blume-20000727-scripts
275    Description:
276    
277    !!!! WARNING !!!!
278    You must recompile the runtime system!
279    !!!! WARNING !!!!
280    
281    This is basically another round of script-enhancements:
282    
283    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
284       and undefine CM preprocessor symbols.
285    
286    2. ml-build avoids generating a new heap image if it finds that the
287       existing one is still ok.  (The condition is that no ML file had to
288       be recompiled and all ML files are found to be older that the heap
289       file.)
290    
291       To make this work smoothly, I also hacked the runtime system as
292       well as SMLofNJ.SysInfo to get access to the heap image suffix
293       (.sparc-solaris, ...) that is currently being used.
294    
295       Moreover, the signature of CM.mk_standalone has changed.  See the
296       CM manual.
297    
298    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
299       CM manual for details.)
300    
301    4. More CM manual updates:
302        - all of the above has been documented.
303        - there is now a section describing the (CM-related) command line
304          arguments that are accepted by the "sml" command
305    
306    ----------------------------------------------------------------------
307    Name: Matthias Blume
308    Date: 2000/07/25 16:20:00 JST
309    Tag: blume-20000725-makedepend
310    Description:
311    
312    Added a script called ml-makedepend.  This can be used in makefiles
313    for Unix' make in a way very similar to the "makedepend" command for
314    C.
315    
316    The script internally uses function CM.sources.
317    
318    Synopsis:
319    
320        ml-makedepend [-f makefile] cmfile targetname
321    
322    The default for the makefile is "makefile" (or "Makefile" should
323    "makefile" not exist).
324    
325    ml-makedepend adds a cmfile/targetname-specific section to this
326    makefile (after removing the previous version of this section).  The
327    section contains a single dependency specification with targetname on
328    the LHS (targetname is an arbitrary name), and a list of files derived
329    from the cmfile on the RHS.  Some of the files on the RHS are
330    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
331    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
332    path names.  The makefile writer is responsible for making sure that
333    these variables have correct at the time "make" is invoked.
334    
335    ----------------------------------------------------------------------
336    Name: Matthias Blume
337    Date: 2000/07/22 23:30:00 JST
338    Tag: blume-20000722-urlupdate
339    Description:
340    
341    Changed BOOT and config/srcarchiveurl to point to BL server:
342    
343        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
344    
345    ----------------------------------------------------------------------
346    Name: Matthias Blume
347    Date: 2000/07/18 18:00:00 JST
348    Tag: blume-20000718-Version_110_29
349    Description:
350    
351    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
352    
353    2. Updated config/version to 110.29
354    
355    3. Updated config/srcarchiveurl
356    
357    3. New boot files!
358       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
359    
360    ----------------------------------------------------------------------
361    Name: Matthias Blume
362    Date: 2000/07/11 13:58:00 JST
363    Tag: blume-20000711-doctypo
364    Description:
365    
366    Fixed a few typos in CM manual.
367    
368    ----------------------------------------------------------------------
369    Name: Allen Leung
370    Date: 2000/06/15 00:38:00
371    Tag: leunga-20000704-sparc-x86
372    
373    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
374    2. fix to x86 RET bug [found by Dan Grossman]
375    3. sparc assembly bug fix for ticc instructions [found by Fermin]
376    
377       Affects c-- and moby only
378    
379    ----------------------------------------------------------------------
380    Name: Matthias Blume
381    Date: 2000/07/04 15:26:00
382    Tag: blume-20000704-trigger
383    Description:
384    
385    1. Improvements to CM manual.
386    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
387       of getting a back-trace.  The function, when called, raises an
388       internal exception which explicitly carries the full back-trace history,
389       so it is unaffected by any intervening handle-raise pairs ("trivial"
390       or not).  The interactive loop will print that history once it arrives
391       at top level.
392       Short of having all exceptions implicitly carry the full history, the
393       recommended way of using this facility is:
394         - compile your program with instrumentation "on"
395         - run it, when it raises an exception, look at the history
396         - if the history is "cut off" because of some handler, go and modify
397           your program so that it explicitly calls BTrace.trigger
398         - recompile (still instrumented), and rerun; look at the full history
399    
400    ----------------------------------------------------------------------
401    Name: Matthias Blume
402    Date: 2000/07/03 15:36:00 JST
403    Tag: blume-20000702-manual
404    Description:
405    
406    Small corrections and updates to CM manual.
407    
408    ----------------------------------------------------------------------
409    Name: Matthias Blume
410    Date: 2000/06/29 16:04:00 JST
411    Tag: blume-20000629-yacctool
412    Description:
413    
414    Changes:
415    
416    1. Class "mlyacc" now takes separate arguments to pass options to
417       generated .sml- and .sig-files independently.
418    2. Corresponding CM manual updates.
419    3. BTrace module now also reports call sites.  (However, for loop clusters
420       it only shows from where the cluster was entered.)  There are associated
421       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
422    
423    ----------------------------------------------------------------------
424    Name: Matthias Blume
425    Date: 2000/06/27 16:51:00 JST
426    Tag: blume-20000627-noweb
427    Description:
428    
429    Changes:
430    
431     1. Implemented "subdir" and "witness" options for noweb tool.
432        This caused some slight internal changes in CM's tool implementation.
433     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
434        some remaining issues from earlier path anchor changes.
435     3. Updated CM manual accordingly.
436    
437     4. Changed implementation of back-tracing so that I now consider it
438        ready for prime-time.
439    
440        In particular, you don't have to explicitly trigger the back-trace
441        anymore.  Instead, if you are running BTrace-instrumented code and
442        there is an uncaught exception (regardless of whether or not it was
443        raised in instrumented code), the top-level evalloop will print
444        the back-trace.
445    
446        Features:
447    
448          - Instrumented and uninstrumented code work together seemlessly.
449            (Of course, uninstrumented code is never mentioned in actual
450             back-traces.)
451    
452          - Asymptotic time- and space-complexity of instrumented code is
453            equal to that of uninstrumented code.  (This means that
454            tail-recursion is preserved by the instrumentation phase.)
455    
456          - Modules whose code has been instrumented in different sessions
457            work together without problem.
458    
459          - There is no penalty whatsoever on uninstrumented code.
460    
461          - There is no penalty on "raise" expressions, even in
462            instrumented code.
463    
464        A potential bug (or perhaps it is a feature, too):
465    
466          A back-trace reaches no further than the outermost instrumented
467          non-trivial "raise".  Here, a "trivial" raise is one that is the
468          sole RHS of a "handle" rule.  Thus, back-traces reach trough
469    
470               <exp> handle e => raise e
471    
472          and even
473    
474               <exp> handle Foo => raise Bar
475    
476          and, of course, through
477    
478               <exp> handle Foo => ...
479    
480         if the exception was not Foo.
481    
482         Back-traces always reach right through any un-instrumented code
483         including any of its "handle" expressions, trivial or not.
484    
485       To try this out, do the following:
486    
487         - Erase all existing binfiles for your program.
488           (You may keep binfiles for those modules where you think you
489            definitely don't need back-tracing.)
490         - Turn on back-trace instrumentation:
491              SMLofNJ.Internals.BTrace.mode (SOME true);
492         - Recompile your program.  (I.e., run "CM.make" or "use".)
493         - You may now turn instrumentation off again (if you want):
494              SMLofNJ.Internals.BTrace.mode (SOME false);
495         - Run your program as usual.  If it raises an exception that
496           reaches the interactive toplevel, then a back-trace will
497           automatically be printed.  After that, the toplevel loop
498           will print the exception history as usual.
499    
500    ----------------------------------------------------------------------
501    Name: Matthias Blume
502    Date: 2000/06/26 09:56:46 JST
503    Tag: blume-20000626-setup
504    Description:
505    
506    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
507          ML code before and after compiling a file (e.g., to set compiler
508          flags)
509    
510    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
511              - associated changes to btrace.sml (BTrace instrumentation pass)
512              - cleaner implementation of btimp.sml (BTrace tracing and report
513                module)
514    
515    CM manual: * new path encoding documented
516               * description of setup-parameter to "sml" added
517    
518    The biggest user-visible change to back-tracing is that it is no
519    longer necessary to compile all traced modules within the same
520    session.  (This was a real limitation.)
521    
522    ----------------------------------------------------------------------
523    Name: Matthias Blume
524    Date: 2000/06/24 12:40:00 JST
525    Tag: blume-20000624-startup
526    Description:
527    
528    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
529    bit too often -- to put it mildly. :)
530    
531    ----------------------------------------------------------------------
532    Name: Matthias Blume
533    Date: 2000/06/23 18:20:00 JST
534    Tag: blume-20000623-btrace
535    Description:
536    
537    This updates adds a backtrace facility to aid programmers in debugging
538    their programs.  This involves the following changes:
539    
540    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
541       keeping track of the current call stack.  When programs are compiled
542       in a special mode, the compiler will insert calls to these hooks
543       into the user program.
544       "Hook" means that it is possible for different implementations of
545       back-tracing to register themselves (at different times).
546    
547    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
548       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
549       It can be turned on using this call:
550         SMLofNJ.Internals.BTrace.mode (SOME true);
551       Turning it off again:
552         SMLofNJ.Internals.BTrace.mode (SOME false);
553       Querying the current status:
554         SMLofNJ.Internals.BTrace.mode NONE;
555       Annotated programs are about twice as big as normal ones, and they
556       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
557       where all hooks do nothing).  The slowdown with a plugin that is
558       actually useful (such as the one supplied by default) is even greater,
559       but in the case of the default plugin it is still only an constant
560       factor (amortized).
561    
562    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
563       with a sub-structure BTrace for controlling back-tracing.  In particular,
564       the above-mentioned function "mode" controls whether the annotation
565       phase is invoked by the compiler.  Another important function is
566       "trigger": when called it aborts the current execution and causes
567       the top-level loop to print a full back-trace.
568    
569    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
570       for back-tracing.  It keeps track of the dynamic call stack and in
571       addition to that it keeps a partial history at each "level" of that
572       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
573       d and b (at separate times, dynamically), then the report will show:
574    
575       GOTO   d
576             /c
577       GOTO  \b
578       CALL   a
579    
580       This shows that there was an initial non-tail call of a, then a
581       tail-call to b or c, looping behavior in a cluster of functions that
582       consist of b and c, and then a goto from that cluster (i.e., either from
583       b or from c) to d.
584    
585       Note that (depending on the user program) the amount of information
586       that the back-trace module has to keep track of at each level is bounded
587       by a constant.  Thus, the whole implementation has the same asymptotical
588       complexity as the original program (both in space and in time).
589    
590    5. compiler/TopLevel/interact/evalloop.sml has been modified to
591       handle the special exception SMLofNJ.Internals.BTrace.BTrace
592       which is raised by the "trigger" function mentioned above.
593    
594    Notes on usage:
595    
596    - Annotated code works well together with unannotated code:
597    Unannotated calls simply do not show up at all in the backtrace.
598    
599    - It is not a good idea to let modules that were annotated during
600    different sessions run at the same time.  This is because the compiler
601    chooses small integers to identify individual functions, and there
602    will be clashes if different modules were compiled in separate sessions.
603    (Nothing will crash, and you will even be told about the clashes, but
604    back-trace information will in general not be useful.)
605    
606    - Back-tracing can be confused by callcc and capture.
607    
608    - The only way of getting a back-trace right now is to explicitly
609    invoke the "trigger" function from your user program.  Eventually, we
610    should make every exception carry back-trace information (if
611    available).  But since this creates more overhead at "raise"-time
612    (similar to the current exnHistory overhead), I have not yet
613    implemented this.  (The implementation will be rather easy.)  With
614    exceptions carrying back-trace information, this facility will be even
615    more useful because users don't need to modify their programs...
616    
617    - While it is possible to compile the compiler with back-trace
618    annotations turned on (I did it to get some confidence in
619    correctness), you must make absolutely sure that core.sml and
620    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
621    be compiled with annotation because there is no core access yet, but
622    if you compile btimp.sml with annotation, then the system will go into
623    an infinite recursion and crash.)
624    Since CM currently does not know about BTrace, the only way to turn
625    annotations on and off for different modules of the compiler is to
626    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
627    this is awkward and clumsy.
628    
629    Sample sessions:
630    
631    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
632    - SMLofNJ.Internals.BTrace.mode (SOME true);
633    [autoloading]
634    [autoloading done]
635    val it = false : bool
636    - structure X = struct
637    -     fun main n = let
638    -         fun a (x, 0) = d x
639    -           | a (x, n) = b (x, n - 1)
640    -         and b (x, n) = c (x, n)
641    -         and c (x, n) = a (x, n)
642    -         and d x = e (x, 3)
643    -         and e (x, 0) = f x
644    -           | e (x, n) = e (x, n - 1)
645    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
646    -           | f n = n * g (n - 1)
647    -         and g n = a (n, 3)
648    -     in
649    -         f n
650    -     end
651    - end;
652    structure X : sig val main : int -> int end
653    - X.main 3;
654    *** BACK-TRACE ***
655    GOTO   stdIn:4.2-13.20: X.main[2].f
656    GOTO-( stdIn:4.2-13.20: X.main[2].e
657    GOTO   stdIn:4.2-13.20: X.main[2].d
658         / stdIn:4.2-13.20: X.main[2].a
659         | stdIn:4.2-13.20: X.main[2].b
660    GOTO-\ stdIn:4.2-13.20: X.main[2].c
661    CALL   stdIn:4.2-13.20: X.main[2].g
662    GOTO   stdIn:4.2-13.20: X.main[2].f
663    GOTO-( stdIn:4.2-13.20: X.main[2].e
664    GOTO   stdIn:4.2-13.20: X.main[2].d
665         / stdIn:4.2-13.20: X.main[2].a
666         | stdIn:4.2-13.20: X.main[2].b
667    GOTO-\ stdIn:4.2-13.20: X.main[2].c
668    CALL   stdIn:4.2-13.20: X.main[2].g
669    GOTO   stdIn:4.2-13.20: X.main[2].f
670    GOTO-( stdIn:4.2-13.20: X.main[2].e
671    GOTO   stdIn:4.2-13.20: X.main[2].d
672         / stdIn:4.2-13.20: X.main[2].a
673         | stdIn:4.2-13.20: X.main[2].b
674    GOTO-\ stdIn:4.2-13.20: X.main[2].c
675    CALL   stdIn:4.2-13.20: X.main[2].g
676    GOTO   stdIn:4.2-13.20: X.main[2].f
677    CALL   stdIn:2.15-17.4: X.main[2]
678    -
679    
680    (Note that because of a FLINt bug the above code currently does not
681    compile without BTrace turned on.)
682    
683    Here is another example, using my modified Tiger compiler:
684    
685    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
686    - SMLofNJ.Internals.BTrace.mode (SOME true);
687    [autoloading]
688    [autoloading done]
689    val it = false : bool
690    - CM.make "sources.cm";
691    [autoloading]
692    ...
693    [autoloading done]
694    [scanning sources.cm]
695    [parsing (sources.cm):parse.sml]
696    [creating directory CM/SKEL ...]
697    [parsing (sources.cm):tiger.lex.sml]
698    ...
699    [wrote CM/sparc-unix/semant.sml]
700    [compiling (sources.cm):main.sml]
701    [wrote CM/sparc-unix/main.sml]
702    [New bindings added.]
703    val it = true : bool
704    - Main.compile ("../testcases/merge.tig", "foo.out");
705    *** BACK-TRACE ***
706    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
707    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
708    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
709    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
710    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
711    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
712    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
713         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
714    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
715    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
716    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
717    CALL   main.sml:1.18-118.4: Main.compile[2]
718    -
719    
720    ----------------------------------------------------------------------
721    Name: Matthias Blumen
722    Date: 2000/06/21 18:00:00 JST
723    Tag: blume-20000621-manual
724    Description:
725    
726    CM manual update: Path environments documented.
727    
728    ----------------------------------------------------------------------
729    Name: Matthias Blume
730    Date: 2000/06/19 13:40:00
731    Tag: blume-20000619-manual
732    Description:
733    
734    CM manual and system/README update.  This only covers the fact that
735    there are no more implicit anchors.  (Path environments and the "bind"
736    option to "cm" have yet to be documented.)
737    
738    ----------------------------------------------------------------------
739    Name: Matthias Blume
740    Date: 2000/06/19 11:05:00 JST
741    Tag: blume-20000619-chdir-bugfix
742    Description:
743    
744    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
745    
746    ----------------------------------------------------------------------
747    Name: Matthias Blume
748    Date: 2000/06/18 22:00:10 JST
749    Tag: blume-20000618-implicit-anchors-really-gone
750    Description:
751    
752    I updates the previous HISTORY entry where I forgot to mention that
753    implicit anchors are no longer with us.
754    
755    The current update also gets rid of the (now useless) controller
756    CM.Control.implicit_anchors.
757    
758    ----------------------------------------------------------------------
759    Name: Matthias Blume
760    Date: 2000/06/16 17:30:00 JST
761    Tag: blume-20000616-anchorenv
762    Description:
763    
764    This patch implements the long anticipated (just kidding :) "anchor
765    environment" mechanism.  In the course of doing this, I also
766    re-implemented CM's internal "SrcPath" module from scratch.  The new
767    one should be more robust in certain boundary cases.  In any case, it
768    is a lot cleaner than its predecessor (IMHO).
769    
770    This time, although there is yet another boot file format change, I
771    kept the unpickler backward-compatible.  As a result, no new bootfiles
772    are necessary and bootstrapping is straightforward.  (You cannot read
773    new bootfiles into an old system, but the other way around is no
774    problem.)
775    
776    Visible changes:
777    
778    ** 0. Implicit path anchors (without the leading $-symbol) are no
779    longer recognized at all. This means that such path names are not
780    illegal either.  For example, the name basis.cm simply refers to a
781    local file called "basis.cm" (i.e, the name is an ordinary path
782    relative to .cm-files directory).  Or, to put it differently, only
783    names that start with $ are anchored paths.
784    
785    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
786    vanished.
787    
788    John (Reppy) had critizised this as soon as I originally proposed and
789    implemented it, but at that time I did not really deeply believe
790    him. :) Now I came full-circle because I need the $<singlearc> syntax
791    in another place where it cannot be seen as an abbreviation for
792    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
793    seems to mean (i.e., it "expands" into the corresponding anchor
794    value).
795    
796    However, when paths are used as members in CM description files, it
797    continues to be true that there must be at least another arc after the
798    anchor.  This is now enforced separately during semantic analysis
799    (i.e., from a lexical/syntactical point of view, the notation is ok.)
800    
801    ** 2. The "cm" class now accepts an option "bind".  The option's value
802    is a sub-option list of precisely two items -- one labeled "anchor"
803    and the other one labeled "value".  As you might expect, "anchor" is
804    used to specify an anchor name to be bound, and "value" specifies what
805    the anchor is being bound to.
806    
807    The value must be a directory name and can be given in either standard
808    syntax (including the possibility that it is itself an anchored path)
809    or native syntax.
810    
811    Examples:
812    
813       foo.cm (bind:(anchor:bar value:$mystuff/bar))
814       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
815    
816    and so on.
817    
818    The meaning of this is that the .cm-file will be processed with an
819    augmented anchor environment where the given anchor(s) is/are bound to
820    the given values(s).
821    
822    The rationale for having this feature is this: Suppose you are trying
823    to use two different (already stable) libraries a.cm and b.cm (that
824    you perhaps didn't write yourself).  Further, suppose each of these
825    two libraries internally uses its own auxiliary library $aux/lib.cm.
826    Normally you would now have a problem because the anchor "lib" can not
827    be bound to more than one value globally.  Therefore, the project that
828    uses both a.cm and b.cm must locally redirect the anchor to some other
829    place:
830    
831       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
832       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
833    
834    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
835    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
836    
837    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
838    you when you do so at the time of CM.stabilize).  Therefore, you can
839    also use an anchored path as the value:
840    
841      a.cm (bind:(anchor:lib value:$a-lib))
842      b.cm (bind:(anchor:lib value:$b-lib))
843    
844    Now you can globally configure (using the usual CM.Anchor.anchor or
845    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
846    itself is always locally bound, setting it globally is no longer
847    meaningful or necessary (but it does not hurt either).  In fact, "lib"
848    can still be used as a global anchor for separate purposes.  As a
849    matter of fact, one can locally define "lib" in terms of a global
850    "lib":
851    
852      a.cm (bind:(anchor:lib value:$lib/a))
853      b.cm (bind:(anchor:lib value:$lib/b))
854    
855    ** 3: The encoding of path names has changed.  This affects the way
856    path names are shown in CM's progress report and also the internal
857    protocol encoding used for parallel make.
858    
859    The encoding now uses one or more ':'-separated segments.  Each
860    segments corresponds to a file that has been specified relative to the
861    file given by its preceding segment.  The first segment is either
862    relative to the CWD, absolute, or anchored.  Each segment itself is
863    basically a Unix pathname; all segments but the first are relative.
864    
865    Example:
866    
867       $foo/bar/baz.cm:a/b/c.sml
868    
869    This path denotes the file bar/a/b/c.sml relative to the directory
870    denoted by anchor "foo".  Notice that the encoding also includes
871    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
872    paths are resolved relative to the .cm-files directory, so baz.cm must
873    be ignored to get the "real" pathname.
874    
875    To make this fact more obvious, CM puts the names of such "virtual
876    arcs" into parentheses when they appear in progress reports. (No
877    parentheses will appear in the internal protocol encoding.)  Thus,
878    what you really see is:
879    
880      $foo/bar/(baz.cm):a/b/c.sml
881    
882    I find this notation to be much more informative than before.
883    
884    Another new feature of the encoding is that special characters
885    including parentheses, colons, (back)slashes, and white space are
886    written as \ddd (where ddd is the decimal encoding of the character).
887    
888    *** The CM manual still needs to be updated.
889    
890    ----------------------------------------------------------------------
891    Name: Allen Leung
892    Date: 2000/06/15 00:38:00
893    Tag: leunga-20000615-x86-peephole
894    
895    x86 Peephole fix by Fermin.  Affects c-- and moby only.
896    
897    ----------------------------------------------------------------------
898    Name: Matthias Blume
899    Date: 2000/06/12 11:40:00
900    Tag: blume-20000612-parmakefix
901    Description:
902    
903    More cleanup after changing the file naming scheme: This time I
904    repaired the parallel make mechanism for CMB.make which I broke earlier.
905    
906    ----------------------------------------------------------------------
907    Name: Allen Leung
908    Date: 2000/06/09 01:25:00
909    Tag: leunga-20000609-various
910    
911    None of these things should affect normal SML/NJ operations
912    
913    1. Peephole improvements provided by Fermin (c--)
914    2. New annotation DEFUSE for adding extra dependence (moby)
915    3. New X86 LOCK instructions (moby)
916    4. New machine description language for reservation tables (scheduling)
917    5. Fixes to various optimization/analysis modules (branch chaining, dominator
918       trees etc.)
919    6. I've changed the CM files so that they can work with versions
920       110.0.6, 110.25 and 110.28
921    
922    ----------------------------------------------------------------------
923    Name: Matthias Blume
924    Date: 2000/06/09 12:40:00
925    Tag: blume-20000609-log
926    Description:
927    
928    - Removed all(?) remaining RCS Log entries from sources.
929    
930    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
931      anchored paths).
932    
933    ----------------------------------------------------------------------
934    Name: Matthias Blume
935    Date: 2000/06/07 17:00:00 JST
936    Tag: blume-20000607-no-implicit-anchors
937    Description:
938    
939    1. This update changes the default setting for
940    CM.Control.implicit_anchors from true to false.  This means that
941    implicit anchors are no longer permitted by default.  I also tried to
942    make sure that nothing else still relies on implicit anchors.
943    (This is the next step on the schedule towards a CM that does not even
944    have the notion of implicit anchors anymore.)
945    
946    2. More CM manual updates.
947    
948    3. I managed to track down and fix the pickling bug I mentioned last
949    time.  Because of the previously existing workaround, this entails no
950    immediate practical changes.
951    
952    ----------------------------------------------------------------------
953    Name: Matthias Blume
954    Date: 2000/06/06 11:15:00 JST
955    Tag: blume-20000606-lazierpickle
956    Description:
957    
958    !!!! NEW BOOT FILES !!!!
959    
960    * The main purpose of this update is to make library pickles lazier in
961    order to reduce the initial space penalty for autoloading a library.
962    As a result, it is now possible to have $smlnj/compiler.cm
963    pre-registered.  This should take care of the many complaints or
964    inquiries about missing structure Compiler.  This required changes to
965    CM's internal data structures and small tweaks to some algorithms.
966    
967    As a neat additional effect, it is no longer necessary (for the sake
968    of lean heap image files) to distinguish between a "minimal" CM and a
969    "full" CM.  Now, there is only one CM (i.e., the "full" version:
970    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
971    interactive top level. ($smlnj/cm/minimal.cm is gone.)
972    
973    To make the life of compiler-hackers easier, "makeml" now also
974    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
975    words, after you bootstrap a new sml for the first time, you will not
976    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
977    around you will still have to do it, though.)
978    
979    * A second change consists of major updates to the CM manual.  There
980    are now several appendices with summary information and also a full
981    specification of the CM description file syntax.
982    
983    * In directory src/system I added the script "allcross".  This script
984    invokes sml and cross-compiles the compiler for all supported
985    architectures.  (Useful when providing a new set of boot files.)
986    
987    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
988    added a small tweak to pickle-util.sml to work around this problem,
989    but it is not a proper fix yet.  I will investigate further.  (The
990    effect of the bug was an inflation of library pickle size.)
991    
992    * Version number increased to 110.28.1 (to avoid compatibility problems).
993    
994    ----------------------------------------------------------------------
995    Name: Allen Leung
996    Date: 2000/05/25 17:28 EDT
997    Tag: leunga-20000525-ra
998    Description:
999    
1000      Fixed a bug in freezing phase of the register allocator.
1001    
1002    ----------------------------------------------------------------------
1003    Name: Allen Leung
1004    Date: 2000/05/15 22:53 EDT
1005    Tag: leunga-20000515-alpha-x86-ra
1006    Description:
1007    
1008      1. Alpha
1009    
1010          Slight cleanup.  Removed the instruction SGNXL
1011    
1012      2. X86
1013    
1014          Added the following instructions to the instruction set:
1015    
1016            ROLx, RORx,
1017            BTx, BTSx, BTLx, BTRx,
1018            XCHGx, and variants with the LOCK prefix
1019    
1020      3. Register Allocation
1021    
1022          The module ra-rewrite-with-renaming has been improved.
1023    
1024      These have no effect on SML/NJ.
1025    
1026    ----------------------------------------------------------------------
1027    Name: Matthias Blume
1028    Date: 2000/05/15 16:20:00 JST
1029    Tag: blume-20000515-lightrebuild
1030    Description:
1031    
1032    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1033       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1034       defined.  In effect, the command will not build any cross-compiler
1035       backends and therefore finish more quickly.
1036    
1037       The "fixpt" script also takes a "-light" switch to be able to use
1038       this new facility while compiling for a fixpoint.
1039    
1040    2. I replaced all mentions of anchored paths in group owner specifications
1041       with simple relative paths (usually starting with "..").
1042       The rationale is that a library's internal workings should not be
1043       compromised by the lack of some anchor.  (An anchor is necessary
1044       for someone who wants to refer to the library by an anchored path,
1045       but it should not be necessary to build the same library in the first
1046       place.)
1047    
1048    3. I changed the way CM's tool mechanism determines the shell command
1049       string used for things like ml-yacc etc. so that it does not break
1050       when CM.Control.implicit_anchors is turned off.
1051    
1052    ----------------------------------------------------------------------
1053    Name: Matthias Blume
1054    Date: 2000/05/12 18:20:00 JST
1055    Tag: blume-20000512-ml-build
1056    Description:
1057    
1058    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1059    getting installed properly (by config/install.sh).
1060    
1061    ----------------------------------------------------------------------
1062    Name: Matthias Blume
1063    Date: 2000/05/12 17:30:00 JST
1064    Tag: blume-20000512-anchors
1065    Description:
1066    
1067    !!! NEW BOOT FILES !!!
1068    
1069    This change is in preparation of fading out support for "implicitly
1070    anchored path names".  I went through all sources and used the
1071    explicit (and relatively new) $-notation.  See system/README and the
1072    CM manual for more info on this.
1073    
1074    I also modified the anchoring scheme for some things such as "smlnj",
1075    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1076    anchors are more expressive: anchor name and first arc do not have to
1077    coincide.  This entails the following user-visible change:
1078    
1079    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1080    particular, when you fire up sml with a command-line argument, say,
1081    e.g.:
1082    
1083       sml '$smlnj/cmb.cm'
1084    
1085    At the ML toplevel prompt:
1086    
1087       CM.autoload "$smlnj/cmb.cm";
1088    
1089    There is also a new controller in CM.Control that can be used to turn
1090    off all remaining support for implicit anchors by saying:
1091    
1092        CM.autoload "$smlnj/
1093        #set CM.Control.implicit_anchors false;
1094    
1095    This causes CM to reject implicitly anchored paths.  This is (for the
1096    time being) less permissive than the "final" version where there will
1097    be no more such implicit anchors and relative paths will be just that:
1098    relative.
1099    
1100    The next step (version after next version?) will be to make the
1101    default for CM.Control.implicit_anchors false.  After the dust has
1102    settled, I can then produce the "final" version of this...
1103    
1104    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1105    
1106    ----------------------------------------------------------------------
1107    Name: Matthias Blume
1108    Date: 2000/05/11 16:30:00 JST
1109    Tag: blume-20000511-sources
1110    Description:
1111    
1112    The main change is that I added function CM.sources as a generalized
1113    version of the earlier CM.makedepend.  This entails the following
1114    additional changes:
1115    
1116      - CM.makedepend has been dropped.
1117    
1118      - CM manual has been updated.
1119    
1120      - TOOLS signature and API have been changed.
1121    
1122    ----------------------------------------------------------------------
1123    Name: Allen Leung
1124    Date: 2000/05/10 21:17 EDT
1125    Tag: leunga-20000510-moby-c--ssa
1126    Description:
1127    
1128      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1129    None of these affect SML/NJ.
1130    
1131    1. Register Allocation
1132    
1133        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1134           This module tries to remove local (i.e. basic block level) redundancies
1135           during spilling.
1136    
1137        b. A new framework for performing region based register allocation.
1138           Not yet entirely functional.
1139    
1140    2. X86
1141    
1142       a. DefUse for POP was missing the stack pointer [found by Lal]
1143       b. Reload for CALL was incorrect in X86Spill [found by John]
1144       c. Various fixes in X86Spill so that it can be used correctly for
1145          the new spilling module.
1146    
1147    3. SSA/IR
1148    
1149       a. New module ir/dj-dataflow.sml implements elimination based
1150          data flow analysis.
1151    
1152    4. MLRiscGen
1153    
1154       a. Fix for gc type annotation
1155    
1156    5. MDGen
1157    
1158       Various fixes for machine description -> ml code translation.  For ssa
1159       only.
1160    
1161    ----------------------------------------------------------------------
1162    Name: Allen Leung
1163    Date: 2000/05/08 22:17 EDT
1164    Tag: leunga-20000508-labexp
1165    Description:
1166    
1167      Fermin has found a few assembly problems with constant expressions
1168      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1169      which choke on dumb assemblers.  This is his fix.
1170    
1171    ----------------------------------------------------------------------
1172    Name: Dave MacQueen
1173    Date: 2000/04/09 14:00 EDT
1174    Tag: dbm-20000502-Version_110_28
1175    Description:
1176    
1177    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1178    
1179    2. Updated config/version to 110.28
1180    
1181    3. Updated config/srcarchiveurl
1182    
1183    3. New boot files!
1184       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1185    
1186    ----------------------------------------------------------------------
1187    Name: Matthias Blume
1188    Date: 2000/05/01 19:05:00 JST
1189    Tag: blume-20000501-noweb
1190    Description:
1191    
1192    A new noweb tool has been added.  The existing system is entirely
1193    unaffected by this, but some CM users have asked for renewed noweb
1194    support.  Everything is documented in the CM manual.
1195    
1196    New (plugin) libraries:
1197    
1198       noweb-tool.cm
1199       nw-ext.cm
1200    
1201    ----------------------------------------------------------------------
1202    Name: Dave MacQueen
1203    Date: 2000/04/30 12:40PM EDT
1204    Tag: dbm-20000430-bug_fixes
1205    Description:
1206    
1207    1. Fix for bug 1498
1208       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1209       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1210         added toRealArray function
1211       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1212         added check for tag Obj.RealArray to array printing case in ppObj
1213    
1214    2. Fix for bug 1510
1215       smlnj/src/compiler/Semant/types/typesutil.sml
1216         fixed definition of dummyargs (used by equalTycon) so that
1217         dummy args are distinct types
1218    
1219    ----------------------------------------------------------------------
1220    Name: Matthias Blume
1221    Date: 2000/04/30 01:00:00 JST
1222    Tag: blume-20000430-versions
1223    Description:
1224    
1225    1. CM version numbering added.  This is an implementation of Lal's
1226       proposal for adding version numbers and version checking to .cm
1227       files.  Lal said that his proposal was just that -- a proposal.
1228       For the time being I went ahead and implemented it so that people
1229       can comment on it.  Everything is completely backward-compatible
1230       (except for the stable library format, i.e., new bootfiles!).
1231    
1232       As usual, see the CM manual for details.
1233    
1234    2. An alternative syntax for anchored paths has been implemented.
1235       Dave has recently voiced the same concerns that I had when I did
1236       this, so there should be some support.  My take is that eventually
1237       I will let support for the current syntax (where anchors are
1238       "implicit") fade out in favor of the new, explicit syntax.
1239       In order to be backward-compatible, both old and new syntax are
1240       currently supported.
1241    
1242       Again, see the CM manual for details.
1243    
1244    3. Parallel make is trying to be slightly smarter:  When the master
1245       process finds a "bottleneck", i.e., when there is only one
1246       compilation unit that can be compiled and everybody else is
1247       waiting on it, then it will simply compile it directly instead
1248       of clumsily telling one of the slaves to do it.
1249    
1250    4. Support for "unsharing" added.  This is necessary in order to be
1251       able to have two different versions of the same library running
1252       at the same time (e.g., for trying out a new MLRISC while still
1253       having the old MLRISC linked into the current compiler, etc.)
1254       See the CM manual.
1255    
1256    5. Simple "makedepend" functionality added for generating Makefile
1257       dependency information.  (This is rather crude at the moment.
1258       Expect some changes here in the future.)
1259    
1260    6. ".fun" added as a recognized suffix for ML files. Also documented
1261       explicitly in the manual that the fallback behavior (unknown suffix
1262       -> ML file) is not an official feature!
1263    
1264    7. Small changes to the pickler for stable libraries.
1265    
1266    8. Several internal changes to CM (for cleanup/improvement).
1267    
1268    
1269    !!!! NEW BINFILES !!!!
1270    
1271    ----------------------------------------------------------------------
1272    Name: Matthias Blume
1273    Date: 2000/04/28 17:30:00 JST
1274    Tag: blume-20000428-pathconfig
1275    Description:
1276    
1277    1. I changed config/install.sh to remove duplicate entries from the
1278       lib/pathconfig file at the end.  Moreover, the final version of
1279       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1280       in src/system/installml.
1281    
1282    2. The config/install.sh script now consistently uses relative
1283       pathnames in lib/pathconfig whenever the anchor is in the lib
1284       directory.  (So far this was true for the libraries that come
1285       pre-compiled and bundled as part of the bootfiles but not for
1286       libraries that are compiled by the script itself.)
1287    
1288    ----------------------------------------------------------------------
1289    Name: Matthias Blume
1290    Date: 2000/04/26 13:10:00 JST
1291    Tag: blume-20000426-fun_suffix
1292    Description:
1293    
1294    Added ".fun" as a recognized file name suffix (for ML code).
1295    
1296    ----------------------------------------------------------------------
1297    Name: Allen Leung
1298    Date: 2000/04/25 17:00:00 EST
1299    Tag: leunga-20000425-alpha-ra
1300    Description:
1301    
1302    1. Alpha
1303    
1304        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1305    in C--.
1306    
1307    2. RA
1308    
1309       Added a flag "ra-dump-size" to print out the size of the flowgraph
1310       and the interference graph.
1311    
1312    ----------------------------------------------------------------------
1313    Name: Dave MacQueen
1314    Date: 2000/04/25/
1315    Tag: dbm-20000425-mlyacc_doc_examples
1316    Description:
1317      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1318      Updated all three examples in src/ml-yacc/examples to run
1319      under 110.* using CM.make.
1320    
1321    ----------------------------------------------------------------------
1322    Name: Allen Leung
1323    Date: 2000/04/20 23:04:00 EST
1324    Tag: leunga-20000420-ssa-c---stuff
1325    Description:
1326    
1327      This update synchronizes my repository with Yale's.  Most of these
1328    changes, however, do not affect SML/NJ at all (the RA is an exception).
1329    
1330    1. Register Allocator
1331    
1332       a. An improvement in the interference graph construction:
1333          Given a copy
1334    
1335                s <- t
1336    
1337          no interference edge between s and t is added for this definition of s.
1338    
1339       b. I've added two new spill heuristic modules that Fermin and I developed
1340          (in the new library RA.cm). These are unused in SML/NJ but maybe
1341          useful for others (Moby?)
1342    
1343    2. X86
1344    
1345       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1346    
1347    3. Alpha
1348    
1349       a. Added the BSR instruction and code generation that goes with it [C--]
1350       b. Other fixes too numerous to recount provided by Fermin [C--]
1351    
1352    4. Regmaps
1353    
1354       a. The regmaps are not initialized with the identity physical bindings
1355          at creation time.  This is unneeded.
1356    
1357    5. MLRISC Optimizations
1358    
1359       a. The DJ-Graph module can now compute the iterated dominance frontiers
1360          intersects with liveness incrementally in linear time! Woohoo!
1361          This is now used in my new SSA construction algorithm.
1362    
1363       b. THe branch reorganization module is now smarter about linear chains of
1364          basic blocks.
1365    
1366    
1367    ----------------------------------------------------------------------
1368    Name: Matthias Blume
1369    Date: 2000/04/12 13:52:00 JST
1370    Tag: blume_main_v110p27_1
1371    Description:
1372    
1373    Changed install.sh script to handle archive files without version number
1374    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1375    name of the boot file archive.
1376    
1377    ----------------------------------------------------------------------
1378    Name: Dave MacQueen
1379    Date: 2000/04/09 14:00 EDT
1380    Tag: dbm-20000410-Version_110_27
1381    Description:
1382    
1383    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1384    
1385    2. Updated src/config/version to 110.27
1386    
1387    3. New boot files!
1388    
1389    ----------------------------------------------------------------------
1390    Name: Allen Leung
1391    Date: 2000/04/09 19:09:00 EST
1392    Tag: leunga-20000409-misc
1393    Description:
1394    
1395    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1396    
1397    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1398    
1399    ----------------------------------------------------------------------
1400    Name: Stefan
1401    Date: 2000/04/07 10:00:00 EDT
1402    Tag: monnier-20000406-branch-handling
1403    Description:
1404    
1405    Improved handling of branches (mostly those generated from
1406    polymorphic equality), removed switchoff and changed the
1407    default optimization settings (more cpsopt and less flintopt).
1408    
1409    ----------------------------------------------------------------------
1410    Name: Allen Leung
1411    Date: 2000/04/06 01:30:00 EST
1412    Tag: leunga-20000406-peephole-x86-SSA-2
1413    Description:
1414    
1415       Forgot a few files.
1416    
1417    ----------------------------------------------------------------------
1418    Name: Allen Leung
1419    Date: 2000/04/06 00:36:00 EST
1420    Tag: leunga-20000406-peephole-x86-SSA
1421    Description:
1422    
1423    1.  New Peephole code
1424    
1425    2.  Minor improvement to X86 instruction selection
1426    
1427    3.  Various fixes to SSA and machine description -> code translator
1428    
1429    ----------------------------------------------------------------------
1430    Name: Matthias Blume
1431    Date: 2000/04/05 12:30:00 JST
1432    Tag: blume_main_v110p26p2_3
1433    Description:
1434    
1435    This update just merges three minor cosmetic updates to CM's sources
1436    to get ready for the 110.27 code freeze on Friday.  No functionality
1437    has changed.
1438    
1439    ----------------------------------------------------------------------
1440    Name: Allen Leung
1441    Date: 2000/04/04 19:39:00 EST
1442    Tag: leunga-20000404-x86-asm
1443    Description:
1444    
1445    1.  Fixed a problem in X86 assembly.
1446    
1447        Things like
1448    
1449           jmp %eax
1450           jmp (%eax)
1451    
1452        should be output as
1453    
1454           jmp *%eax
1455           jmp *(%eax)
1456    
1457    2.  Assembly output
1458    
1459          Added a new flag
1460    
1461              "asm-indent-copies" (default to false)
1462    
1463          When this flag is on, parallel copies will be indented an extra level.
1464    
1465    ----------------------------------------------------------------------
1466    Name: Allen Leung
1467    Date: 2000/04/04 03:18:00 EST
1468    Tag: leunga-20000404-C--Moby
1469    Description:
1470    
1471        All of these fixes are related to C--, Moby, and my own optimization
1472        stuff; so they shouldn't affect SML/NJ.
1473    
1474    1.  X86
1475    
1476        Various fixes related floating point, and extensions.
1477    
1478    2.  Alpha
1479    
1480        Some extra patterns related to loads with signed/zero extension
1481        provided by Fermin.
1482    
1483    3.  Assembly
1484    
1485        When generating assembly, resolve the value of client defined constants,
1486        instead of generating symbolic values.  This is controlled by the
1487        new flag "asm-resolve-constants", which is default to true.
1488    
1489    4.  Machine Descriptions
1490    
1491        a. The precedence parser was slightly broken when parsing infixr symbols.
1492        b. The type generalizing code had the bound variables reversed, resulting
1493           in a problem during arity raising.
1494        c. Various fixes in machine descriptions.
1495    
1496    ----------------------------------------------------------------------
1497    Name: Matthias Blume
1498    Date: 2000/04/03 16:05:00 JST
1499    Tag: blume_main_v110p26p2_2
1500    Description:
1501    
1502    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1503    now done via the ordinary static environment that is context to each
1504    compilation unit.
1505    
1506    To this end, I arranged that instead of "structure Core" as "structure
1507    _Core" is bound in the pervasive environment.  Core access is done via
1508    _Core (which can never be accidentally rebound because _Core is not a
1509    legal surface-syntax symbol).
1510    
1511    The current solution is much cleaner because the core environment is
1512    now simply part of the pervasive environment which is part of every
1513    compilation unit's context anyway.  In particular, this eliminates all
1514    special-case handling that was necessary until now in order to deal
1515    with dynamic and symbolic parts of the core environment.
1516    
1517    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1518    compilation manager's bootstrap compiler (actually: in the "init group"
1519    handling).  See the comments in src/system/smlnj/init/init.cmi for
1520    more details.
1521    
1522    I also tried to track down all mentions of "Core" (as string argument
1523    to Symbol.strSymbol) in the compiler and replaced them with a
1524    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1525    name appears in one place only.
1526    
1527    Binfile and bootfile format have not changed, but the switchover from
1528    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1529    bootfiles anyway.
1530    
1531    ----------------------------------------------------------------------
1532    Name: Allen Leung
1533    Date: 2000/04/02 21:17:00 EST
1534    Tag: leunga-20000402-mltree
1535    Description:
1536    
1537       1. Renamed the constructor CALL in MLTREE by popular demand.
1538       2. Added a bunch of files from my repository.  These are currently
1539          used by other non-SMLNJ backends.
1540    
1541    ----------------------------------------------------------------------
1542    Name: Allen Leung
1543    Date: 2000/03/31 21:15:00 EST
1544    Tag: leunga-20000331-aliasing
1545    Description:
1546    
1547    This update contains a rewritten (and hopefully more correct) module
1548    for extracting aliasing information from CPS.
1549    
1550       To turn on this feature:
1551    
1552            Compiler.Control.CG.memDisambiguate := true
1553    
1554       To pretty print the region information with assembly
1555    
1556           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1557    
1558       To control how many levels of aliasing information are printed, use:
1559    
1560           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1561    
1562       The default of n is 3.
1563    
1564    ----------------------------------------------------------------------
1565    Name: David MacQueen
1566    Date: 2000/03/31 11:15:00 EST
1567    Tag: dbm-20000331-runtime_fix
1568    Description:
1569    
1570    This update contains:
1571    
1572    1. runtime/c-lib/c-libraries.c
1573       includes added in revision 1.2 caused compilation errors on hppa-hpux
1574    
1575    2. fix for bug 1556
1576       system/Basis/Implementation/NJ/internal-signals.sml
1577    
1578    ----------------------------------------------------------------------
1579    Name: Matthias Blume
1580    Date: 2000/03/31 18:00:00 JST
1581    Tag: blume_main_v110p26p2_1
1582    Description:
1583    
1584    This update contains:
1585    
1586    1. A small change to CM's handling of stable libraries:
1587       CM now maintains one "global" modmap that is used for all stable
1588       libraries.  The use of such a global modmap maximizes sharing and
1589       minimizes the need for re-traversing parts of environments during
1590       modmap construction.  (However, this has minor impact since modmap
1591       construction seems to account for just one percent or less of total
1592       compile time.)
1593    
1594    2. I added a "genmap" phase to the statistics.  This is where I got the
1595       "one percent" number (see above).
1596    
1597    3. CM's new tool parameter mechanism just became _even_ better. :)
1598       - The parser understands named parameters and recursive options.
1599       - The "make" and "shell" tools use these new features.
1600         (This makes it a lot easier to cascade these tools.)
1601       - There is a small syntax change: named parameters use a
1602    
1603           <name> : ( <option> ... )            or
1604           <name> : <string>
1605    
1606         syntax.  Previously, named parameters were implemented in an
1607         ad-hoc fashion by each tool individually (by parsing strings)
1608         and had the form
1609    
1610           <name>=<string>
1611    
1612       See the CM manual for a full description of these issues.
1613    
1614    ----------------------------------------------------------------------
1615    Name: Matthias Blume
1616    Date: 2000/03/30 18:00:00 JST
1617    Tag: blume_main_v110p26p2_0
1618    Description:
1619    
1620    !!!!! WARNING !!!!!!
1621    !!  New binfiles  !!
1622    !!!!!!!!!!!!!!!!!!!!
1623    
1624    This update contains:
1625    
1626    1. Moderate changes to CM:
1627    
1628       - Changes to CM's tools mechanism.  In particular, it is now possible
1629       to have tools that accept additional "command line" parameters
1630       (specified in the .cm file at each instance where the tool's class is
1631       used).
1632    
1633       This was done to accommodate the new "make" and "shell" tools which
1634       facilitate fairly seamless hookup to portions of code managed using
1635       Makefiles or Shell scripts.
1636    
1637       There are no classes "shared" or "private" anymore.  Instead, the
1638       sharing annotation is now a parameter to the "sml" class.
1639    
1640       There is a bit of generic machinery for implementing one's own
1641       tools that accept command-line parameters.  However, I am not yet fully
1642       satisfied with that part, so expect changes here in the future.
1643    
1644       All existing tools are described in the CM manual.
1645    
1646       - Slightly better error handling.  (CM now suppresses many followup
1647       error messages that tended to be more annoying than helpful.)
1648    
1649    2. Major changes to the compiler's static environment data structures.
1650    
1651       - no CMStaticEnv anymore.
1652            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
1653              but it is called Environment), no conversions between different
1654              kinds of static environments
1655    
1656       - There is still a notion of a "modmap", but such modmaps are generated
1657         on demand at the time when they are needed.  This sounds slow, but I
1658         sped up the code that generates modmaps enough for this not to lead to
1659         a slowdown of the compiler (at least I didn't detect any).
1660    
1661       - To facilitate rapid modmap generation, static environments now
1662         contain an (optional) "modtree" structure.  Modtree annotations are
1663         constructed by the unpickler during unpickling.  (This means that
1664         the elaborator does not have to worry about modtrees at all.)
1665         Modtrees have the advantage that they are compositional in the same
1666         way as the environment data structure itself is compositional.
1667         As a result, modtrees never hang on to parts of an environment that
1668         has already been rendered "stale" by filtering or rebinding.
1669    
1670       - I went through many, many trials and errors before arriving at the
1671         current solution.  (The initial idea of "linkpaths" did not work.)
1672         But the result of all this is that I have touched a lot of files that
1673         depend on the "modules" and "types" data structures (most of the
1674         elaborator). There were a lot of changes during my "linkpath" trials
1675         that could have been reverted to their original state but weren't.
1676         Please, don't be too harsh on me for messing with this code a bit more
1677         than what was strictly necessary...  (I _did_ resist the tempation
1678         of doing any "global reformatting" to avoid an untimely death at
1679         Dave's hands. :)
1680    
1681       - One positive aspect of the previous point:  At least I made sure that
1682         all files that I touched now compile without warnings (other than
1683         "polyEqual").
1684    
1685       - compiler now tends to run "leaner" (i.e., ties up less memory in
1686         redundant modmaps)
1687    
1688    ----------------------------------------------------------------------
1689  Name: Allen Leung  Name: Allen Leung
1690  Date: 2000/03/29 18:00:00  Date: 2000/03/29 18:00:00
1691  Tag: leunga-20000327-mlriscGen_hppa_alpha_x86  Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
1692    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
1693  Description:  Description:
1694    
1695     This update contains *MAJOR* changes to the way code is generated from CPS     This update contains *MAJOR* changes to the way code is generated from CPS
# Line 103  Line 1777 
1777    
1778       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1779       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1780       testl is perferable to andl.       testl is preferable to andl.
1781    
1782  7. RA (x86 only)  7. RA (x86 only)
1783    
# Line 285  Line 1959 
1959    
1960  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
1961  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
1962  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
1963  explanation.  explanation.
1964    
1965  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
1966  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
1967  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
1968  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 332  Line 2006 
2006  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
2007  <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,
2008  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
2009  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2010  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2011  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2012    
# Line 382  Line 2056 
2056    it from that remote directory.    it from that remote directory.
2057    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2058    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2059    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2060    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2061    
2062    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.585  
changed lines
  Added in v.737

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