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

Legend:
Removed from v.572  
changed lines
  Added in v.729

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