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 589, Fri Mar 31 16:14:36 2000 UTC revision 775, Fri Jan 12 01:17:51 2001 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    ----------------------------------------------------------------------
15    Name: Allen Leung
16    Date: Thu Jan 11 21:03:00 EST 2001
17    Tag: leunga-20010111-labexp=mltree
18    Description:
19    
20    1.  Removed the type LabelExp and replace it by MLTree.
21    2.  Rewritten mltree-simplify with the pattern matcher tool.
22    3.  There were some bugs in alpha code generator which would break
23        64-bit code generation.
24    4.  Redo the tools to generate code with the
25    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
26        are now generated by perl scripts.
27    
28    ----------------------------------------------------------------------
29    Name: Matthias Blume
30    Date: 2001/01/10 21:55:00 JST
31    Tag: blume-20010110-rcc
32    Description:
33    
34    The RCC stuff now seems to work (but only on the x86).
35    This required hacking of the c-calls interface (and -implementation) in
36    MLRISC.
37    
38    Normal compiler users should be unaffected.
39    
40    ----------------------------------------------------------------------
41    Name: Matthias Blume
42    Date: 2001/01/09 01:20:00 JST
43    Tag: blume-20010109-rcc
44    Description:
45    
46    This is a fairly big patch, flushing out a large number of pending
47    changes that I made to my development copy over the last couple of days.
48    
49    Of practical relevance at this moment is a workaround for a pickling
50    bug that Allen ran into the other day.  The cause of the bug itself is
51    still unknown and it might be hard to fix it properly, but the
52    workaround has some merits of its own (namely somewhat reducing pickling
53    overhead for certain libraries).  Therefore, I think this solution should
54    be satisfactory at this time.
55    
56    The rest of the changes (i.e., the vast majority) has to do with my
57    ongoing efforts of providing direct support for C function calls from
58    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
59    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
60    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
61    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
62    handle RCC.
63    
64    The last part (the changes to mlriscGen.sml) are still known to be
65    wrong on the x86 and not implemented on all other architectures.  But
66    the infrastructure is in place. I had to change a few functor
67    signatures in the backend to be able to route the CCalls interface
68    from MLRISC there, and I had to specialize the mltree type (on the
69    x86) to include the necessary extensions. (The extensions themselves
70    were already there and redy to go in MLRISC/x86).
71    
72    Everything should be very happy as soon as someone helps me with
73    mlriscGen.sml...
74    
75    In any case, nothing of this should matter to anyone as long as the
76    new primop is not being used (which is going to be the case unless you
77    find it where I hid it :). The rest of the compiler is completely
78    unaffected.
79    
80    ----------------------------------------------------------------------
81    Name: Matthias Blume
82    Date: 2001/01/05 00:30:00 JST
83    Tag: blume-20010105-primops
84    Description:
85    
86    Added some experimental support for work that I am doing right now.
87    These changes mostly concern added primops, but there is also a new
88    experimental C library in the runtime system (but currently not enabled
89    anywhere except on Linux/X86).
90    
91    In the course of adding primops (and playing with them), I discovered that
92    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
93    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
94    To recover, I made type information in INL_PRIM mandatory and changed
95    prim.sml as well as built-in.sml accordingly.  The InLine structure now
96    has complete, correct type information (i.e., no bottom types).
97    
98    Since all these changes mean that we need new binfiles, I also bumped the
99    version number to 110.32.1.
100    
101    ----------------------------------------------------------------------
102    Name: Matthias Blume
103    Date: 2000/12/30 22:10:00 JST
104    Tag: blume-20001230-various
105    Description:
106    
107    Added proxy libraries for MLRISC and let MLRISC libraries refer
108    to each other using path anchors.  (See CM manual for explanation.)
109    
110    Updated CM documentation.
111    
112    Fixed some bugs in CM.
113    
114    Implemented "proxy" libraries (= syntactic sugar for CM).
115    
116    Added "-quiet" option to makeml and changed runtime system accordingly.
117    
118    Added cleanup handler for exportML to reset timers and compiler stats.
119    
120    ----------------------------------------------------------------------
121    Name: Lal George
122    Date: 2000/12/22 22:22:58 EST 2000
123    Tag: Release_110_32
124    Description:
125    
126            Infinite precision used throughout MLRISC.
127            see MLRISC/mltree/machine-int.sig
128    
129    ----------------------------------------------------------------------
130    Name: Matthias Blume
131    Date: 2000/12/22 23:16:00 JST
132    Tag: blume-20001222-warn
133    Description:
134    
135    Corrected wording and formatting of some CM warning message which I
136    broke in my previous patch.
137    
138    ----------------------------------------------------------------------
139    Name: Matthias Blume
140    Date: 2000/12/22 21:20:00 JST
141    Tag: blume-20001222-anchorenv
142    Description:
143    
144    Fixed CM's handling of anchor environments in connection with CMB.make.
145    
146    ----------------------------------------------------------------------
147    Name: Matthias Blume
148    Date: 2000/12/22 13:15:00 JST
149    Tag: blume-20001222-cleanup
150    Description:
151    
152    Removed src/cm/ffi which does not (and did not) belong here.
153    
154    ----------------------------------------------------------------------
155    Name: Matthias Blume
156    Date: 2000/12/21 23:55:00 JST
157    Tag: blume-20001221-exn
158    Description:
159    
160    Probably most important: CM no longer silently swallows all exceptions
161    in the compiler.
162    Plus: some other minor CM changes.  For example, CM now reports some
163    sizes for generated binfiles (code, data, envpickle, lambdapickle).
164    
165    ----------------------------------------------------------------------
166    Name: Matthias Blume
167    Date: 2000/12/15 00:01:05 JST
168    Tag: blume-20001215-dirtool
169    Description:
170    
171    - "dir" tool added.
172    - improvements and cleanup to Tools structure
173    - documentation updates
174    
175    ----------------------------------------------------------------------
176    Name: Allen Leung
177    Date: Thu Dec 14 03:45:24 EST 2000
178    Description:
179    Tag:  leunga-20001214-int-inf
180    Description:
181    
182       In IntInf, added these standard functions, which are missing from our
183    implementation:
184    
185        andb : int * int -> int
186        xorb : int * int -> int
187        orb  : int * int -> int
188        notb : int -> int
189         <<   : int * word -> int
190        ~>>  : int * word -> int
191    
192       Not tested, I hope they are correct.
193    
194    ----------------------------------------------------------------------
195    Name: Allen Leung
196    Date: Fri Dec  8 19:23:26 EST 2000
197    Description:
198    Tag:  leunga-20001208-nowhere
199    Description:
200    
201      Slight improvements to the 'nowhere' tool to handle OR-patterns,
202    to generate better error messages etc.  Plus a brief manual.
203    
204    ----------------------------------------------------------------------
205    Name: Lal George
206    Date: 2000/12/08 09:54:02 EST 2000
207    Tag: Release_110_31
208    Description:
209    
210    - Version 110.31
211    ----------------------------------------------------------------------
212    Name: Allen Leung
213    Date: Thu Dec  7 22:01:04 EST 2000
214    Tag:  leunga-20001207-cell-monster-hack
215    Description:
216    
217    Major MLRISC internal changes.  Affect all clients.
218    Summary:
219    
220    1.  Type CELLS.cell = int is now replaced by a datatype.
221        As a result, the old regmap is now gone.  Almost all interfaces
222        in MLRISC change as a consequence.
223    
224    2.  A new brand version of machine description tool (v3.0) that generates
225        modules expecting the new interface.  The old version is removed.
226    
227    3.  The RA interface has been further abstracted into two new functors.
228        RISC_RA and X86RA.  These functors have much simpler interfaces.
229        [See also directory MLRISC/demo.]
230    
231    4.  Some other new source->source code generation tools are available:
232    
233        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
234        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
235           I use this tool to generate the peephole optimizers---with the new
236           cell type changes, peephole rules are becoming difficult to write
237           without conditional pattern matching.
238    
239    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
240        cover the entire MLRISC source tree so many things broke.
241    
242    6.  CM files have been moved to the subdirectory MLRISC/cm.
243        They are moved because there are a lot of them and they clutter up the
244        root dir.
245    
246    7.  More detailed documentation to come...
247    
248        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
249        a makeml -rebuild first.  This is because of other other
250        changes that Matthias has made (see below).
251    
252    
253    ----------------------------------------------------------------------
254    Name: Matthias Blume
255    Date: 2000/11/30 23:12:00 JST
256    Tag: blume-20001130-filereorg
257    Description:
258    
259    Some manual updates and some file reorganizations in CM.
260    
261    ----------------------------------------------------------------------
262    Name: Matthias Blume
263    Date: 2000/11/24 17:45:00 JST
264    Tag: blume-20001124-link
265    Description:
266    
267    Drastically improved link traversal code for the case that the dynamic
268    value was already loaded at bootstrap time.  As a result, CM and CMB
269    now both load blazingly fast -- even on a very slow machine.  Also,
270    memory consumption has been further reduced by this.
271    
272    Warning: The format of the PIDMAP file has changed.  THerefore, to
273    bootstrap you have to do this:
274    
275    1. Run CMB.make
276    2. Make a symbolic link for the boot directory:
277         ln -s sml.boot.ARCH-OS xxx
278    3. "Rebuild" the boot directory:
279         ./makeml -boot xxx -rebuild sml ; rm xxx
280    4. Boot normally:
281          ./makeml
282    
283    ----------------------------------------------------------------------
284    Name: Matthias Blume
285    Date: 2000/11/21 21:20:00 JST
286    Tag: blume-20001121-tools
287    Description:
288    
289    Continued hacking on autoloading problem -- with success this time.
290    Also changed tool-plugin mechanism.  See new CM manual.
291    
292    ----------------------------------------------------------------------
293    Name: Matthias Blume
294    Date: 2000/11/19 14:30:00 JST
295    Tag:  blume-20001119-autoload
296    Description:
297    
298    Some hacking to make autoloading faster.  Success for CMB, no success
299    so far for CM.  There is a reduced structure CM' that autoloads faster.
300    (This is a temporary, non-documented hack to be eliminated again when
301    the general problem is solved.)
302    
303    ----------------------------------------------------------------------
304    Name: Matthias Blume
305    Date: 2000/11/17 14:10:00 JST
306    Tag: blume-20001117-pickle-lib
307    Description:
308    
309    1. Eliminated comp-lib.cm
310    2. Made pickle-lib.cm
311    3. Eliminated all uses of intset.sml (from comp-lib.cm)
312    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
313       equivalent constructs from smlnj-lib.cm (INtHashTable).
314    5. Point 4. also goes for those uses of intmap.* in MLRISC.
315       Duplicated intmap modules thrown out.
316    6. Hunted down all duplicated SCC code and replaced it with
317       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
318    7. Rewrote Feedback module.
319    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
320       should be thrown out and equivalent modules from smlnj-lib.cm
321       should be used (IntRedBlackSet, IntListSet, ...).
322    
323    Confirmed that compiler compiles to fixpoint.
324    
325    ----------------------------------------------------------------------
326    Name: Allen Leung
327    Date: 2000/11/10 18:00:00
328    Tag: leunga-20001110-new-x86-fp
329    
330    A new x86 floating point code generator has been added.
331    By default this is turned off.  To turn this on, do:
332    
333        CM.autoload "$smlnj/compiler.cm";
334        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
335    
336    Changes:
337    
338    1.  Changed FTAN to FPTAN so that the assembly output is correct.
339    2.  Changed the extension callback for FTANGENT to generate:
340    
341              fptan
342              fstp  %st(0)
343        instead of
344              fptan
345              fstpl ftempmem
346    
347    3.  Numerous assembly fixes for x86.
348    
349    5.  Cleaned up the machine code output module x86/x86MC.sml and added
350        support for a whole bunch of instructions and addressing modes:
351    
352          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
353          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
354          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
355          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
356          fxch %st(n)
357          fld %st(n)
358          fst %st(n)
359          fst mem
360          fstp %st(n)
361          fucom %st(n)
362          fucomp %st(n)
363    
364        All these are now generated when the fast fp mode is turned on.
365    
366    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
367    
368    ----------------------------------------------------------------------
369    Name: Matthias Blume
370    Date: 2000/11/09 11:20:00 JST
371    Tag: blume-20001109-scc
372    Description:
373    
374    Eliminated some code duplication:
375    
376    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
377       (Otherwise the functor is useless.)
378    2. Used GraphSCCFn where SCCUtilFun was used previously.
379    3. Got rid of SCCUtilFun (in comp-lib.cm).
380    
381    ----------------------------------------------------------------------
382    Name: Lal George
383    Date: 2000/11/06 09:02:21 EST 2000
384    Tag: Release_110_30
385    Description:
386    
387    - Version 110.30
388    ----------------------------------------------------------------------
389    Name: Matthias Blume
390    Date: 2000/11/04 14:45:00
391    Tag: blume-20001104-mlbuild
392    Description:
393    
394    - Made ml-build faster on startup.
395    - Documentation fixes.
396    
397    ----------------------------------------------------------------------
398    Name: Matthias Blume
399    Date: 2000/11/02 17:00:00 JST
400    Tag: blume-20001102-condcomp
401    Description:
402    
403    - Small tweaks to pickler -- new BOOTFILES!
404    - Version bumped to 110.29.2.
405    - Added conditional compilation facility to init.cmi (see comment there).
406    ----------------------------------------------------------------------
407    Name: Allen Leung
408    Date: 2000/10/23 19:31:00
409    Tag: leunga-20001023-demo-ra
410    
411    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
412    2. Test programs for the graph library updated
413    3. Some new MLRISC demo programs added
414    
415    ----------------------------------------------------------------------
416    Name: Matthias Blume
417    Date: 2000/08/31 22:15:00 JST
418    Tag: blume-20001017-errmsg
419    Description:
420    
421    More error message grief: Where there used to be no messages, there
422    now were some that had bogus error regions.  Fixed.
423    
424    ----------------------------------------------------------------------
425    Name: Matthias Blume
426    Date: 2000/08/31 17:30:00 JST
427    Tag: blume-20001017-v110p29p1
428    Description:
429    
430    I made a version 110.29.1 with new bootfiles.
431    
432    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
433              CM documentation changes and a small bugfix in CM's error reporting.
434    
435    ----------------------------------------------------------------------
436    Name: Lal George
437    Date: 2000/09/27 14:42:35 EDT
438    Tag: george-20000927-nodestatus
439    Description:
440    
441    Changed the type of the nodestatus, so that:
442    
443            SPILLED(~1)             is now SPILLED
444            SPILLED(m) where m>=0   is now MEMREG(m)
445            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
446    
447    ----------------------------------------------------------------------
448    Name: Matthias Blume
449    Date: 2000/09/07 14:45:00 JST
450    Tag: blume-20000907-cmerrmsg
451    Description:
452    
453    Small tweak to CM to avoid getting ML syntax error messages twice.
454    
455    ----------------------------------------------------------------------
456    Name: Matthias Blume
457    Date: 2000/08/31 18:00:00 JST
458    Tag: blume-20000831-cvsbootfiles
459    Description:
460    
461    New URL for boot files (because the 110.29 files on the BL server do
462    now work correctly with my updated install scripts for yacc and lex).
463    
464    ----------------------------------------------------------------------
465    Name: Matthias Blume
466    Date: 2000/08/08 12:33:00 JST
467    Tag: blume-20000808-manual
468    Description:
469    
470    Tiny update to CM manual.
471    
472    ----------------------------------------------------------------------
473    Name: Allen Leung
474    Date: 2000/08/7 19:31:00
475    Tag: leunga-20000807-a-whole-bunch-of-stuff
476    
477      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
478    and Mobdy.
479    
480    1.  x86
481    
482       a.  Fixes to peephole module by John and Dan.
483       b.  Assembly fix to SETcc by Allen.
484       c.  Fix to c-call by John.
485       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
486       e.  Instruction selection optimization to SETcc as suggested by John.
487    
488           For example,
489    
490            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
491    
492           should generate:
493    
494            MOVL a, x
495            SUBL b, x
496            SHRL 31, x
497    
498    2.  IR stuff
499    
500         A bunch of new DJ-graph related algorithms added.  These
501         speed up SSA construction.
502    
503    3.  SSA + Scheduling
504    
505         Added code for SSA and scheduling to the repository
506    
507    ----------------------------------------------------------------------
508    Name: Lal George
509    Date: 2000/07/27 11:53:14 EDT
510    
511    Tag: lal-20000727-linux-ppc
512    Description:
513    
514     Made changes to support Linux PPC.
515     p.s. I have confirmation that the 110.29 boot files work fine.
516    
517    ----------------------------------------------------------------------
518    Name: Matthias Blume
519    Date: 2000/07/27 17:40:00 JST
520    Tag: blume-20000727-scripts
521    Description:
522    
523    !!!! WARNING !!!!
524    You must recompile the runtime system!
525    !!!! WARNING !!!!
526    
527    This is basically another round of script-enhancements:
528    
529    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
530       and undefine CM preprocessor symbols.
531    
532    2. ml-build avoids generating a new heap image if it finds that the
533       existing one is still ok.  (The condition is that no ML file had to
534       be recompiled and all ML files are found to be older that the heap
535       file.)
536    
537       To make this work smoothly, I also hacked the runtime system as
538       well as SMLofNJ.SysInfo to get access to the heap image suffix
539       (.sparc-solaris, ...) that is currently being used.
540    
541       Moreover, the signature of CM.mk_standalone has changed.  See the
542       CM manual.
543    
544    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
545       CM manual for details.)
546    
547    4. More CM manual updates:
548        - all of the above has been documented.
549        - there is now a section describing the (CM-related) command line
550          arguments that are accepted by the "sml" command
551    
552    ----------------------------------------------------------------------
553    Name: Matthias Blume
554    Date: 2000/07/25 16:20:00 JST
555    Tag: blume-20000725-makedepend
556    Description:
557    
558    Added a script called ml-makedepend.  This can be used in makefiles
559    for Unix' make in a way very similar to the "makedepend" command for
560    C.
561    
562    The script internally uses function CM.sources.
563    
564    Synopsis:
565    
566        ml-makedepend [-f makefile] cmfile targetname
567    
568    The default for the makefile is "makefile" (or "Makefile" should
569    "makefile" not exist).
570    
571    ml-makedepend adds a cmfile/targetname-specific section to this
572    makefile (after removing the previous version of this section).  The
573    section contains a single dependency specification with targetname on
574    the LHS (targetname is an arbitrary name), and a list of files derived
575    from the cmfile on the RHS.  Some of the files on the RHS are
576    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
577    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
578    path names.  The makefile writer is responsible for making sure that
579    these variables have correct at the time "make" is invoked.
580    
581    ----------------------------------------------------------------------
582    Name: Matthias Blume
583    Date: 2000/07/22 23:30:00 JST
584    Tag: blume-20000722-urlupdate
585    Description:
586    
587    Changed BOOT and config/srcarchiveurl to point to BL server:
588    
589        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
590    
591    ----------------------------------------------------------------------
592    Name: Matthias Blume
593    Date: 2000/07/18 18:00:00 JST
594    Tag: blume-20000718-Version_110_29
595    Description:
596    
597    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
598    
599    2. Updated config/version to 110.29
600    
601    3. Updated config/srcarchiveurl
602    
603    3. New boot files!
604       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
605    
606    ----------------------------------------------------------------------
607    Name: Matthias Blume
608    Date: 2000/07/11 13:58:00 JST
609    Tag: blume-20000711-doctypo
610    Description:
611    
612    Fixed a few typos in CM manual.
613    
614    ----------------------------------------------------------------------
615    Name: Allen Leung
616    Date: 2000/06/15 00:38:00
617    Tag: leunga-20000704-sparc-x86
618    
619    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
620    2. fix to x86 RET bug [found by Dan Grossman]
621    3. sparc assembly bug fix for ticc instructions [found by Fermin]
622    
623       Affects c-- and moby only
624    
625    ----------------------------------------------------------------------
626    Name: Matthias Blume
627    Date: 2000/07/04 15:26:00
628    Tag: blume-20000704-trigger
629    Description:
630    
631    1. Improvements to CM manual.
632    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
633       of getting a back-trace.  The function, when called, raises an
634       internal exception which explicitly carries the full back-trace history,
635       so it is unaffected by any intervening handle-raise pairs ("trivial"
636       or not).  The interactive loop will print that history once it arrives
637       at top level.
638       Short of having all exceptions implicitly carry the full history, the
639       recommended way of using this facility is:
640         - compile your program with instrumentation "on"
641         - run it, when it raises an exception, look at the history
642         - if the history is "cut off" because of some handler, go and modify
643           your program so that it explicitly calls BTrace.trigger
644         - recompile (still instrumented), and rerun; look at the full history
645    
646    ----------------------------------------------------------------------
647    Name: Matthias Blume
648    Date: 2000/07/03 15:36:00 JST
649    Tag: blume-20000702-manual
650    Description:
651    
652    Small corrections and updates to CM manual.
653    
654    ----------------------------------------------------------------------
655    Name: Matthias Blume
656    Date: 2000/06/29 16:04:00 JST
657    Tag: blume-20000629-yacctool
658    Description:
659    
660    Changes:
661    
662    1. Class "mlyacc" now takes separate arguments to pass options to
663       generated .sml- and .sig-files independently.
664    2. Corresponding CM manual updates.
665    3. BTrace module now also reports call sites.  (However, for loop clusters
666       it only shows from where the cluster was entered.)  There are associated
667       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
668    
669    ----------------------------------------------------------------------
670    Name: Matthias Blume
671    Date: 2000/06/27 16:51:00 JST
672    Tag: blume-20000627-noweb
673    Description:
674    
675    Changes:
676    
677     1. Implemented "subdir" and "witness" options for noweb tool.
678        This caused some slight internal changes in CM's tool implementation.
679     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
680        some remaining issues from earlier path anchor changes.
681     3. Updated CM manual accordingly.
682    
683     4. Changed implementation of back-tracing so that I now consider it
684        ready for prime-time.
685    
686        In particular, you don't have to explicitly trigger the back-trace
687        anymore.  Instead, if you are running BTrace-instrumented code and
688        there is an uncaught exception (regardless of whether or not it was
689        raised in instrumented code), the top-level evalloop will print
690        the back-trace.
691    
692        Features:
693    
694          - Instrumented and uninstrumented code work together seemlessly.
695            (Of course, uninstrumented code is never mentioned in actual
696             back-traces.)
697    
698          - Asymptotic time- and space-complexity of instrumented code is
699            equal to that of uninstrumented code.  (This means that
700            tail-recursion is preserved by the instrumentation phase.)
701    
702          - Modules whose code has been instrumented in different sessions
703            work together without problem.
704    
705          - There is no penalty whatsoever on uninstrumented code.
706    
707          - There is no penalty on "raise" expressions, even in
708            instrumented code.
709    
710        A potential bug (or perhaps it is a feature, too):
711    
712          A back-trace reaches no further than the outermost instrumented
713          non-trivial "raise".  Here, a "trivial" raise is one that is the
714          sole RHS of a "handle" rule.  Thus, back-traces reach trough
715    
716               <exp> handle e => raise e
717    
718          and even
719    
720               <exp> handle Foo => raise Bar
721    
722          and, of course, through
723    
724               <exp> handle Foo => ...
725    
726         if the exception was not Foo.
727    
728         Back-traces always reach right through any un-instrumented code
729         including any of its "handle" expressions, trivial or not.
730    
731       To try this out, do the following:
732    
733         - Erase all existing binfiles for your program.
734           (You may keep binfiles for those modules where you think you
735            definitely don't need back-tracing.)
736         - Turn on back-trace instrumentation:
737              SMLofNJ.Internals.BTrace.mode (SOME true);
738         - Recompile your program.  (I.e., run "CM.make" or "use".)
739         - You may now turn instrumentation off again (if you want):
740              SMLofNJ.Internals.BTrace.mode (SOME false);
741         - Run your program as usual.  If it raises an exception that
742           reaches the interactive toplevel, then a back-trace will
743           automatically be printed.  After that, the toplevel loop
744           will print the exception history as usual.
745    
746    ----------------------------------------------------------------------
747    Name: Matthias Blume
748    Date: 2000/06/26 09:56:46 JST
749    Tag: blume-20000626-setup
750    Description:
751    
752    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
753          ML code before and after compiling a file (e.g., to set compiler
754          flags)
755    
756    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
757              - associated changes to btrace.sml (BTrace instrumentation pass)
758              - cleaner implementation of btimp.sml (BTrace tracing and report
759                module)
760    
761    CM manual: * new path encoding documented
762               * description of setup-parameter to "sml" added
763    
764    The biggest user-visible change to back-tracing is that it is no
765    longer necessary to compile all traced modules within the same
766    session.  (This was a real limitation.)
767    
768    ----------------------------------------------------------------------
769    Name: Matthias Blume
770    Date: 2000/06/24 12:40:00 JST
771    Tag: blume-20000624-startup
772    Description:
773    
774    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
775    bit too often -- to put it mildly. :)
776    
777    ----------------------------------------------------------------------
778    Name: Matthias Blume
779    Date: 2000/06/23 18:20:00 JST
780    Tag: blume-20000623-btrace
781    Description:
782    
783    This updates adds a backtrace facility to aid programmers in debugging
784    their programs.  This involves the following changes:
785    
786    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
787       keeping track of the current call stack.  When programs are compiled
788       in a special mode, the compiler will insert calls to these hooks
789       into the user program.
790       "Hook" means that it is possible for different implementations of
791       back-tracing to register themselves (at different times).
792    
793    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
794       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
795       It can be turned on using this call:
796         SMLofNJ.Internals.BTrace.mode (SOME true);
797       Turning it off again:
798         SMLofNJ.Internals.BTrace.mode (SOME false);
799       Querying the current status:
800         SMLofNJ.Internals.BTrace.mode NONE;
801       Annotated programs are about twice as big as normal ones, and they
802       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
803       where all hooks do nothing).  The slowdown with a plugin that is
804       actually useful (such as the one supplied by default) is even greater,
805       but in the case of the default plugin it is still only an constant
806       factor (amortized).
807    
808    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
809       with a sub-structure BTrace for controlling back-tracing.  In particular,
810       the above-mentioned function "mode" controls whether the annotation
811       phase is invoked by the compiler.  Another important function is
812       "trigger": when called it aborts the current execution and causes
813       the top-level loop to print a full back-trace.
814    
815    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
816       for back-tracing.  It keeps track of the dynamic call stack and in
817       addition to that it keeps a partial history at each "level" of that
818       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
819       d and b (at separate times, dynamically), then the report will show:
820    
821       GOTO   d
822             /c
823       GOTO  \b
824       CALL   a
825    
826       This shows that there was an initial non-tail call of a, then a
827       tail-call to b or c, looping behavior in a cluster of functions that
828       consist of b and c, and then a goto from that cluster (i.e., either from
829       b or from c) to d.
830    
831       Note that (depending on the user program) the amount of information
832       that the back-trace module has to keep track of at each level is bounded
833       by a constant.  Thus, the whole implementation has the same asymptotical
834       complexity as the original program (both in space and in time).
835    
836    5. compiler/TopLevel/interact/evalloop.sml has been modified to
837       handle the special exception SMLofNJ.Internals.BTrace.BTrace
838       which is raised by the "trigger" function mentioned above.
839    
840    Notes on usage:
841    
842    - Annotated code works well together with unannotated code:
843    Unannotated calls simply do not show up at all in the backtrace.
844    
845    - It is not a good idea to let modules that were annotated during
846    different sessions run at the same time.  This is because the compiler
847    chooses small integers to identify individual functions, and there
848    will be clashes if different modules were compiled in separate sessions.
849    (Nothing will crash, and you will even be told about the clashes, but
850    back-trace information will in general not be useful.)
851    
852    - Back-tracing can be confused by callcc and capture.
853    
854    - The only way of getting a back-trace right now is to explicitly
855    invoke the "trigger" function from your user program.  Eventually, we
856    should make every exception carry back-trace information (if
857    available).  But since this creates more overhead at "raise"-time
858    (similar to the current exnHistory overhead), I have not yet
859    implemented this.  (The implementation will be rather easy.)  With
860    exceptions carrying back-trace information, this facility will be even
861    more useful because users don't need to modify their programs...
862    
863    - While it is possible to compile the compiler with back-trace
864    annotations turned on (I did it to get some confidence in
865    correctness), you must make absolutely sure that core.sml and
866    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
867    be compiled with annotation because there is no core access yet, but
868    if you compile btimp.sml with annotation, then the system will go into
869    an infinite recursion and crash.)
870    Since CM currently does not know about BTrace, the only way to turn
871    annotations on and off for different modules of the compiler is to
872    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
873    this is awkward and clumsy.
874    
875    Sample sessions:
876    
877    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
878    - SMLofNJ.Internals.BTrace.mode (SOME true);
879    [autoloading]
880    [autoloading done]
881    val it = false : bool
882    - structure X = struct
883    -     fun main n = let
884    -         fun a (x, 0) = d x
885    -           | a (x, n) = b (x, n - 1)
886    -         and b (x, n) = c (x, n)
887    -         and c (x, n) = a (x, n)
888    -         and d x = e (x, 3)
889    -         and e (x, 0) = f x
890    -           | e (x, n) = e (x, n - 1)
891    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
892    -           | f n = n * g (n - 1)
893    -         and g n = a (n, 3)
894    -     in
895    -         f n
896    -     end
897    - end;
898    structure X : sig val main : int -> int end
899    - X.main 3;
900    *** BACK-TRACE ***
901    GOTO   stdIn:4.2-13.20: X.main[2].f
902    GOTO-( stdIn:4.2-13.20: X.main[2].e
903    GOTO   stdIn:4.2-13.20: X.main[2].d
904         / stdIn:4.2-13.20: X.main[2].a
905         | stdIn:4.2-13.20: X.main[2].b
906    GOTO-\ stdIn:4.2-13.20: X.main[2].c
907    CALL   stdIn:4.2-13.20: X.main[2].g
908    GOTO   stdIn:4.2-13.20: X.main[2].f
909    GOTO-( stdIn:4.2-13.20: X.main[2].e
910    GOTO   stdIn:4.2-13.20: X.main[2].d
911         / stdIn:4.2-13.20: X.main[2].a
912         | stdIn:4.2-13.20: X.main[2].b
913    GOTO-\ stdIn:4.2-13.20: X.main[2].c
914    CALL   stdIn:4.2-13.20: X.main[2].g
915    GOTO   stdIn:4.2-13.20: X.main[2].f
916    GOTO-( stdIn:4.2-13.20: X.main[2].e
917    GOTO   stdIn:4.2-13.20: X.main[2].d
918         / stdIn:4.2-13.20: X.main[2].a
919         | stdIn:4.2-13.20: X.main[2].b
920    GOTO-\ stdIn:4.2-13.20: X.main[2].c
921    CALL   stdIn:4.2-13.20: X.main[2].g
922    GOTO   stdIn:4.2-13.20: X.main[2].f
923    CALL   stdIn:2.15-17.4: X.main[2]
924    -
925    
926    (Note that because of a FLINt bug the above code currently does not
927    compile without BTrace turned on.)
928    
929    Here is another example, using my modified Tiger compiler:
930    
931    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
932    - SMLofNJ.Internals.BTrace.mode (SOME true);
933    [autoloading]
934    [autoloading done]
935    val it = false : bool
936    - CM.make "sources.cm";
937    [autoloading]
938    ...
939    [autoloading done]
940    [scanning sources.cm]
941    [parsing (sources.cm):parse.sml]
942    [creating directory CM/SKEL ...]
943    [parsing (sources.cm):tiger.lex.sml]
944    ...
945    [wrote CM/sparc-unix/semant.sml]
946    [compiling (sources.cm):main.sml]
947    [wrote CM/sparc-unix/main.sml]
948    [New bindings added.]
949    val it = true : bool
950    - Main.compile ("../testcases/merge.tig", "foo.out");
951    *** BACK-TRACE ***
952    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
953    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
954    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
955    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
956    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
957    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
958    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
959         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
960    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
961    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
962    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
963    CALL   main.sml:1.18-118.4: Main.compile[2]
964    -
965    
966    ----------------------------------------------------------------------
967    Name: Matthias Blumen
968    Date: 2000/06/21 18:00:00 JST
969    Tag: blume-20000621-manual
970    Description:
971    
972    CM manual update: Path environments documented.
973    
974    ----------------------------------------------------------------------
975    Name: Matthias Blume
976    Date: 2000/06/19 13:40:00
977    Tag: blume-20000619-manual
978    Description:
979    
980    CM manual and system/README update.  This only covers the fact that
981    there are no more implicit anchors.  (Path environments and the "bind"
982    option to "cm" have yet to be documented.)
983    
984    ----------------------------------------------------------------------
985    Name: Matthias Blume
986    Date: 2000/06/19 11:05:00 JST
987    Tag: blume-20000619-chdir-bugfix
988    Description:
989    
990    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
991    
992    ----------------------------------------------------------------------
993    Name: Matthias Blume
994    Date: 2000/06/18 22:00:10 JST
995    Tag: blume-20000618-implicit-anchors-really-gone
996    Description:
997    
998    I updates the previous HISTORY entry where I forgot to mention that
999    implicit anchors are no longer with us.
1000    
1001    The current update also gets rid of the (now useless) controller
1002    CM.Control.implicit_anchors.
1003    
1004    ----------------------------------------------------------------------
1005    Name: Matthias Blume
1006    Date: 2000/06/16 17:30:00 JST
1007    Tag: blume-20000616-anchorenv
1008    Description:
1009    
1010    This patch implements the long anticipated (just kidding :) "anchor
1011    environment" mechanism.  In the course of doing this, I also
1012    re-implemented CM's internal "SrcPath" module from scratch.  The new
1013    one should be more robust in certain boundary cases.  In any case, it
1014    is a lot cleaner than its predecessor (IMHO).
1015    
1016    This time, although there is yet another boot file format change, I
1017    kept the unpickler backward-compatible.  As a result, no new bootfiles
1018    are necessary and bootstrapping is straightforward.  (You cannot read
1019    new bootfiles into an old system, but the other way around is no
1020    problem.)
1021    
1022    Visible changes:
1023    
1024    ** 0. Implicit path anchors (without the leading $-symbol) are no
1025    longer recognized at all. This means that such path names are not
1026    illegal either.  For example, the name basis.cm simply refers to a
1027    local file called "basis.cm" (i.e, the name is an ordinary path
1028    relative to .cm-files directory).  Or, to put it differently, only
1029    names that start with $ are anchored paths.
1030    
1031    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1032    vanished.
1033    
1034    John (Reppy) had critizised this as soon as I originally proposed and
1035    implemented it, but at that time I did not really deeply believe
1036    him. :) Now I came full-circle because I need the $<singlearc> syntax
1037    in another place where it cannot be seen as an abbreviation for
1038    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1039    seems to mean (i.e., it "expands" into the corresponding anchor
1040    value).
1041    
1042    However, when paths are used as members in CM description files, it
1043    continues to be true that there must be at least another arc after the
1044    anchor.  This is now enforced separately during semantic analysis
1045    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1046    
1047    ** 2. The "cm" class now accepts an option "bind".  The option's value
1048    is a sub-option list of precisely two items -- one labeled "anchor"
1049    and the other one labeled "value".  As you might expect, "anchor" is
1050    used to specify an anchor name to be bound, and "value" specifies what
1051    the anchor is being bound to.
1052    
1053    The value must be a directory name and can be given in either standard
1054    syntax (including the possibility that it is itself an anchored path)
1055    or native syntax.
1056    
1057    Examples:
1058    
1059       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1060       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1061    
1062    and so on.
1063    
1064    The meaning of this is that the .cm-file will be processed with an
1065    augmented anchor environment where the given anchor(s) is/are bound to
1066    the given values(s).
1067    
1068    The rationale for having this feature is this: Suppose you are trying
1069    to use two different (already stable) libraries a.cm and b.cm (that
1070    you perhaps didn't write yourself).  Further, suppose each of these
1071    two libraries internally uses its own auxiliary library $aux/lib.cm.
1072    Normally you would now have a problem because the anchor "lib" can not
1073    be bound to more than one value globally.  Therefore, the project that
1074    uses both a.cm and b.cm must locally redirect the anchor to some other
1075    place:
1076    
1077       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1078       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1079    
1080    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1081    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1082    
1083    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1084    you when you do so at the time of CM.stabilize).  Therefore, you can
1085    also use an anchored path as the value:
1086    
1087      a.cm (bind:(anchor:lib value:$a-lib))
1088      b.cm (bind:(anchor:lib value:$b-lib))
1089    
1090    Now you can globally configure (using the usual CM.Anchor.anchor or
1091    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1092    itself is always locally bound, setting it globally is no longer
1093    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1094    can still be used as a global anchor for separate purposes.  As a
1095    matter of fact, one can locally define "lib" in terms of a global
1096    "lib":
1097    
1098      a.cm (bind:(anchor:lib value:$lib/a))
1099      b.cm (bind:(anchor:lib value:$lib/b))
1100    
1101    ** 3: The encoding of path names has changed.  This affects the way
1102    path names are shown in CM's progress report and also the internal
1103    protocol encoding used for parallel make.
1104    
1105    The encoding now uses one or more ':'-separated segments.  Each
1106    segments corresponds to a file that has been specified relative to the
1107    file given by its preceding segment.  The first segment is either
1108    relative to the CWD, absolute, or anchored.  Each segment itself is
1109    basically a Unix pathname; all segments but the first are relative.
1110    
1111    Example:
1112    
1113       $foo/bar/baz.cm:a/b/c.sml
1114    
1115    This path denotes the file bar/a/b/c.sml relative to the directory
1116    denoted by anchor "foo".  Notice that the encoding also includes
1117    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1118    paths are resolved relative to the .cm-files directory, so baz.cm must
1119    be ignored to get the "real" pathname.
1120    
1121    To make this fact more obvious, CM puts the names of such "virtual
1122    arcs" into parentheses when they appear in progress reports. (No
1123    parentheses will appear in the internal protocol encoding.)  Thus,
1124    what you really see is:
1125    
1126      $foo/bar/(baz.cm):a/b/c.sml
1127    
1128    I find this notation to be much more informative than before.
1129    
1130    Another new feature of the encoding is that special characters
1131    including parentheses, colons, (back)slashes, and white space are
1132    written as \ddd (where ddd is the decimal encoding of the character).
1133    
1134    *** The CM manual still needs to be updated.
1135    
1136    ----------------------------------------------------------------------
1137    Name: Allen Leung
1138    Date: 2000/06/15 00:38:00
1139    Tag: leunga-20000615-x86-peephole
1140    
1141    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1142    
1143    ----------------------------------------------------------------------
1144    Name: Matthias Blume
1145    Date: 2000/06/12 11:40:00
1146    Tag: blume-20000612-parmakefix
1147    Description:
1148    
1149    More cleanup after changing the file naming scheme: This time I
1150    repaired the parallel make mechanism for CMB.make which I broke earlier.
1151    
1152    ----------------------------------------------------------------------
1153    Name: Allen Leung
1154    Date: 2000/06/09 01:25:00
1155    Tag: leunga-20000609-various
1156    
1157    None of these things should affect normal SML/NJ operations
1158    
1159    1. Peephole improvements provided by Fermin (c--)
1160    2. New annotation DEFUSE for adding extra dependence (moby)
1161    3. New X86 LOCK instructions (moby)
1162    4. New machine description language for reservation tables (scheduling)
1163    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1164       trees etc.)
1165    6. I've changed the CM files so that they can work with versions
1166       110.0.6, 110.25 and 110.28
1167    
1168    ----------------------------------------------------------------------
1169    Name: Matthias Blume
1170    Date: 2000/06/09 12:40:00
1171    Tag: blume-20000609-log
1172    Description:
1173    
1174    - Removed all(?) remaining RCS Log entries from sources.
1175    
1176    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1177      anchored paths).
1178    
1179    ----------------------------------------------------------------------
1180    Name: Matthias Blume
1181    Date: 2000/06/07 17:00:00 JST
1182    Tag: blume-20000607-no-implicit-anchors
1183    Description:
1184    
1185    1. This update changes the default setting for
1186    CM.Control.implicit_anchors from true to false.  This means that
1187    implicit anchors are no longer permitted by default.  I also tried to
1188    make sure that nothing else still relies on implicit anchors.
1189    (This is the next step on the schedule towards a CM that does not even
1190    have the notion of implicit anchors anymore.)
1191    
1192    2. More CM manual updates.
1193    
1194    3. I managed to track down and fix the pickling bug I mentioned last
1195    time.  Because of the previously existing workaround, this entails no
1196    immediate practical changes.
1197    
1198    ----------------------------------------------------------------------
1199    Name: Matthias Blume
1200    Date: 2000/06/06 11:15:00 JST
1201    Tag: blume-20000606-lazierpickle
1202    Description:
1203    
1204    !!!! NEW BOOT FILES !!!!
1205    
1206    * The main purpose of this update is to make library pickles lazier in
1207    order to reduce the initial space penalty for autoloading a library.
1208    As a result, it is now possible to have $smlnj/compiler.cm
1209    pre-registered.  This should take care of the many complaints or
1210    inquiries about missing structure Compiler.  This required changes to
1211    CM's internal data structures and small tweaks to some algorithms.
1212    
1213    As a neat additional effect, it is no longer necessary (for the sake
1214    of lean heap image files) to distinguish between a "minimal" CM and a
1215    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1216    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1217    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1218    
1219    To make the life of compiler-hackers easier, "makeml" now also
1220    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1221    words, after you bootstrap a new sml for the first time, you will not
1222    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1223    around you will still have to do it, though.)
1224    
1225    * A second change consists of major updates to the CM manual.  There
1226    are now several appendices with summary information and also a full
1227    specification of the CM description file syntax.
1228    
1229    * In directory src/system I added the script "allcross".  This script
1230    invokes sml and cross-compiles the compiler for all supported
1231    architectures.  (Useful when providing a new set of boot files.)
1232    
1233    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1234    added a small tweak to pickle-util.sml to work around this problem,
1235    but it is not a proper fix yet.  I will investigate further.  (The
1236    effect of the bug was an inflation of library pickle size.)
1237    
1238    * Version number increased to 110.28.1 (to avoid compatibility problems).
1239    
1240    ----------------------------------------------------------------------
1241    Name: Allen Leung
1242    Date: 2000/05/25 17:28 EDT
1243    Tag: leunga-20000525-ra
1244    Description:
1245    
1246      Fixed a bug in freezing phase of the register allocator.
1247    
1248    ----------------------------------------------------------------------
1249    Name: Allen Leung
1250    Date: 2000/05/15 22:53 EDT
1251    Tag: leunga-20000515-alpha-x86-ra
1252    Description:
1253    
1254      1. Alpha
1255    
1256          Slight cleanup.  Removed the instruction SGNXL
1257    
1258      2. X86
1259    
1260          Added the following instructions to the instruction set:
1261    
1262            ROLx, RORx,
1263            BTx, BTSx, BTLx, BTRx,
1264            XCHGx, and variants with the LOCK prefix
1265    
1266      3. Register Allocation
1267    
1268          The module ra-rewrite-with-renaming has been improved.
1269    
1270      These have no effect on SML/NJ.
1271    
1272    ----------------------------------------------------------------------
1273    Name: Matthias Blume
1274    Date: 2000/05/15 16:20:00 JST
1275    Tag: blume-20000515-lightrebuild
1276    Description:
1277    
1278    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1279       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1280       defined.  In effect, the command will not build any cross-compiler
1281       backends and therefore finish more quickly.
1282    
1283       The "fixpt" script also takes a "-light" switch to be able to use
1284       this new facility while compiling for a fixpoint.
1285    
1286    2. I replaced all mentions of anchored paths in group owner specifications
1287       with simple relative paths (usually starting with "..").
1288       The rationale is that a library's internal workings should not be
1289       compromised by the lack of some anchor.  (An anchor is necessary
1290       for someone who wants to refer to the library by an anchored path,
1291       but it should not be necessary to build the same library in the first
1292       place.)
1293    
1294    3. I changed the way CM's tool mechanism determines the shell command
1295       string used for things like ml-yacc etc. so that it does not break
1296       when CM.Control.implicit_anchors is turned off.
1297    
1298    ----------------------------------------------------------------------
1299    Name: Matthias Blume
1300    Date: 2000/05/12 18:20:00 JST
1301    Tag: blume-20000512-ml-build
1302    Description:
1303    
1304    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1305    getting installed properly (by config/install.sh).
1306    
1307    ----------------------------------------------------------------------
1308    Name: Matthias Blume
1309    Date: 2000/05/12 17:30:00 JST
1310    Tag: blume-20000512-anchors
1311    Description:
1312    
1313    !!! NEW BOOT FILES !!!
1314    
1315    This change is in preparation of fading out support for "implicitly
1316    anchored path names".  I went through all sources and used the
1317    explicit (and relatively new) $-notation.  See system/README and the
1318    CM manual for more info on this.
1319    
1320    I also modified the anchoring scheme for some things such as "smlnj",
1321    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1322    anchors are more expressive: anchor name and first arc do not have to
1323    coincide.  This entails the following user-visible change:
1324    
1325    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1326    particular, when you fire up sml with a command-line argument, say,
1327    e.g.:
1328    
1329       sml '$smlnj/cmb.cm'
1330    
1331    At the ML toplevel prompt:
1332    
1333       CM.autoload "$smlnj/cmb.cm";
1334    
1335    There is also a new controller in CM.Control that can be used to turn
1336    off all remaining support for implicit anchors by saying:
1337    
1338        CM.autoload "$smlnj/
1339        #set CM.Control.implicit_anchors false;
1340    
1341    This causes CM to reject implicitly anchored paths.  This is (for the
1342    time being) less permissive than the "final" version where there will
1343    be no more such implicit anchors and relative paths will be just that:
1344    relative.
1345    
1346    The next step (version after next version?) will be to make the
1347    default for CM.Control.implicit_anchors false.  After the dust has
1348    settled, I can then produce the "final" version of this...
1349    
1350    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1351    
1352    ----------------------------------------------------------------------
1353    Name: Matthias Blume
1354    Date: 2000/05/11 16:30:00 JST
1355    Tag: blume-20000511-sources
1356    Description:
1357    
1358    The main change is that I added function CM.sources as a generalized
1359    version of the earlier CM.makedepend.  This entails the following
1360    additional changes:
1361    
1362      - CM.makedepend has been dropped.
1363    
1364      - CM manual has been updated.
1365    
1366      - TOOLS signature and API have been changed.
1367    
1368    ----------------------------------------------------------------------
1369    Name: Allen Leung
1370    Date: 2000/05/10 21:17 EDT
1371    Tag: leunga-20000510-moby-c--ssa
1372    Description:
1373    
1374      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1375    None of these affect SML/NJ.
1376    
1377    1. Register Allocation
1378    
1379        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1380           This module tries to remove local (i.e. basic block level) redundancies
1381           during spilling.
1382    
1383        b. A new framework for performing region based register allocation.
1384           Not yet entirely functional.
1385    
1386    2. X86
1387    
1388       a. DefUse for POP was missing the stack pointer [found by Lal]
1389       b. Reload for CALL was incorrect in X86Spill [found by John]
1390       c. Various fixes in X86Spill so that it can be used correctly for
1391          the new spilling module.
1392    
1393    3. SSA/IR
1394    
1395       a. New module ir/dj-dataflow.sml implements elimination based
1396          data flow analysis.
1397    
1398    4. MLRiscGen
1399    
1400       a. Fix for gc type annotation
1401    
1402    5. MDGen
1403    
1404       Various fixes for machine description -> ml code translation.  For ssa
1405       only.
1406    
1407    ----------------------------------------------------------------------
1408    Name: Allen Leung
1409    Date: 2000/05/08 22:17 EDT
1410    Tag: leunga-20000508-labexp
1411    Description:
1412    
1413      Fermin has found a few assembly problems with constant expressions
1414      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1415      which choke on dumb assemblers.  This is his fix.
1416    
1417    ----------------------------------------------------------------------
1418    Name: Dave MacQueen
1419    Date: 2000/04/09 14:00 EDT
1420    Tag: dbm-20000502-Version_110_28
1421    Description:
1422    
1423    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1424    
1425    2. Updated config/version to 110.28
1426    
1427    3. Updated config/srcarchiveurl
1428    
1429    3. New boot files!
1430       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1431    
1432    ----------------------------------------------------------------------
1433    Name: Matthias Blume
1434    Date: 2000/05/01 19:05:00 JST
1435    Tag: blume-20000501-noweb
1436    Description:
1437    
1438    A new noweb tool has been added.  The existing system is entirely
1439    unaffected by this, but some CM users have asked for renewed noweb
1440    support.  Everything is documented in the CM manual.
1441    
1442    New (plugin) libraries:
1443    
1444       noweb-tool.cm
1445       nw-ext.cm
1446    
1447    ----------------------------------------------------------------------
1448    Name: Dave MacQueen
1449    Date: 2000/04/30 12:40PM EDT
1450    Tag: dbm-20000430-bug_fixes
1451    Description:
1452    
1453    1. Fix for bug 1498
1454       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1455       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1456         added toRealArray function
1457       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1458         added check for tag Obj.RealArray to array printing case in ppObj
1459    
1460    2. Fix for bug 1510
1461       smlnj/src/compiler/Semant/types/typesutil.sml
1462         fixed definition of dummyargs (used by equalTycon) so that
1463         dummy args are distinct types
1464    
1465    ----------------------------------------------------------------------
1466    Name: Matthias Blume
1467    Date: 2000/04/30 01:00:00 JST
1468    Tag: blume-20000430-versions
1469    Description:
1470    
1471    1. CM version numbering added.  This is an implementation of Lal's
1472       proposal for adding version numbers and version checking to .cm
1473       files.  Lal said that his proposal was just that -- a proposal.
1474       For the time being I went ahead and implemented it so that people
1475       can comment on it.  Everything is completely backward-compatible
1476       (except for the stable library format, i.e., new bootfiles!).
1477    
1478       As usual, see the CM manual for details.
1479    
1480    2. An alternative syntax for anchored paths has been implemented.
1481       Dave has recently voiced the same concerns that I had when I did
1482       this, so there should be some support.  My take is that eventually
1483       I will let support for the current syntax (where anchors are
1484       "implicit") fade out in favor of the new, explicit syntax.
1485       In order to be backward-compatible, both old and new syntax are
1486       currently supported.
1487    
1488       Again, see the CM manual for details.
1489    
1490    3. Parallel make is trying to be slightly smarter:  When the master
1491       process finds a "bottleneck", i.e., when there is only one
1492       compilation unit that can be compiled and everybody else is
1493       waiting on it, then it will simply compile it directly instead
1494       of clumsily telling one of the slaves to do it.
1495    
1496    4. Support for "unsharing" added.  This is necessary in order to be
1497       able to have two different versions of the same library running
1498       at the same time (e.g., for trying out a new MLRISC while still
1499       having the old MLRISC linked into the current compiler, etc.)
1500       See the CM manual.
1501    
1502    5. Simple "makedepend" functionality added for generating Makefile
1503       dependency information.  (This is rather crude at the moment.
1504       Expect some changes here in the future.)
1505    
1506    6. ".fun" added as a recognized suffix for ML files. Also documented
1507       explicitly in the manual that the fallback behavior (unknown suffix
1508       -> ML file) is not an official feature!
1509    
1510    7. Small changes to the pickler for stable libraries.
1511    
1512    8. Several internal changes to CM (for cleanup/improvement).
1513    
1514    
1515    !!!! NEW BINFILES !!!!
1516    
1517    ----------------------------------------------------------------------
1518    Name: Matthias Blume
1519    Date: 2000/04/28 17:30:00 JST
1520    Tag: blume-20000428-pathconfig
1521    Description:
1522    
1523    1. I changed config/install.sh to remove duplicate entries from the
1524       lib/pathconfig file at the end.  Moreover, the final version of
1525       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1526       in src/system/installml.
1527    
1528    2. The config/install.sh script now consistently uses relative
1529       pathnames in lib/pathconfig whenever the anchor is in the lib
1530       directory.  (So far this was true for the libraries that come
1531       pre-compiled and bundled as part of the bootfiles but not for
1532       libraries that are compiled by the script itself.)
1533    
1534    ----------------------------------------------------------------------
1535    Name: Matthias Blume
1536    Date: 2000/04/26 13:10:00 JST
1537    Tag: blume-20000426-fun_suffix
1538    Description:
1539    
1540    Added ".fun" as a recognized file name suffix (for ML code).
1541    
1542    ----------------------------------------------------------------------
1543    Name: Allen Leung
1544    Date: 2000/04/25 17:00:00 EST
1545    Tag: leunga-20000425-alpha-ra
1546    Description:
1547    
1548    1. Alpha
1549    
1550        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1551    in C--.
1552    
1553    2. RA
1554    
1555       Added a flag "ra-dump-size" to print out the size of the flowgraph
1556       and the interference graph.
1557    
1558    ----------------------------------------------------------------------
1559    Name: Dave MacQueen
1560    Date: 2000/04/25/
1561    Tag: dbm-20000425-mlyacc_doc_examples
1562    Description:
1563      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1564      Updated all three examples in src/ml-yacc/examples to run
1565      under 110.* using CM.make.
1566    
1567    ----------------------------------------------------------------------
1568    Name: Allen Leung
1569    Date: 2000/04/20 23:04:00 EST
1570    Tag: leunga-20000420-ssa-c---stuff
1571    Description:
1572    
1573      This update synchronizes my repository with Yale's.  Most of these
1574    changes, however, do not affect SML/NJ at all (the RA is an exception).
1575    
1576    1. Register Allocator
1577    
1578       a. An improvement in the interference graph construction:
1579          Given a copy
1580    
1581                s <- t
1582    
1583          no interference edge between s and t is added for this definition of s.
1584    
1585       b. I've added two new spill heuristic modules that Fermin and I developed
1586          (in the new library RA.cm). These are unused in SML/NJ but maybe
1587          useful for others (Moby?)
1588    
1589    2. X86
1590    
1591       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1592    
1593    3. Alpha
1594    
1595       a. Added the BSR instruction and code generation that goes with it [C--]
1596       b. Other fixes too numerous to recount provided by Fermin [C--]
1597    
1598    4. Regmaps
1599    
1600       a. The regmaps are not initialized with the identity physical bindings
1601          at creation time.  This is unneeded.
1602    
1603    5. MLRISC Optimizations
1604    
1605       a. The DJ-Graph module can now compute the iterated dominance frontiers
1606          intersects with liveness incrementally in linear time! Woohoo!
1607          This is now used in my new SSA construction algorithm.
1608    
1609       b. THe branch reorganization module is now smarter about linear chains of
1610          basic blocks.
1611    
1612    
1613    ----------------------------------------------------------------------
1614    Name: Matthias Blume
1615    Date: 2000/04/12 13:52:00 JST
1616    Tag: blume_main_v110p27_1
1617    Description:
1618    
1619    Changed install.sh script to handle archive files without version number
1620    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1621    name of the boot file archive.
1622    
1623    ----------------------------------------------------------------------
1624    Name: Dave MacQueen
1625    Date: 2000/04/09 14:00 EDT
1626    Tag: dbm-20000410-Version_110_27
1627    Description:
1628    
1629    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1630    
1631    2. Updated src/config/version to 110.27
1632    
1633    3. New boot files!
1634    
1635    ----------------------------------------------------------------------
1636    Name: Allen Leung
1637    Date: 2000/04/09 19:09:00 EST
1638    Tag: leunga-20000409-misc
1639    Description:
1640    
1641    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1642    
1643    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1644    
1645    ----------------------------------------------------------------------
1646    Name: Stefan
1647    Date: 2000/04/07 10:00:00 EDT
1648    Tag: monnier-20000406-branch-handling
1649    Description:
1650    
1651    Improved handling of branches (mostly those generated from
1652    polymorphic equality), removed switchoff and changed the
1653    default optimization settings (more cpsopt and less flintopt).
1654    
1655    ----------------------------------------------------------------------
1656    Name: Allen Leung
1657    Date: 2000/04/06 01:30:00 EST
1658    Tag: leunga-20000406-peephole-x86-SSA-2
1659    Description:
1660    
1661       Forgot a few files.
1662    
1663    ----------------------------------------------------------------------
1664    Name: Allen Leung
1665    Date: 2000/04/06 00:36:00 EST
1666    Tag: leunga-20000406-peephole-x86-SSA
1667    Description:
1668    
1669    1.  New Peephole code
1670    
1671    2.  Minor improvement to X86 instruction selection
1672    
1673    3.  Various fixes to SSA and machine description -> code translator
1674    
1675    ----------------------------------------------------------------------
1676    Name: Matthias Blume
1677    Date: 2000/04/05 12:30:00 JST
1678    Tag: blume_main_v110p26p2_3
1679    Description:
1680    
1681    This update just merges three minor cosmetic updates to CM's sources
1682    to get ready for the 110.27 code freeze on Friday.  No functionality
1683    has changed.
1684    
1685    ----------------------------------------------------------------------
1686    Name: Allen Leung
1687    Date: 2000/04/04 19:39:00 EST
1688    Tag: leunga-20000404-x86-asm
1689    Description:
1690    
1691    1.  Fixed a problem in X86 assembly.
1692    
1693        Things like
1694    
1695           jmp %eax
1696           jmp (%eax)
1697    
1698        should be output as
1699    
1700           jmp *%eax
1701           jmp *(%eax)
1702    
1703    2.  Assembly output
1704    
1705          Added a new flag
1706    
1707              "asm-indent-copies" (default to false)
1708    
1709          When this flag is on, parallel copies will be indented an extra level.
1710    
1711    ----------------------------------------------------------------------
1712    Name: Allen Leung
1713    Date: 2000/04/04 03:18:00 EST
1714    Tag: leunga-20000404-C--Moby
1715    Description:
1716    
1717        All of these fixes are related to C--, Moby, and my own optimization
1718        stuff; so they shouldn't affect SML/NJ.
1719    
1720    1.  X86
1721    
1722        Various fixes related floating point, and extensions.
1723    
1724    2.  Alpha
1725    
1726        Some extra patterns related to loads with signed/zero extension
1727        provided by Fermin.
1728    
1729    3.  Assembly
1730    
1731        When generating assembly, resolve the value of client defined constants,
1732        instead of generating symbolic values.  This is controlled by the
1733        new flag "asm-resolve-constants", which is default to true.
1734    
1735    4.  Machine Descriptions
1736    
1737        a. The precedence parser was slightly broken when parsing infixr symbols.
1738        b. The type generalizing code had the bound variables reversed, resulting
1739           in a problem during arity raising.
1740        c. Various fixes in machine descriptions.
1741    
1742    ----------------------------------------------------------------------
1743    Name: Matthias Blume
1744    Date: 2000/04/03 16:05:00 JST
1745    Tag: blume_main_v110p26p2_2
1746    Description:
1747    
1748    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1749    now done via the ordinary static environment that is context to each
1750    compilation unit.
1751    
1752    To this end, I arranged that instead of "structure Core" as "structure
1753    _Core" is bound in the pervasive environment.  Core access is done via
1754    _Core (which can never be accidentally rebound because _Core is not a
1755    legal surface-syntax symbol).
1756    
1757    The current solution is much cleaner because the core environment is
1758    now simply part of the pervasive environment which is part of every
1759    compilation unit's context anyway.  In particular, this eliminates all
1760    special-case handling that was necessary until now in order to deal
1761    with dynamic and symbolic parts of the core environment.
1762    
1763    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1764    compilation manager's bootstrap compiler (actually: in the "init group"
1765    handling).  See the comments in src/system/smlnj/init/init.cmi for
1766    more details.
1767    
1768    I also tried to track down all mentions of "Core" (as string argument
1769    to Symbol.strSymbol) in the compiler and replaced them with a
1770    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1771    name appears in one place only.
1772    
1773    Binfile and bootfile format have not changed, but the switchover from
1774    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1775    bootfiles anyway.
1776    
1777    ----------------------------------------------------------------------
1778    Name: Allen Leung
1779    Date: 2000/04/02 21:17:00 EST
1780    Tag: leunga-20000402-mltree
1781    Description:
1782    
1783       1. Renamed the constructor CALL in MLTREE by popular demand.
1784       2. Added a bunch of files from my repository.  These are currently
1785          used by other non-SMLNJ backends.
1786    
1787    ----------------------------------------------------------------------
1788    Name: Allen Leung
1789    Date: 2000/03/31 21:15:00 EST
1790    Tag: leunga-20000331-aliasing
1791    Description:
1792    
1793    This update contains a rewritten (and hopefully more correct) module
1794    for extracting aliasing information from CPS.
1795    
1796       To turn on this feature:
1797    
1798            Compiler.Control.CG.memDisambiguate := true
1799    
1800       To pretty print the region information with assembly
1801    
1802           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1803    
1804       To control how many levels of aliasing information are printed, use:
1805    
1806           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1807    
1808       The default of n is 3.
1809    
1810  ----------------------------------------------------------------------  ----------------------------------------------------------------------
1811  Name: David MacQueen  Name: David MacQueen
# Line 81  Line 1876 
1876     (specified in the .cm file at each instance where the tool's class is     (specified in the .cm file at each instance where the tool's class is
1877     used).     used).
1878    
1879     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1880     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1881     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1882    
1883     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 94  Line 1889 
1889    
1890     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1891    
1892     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1893     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1894    
1895  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 228  Line 2023 
2023    
2024       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2025       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2026       testl is perferable to andl.       testl is preferable to andl.
2027    
2028  7. RA (x86 only)  7. RA (x86 only)
2029    
# Line 410  Line 2205 
2205    
2206  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2207  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2208  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
2209  explanation.  explanation.
2210    
2211  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2212  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2213  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2214  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 457  Line 2252 
2252  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
2253  <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,
2254  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
2255  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2256  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2257  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2258    
# Line 507  Line 2302 
2302    it from that remote directory.    it from that remote directory.
2303    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2304    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2305    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2306    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2307    
2308    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.589  
changed lines
  Added in v.775

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