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 593, Tue Apr 4 07:26:33 2000 UTC revision 761, Sat Dec 23 05:37:37 2000 UTC
# Line 12  Line 12 
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15    Name: Lal George
16    Date: 2000/12/22 22:22:58 EST 2000
17    Tag: Release_110_32
18    Description:
19    
20            Infinite precision used throughout MLRISC.
21            see MLRISC/mltree/machine-int.sig
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume
25    Date: 2000/12/22 21:20:00 JST
26    Tag: blume-20001222-anchorenv
27    Description:
28    ----------------------------------------------------------------------
29    Name: Matthias Blume
30    Date: 2000/12/22 23:16:00 JST
31    Tag: blume-20001222-warn
32    Description:
33    
34    Corrected wording and formatting of some CM warning message which I
35    broke in my previous patch.
36    
37    ----------------------------------------------------------------------
38    Name: Matthias Blume
39    Date: 2000/12/22 13:15:00 JST
40    Tag: blume-20001222-cleanup
41    Description:
42    
43    Removed src/cm/ffi which does not (and did not) belong here.
44    
45    ----------------------------------------------------------------------
46    Name: Matthias Blume
47    Date: 2000/12/21 23:55:00 JST
48    Tag: blume-20001221-exn
49    Description:
50    
51    Probably most important: CM no longer silently swallows all exceptions
52    in the compiler.
53    Plus: some other minor CM changes.  For example, CM now reports some
54    sizes for generated binfiles (code, data, envpickle, lambdapickle).
55    
56    ----------------------------------------------------------------------
57    Name: Matthias Blume
58    Date: 2000/12/15 00:01:05 JST
59    Tag: blume-20001215-dirtool
60    Description:
61    
62    - "dir" tool added.
63    - improvements and cleanup to Tools structure
64    - documentation updates
65    
66    ----------------------------------------------------------------------
67    Name: Allen Leung
68    Date: Thu Dec 14 03:45:24 EST 2000
69    Description:
70    Tag:  leunga-20001214-int-inf
71    Description:
72    
73       In IntInf, added these standard functions, which are missing from our
74    implementation:
75    
76        andb : int * int -> int
77        xorb : int * int -> int
78        orb  : int * int -> int
79        notb : int -> int
80         <<   : int * word -> int
81        ~>>  : int * word -> int
82    
83       Not tested, I hope they are correct.
84    
85    ----------------------------------------------------------------------
86    Name: Allen Leung
87    Date: Fri Dec  8 19:23:26 EST 2000
88    Description:
89    Tag:  leunga-20001208-nowhere
90    Description:
91    
92      Slight improvements to the 'nowhere' tool to handle OR-patterns,
93    to generate better error messages etc.  Plus a brief manual.
94    
95    ----------------------------------------------------------------------
96    Name: Lal George
97    Date: 2000/12/08 09:54:02 EST 2000
98    Tag: Release_110_31
99    Description:
100    
101    - Version 110.31
102    ----------------------------------------------------------------------
103    Name: Allen Leung
104    Date: Thu Dec  7 22:01:04 EST 2000
105    Tag:  leunga-20001207-cell-monster-hack
106    Description:
107    
108    Major MLRISC internal changes.  Affect all clients.
109    Summary:
110    
111    1.  Type CELLS.cell = int is now replaced by a datatype.
112        As a result, the old regmap is now gone.  Almost all interfaces
113        in MLRISC change as a consequence.
114    
115    2.  A new brand version of machine description tool (v3.0) that generates
116        modules expecting the new interface.  The old version is removed.
117    
118    3.  The RA interface has been further abstracted into two new functors.
119        RISC_RA and X86RA.  These functors have much simpler interfaces.
120        [See also directory MLRISC/demo.]
121    
122    4.  Some other new source->source code generation tools are available:
123    
124        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
125        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
126           I use this tool to generate the peephole optimizers---with the new
127           cell type changes, peephole rules are becoming difficult to write
128           without conditional pattern matching.
129    
130    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
131        cover the entire MLRISC source tree so many things broke.
132    
133    6.  CM files have been moved to the subdirectory MLRISC/cm.
134        They are moved because there are a lot of them and they clutter up the
135        root dir.
136    
137    7.  More detailed documentation to come...
138    
139        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
140        a makeml -rebuild first.  This is because of other other
141        changes that Matthias has made (see below).
142    
143    
144    ----------------------------------------------------------------------
145    Name: Matthias Blume
146    Date: 2000/11/30 23:12:00 JST
147    Tag: blume-20001130-filereorg
148    Description:
149    
150    Some manual updates and some file reorganizations in CM.
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume
154    Date: 2000/11/24 17:45:00 JST
155    Tag: blume-20001124-link
156    Description:
157    
158    Drastically improved link traversal code for the case that the dynamic
159    value was already loaded at bootstrap time.  As a result, CM and CMB
160    now both load blazingly fast -- even on a very slow machine.  Also,
161    memory consumption has been further reduced by this.
162    
163    Warning: The format of the PIDMAP file has changed.  THerefore, to
164    bootstrap you have to do this:
165    
166    1. Run CMB.make
167    2. Make a symbolic link for the boot directory:
168         ln -s sml.boot.ARCH-OS xxx
169    3. "Rebuild" the boot directory:
170         ./makeml -boot xxx -rebuild sml ; rm xxx
171    4. Boot normally:
172          ./makeml
173    
174    ----------------------------------------------------------------------
175    Name: Matthias Blume
176    Date: 2000/11/21 21:20:00 JST
177    Tag: blume-20001121-tools
178    Description:
179    
180    Continued hacking on autoloading problem -- with success this time.
181    Also changed tool-plugin mechanism.  See new CM manual.
182    
183    ----------------------------------------------------------------------
184    Name: Matthias Blume
185    Date: 2000/11/19 14:30:00 JST
186    Tag:  blume-20001119-autoload
187    Description:
188    
189    Some hacking to make autoloading faster.  Success for CMB, no success
190    so far for CM.  There is a reduced structure CM' that autoloads faster.
191    (This is a temporary, non-documented hack to be eliminated again when
192    the general problem is solved.)
193    
194    ----------------------------------------------------------------------
195    Name: Matthias Blume
196    Date: 2000/11/17 14:10:00 JST
197    Tag: blume-20001117-pickle-lib
198    Description:
199    
200    1. Eliminated comp-lib.cm
201    2. Made pickle-lib.cm
202    3. Eliminated all uses of intset.sml (from comp-lib.cm)
203    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
204       equivalent constructs from smlnj-lib.cm (INtHashTable).
205    5. Point 4. also goes for those uses of intmap.* in MLRISC.
206       Duplicated intmap modules thrown out.
207    6. Hunted down all duplicated SCC code and replaced it with
208       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
209    7. Rewrote Feedback module.
210    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
211       should be thrown out and equivalent modules from smlnj-lib.cm
212       should be used (IntRedBlackSet, IntListSet, ...).
213    
214    Confirmed that compiler compiles to fixpoint.
215    
216    ----------------------------------------------------------------------
217    Name: Allen Leung
218    Date: 2000/11/10 18:00:00
219    Tag: leunga-20001110-new-x86-fp
220    
221    A new x86 floating point code generator has been added.
222    By default this is turned off.  To turn this on, do:
223    
224        CM.autoload "$smlnj/compiler.cm";
225        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
226    
227    Changes:
228    
229    1.  Changed FTAN to FPTAN so that the assembly output is correct.
230    2.  Changed the extension callback for FTANGENT to generate:
231    
232              fptan
233              fstp  %st(0)
234        instead of
235              fptan
236              fstpl ftempmem
237    
238    3.  Numerous assembly fixes for x86.
239    
240    5.  Cleaned up the machine code output module x86/x86MC.sml and added
241        support for a whole bunch of instructions and addressing modes:
242    
243          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
244          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
245          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
246          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
247          fxch %st(n)
248          fld %st(n)
249          fst %st(n)
250          fst mem
251          fstp %st(n)
252          fucom %st(n)
253          fucomp %st(n)
254    
255        All these are now generated when the fast fp mode is turned on.
256    
257    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
258    
259    ----------------------------------------------------------------------
260    Name: Matthias Blume
261    Date: 2000/11/09 11:20:00 JST
262    Tag: blume-20001109-scc
263    Description:
264    
265    Eliminated some code duplication:
266    
267    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
268       (Otherwise the functor is useless.)
269    2. Used GraphSCCFn where SCCUtilFun was used previously.
270    3. Got rid of SCCUtilFun (in comp-lib.cm).
271    
272    ----------------------------------------------------------------------
273    Name: Lal George
274    Date: 2000/11/06 09:02:21 EST 2000
275    Tag: Release_110_30
276    Description:
277    
278    - Version 110.30
279    ----------------------------------------------------------------------
280    Name: Matthias Blume
281    Date: 2000/11/04 14:45:00
282    Tag: blume-20001104-mlbuild
283    Description:
284    
285    - Made ml-build faster on startup.
286    - Documentation fixes.
287    
288    ----------------------------------------------------------------------
289    Name: Matthias Blume
290    Date: 2000/11/02 17:00:00 JST
291    Tag: blume-20001102-condcomp
292    Description:
293    
294    - Small tweaks to pickler -- new BOOTFILES!
295    - Version bumped to 110.29.2.
296    - Added conditional compilation facility to init.cmi (see comment there).
297    ----------------------------------------------------------------------
298    Name: Allen Leung
299    Date: 2000/10/23 19:31:00
300    Tag: leunga-20001023-demo-ra
301    
302    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
303    2. Test programs for the graph library updated
304    3. Some new MLRISC demo programs added
305    
306    ----------------------------------------------------------------------
307    Name: Matthias Blume
308    Date: 2000/08/31 22:15:00 JST
309    Tag: blume-20001017-errmsg
310    Description:
311    
312    More error message grief: Where there used to be no messages, there
313    now were some that had bogus error regions.  Fixed.
314    
315    ----------------------------------------------------------------------
316    Name: Matthias Blume
317    Date: 2000/08/31 17:30:00 JST
318    Tag: blume-20001017-v110p29p1
319    Description:
320    
321    I made a version 110.29.1 with new bootfiles.
322    
323    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
324              CM documentation changes and a small bugfix in CM's error reporting.
325    
326    ----------------------------------------------------------------------
327    Name: Lal George
328    Date: 2000/09/27 14:42:35 EDT
329    Tag: george-20000927-nodestatus
330    Description:
331    
332    Changed the type of the nodestatus, so that:
333    
334            SPILLED(~1)             is now SPILLED
335            SPILLED(m) where m>=0   is now MEMREG(m)
336            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
337    
338    ----------------------------------------------------------------------
339    Name: Matthias Blume
340    Date: 2000/09/07 14:45:00 JST
341    Tag: blume-20000907-cmerrmsg
342    Description:
343    
344    Small tweak to CM to avoid getting ML syntax error messages twice.
345    
346    ----------------------------------------------------------------------
347    Name: Matthias Blume
348    Date: 2000/08/31 18:00:00 JST
349    Tag: blume-20000831-cvsbootfiles
350    Description:
351    
352    New URL for boot files (because the 110.29 files on the BL server do
353    now work correctly with my updated install scripts for yacc and lex).
354    
355    ----------------------------------------------------------------------
356    Name: Matthias Blume
357    Date: 2000/08/08 12:33:00 JST
358    Tag: blume-20000808-manual
359    Description:
360    
361    Tiny update to CM manual.
362    
363    ----------------------------------------------------------------------
364    Name: Allen Leung
365    Date: 2000/08/7 19:31:00
366    Tag: leunga-20000807-a-whole-bunch-of-stuff
367    
368      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
369    and Mobdy.
370    
371    1.  x86
372    
373       a.  Fixes to peephole module by John and Dan.
374       b.  Assembly fix to SETcc by Allen.
375       c.  Fix to c-call by John.
376       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
377       e.  Instruction selection optimization to SETcc as suggested by John.
378    
379           For example,
380    
381            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
382    
383           should generate:
384    
385            MOVL a, x
386            SUBL b, x
387            SHRL 31, x
388    
389    2.  IR stuff
390    
391         A bunch of new DJ-graph related algorithms added.  These
392         speed up SSA construction.
393    
394    3.  SSA + Scheduling
395    
396         Added code for SSA and scheduling to the repository
397    
398    ----------------------------------------------------------------------
399    Name: Lal George
400    Date: 2000/07/27 11:53:14 EDT
401    
402    Tag: lal-20000727-linux-ppc
403    Description:
404    
405     Made changes to support Linux PPC.
406     p.s. I have confirmation that the 110.29 boot files work fine.
407    
408    ----------------------------------------------------------------------
409    Name: Matthias Blume
410    Date: 2000/07/27 17:40:00 JST
411    Tag: blume-20000727-scripts
412    Description:
413    
414    !!!! WARNING !!!!
415    You must recompile the runtime system!
416    !!!! WARNING !!!!
417    
418    This is basically another round of script-enhancements:
419    
420    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
421       and undefine CM preprocessor symbols.
422    
423    2. ml-build avoids generating a new heap image if it finds that the
424       existing one is still ok.  (The condition is that no ML file had to
425       be recompiled and all ML files are found to be older that the heap
426       file.)
427    
428       To make this work smoothly, I also hacked the runtime system as
429       well as SMLofNJ.SysInfo to get access to the heap image suffix
430       (.sparc-solaris, ...) that is currently being used.
431    
432       Moreover, the signature of CM.mk_standalone has changed.  See the
433       CM manual.
434    
435    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
436       CM manual for details.)
437    
438    4. More CM manual updates:
439        - all of the above has been documented.
440        - there is now a section describing the (CM-related) command line
441          arguments that are accepted by the "sml" command
442    
443    ----------------------------------------------------------------------
444    Name: Matthias Blume
445    Date: 2000/07/25 16:20:00 JST
446    Tag: blume-20000725-makedepend
447    Description:
448    
449    Added a script called ml-makedepend.  This can be used in makefiles
450    for Unix' make in a way very similar to the "makedepend" command for
451    C.
452    
453    The script internally uses function CM.sources.
454    
455    Synopsis:
456    
457        ml-makedepend [-f makefile] cmfile targetname
458    
459    The default for the makefile is "makefile" (or "Makefile" should
460    "makefile" not exist).
461    
462    ml-makedepend adds a cmfile/targetname-specific section to this
463    makefile (after removing the previous version of this section).  The
464    section contains a single dependency specification with targetname on
465    the LHS (targetname is an arbitrary name), and a list of files derived
466    from the cmfile on the RHS.  Some of the files on the RHS are
467    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
468    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
469    path names.  The makefile writer is responsible for making sure that
470    these variables have correct at the time "make" is invoked.
471    
472    ----------------------------------------------------------------------
473    Name: Matthias Blume
474    Date: 2000/07/22 23:30:00 JST
475    Tag: blume-20000722-urlupdate
476    Description:
477    
478    Changed BOOT and config/srcarchiveurl to point to BL server:
479    
480        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
481    
482    ----------------------------------------------------------------------
483    Name: Matthias Blume
484    Date: 2000/07/18 18:00:00 JST
485    Tag: blume-20000718-Version_110_29
486    Description:
487    
488    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
489    
490    2. Updated config/version to 110.29
491    
492    3. Updated config/srcarchiveurl
493    
494    3. New boot files!
495       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
496    
497    ----------------------------------------------------------------------
498    Name: Matthias Blume
499    Date: 2000/07/11 13:58:00 JST
500    Tag: blume-20000711-doctypo
501    Description:
502    
503    Fixed a few typos in CM manual.
504    
505    ----------------------------------------------------------------------
506    Name: Allen Leung
507    Date: 2000/06/15 00:38:00
508    Tag: leunga-20000704-sparc-x86
509    
510    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
511    2. fix to x86 RET bug [found by Dan Grossman]
512    3. sparc assembly bug fix for ticc instructions [found by Fermin]
513    
514       Affects c-- and moby only
515    
516    ----------------------------------------------------------------------
517    Name: Matthias Blume
518    Date: 2000/07/04 15:26:00
519    Tag: blume-20000704-trigger
520    Description:
521    
522    1. Improvements to CM manual.
523    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
524       of getting a back-trace.  The function, when called, raises an
525       internal exception which explicitly carries the full back-trace history,
526       so it is unaffected by any intervening handle-raise pairs ("trivial"
527       or not).  The interactive loop will print that history once it arrives
528       at top level.
529       Short of having all exceptions implicitly carry the full history, the
530       recommended way of using this facility is:
531         - compile your program with instrumentation "on"
532         - run it, when it raises an exception, look at the history
533         - if the history is "cut off" because of some handler, go and modify
534           your program so that it explicitly calls BTrace.trigger
535         - recompile (still instrumented), and rerun; look at the full history
536    
537    ----------------------------------------------------------------------
538    Name: Matthias Blume
539    Date: 2000/07/03 15:36:00 JST
540    Tag: blume-20000702-manual
541    Description:
542    
543    Small corrections and updates to CM manual.
544    
545    ----------------------------------------------------------------------
546    Name: Matthias Blume
547    Date: 2000/06/29 16:04:00 JST
548    Tag: blume-20000629-yacctool
549    Description:
550    
551    Changes:
552    
553    1. Class "mlyacc" now takes separate arguments to pass options to
554       generated .sml- and .sig-files independently.
555    2. Corresponding CM manual updates.
556    3. BTrace module now also reports call sites.  (However, for loop clusters
557       it only shows from where the cluster was entered.)  There are associated
558       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
559    
560    ----------------------------------------------------------------------
561    Name: Matthias Blume
562    Date: 2000/06/27 16:51:00 JST
563    Tag: blume-20000627-noweb
564    Description:
565    
566    Changes:
567    
568     1. Implemented "subdir" and "witness" options for noweb tool.
569        This caused some slight internal changes in CM's tool implementation.
570     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
571        some remaining issues from earlier path anchor changes.
572     3. Updated CM manual accordingly.
573    
574     4. Changed implementation of back-tracing so that I now consider it
575        ready for prime-time.
576    
577        In particular, you don't have to explicitly trigger the back-trace
578        anymore.  Instead, if you are running BTrace-instrumented code and
579        there is an uncaught exception (regardless of whether or not it was
580        raised in instrumented code), the top-level evalloop will print
581        the back-trace.
582    
583        Features:
584    
585          - Instrumented and uninstrumented code work together seemlessly.
586            (Of course, uninstrumented code is never mentioned in actual
587             back-traces.)
588    
589          - Asymptotic time- and space-complexity of instrumented code is
590            equal to that of uninstrumented code.  (This means that
591            tail-recursion is preserved by the instrumentation phase.)
592    
593          - Modules whose code has been instrumented in different sessions
594            work together without problem.
595    
596          - There is no penalty whatsoever on uninstrumented code.
597    
598          - There is no penalty on "raise" expressions, even in
599            instrumented code.
600    
601        A potential bug (or perhaps it is a feature, too):
602    
603          A back-trace reaches no further than the outermost instrumented
604          non-trivial "raise".  Here, a "trivial" raise is one that is the
605          sole RHS of a "handle" rule.  Thus, back-traces reach trough
606    
607               <exp> handle e => raise e
608    
609          and even
610    
611               <exp> handle Foo => raise Bar
612    
613          and, of course, through
614    
615               <exp> handle Foo => ...
616    
617         if the exception was not Foo.
618    
619         Back-traces always reach right through any un-instrumented code
620         including any of its "handle" expressions, trivial or not.
621    
622       To try this out, do the following:
623    
624         - Erase all existing binfiles for your program.
625           (You may keep binfiles for those modules where you think you
626            definitely don't need back-tracing.)
627         - Turn on back-trace instrumentation:
628              SMLofNJ.Internals.BTrace.mode (SOME true);
629         - Recompile your program.  (I.e., run "CM.make" or "use".)
630         - You may now turn instrumentation off again (if you want):
631              SMLofNJ.Internals.BTrace.mode (SOME false);
632         - Run your program as usual.  If it raises an exception that
633           reaches the interactive toplevel, then a back-trace will
634           automatically be printed.  After that, the toplevel loop
635           will print the exception history as usual.
636    
637    ----------------------------------------------------------------------
638    Name: Matthias Blume
639    Date: 2000/06/26 09:56:46 JST
640    Tag: blume-20000626-setup
641    Description:
642    
643    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
644          ML code before and after compiling a file (e.g., to set compiler
645          flags)
646    
647    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
648              - associated changes to btrace.sml (BTrace instrumentation pass)
649              - cleaner implementation of btimp.sml (BTrace tracing and report
650                module)
651    
652    CM manual: * new path encoding documented
653               * description of setup-parameter to "sml" added
654    
655    The biggest user-visible change to back-tracing is that it is no
656    longer necessary to compile all traced modules within the same
657    session.  (This was a real limitation.)
658    
659    ----------------------------------------------------------------------
660    Name: Matthias Blume
661    Date: 2000/06/24 12:40:00 JST
662    Tag: blume-20000624-startup
663    Description:
664    
665    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
666    bit too often -- to put it mildly. :)
667    
668    ----------------------------------------------------------------------
669    Name: Matthias Blume
670    Date: 2000/06/23 18:20:00 JST
671    Tag: blume-20000623-btrace
672    Description:
673    
674    This updates adds a backtrace facility to aid programmers in debugging
675    their programs.  This involves the following changes:
676    
677    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
678       keeping track of the current call stack.  When programs are compiled
679       in a special mode, the compiler will insert calls to these hooks
680       into the user program.
681       "Hook" means that it is possible for different implementations of
682       back-tracing to register themselves (at different times).
683    
684    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
685       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
686       It can be turned on using this call:
687         SMLofNJ.Internals.BTrace.mode (SOME true);
688       Turning it off again:
689         SMLofNJ.Internals.BTrace.mode (SOME false);
690       Querying the current status:
691         SMLofNJ.Internals.BTrace.mode NONE;
692       Annotated programs are about twice as big as normal ones, and they
693       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
694       where all hooks do nothing).  The slowdown with a plugin that is
695       actually useful (such as the one supplied by default) is even greater,
696       but in the case of the default plugin it is still only an constant
697       factor (amortized).
698    
699    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
700       with a sub-structure BTrace for controlling back-tracing.  In particular,
701       the above-mentioned function "mode" controls whether the annotation
702       phase is invoked by the compiler.  Another important function is
703       "trigger": when called it aborts the current execution and causes
704       the top-level loop to print a full back-trace.
705    
706    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
707       for back-tracing.  It keeps track of the dynamic call stack and in
708       addition to that it keeps a partial history at each "level" of that
709       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
710       d and b (at separate times, dynamically), then the report will show:
711    
712       GOTO   d
713             /c
714       GOTO  \b
715       CALL   a
716    
717       This shows that there was an initial non-tail call of a, then a
718       tail-call to b or c, looping behavior in a cluster of functions that
719       consist of b and c, and then a goto from that cluster (i.e., either from
720       b or from c) to d.
721    
722       Note that (depending on the user program) the amount of information
723       that the back-trace module has to keep track of at each level is bounded
724       by a constant.  Thus, the whole implementation has the same asymptotical
725       complexity as the original program (both in space and in time).
726    
727    5. compiler/TopLevel/interact/evalloop.sml has been modified to
728       handle the special exception SMLofNJ.Internals.BTrace.BTrace
729       which is raised by the "trigger" function mentioned above.
730    
731    Notes on usage:
732    
733    - Annotated code works well together with unannotated code:
734    Unannotated calls simply do not show up at all in the backtrace.
735    
736    - It is not a good idea to let modules that were annotated during
737    different sessions run at the same time.  This is because the compiler
738    chooses small integers to identify individual functions, and there
739    will be clashes if different modules were compiled in separate sessions.
740    (Nothing will crash, and you will even be told about the clashes, but
741    back-trace information will in general not be useful.)
742    
743    - Back-tracing can be confused by callcc and capture.
744    
745    - The only way of getting a back-trace right now is to explicitly
746    invoke the "trigger" function from your user program.  Eventually, we
747    should make every exception carry back-trace information (if
748    available).  But since this creates more overhead at "raise"-time
749    (similar to the current exnHistory overhead), I have not yet
750    implemented this.  (The implementation will be rather easy.)  With
751    exceptions carrying back-trace information, this facility will be even
752    more useful because users don't need to modify their programs...
753    
754    - While it is possible to compile the compiler with back-trace
755    annotations turned on (I did it to get some confidence in
756    correctness), you must make absolutely sure that core.sml and
757    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
758    be compiled with annotation because there is no core access yet, but
759    if you compile btimp.sml with annotation, then the system will go into
760    an infinite recursion and crash.)
761    Since CM currently does not know about BTrace, the only way to turn
762    annotations on and off for different modules of the compiler is to
763    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
764    this is awkward and clumsy.
765    
766    Sample sessions:
767    
768    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
769    - SMLofNJ.Internals.BTrace.mode (SOME true);
770    [autoloading]
771    [autoloading done]
772    val it = false : bool
773    - structure X = struct
774    -     fun main n = let
775    -         fun a (x, 0) = d x
776    -           | a (x, n) = b (x, n - 1)
777    -         and b (x, n) = c (x, n)
778    -         and c (x, n) = a (x, n)
779    -         and d x = e (x, 3)
780    -         and e (x, 0) = f x
781    -           | e (x, n) = e (x, n - 1)
782    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
783    -           | f n = n * g (n - 1)
784    -         and g n = a (n, 3)
785    -     in
786    -         f n
787    -     end
788    - end;
789    structure X : sig val main : int -> int end
790    - X.main 3;
791    *** BACK-TRACE ***
792    GOTO   stdIn:4.2-13.20: X.main[2].f
793    GOTO-( stdIn:4.2-13.20: X.main[2].e
794    GOTO   stdIn:4.2-13.20: X.main[2].d
795         / stdIn:4.2-13.20: X.main[2].a
796         | stdIn:4.2-13.20: X.main[2].b
797    GOTO-\ stdIn:4.2-13.20: X.main[2].c
798    CALL   stdIn:4.2-13.20: X.main[2].g
799    GOTO   stdIn:4.2-13.20: X.main[2].f
800    GOTO-( stdIn:4.2-13.20: X.main[2].e
801    GOTO   stdIn:4.2-13.20: X.main[2].d
802         / stdIn:4.2-13.20: X.main[2].a
803         | stdIn:4.2-13.20: X.main[2].b
804    GOTO-\ stdIn:4.2-13.20: X.main[2].c
805    CALL   stdIn:4.2-13.20: X.main[2].g
806    GOTO   stdIn:4.2-13.20: X.main[2].f
807    GOTO-( stdIn:4.2-13.20: X.main[2].e
808    GOTO   stdIn:4.2-13.20: X.main[2].d
809         / stdIn:4.2-13.20: X.main[2].a
810         | stdIn:4.2-13.20: X.main[2].b
811    GOTO-\ stdIn:4.2-13.20: X.main[2].c
812    CALL   stdIn:4.2-13.20: X.main[2].g
813    GOTO   stdIn:4.2-13.20: X.main[2].f
814    CALL   stdIn:2.15-17.4: X.main[2]
815    -
816    
817    (Note that because of a FLINt bug the above code currently does not
818    compile without BTrace turned on.)
819    
820    Here is another example, using my modified Tiger compiler:
821    
822    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
823    - SMLofNJ.Internals.BTrace.mode (SOME true);
824    [autoloading]
825    [autoloading done]
826    val it = false : bool
827    - CM.make "sources.cm";
828    [autoloading]
829    ...
830    [autoloading done]
831    [scanning sources.cm]
832    [parsing (sources.cm):parse.sml]
833    [creating directory CM/SKEL ...]
834    [parsing (sources.cm):tiger.lex.sml]
835    ...
836    [wrote CM/sparc-unix/semant.sml]
837    [compiling (sources.cm):main.sml]
838    [wrote CM/sparc-unix/main.sml]
839    [New bindings added.]
840    val it = true : bool
841    - Main.compile ("../testcases/merge.tig", "foo.out");
842    *** BACK-TRACE ***
843    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
844    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
845    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
846    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
847    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
848    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
849    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
850         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
851    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
852    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
853    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
854    CALL   main.sml:1.18-118.4: Main.compile[2]
855    -
856    
857    ----------------------------------------------------------------------
858    Name: Matthias Blumen
859    Date: 2000/06/21 18:00:00 JST
860    Tag: blume-20000621-manual
861    Description:
862    
863    CM manual update: Path environments documented.
864    
865    ----------------------------------------------------------------------
866    Name: Matthias Blume
867    Date: 2000/06/19 13:40:00
868    Tag: blume-20000619-manual
869    Description:
870    
871    CM manual and system/README update.  This only covers the fact that
872    there are no more implicit anchors.  (Path environments and the "bind"
873    option to "cm" have yet to be documented.)
874    
875    ----------------------------------------------------------------------
876    Name: Matthias Blume
877    Date: 2000/06/19 11:05:00 JST
878    Tag: blume-20000619-chdir-bugfix
879    Description:
880    
881    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
882    
883    ----------------------------------------------------------------------
884    Name: Matthias Blume
885    Date: 2000/06/18 22:00:10 JST
886    Tag: blume-20000618-implicit-anchors-really-gone
887    Description:
888    
889    I updates the previous HISTORY entry where I forgot to mention that
890    implicit anchors are no longer with us.
891    
892    The current update also gets rid of the (now useless) controller
893    CM.Control.implicit_anchors.
894    
895    ----------------------------------------------------------------------
896    Name: Matthias Blume
897    Date: 2000/06/16 17:30:00 JST
898    Tag: blume-20000616-anchorenv
899    Description:
900    
901    This patch implements the long anticipated (just kidding :) "anchor
902    environment" mechanism.  In the course of doing this, I also
903    re-implemented CM's internal "SrcPath" module from scratch.  The new
904    one should be more robust in certain boundary cases.  In any case, it
905    is a lot cleaner than its predecessor (IMHO).
906    
907    This time, although there is yet another boot file format change, I
908    kept the unpickler backward-compatible.  As a result, no new bootfiles
909    are necessary and bootstrapping is straightforward.  (You cannot read
910    new bootfiles into an old system, but the other way around is no
911    problem.)
912    
913    Visible changes:
914    
915    ** 0. Implicit path anchors (without the leading $-symbol) are no
916    longer recognized at all. This means that such path names are not
917    illegal either.  For example, the name basis.cm simply refers to a
918    local file called "basis.cm" (i.e, the name is an ordinary path
919    relative to .cm-files directory).  Or, to put it differently, only
920    names that start with $ are anchored paths.
921    
922    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
923    vanished.
924    
925    John (Reppy) had critizised this as soon as I originally proposed and
926    implemented it, but at that time I did not really deeply believe
927    him. :) Now I came full-circle because I need the $<singlearc> syntax
928    in another place where it cannot be seen as an abbreviation for
929    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
930    seems to mean (i.e., it "expands" into the corresponding anchor
931    value).
932    
933    However, when paths are used as members in CM description files, it
934    continues to be true that there must be at least another arc after the
935    anchor.  This is now enforced separately during semantic analysis
936    (i.e., from a lexical/syntactical point of view, the notation is ok.)
937    
938    ** 2. The "cm" class now accepts an option "bind".  The option's value
939    is a sub-option list of precisely two items -- one labeled "anchor"
940    and the other one labeled "value".  As you might expect, "anchor" is
941    used to specify an anchor name to be bound, and "value" specifies what
942    the anchor is being bound to.
943    
944    The value must be a directory name and can be given in either standard
945    syntax (including the possibility that it is itself an anchored path)
946    or native syntax.
947    
948    Examples:
949    
950       foo.cm (bind:(anchor:bar value:$mystuff/bar))
951       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
952    
953    and so on.
954    
955    The meaning of this is that the .cm-file will be processed with an
956    augmented anchor environment where the given anchor(s) is/are bound to
957    the given values(s).
958    
959    The rationale for having this feature is this: Suppose you are trying
960    to use two different (already stable) libraries a.cm and b.cm (that
961    you perhaps didn't write yourself).  Further, suppose each of these
962    two libraries internally uses its own auxiliary library $aux/lib.cm.
963    Normally you would now have a problem because the anchor "lib" can not
964    be bound to more than one value globally.  Therefore, the project that
965    uses both a.cm and b.cm must locally redirect the anchor to some other
966    place:
967    
968       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
969       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
970    
971    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
972    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
973    
974    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
975    you when you do so at the time of CM.stabilize).  Therefore, you can
976    also use an anchored path as the value:
977    
978      a.cm (bind:(anchor:lib value:$a-lib))
979      b.cm (bind:(anchor:lib value:$b-lib))
980    
981    Now you can globally configure (using the usual CM.Anchor.anchor or
982    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
983    itself is always locally bound, setting it globally is no longer
984    meaningful or necessary (but it does not hurt either).  In fact, "lib"
985    can still be used as a global anchor for separate purposes.  As a
986    matter of fact, one can locally define "lib" in terms of a global
987    "lib":
988    
989      a.cm (bind:(anchor:lib value:$lib/a))
990      b.cm (bind:(anchor:lib value:$lib/b))
991    
992    ** 3: The encoding of path names has changed.  This affects the way
993    path names are shown in CM's progress report and also the internal
994    protocol encoding used for parallel make.
995    
996    The encoding now uses one or more ':'-separated segments.  Each
997    segments corresponds to a file that has been specified relative to the
998    file given by its preceding segment.  The first segment is either
999    relative to the CWD, absolute, or anchored.  Each segment itself is
1000    basically a Unix pathname; all segments but the first are relative.
1001    
1002    Example:
1003    
1004       $foo/bar/baz.cm:a/b/c.sml
1005    
1006    This path denotes the file bar/a/b/c.sml relative to the directory
1007    denoted by anchor "foo".  Notice that the encoding also includes
1008    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1009    paths are resolved relative to the .cm-files directory, so baz.cm must
1010    be ignored to get the "real" pathname.
1011    
1012    To make this fact more obvious, CM puts the names of such "virtual
1013    arcs" into parentheses when they appear in progress reports. (No
1014    parentheses will appear in the internal protocol encoding.)  Thus,
1015    what you really see is:
1016    
1017      $foo/bar/(baz.cm):a/b/c.sml
1018    
1019    I find this notation to be much more informative than before.
1020    
1021    Another new feature of the encoding is that special characters
1022    including parentheses, colons, (back)slashes, and white space are
1023    written as \ddd (where ddd is the decimal encoding of the character).
1024    
1025    *** The CM manual still needs to be updated.
1026    
1027    ----------------------------------------------------------------------
1028    Name: Allen Leung
1029    Date: 2000/06/15 00:38:00
1030    Tag: leunga-20000615-x86-peephole
1031    
1032    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1033    
1034    ----------------------------------------------------------------------
1035    Name: Matthias Blume
1036    Date: 2000/06/12 11:40:00
1037    Tag: blume-20000612-parmakefix
1038    Description:
1039    
1040    More cleanup after changing the file naming scheme: This time I
1041    repaired the parallel make mechanism for CMB.make which I broke earlier.
1042    
1043    ----------------------------------------------------------------------
1044    Name: Allen Leung
1045    Date: 2000/06/09 01:25:00
1046    Tag: leunga-20000609-various
1047    
1048    None of these things should affect normal SML/NJ operations
1049    
1050    1. Peephole improvements provided by Fermin (c--)
1051    2. New annotation DEFUSE for adding extra dependence (moby)
1052    3. New X86 LOCK instructions (moby)
1053    4. New machine description language for reservation tables (scheduling)
1054    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1055       trees etc.)
1056    6. I've changed the CM files so that they can work with versions
1057       110.0.6, 110.25 and 110.28
1058    
1059    ----------------------------------------------------------------------
1060    Name: Matthias Blume
1061    Date: 2000/06/09 12:40:00
1062    Tag: blume-20000609-log
1063    Description:
1064    
1065    - Removed all(?) remaining RCS Log entries from sources.
1066    
1067    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1068      anchored paths).
1069    
1070    ----------------------------------------------------------------------
1071    Name: Matthias Blume
1072    Date: 2000/06/07 17:00:00 JST
1073    Tag: blume-20000607-no-implicit-anchors
1074    Description:
1075    
1076    1. This update changes the default setting for
1077    CM.Control.implicit_anchors from true to false.  This means that
1078    implicit anchors are no longer permitted by default.  I also tried to
1079    make sure that nothing else still relies on implicit anchors.
1080    (This is the next step on the schedule towards a CM that does not even
1081    have the notion of implicit anchors anymore.)
1082    
1083    2. More CM manual updates.
1084    
1085    3. I managed to track down and fix the pickling bug I mentioned last
1086    time.  Because of the previously existing workaround, this entails no
1087    immediate practical changes.
1088    
1089    ----------------------------------------------------------------------
1090    Name: Matthias Blume
1091    Date: 2000/06/06 11:15:00 JST
1092    Tag: blume-20000606-lazierpickle
1093    Description:
1094    
1095    !!!! NEW BOOT FILES !!!!
1096    
1097    * The main purpose of this update is to make library pickles lazier in
1098    order to reduce the initial space penalty for autoloading a library.
1099    As a result, it is now possible to have $smlnj/compiler.cm
1100    pre-registered.  This should take care of the many complaints or
1101    inquiries about missing structure Compiler.  This required changes to
1102    CM's internal data structures and small tweaks to some algorithms.
1103    
1104    As a neat additional effect, it is no longer necessary (for the sake
1105    of lean heap image files) to distinguish between a "minimal" CM and a
1106    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1107    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1108    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1109    
1110    To make the life of compiler-hackers easier, "makeml" now also
1111    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1112    words, after you bootstrap a new sml for the first time, you will not
1113    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1114    around you will still have to do it, though.)
1115    
1116    * A second change consists of major updates to the CM manual.  There
1117    are now several appendices with summary information and also a full
1118    specification of the CM description file syntax.
1119    
1120    * In directory src/system I added the script "allcross".  This script
1121    invokes sml and cross-compiles the compiler for all supported
1122    architectures.  (Useful when providing a new set of boot files.)
1123    
1124    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1125    added a small tweak to pickle-util.sml to work around this problem,
1126    but it is not a proper fix yet.  I will investigate further.  (The
1127    effect of the bug was an inflation of library pickle size.)
1128    
1129    * Version number increased to 110.28.1 (to avoid compatibility problems).
1130    
1131    ----------------------------------------------------------------------
1132    Name: Allen Leung
1133    Date: 2000/05/25 17:28 EDT
1134    Tag: leunga-20000525-ra
1135    Description:
1136    
1137      Fixed a bug in freezing phase of the register allocator.
1138    
1139    ----------------------------------------------------------------------
1140    Name: Allen Leung
1141    Date: 2000/05/15 22:53 EDT
1142    Tag: leunga-20000515-alpha-x86-ra
1143    Description:
1144    
1145      1. Alpha
1146    
1147          Slight cleanup.  Removed the instruction SGNXL
1148    
1149      2. X86
1150    
1151          Added the following instructions to the instruction set:
1152    
1153            ROLx, RORx,
1154            BTx, BTSx, BTLx, BTRx,
1155            XCHGx, and variants with the LOCK prefix
1156    
1157      3. Register Allocation
1158    
1159          The module ra-rewrite-with-renaming has been improved.
1160    
1161      These have no effect on SML/NJ.
1162    
1163    ----------------------------------------------------------------------
1164    Name: Matthias Blume
1165    Date: 2000/05/15 16:20:00 JST
1166    Tag: blume-20000515-lightrebuild
1167    Description:
1168    
1169    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1170       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1171       defined.  In effect, the command will not build any cross-compiler
1172       backends and therefore finish more quickly.
1173    
1174       The "fixpt" script also takes a "-light" switch to be able to use
1175       this new facility while compiling for a fixpoint.
1176    
1177    2. I replaced all mentions of anchored paths in group owner specifications
1178       with simple relative paths (usually starting with "..").
1179       The rationale is that a library's internal workings should not be
1180       compromised by the lack of some anchor.  (An anchor is necessary
1181       for someone who wants to refer to the library by an anchored path,
1182       but it should not be necessary to build the same library in the first
1183       place.)
1184    
1185    3. I changed the way CM's tool mechanism determines the shell command
1186       string used for things like ml-yacc etc. so that it does not break
1187       when CM.Control.implicit_anchors is turned off.
1188    
1189    ----------------------------------------------------------------------
1190    Name: Matthias Blume
1191    Date: 2000/05/12 18:20:00 JST
1192    Tag: blume-20000512-ml-build
1193    Description:
1194    
1195    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1196    getting installed properly (by config/install.sh).
1197    
1198    ----------------------------------------------------------------------
1199    Name: Matthias Blume
1200    Date: 2000/05/12 17:30:00 JST
1201    Tag: blume-20000512-anchors
1202    Description:
1203    
1204    !!! NEW BOOT FILES !!!
1205    
1206    This change is in preparation of fading out support for "implicitly
1207    anchored path names".  I went through all sources and used the
1208    explicit (and relatively new) $-notation.  See system/README and the
1209    CM manual for more info on this.
1210    
1211    I also modified the anchoring scheme for some things such as "smlnj",
1212    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1213    anchors are more expressive: anchor name and first arc do not have to
1214    coincide.  This entails the following user-visible change:
1215    
1216    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1217    particular, when you fire up sml with a command-line argument, say,
1218    e.g.:
1219    
1220       sml '$smlnj/cmb.cm'
1221    
1222    At the ML toplevel prompt:
1223    
1224       CM.autoload "$smlnj/cmb.cm";
1225    
1226    There is also a new controller in CM.Control that can be used to turn
1227    off all remaining support for implicit anchors by saying:
1228    
1229        CM.autoload "$smlnj/
1230        #set CM.Control.implicit_anchors false;
1231    
1232    This causes CM to reject implicitly anchored paths.  This is (for the
1233    time being) less permissive than the "final" version where there will
1234    be no more such implicit anchors and relative paths will be just that:
1235    relative.
1236    
1237    The next step (version after next version?) will be to make the
1238    default for CM.Control.implicit_anchors false.  After the dust has
1239    settled, I can then produce the "final" version of this...
1240    
1241    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1242    
1243    ----------------------------------------------------------------------
1244    Name: Matthias Blume
1245    Date: 2000/05/11 16:30:00 JST
1246    Tag: blume-20000511-sources
1247    Description:
1248    
1249    The main change is that I added function CM.sources as a generalized
1250    version of the earlier CM.makedepend.  This entails the following
1251    additional changes:
1252    
1253      - CM.makedepend has been dropped.
1254    
1255      - CM manual has been updated.
1256    
1257      - TOOLS signature and API have been changed.
1258    
1259    ----------------------------------------------------------------------
1260    Name: Allen Leung
1261    Date: 2000/05/10 21:17 EDT
1262    Tag: leunga-20000510-moby-c--ssa
1263    Description:
1264    
1265      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1266    None of these affect SML/NJ.
1267    
1268    1. Register Allocation
1269    
1270        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1271           This module tries to remove local (i.e. basic block level) redundancies
1272           during spilling.
1273    
1274        b. A new framework for performing region based register allocation.
1275           Not yet entirely functional.
1276    
1277    2. X86
1278    
1279       a. DefUse for POP was missing the stack pointer [found by Lal]
1280       b. Reload for CALL was incorrect in X86Spill [found by John]
1281       c. Various fixes in X86Spill so that it can be used correctly for
1282          the new spilling module.
1283    
1284    3. SSA/IR
1285    
1286       a. New module ir/dj-dataflow.sml implements elimination based
1287          data flow analysis.
1288    
1289    4. MLRiscGen
1290    
1291       a. Fix for gc type annotation
1292    
1293    5. MDGen
1294    
1295       Various fixes for machine description -> ml code translation.  For ssa
1296       only.
1297    
1298    ----------------------------------------------------------------------
1299    Name: Allen Leung
1300    Date: 2000/05/08 22:17 EDT
1301    Tag: leunga-20000508-labexp
1302    Description:
1303    
1304      Fermin has found a few assembly problems with constant expressions
1305      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1306      which choke on dumb assemblers.  This is his fix.
1307    
1308    ----------------------------------------------------------------------
1309    Name: Dave MacQueen
1310    Date: 2000/04/09 14:00 EDT
1311    Tag: dbm-20000502-Version_110_28
1312    Description:
1313    
1314    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1315    
1316    2. Updated config/version to 110.28
1317    
1318    3. Updated config/srcarchiveurl
1319    
1320    3. New boot files!
1321       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1322    
1323    ----------------------------------------------------------------------
1324    Name: Matthias Blume
1325    Date: 2000/05/01 19:05:00 JST
1326    Tag: blume-20000501-noweb
1327    Description:
1328    
1329    A new noweb tool has been added.  The existing system is entirely
1330    unaffected by this, but some CM users have asked for renewed noweb
1331    support.  Everything is documented in the CM manual.
1332    
1333    New (plugin) libraries:
1334    
1335       noweb-tool.cm
1336       nw-ext.cm
1337    
1338    ----------------------------------------------------------------------
1339    Name: Dave MacQueen
1340    Date: 2000/04/30 12:40PM EDT
1341    Tag: dbm-20000430-bug_fixes
1342    Description:
1343    
1344    1. Fix for bug 1498
1345       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1346       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1347         added toRealArray function
1348       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1349         added check for tag Obj.RealArray to array printing case in ppObj
1350    
1351    2. Fix for bug 1510
1352       smlnj/src/compiler/Semant/types/typesutil.sml
1353         fixed definition of dummyargs (used by equalTycon) so that
1354         dummy args are distinct types
1355    
1356    ----------------------------------------------------------------------
1357    Name: Matthias Blume
1358    Date: 2000/04/30 01:00:00 JST
1359    Tag: blume-20000430-versions
1360    Description:
1361    
1362    1. CM version numbering added.  This is an implementation of Lal's
1363       proposal for adding version numbers and version checking to .cm
1364       files.  Lal said that his proposal was just that -- a proposal.
1365       For the time being I went ahead and implemented it so that people
1366       can comment on it.  Everything is completely backward-compatible
1367       (except for the stable library format, i.e., new bootfiles!).
1368    
1369       As usual, see the CM manual for details.
1370    
1371    2. An alternative syntax for anchored paths has been implemented.
1372       Dave has recently voiced the same concerns that I had when I did
1373       this, so there should be some support.  My take is that eventually
1374       I will let support for the current syntax (where anchors are
1375       "implicit") fade out in favor of the new, explicit syntax.
1376       In order to be backward-compatible, both old and new syntax are
1377       currently supported.
1378    
1379       Again, see the CM manual for details.
1380    
1381    3. Parallel make is trying to be slightly smarter:  When the master
1382       process finds a "bottleneck", i.e., when there is only one
1383       compilation unit that can be compiled and everybody else is
1384       waiting on it, then it will simply compile it directly instead
1385       of clumsily telling one of the slaves to do it.
1386    
1387    4. Support for "unsharing" added.  This is necessary in order to be
1388       able to have two different versions of the same library running
1389       at the same time (e.g., for trying out a new MLRISC while still
1390       having the old MLRISC linked into the current compiler, etc.)
1391       See the CM manual.
1392    
1393    5. Simple "makedepend" functionality added for generating Makefile
1394       dependency information.  (This is rather crude at the moment.
1395       Expect some changes here in the future.)
1396    
1397    6. ".fun" added as a recognized suffix for ML files. Also documented
1398       explicitly in the manual that the fallback behavior (unknown suffix
1399       -> ML file) is not an official feature!
1400    
1401    7. Small changes to the pickler for stable libraries.
1402    
1403    8. Several internal changes to CM (for cleanup/improvement).
1404    
1405    
1406    !!!! NEW BINFILES !!!!
1407    
1408    ----------------------------------------------------------------------
1409    Name: Matthias Blume
1410    Date: 2000/04/28 17:30:00 JST
1411    Tag: blume-20000428-pathconfig
1412    Description:
1413    
1414    1. I changed config/install.sh to remove duplicate entries from the
1415       lib/pathconfig file at the end.  Moreover, the final version of
1416       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1417       in src/system/installml.
1418    
1419    2. The config/install.sh script now consistently uses relative
1420       pathnames in lib/pathconfig whenever the anchor is in the lib
1421       directory.  (So far this was true for the libraries that come
1422       pre-compiled and bundled as part of the bootfiles but not for
1423       libraries that are compiled by the script itself.)
1424    
1425    ----------------------------------------------------------------------
1426    Name: Matthias Blume
1427    Date: 2000/04/26 13:10:00 JST
1428    Tag: blume-20000426-fun_suffix
1429    Description:
1430    
1431    Added ".fun" as a recognized file name suffix (for ML code).
1432    
1433    ----------------------------------------------------------------------
1434    Name: Allen Leung
1435    Date: 2000/04/25 17:00:00 EST
1436    Tag: leunga-20000425-alpha-ra
1437    Description:
1438    
1439    1. Alpha
1440    
1441        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1442    in C--.
1443    
1444    2. RA
1445    
1446       Added a flag "ra-dump-size" to print out the size of the flowgraph
1447       and the interference graph.
1448    
1449    ----------------------------------------------------------------------
1450    Name: Dave MacQueen
1451    Date: 2000/04/25/
1452    Tag: dbm-20000425-mlyacc_doc_examples
1453    Description:
1454      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1455      Updated all three examples in src/ml-yacc/examples to run
1456      under 110.* using CM.make.
1457    
1458    ----------------------------------------------------------------------
1459    Name: Allen Leung
1460    Date: 2000/04/20 23:04:00 EST
1461    Tag: leunga-20000420-ssa-c---stuff
1462    Description:
1463    
1464      This update synchronizes my repository with Yale's.  Most of these
1465    changes, however, do not affect SML/NJ at all (the RA is an exception).
1466    
1467    1. Register Allocator
1468    
1469       a. An improvement in the interference graph construction:
1470          Given a copy
1471    
1472                s <- t
1473    
1474          no interference edge between s and t is added for this definition of s.
1475    
1476       b. I've added two new spill heuristic modules that Fermin and I developed
1477          (in the new library RA.cm). These are unused in SML/NJ but maybe
1478          useful for others (Moby?)
1479    
1480    2. X86
1481    
1482       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1483    
1484    3. Alpha
1485    
1486       a. Added the BSR instruction and code generation that goes with it [C--]
1487       b. Other fixes too numerous to recount provided by Fermin [C--]
1488    
1489    4. Regmaps
1490    
1491       a. The regmaps are not initialized with the identity physical bindings
1492          at creation time.  This is unneeded.
1493    
1494    5. MLRISC Optimizations
1495    
1496       a. The DJ-Graph module can now compute the iterated dominance frontiers
1497          intersects with liveness incrementally in linear time! Woohoo!
1498          This is now used in my new SSA construction algorithm.
1499    
1500       b. THe branch reorganization module is now smarter about linear chains of
1501          basic blocks.
1502    
1503    
1504    ----------------------------------------------------------------------
1505    Name: Matthias Blume
1506    Date: 2000/04/12 13:52:00 JST
1507    Tag: blume_main_v110p27_1
1508    Description:
1509    
1510    Changed install.sh script to handle archive files without version number
1511    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1512    name of the boot file archive.
1513    
1514    ----------------------------------------------------------------------
1515    Name: Dave MacQueen
1516    Date: 2000/04/09 14:00 EDT
1517    Tag: dbm-20000410-Version_110_27
1518    Description:
1519    
1520    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1521    
1522    2. Updated src/config/version to 110.27
1523    
1524    3. New boot files!
1525    
1526    ----------------------------------------------------------------------
1527    Name: Allen Leung
1528    Date: 2000/04/09 19:09:00 EST
1529    Tag: leunga-20000409-misc
1530    Description:
1531    
1532    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1533    
1534    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1535    
1536    ----------------------------------------------------------------------
1537    Name: Stefan
1538    Date: 2000/04/07 10:00:00 EDT
1539    Tag: monnier-20000406-branch-handling
1540    Description:
1541    
1542    Improved handling of branches (mostly those generated from
1543    polymorphic equality), removed switchoff and changed the
1544    default optimization settings (more cpsopt and less flintopt).
1545    
1546    ----------------------------------------------------------------------
1547    Name: Allen Leung
1548    Date: 2000/04/06 01:30:00 EST
1549    Tag: leunga-20000406-peephole-x86-SSA-2
1550    Description:
1551    
1552       Forgot a few files.
1553    
1554    ----------------------------------------------------------------------
1555    Name: Allen Leung
1556    Date: 2000/04/06 00:36:00 EST
1557    Tag: leunga-20000406-peephole-x86-SSA
1558    Description:
1559    
1560    1.  New Peephole code
1561    
1562    2.  Minor improvement to X86 instruction selection
1563    
1564    3.  Various fixes to SSA and machine description -> code translator
1565    
1566    ----------------------------------------------------------------------
1567    Name: Matthias Blume
1568    Date: 2000/04/05 12:30:00 JST
1569    Tag: blume_main_v110p26p2_3
1570    Description:
1571    
1572    This update just merges three minor cosmetic updates to CM's sources
1573    to get ready for the 110.27 code freeze on Friday.  No functionality
1574    has changed.
1575    
1576    ----------------------------------------------------------------------
1577    Name: Allen Leung
1578    Date: 2000/04/04 19:39:00 EST
1579    Tag: leunga-20000404-x86-asm
1580    Description:
1581    
1582    1.  Fixed a problem in X86 assembly.
1583    
1584        Things like
1585    
1586           jmp %eax
1587           jmp (%eax)
1588    
1589        should be output as
1590    
1591           jmp *%eax
1592           jmp *(%eax)
1593    
1594    2.  Assembly output
1595    
1596          Added a new flag
1597    
1598              "asm-indent-copies" (default to false)
1599    
1600          When this flag is on, parallel copies will be indented an extra level.
1601    
1602    ----------------------------------------------------------------------
1603  Name: Allen Leung  Name: Allen Leung
1604  Date: 2000/04/04 03:18:00 EST  Date: 2000/04/04 03:18:00 EST
1605  Tag: leunga-20000404-C--Moby  Tag: leunga-20000404-C--Moby
# Line 31  Line 1619 
1619    
1620  3.  Assembly  3.  Assembly
1621    
1622      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
1623      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
1624      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
1625    
# Line 54  Line 1642 
1642    
1643  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
1644  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
1645  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
1646  legal surface-syntax symbol).  legal surface-syntax symbol).
1647    
1648  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 64  Line 1652 
1652  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
1653    
1654  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1655  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
1656  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
1657  more details.  more details.
1658    
# Line 179  Line 1767 
1767     (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
1768     used).     used).
1769    
1770     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1771     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1772     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1773    
1774     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 192  Line 1780 
1780    
1781     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1782    
1783     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1784     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1785    
1786  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 326  Line 1914 
1914    
1915       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1916       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1917       testl is perferable to andl.       testl is preferable to andl.
1918    
1919  7. RA (x86 only)  7. RA (x86 only)
1920    
# Line 508  Line 2096 
2096    
2097  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2098  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2099  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
2100  explanation.  explanation.
2101    
2102  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2103  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2104  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2105  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 555  Line 2143 
2143  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
2144  <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,
2145  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
2146  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2147  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2148  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2149    
# Line 605  Line 2193 
2193    it from that remote directory.    it from that remote directory.
2194    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2195    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2196    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2197    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2198    
2199    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.593  
changed lines
  Added in v.761

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