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 615, Wed Apr 12 04:52:09 2000 UTC revision 758, Fri Dec 22 04:12:36 2000 UTC
# Line 13  Line 13 
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15  Name: Matthias Blume  Name: Matthias Blume
16    Date: 2000/12/22 13:15:00 JST
17    Tag: blume-20001222-cleanup
18    Description:
19    
20    Removed src/cm/ffi which does not (and did not) belong here.
21    
22    ----------------------------------------------------------------------
23    Name: Matthias Blume
24    Date: 2000/12/21 23:55:00 JST
25    Tag: blume-20001221-exn
26    Description:
27    
28    Probably most important: CM no longer silently swallows all exceptions
29    in the compiler.
30    Plus: some other minor CM changes.  For example, CM now reports some
31    sizes for generated binfiles (code, data, envpickle, lambdapickle).
32    
33    ----------------------------------------------------------------------
34    Name: Matthias Blume
35    Date: 2000/12/15 00:01:05 JST
36    Tag: blume-20001215-dirtool
37    Description:
38    
39    - "dir" tool added.
40    - improvements and cleanup to Tools structure
41    - documentation updates
42    
43    ----------------------------------------------------------------------
44    Name: Allen Leung
45    Date: Thu Dec 14 03:45:24 EST 2000
46    Description:
47    Tag:  leunga-20001214-int-inf
48    Description:
49    
50       In IntInf, added these standard functions, which are missing from our
51    implementation:
52    
53        andb : int * int -> int
54        xorb : int * int -> int
55        orb  : int * int -> int
56        notb : int -> int
57         <<   : int * word -> int
58        ~>>  : int * word -> int
59    
60       Not tested, I hope they are correct.
61    
62    ----------------------------------------------------------------------
63    Name: Allen Leung
64    Date: Fri Dec  8 19:23:26 EST 2000
65    Description:
66    Tag:  leunga-20001208-nowhere
67    Description:
68    
69      Slight improvements to the 'nowhere' tool to handle OR-patterns,
70    to generate better error messages etc.  Plus a brief manual.
71    
72    ----------------------------------------------------------------------
73    Name: Lal George
74    Date: 2000/12/08 09:54:02 EST 2000
75    Tag: Release_110_31
76    Description:
77    
78    - Version 110.31
79    ----------------------------------------------------------------------
80    Name: Allen Leung
81    Date: Thu Dec  7 22:01:04 EST 2000
82    Tag:  leunga-20001207-cell-monster-hack
83    Description:
84    
85    Major MLRISC internal changes.  Affect all clients.
86    Summary:
87    
88    1.  Type CELLS.cell = int is now replaced by a datatype.
89        As a result, the old regmap is now gone.  Almost all interfaces
90        in MLRISC change as a consequence.
91    
92    2.  A new brand version of machine description tool (v3.0) that generates
93        modules expecting the new interface.  The old version is removed.
94    
95    3.  The RA interface has been further abstracted into two new functors.
96        RISC_RA and X86RA.  These functors have much simpler interfaces.
97        [See also directory MLRISC/demo.]
98    
99    4.  Some other new source->source code generation tools are available:
100    
101        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
102        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
103           I use this tool to generate the peephole optimizers---with the new
104           cell type changes, peephole rules are becoming difficult to write
105           without conditional pattern matching.
106    
107    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
108        cover the entire MLRISC source tree so many things broke.
109    
110    6.  CM files have been moved to the subdirectory MLRISC/cm.
111        They are moved because there are a lot of them and they clutter up the
112        root dir.
113    
114    7.  More detailed documentation to come...
115    
116        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
117        a makeml -rebuild first.  This is because of other other
118        changes that Matthias has made (see below).
119    
120    
121    ----------------------------------------------------------------------
122    Name: Matthias Blume
123    Date: 2000/11/30 23:12:00 JST
124    Tag: blume-20001130-filereorg
125    Description:
126    
127    Some manual updates and some file reorganizations in CM.
128    
129    ----------------------------------------------------------------------
130    Name: Matthias Blume
131    Date: 2000/11/24 17:45:00 JST
132    Tag: blume-20001124-link
133    Description:
134    
135    Drastically improved link traversal code for the case that the dynamic
136    value was already loaded at bootstrap time.  As a result, CM and CMB
137    now both load blazingly fast -- even on a very slow machine.  Also,
138    memory consumption has been further reduced by this.
139    
140    Warning: The format of the PIDMAP file has changed.  THerefore, to
141    bootstrap you have to do this:
142    
143    1. Run CMB.make
144    2. Make a symbolic link for the boot directory:
145         ln -s sml.boot.ARCH-OS xxx
146    3. "Rebuild" the boot directory:
147         ./makeml -boot xxx -rebuild sml ; rm xxx
148    4. Boot normally:
149          ./makeml
150    
151    ----------------------------------------------------------------------
152    Name: Matthias Blume
153    Date: 2000/11/21 21:20:00 JST
154    Tag: blume-20001121-tools
155    Description:
156    
157    Continued hacking on autoloading problem -- with success this time.
158    Also changed tool-plugin mechanism.  See new CM manual.
159    
160    ----------------------------------------------------------------------
161    Name: Matthias Blume
162    Date: 2000/11/19 14:30:00 JST
163    Tag:  blume-20001119-autoload
164    Description:
165    
166    Some hacking to make autoloading faster.  Success for CMB, no success
167    so far for CM.  There is a reduced structure CM' that autoloads faster.
168    (This is a temporary, non-documented hack to be eliminated again when
169    the general problem is solved.)
170    
171    ----------------------------------------------------------------------
172    Name: Matthias Blume
173    Date: 2000/11/17 14:10:00 JST
174    Tag: blume-20001117-pickle-lib
175    Description:
176    
177    1. Eliminated comp-lib.cm
178    2. Made pickle-lib.cm
179    3. Eliminated all uses of intset.sml (from comp-lib.cm)
180    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
181       equivalent constructs from smlnj-lib.cm (INtHashTable).
182    5. Point 4. also goes for those uses of intmap.* in MLRISC.
183       Duplicated intmap modules thrown out.
184    6. Hunted down all duplicated SCC code and replaced it with
185       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
186    7. Rewrote Feedback module.
187    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
188       should be thrown out and equivalent modules from smlnj-lib.cm
189       should be used (IntRedBlackSet, IntListSet, ...).
190    
191    Confirmed that compiler compiles to fixpoint.
192    
193    ----------------------------------------------------------------------
194    Name: Allen Leung
195    Date: 2000/11/10 18:00:00
196    Tag: leunga-20001110-new-x86-fp
197    
198    A new x86 floating point code generator has been added.
199    By default this is turned off.  To turn this on, do:
200    
201        CM.autoload "$smlnj/compiler.cm";
202        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
203    
204    Changes:
205    
206    1.  Changed FTAN to FPTAN so that the assembly output is correct.
207    2.  Changed the extension callback for FTANGENT to generate:
208    
209              fptan
210              fstp  %st(0)
211        instead of
212              fptan
213              fstpl ftempmem
214    
215    3.  Numerous assembly fixes for x86.
216    
217    5.  Cleaned up the machine code output module x86/x86MC.sml and added
218        support for a whole bunch of instructions and addressing modes:
219    
220          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
221          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
222          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
223          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
224          fxch %st(n)
225          fld %st(n)
226          fst %st(n)
227          fst mem
228          fstp %st(n)
229          fucom %st(n)
230          fucomp %st(n)
231    
232        All these are now generated when the fast fp mode is turned on.
233    
234    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
235    
236    ----------------------------------------------------------------------
237    Name: Matthias Blume
238    Date: 2000/11/09 11:20:00 JST
239    Tag: blume-20001109-scc
240    Description:
241    
242    Eliminated some code duplication:
243    
244    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
245       (Otherwise the functor is useless.)
246    2. Used GraphSCCFn where SCCUtilFun was used previously.
247    3. Got rid of SCCUtilFun (in comp-lib.cm).
248    
249    ----------------------------------------------------------------------
250    Name: Lal George
251    Date: 2000/11/06 09:02:21 EST 2000
252    Tag: Release_110_30
253    Description:
254    
255    - Version 110.30
256    ----------------------------------------------------------------------
257    Name: Matthias Blume
258    Date: 2000/11/04 14:45:00
259    Tag: blume-20001104-mlbuild
260    Description:
261    
262    - Made ml-build faster on startup.
263    - Documentation fixes.
264    
265    ----------------------------------------------------------------------
266    Name: Matthias Blume
267    Date: 2000/11/02 17:00:00 JST
268    Tag: blume-20001102-condcomp
269    Description:
270    
271    - Small tweaks to pickler -- new BOOTFILES!
272    - Version bumped to 110.29.2.
273    - Added conditional compilation facility to init.cmi (see comment there).
274    ----------------------------------------------------------------------
275    Name: Allen Leung
276    Date: 2000/10/23 19:31:00
277    Tag: leunga-20001023-demo-ra
278    
279    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
280    2. Test programs for the graph library updated
281    3. Some new MLRISC demo programs added
282    
283    ----------------------------------------------------------------------
284    Name: Matthias Blume
285    Date: 2000/08/31 22:15:00 JST
286    Tag: blume-20001017-errmsg
287    Description:
288    
289    More error message grief: Where there used to be no messages, there
290    now were some that had bogus error regions.  Fixed.
291    
292    ----------------------------------------------------------------------
293    Name: Matthias Blume
294    Date: 2000/08/31 17:30:00 JST
295    Tag: blume-20001017-v110p29p1
296    Description:
297    
298    I made a version 110.29.1 with new bootfiles.
299    
300    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
301              CM documentation changes and a small bugfix in CM's error reporting.
302    
303    ----------------------------------------------------------------------
304    Name: Lal George
305    Date: 2000/09/27 14:42:35 EDT
306    Tag: george-20000927-nodestatus
307    Description:
308    
309    Changed the type of the nodestatus, so that:
310    
311            SPILLED(~1)             is now SPILLED
312            SPILLED(m) where m>=0   is now MEMREG(m)
313            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
314    
315    ----------------------------------------------------------------------
316    Name: Matthias Blume
317    Date: 2000/09/07 14:45:00 JST
318    Tag: blume-20000907-cmerrmsg
319    Description:
320    
321    Small tweak to CM to avoid getting ML syntax error messages twice.
322    
323    ----------------------------------------------------------------------
324    Name: Matthias Blume
325    Date: 2000/08/31 18:00:00 JST
326    Tag: blume-20000831-cvsbootfiles
327    Description:
328    
329    New URL for boot files (because the 110.29 files on the BL server do
330    now work correctly with my updated install scripts for yacc and lex).
331    
332    ----------------------------------------------------------------------
333    Name: Matthias Blume
334    Date: 2000/08/08 12:33:00 JST
335    Tag: blume-20000808-manual
336    Description:
337    
338    Tiny update to CM manual.
339    
340    ----------------------------------------------------------------------
341    Name: Allen Leung
342    Date: 2000/08/7 19:31:00
343    Tag: leunga-20000807-a-whole-bunch-of-stuff
344    
345      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
346    and Mobdy.
347    
348    1.  x86
349    
350       a.  Fixes to peephole module by John and Dan.
351       b.  Assembly fix to SETcc by Allen.
352       c.  Fix to c-call by John.
353       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
354       e.  Instruction selection optimization to SETcc as suggested by John.
355    
356           For example,
357    
358            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
359    
360           should generate:
361    
362            MOVL a, x
363            SUBL b, x
364            SHRL 31, x
365    
366    2.  IR stuff
367    
368         A bunch of new DJ-graph related algorithms added.  These
369         speed up SSA construction.
370    
371    3.  SSA + Scheduling
372    
373         Added code for SSA and scheduling to the repository
374    
375    ----------------------------------------------------------------------
376    Name: Lal George
377    Date: 2000/07/27 11:53:14 EDT
378    
379    Tag: lal-20000727-linux-ppc
380    Description:
381    
382     Made changes to support Linux PPC.
383     p.s. I have confirmation that the 110.29 boot files work fine.
384    
385    ----------------------------------------------------------------------
386    Name: Matthias Blume
387    Date: 2000/07/27 17:40:00 JST
388    Tag: blume-20000727-scripts
389    Description:
390    
391    !!!! WARNING !!!!
392    You must recompile the runtime system!
393    !!!! WARNING !!!!
394    
395    This is basically another round of script-enhancements:
396    
397    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
398       and undefine CM preprocessor symbols.
399    
400    2. ml-build avoids generating a new heap image if it finds that the
401       existing one is still ok.  (The condition is that no ML file had to
402       be recompiled and all ML files are found to be older that the heap
403       file.)
404    
405       To make this work smoothly, I also hacked the runtime system as
406       well as SMLofNJ.SysInfo to get access to the heap image suffix
407       (.sparc-solaris, ...) that is currently being used.
408    
409       Moreover, the signature of CM.mk_standalone has changed.  See the
410       CM manual.
411    
412    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
413       CM manual for details.)
414    
415    4. More CM manual updates:
416        - all of the above has been documented.
417        - there is now a section describing the (CM-related) command line
418          arguments that are accepted by the "sml" command
419    
420    ----------------------------------------------------------------------
421    Name: Matthias Blume
422    Date: 2000/07/25 16:20:00 JST
423    Tag: blume-20000725-makedepend
424    Description:
425    
426    Added a script called ml-makedepend.  This can be used in makefiles
427    for Unix' make in a way very similar to the "makedepend" command for
428    C.
429    
430    The script internally uses function CM.sources.
431    
432    Synopsis:
433    
434        ml-makedepend [-f makefile] cmfile targetname
435    
436    The default for the makefile is "makefile" (or "Makefile" should
437    "makefile" not exist).
438    
439    ml-makedepend adds a cmfile/targetname-specific section to this
440    makefile (after removing the previous version of this section).  The
441    section contains a single dependency specification with targetname on
442    the LHS (targetname is an arbitrary name), and a list of files derived
443    from the cmfile on the RHS.  Some of the files on the RHS are
444    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
445    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
446    path names.  The makefile writer is responsible for making sure that
447    these variables have correct at the time "make" is invoked.
448    
449    ----------------------------------------------------------------------
450    Name: Matthias Blume
451    Date: 2000/07/22 23:30:00 JST
452    Tag: blume-20000722-urlupdate
453    Description:
454    
455    Changed BOOT and config/srcarchiveurl to point to BL server:
456    
457        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
458    
459    ----------------------------------------------------------------------
460    Name: Matthias Blume
461    Date: 2000/07/18 18:00:00 JST
462    Tag: blume-20000718-Version_110_29
463    Description:
464    
465    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
466    
467    2. Updated config/version to 110.29
468    
469    3. Updated config/srcarchiveurl
470    
471    3. New boot files!
472       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
473    
474    ----------------------------------------------------------------------
475    Name: Matthias Blume
476    Date: 2000/07/11 13:58:00 JST
477    Tag: blume-20000711-doctypo
478    Description:
479    
480    Fixed a few typos in CM manual.
481    
482    ----------------------------------------------------------------------
483    Name: Allen Leung
484    Date: 2000/06/15 00:38:00
485    Tag: leunga-20000704-sparc-x86
486    
487    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
488    2. fix to x86 RET bug [found by Dan Grossman]
489    3. sparc assembly bug fix for ticc instructions [found by Fermin]
490    
491       Affects c-- and moby only
492    
493    ----------------------------------------------------------------------
494    Name: Matthias Blume
495    Date: 2000/07/04 15:26:00
496    Tag: blume-20000704-trigger
497    Description:
498    
499    1. Improvements to CM manual.
500    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
501       of getting a back-trace.  The function, when called, raises an
502       internal exception which explicitly carries the full back-trace history,
503       so it is unaffected by any intervening handle-raise pairs ("trivial"
504       or not).  The interactive loop will print that history once it arrives
505       at top level.
506       Short of having all exceptions implicitly carry the full history, the
507       recommended way of using this facility is:
508         - compile your program with instrumentation "on"
509         - run it, when it raises an exception, look at the history
510         - if the history is "cut off" because of some handler, go and modify
511           your program so that it explicitly calls BTrace.trigger
512         - recompile (still instrumented), and rerun; look at the full history
513    
514    ----------------------------------------------------------------------
515    Name: Matthias Blume
516    Date: 2000/07/03 15:36:00 JST
517    Tag: blume-20000702-manual
518    Description:
519    
520    Small corrections and updates to CM manual.
521    
522    ----------------------------------------------------------------------
523    Name: Matthias Blume
524    Date: 2000/06/29 16:04:00 JST
525    Tag: blume-20000629-yacctool
526    Description:
527    
528    Changes:
529    
530    1. Class "mlyacc" now takes separate arguments to pass options to
531       generated .sml- and .sig-files independently.
532    2. Corresponding CM manual updates.
533    3. BTrace module now also reports call sites.  (However, for loop clusters
534       it only shows from where the cluster was entered.)  There are associated
535       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
536    
537    ----------------------------------------------------------------------
538    Name: Matthias Blume
539    Date: 2000/06/27 16:51:00 JST
540    Tag: blume-20000627-noweb
541    Description:
542    
543    Changes:
544    
545     1. Implemented "subdir" and "witness" options for noweb tool.
546        This caused some slight internal changes in CM's tool implementation.
547     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
548        some remaining issues from earlier path anchor changes.
549     3. Updated CM manual accordingly.
550    
551     4. Changed implementation of back-tracing so that I now consider it
552        ready for prime-time.
553    
554        In particular, you don't have to explicitly trigger the back-trace
555        anymore.  Instead, if you are running BTrace-instrumented code and
556        there is an uncaught exception (regardless of whether or not it was
557        raised in instrumented code), the top-level evalloop will print
558        the back-trace.
559    
560        Features:
561    
562          - Instrumented and uninstrumented code work together seemlessly.
563            (Of course, uninstrumented code is never mentioned in actual
564             back-traces.)
565    
566          - Asymptotic time- and space-complexity of instrumented code is
567            equal to that of uninstrumented code.  (This means that
568            tail-recursion is preserved by the instrumentation phase.)
569    
570          - Modules whose code has been instrumented in different sessions
571            work together without problem.
572    
573          - There is no penalty whatsoever on uninstrumented code.
574    
575          - There is no penalty on "raise" expressions, even in
576            instrumented code.
577    
578        A potential bug (or perhaps it is a feature, too):
579    
580          A back-trace reaches no further than the outermost instrumented
581          non-trivial "raise".  Here, a "trivial" raise is one that is the
582          sole RHS of a "handle" rule.  Thus, back-traces reach trough
583    
584               <exp> handle e => raise e
585    
586          and even
587    
588               <exp> handle Foo => raise Bar
589    
590          and, of course, through
591    
592               <exp> handle Foo => ...
593    
594         if the exception was not Foo.
595    
596         Back-traces always reach right through any un-instrumented code
597         including any of its "handle" expressions, trivial or not.
598    
599       To try this out, do the following:
600    
601         - Erase all existing binfiles for your program.
602           (You may keep binfiles for those modules where you think you
603            definitely don't need back-tracing.)
604         - Turn on back-trace instrumentation:
605              SMLofNJ.Internals.BTrace.mode (SOME true);
606         - Recompile your program.  (I.e., run "CM.make" or "use".)
607         - You may now turn instrumentation off again (if you want):
608              SMLofNJ.Internals.BTrace.mode (SOME false);
609         - Run your program as usual.  If it raises an exception that
610           reaches the interactive toplevel, then a back-trace will
611           automatically be printed.  After that, the toplevel loop
612           will print the exception history as usual.
613    
614    ----------------------------------------------------------------------
615    Name: Matthias Blume
616    Date: 2000/06/26 09:56:46 JST
617    Tag: blume-20000626-setup
618    Description:
619    
620    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
621          ML code before and after compiling a file (e.g., to set compiler
622          flags)
623    
624    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
625              - associated changes to btrace.sml (BTrace instrumentation pass)
626              - cleaner implementation of btimp.sml (BTrace tracing and report
627                module)
628    
629    CM manual: * new path encoding documented
630               * description of setup-parameter to "sml" added
631    
632    The biggest user-visible change to back-tracing is that it is no
633    longer necessary to compile all traced modules within the same
634    session.  (This was a real limitation.)
635    
636    ----------------------------------------------------------------------
637    Name: Matthias Blume
638    Date: 2000/06/24 12:40:00 JST
639    Tag: blume-20000624-startup
640    Description:
641    
642    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
643    bit too often -- to put it mildly. :)
644    
645    ----------------------------------------------------------------------
646    Name: Matthias Blume
647    Date: 2000/06/23 18:20:00 JST
648    Tag: blume-20000623-btrace
649    Description:
650    
651    This updates adds a backtrace facility to aid programmers in debugging
652    their programs.  This involves the following changes:
653    
654    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
655       keeping track of the current call stack.  When programs are compiled
656       in a special mode, the compiler will insert calls to these hooks
657       into the user program.
658       "Hook" means that it is possible for different implementations of
659       back-tracing to register themselves (at different times).
660    
661    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
662       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
663       It can be turned on using this call:
664         SMLofNJ.Internals.BTrace.mode (SOME true);
665       Turning it off again:
666         SMLofNJ.Internals.BTrace.mode (SOME false);
667       Querying the current status:
668         SMLofNJ.Internals.BTrace.mode NONE;
669       Annotated programs are about twice as big as normal ones, and they
670       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
671       where all hooks do nothing).  The slowdown with a plugin that is
672       actually useful (such as the one supplied by default) is even greater,
673       but in the case of the default plugin it is still only an constant
674       factor (amortized).
675    
676    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
677       with a sub-structure BTrace for controlling back-tracing.  In particular,
678       the above-mentioned function "mode" controls whether the annotation
679       phase is invoked by the compiler.  Another important function is
680       "trigger": when called it aborts the current execution and causes
681       the top-level loop to print a full back-trace.
682    
683    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
684       for back-tracing.  It keeps track of the dynamic call stack and in
685       addition to that it keeps a partial history at each "level" of that
686       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
687       d and b (at separate times, dynamically), then the report will show:
688    
689       GOTO   d
690             /c
691       GOTO  \b
692       CALL   a
693    
694       This shows that there was an initial non-tail call of a, then a
695       tail-call to b or c, looping behavior in a cluster of functions that
696       consist of b and c, and then a goto from that cluster (i.e., either from
697       b or from c) to d.
698    
699       Note that (depending on the user program) the amount of information
700       that the back-trace module has to keep track of at each level is bounded
701       by a constant.  Thus, the whole implementation has the same asymptotical
702       complexity as the original program (both in space and in time).
703    
704    5. compiler/TopLevel/interact/evalloop.sml has been modified to
705       handle the special exception SMLofNJ.Internals.BTrace.BTrace
706       which is raised by the "trigger" function mentioned above.
707    
708    Notes on usage:
709    
710    - Annotated code works well together with unannotated code:
711    Unannotated calls simply do not show up at all in the backtrace.
712    
713    - It is not a good idea to let modules that were annotated during
714    different sessions run at the same time.  This is because the compiler
715    chooses small integers to identify individual functions, and there
716    will be clashes if different modules were compiled in separate sessions.
717    (Nothing will crash, and you will even be told about the clashes, but
718    back-trace information will in general not be useful.)
719    
720    - Back-tracing can be confused by callcc and capture.
721    
722    - The only way of getting a back-trace right now is to explicitly
723    invoke the "trigger" function from your user program.  Eventually, we
724    should make every exception carry back-trace information (if
725    available).  But since this creates more overhead at "raise"-time
726    (similar to the current exnHistory overhead), I have not yet
727    implemented this.  (The implementation will be rather easy.)  With
728    exceptions carrying back-trace information, this facility will be even
729    more useful because users don't need to modify their programs...
730    
731    - While it is possible to compile the compiler with back-trace
732    annotations turned on (I did it to get some confidence in
733    correctness), you must make absolutely sure that core.sml and
734    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
735    be compiled with annotation because there is no core access yet, but
736    if you compile btimp.sml with annotation, then the system will go into
737    an infinite recursion and crash.)
738    Since CM currently does not know about BTrace, the only way to turn
739    annotations on and off for different modules of the compiler is to
740    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
741    this is awkward and clumsy.
742    
743    Sample sessions:
744    
745    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
746    - SMLofNJ.Internals.BTrace.mode (SOME true);
747    [autoloading]
748    [autoloading done]
749    val it = false : bool
750    - structure X = struct
751    -     fun main n = let
752    -         fun a (x, 0) = d x
753    -           | a (x, n) = b (x, n - 1)
754    -         and b (x, n) = c (x, n)
755    -         and c (x, n) = a (x, n)
756    -         and d x = e (x, 3)
757    -         and e (x, 0) = f x
758    -           | e (x, n) = e (x, n - 1)
759    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
760    -           | f n = n * g (n - 1)
761    -         and g n = a (n, 3)
762    -     in
763    -         f n
764    -     end
765    - end;
766    structure X : sig val main : int -> int end
767    - X.main 3;
768    *** BACK-TRACE ***
769    GOTO   stdIn:4.2-13.20: X.main[2].f
770    GOTO-( stdIn:4.2-13.20: X.main[2].e
771    GOTO   stdIn:4.2-13.20: X.main[2].d
772         / stdIn:4.2-13.20: X.main[2].a
773         | stdIn:4.2-13.20: X.main[2].b
774    GOTO-\ stdIn:4.2-13.20: X.main[2].c
775    CALL   stdIn:4.2-13.20: X.main[2].g
776    GOTO   stdIn:4.2-13.20: X.main[2].f
777    GOTO-( stdIn:4.2-13.20: X.main[2].e
778    GOTO   stdIn:4.2-13.20: X.main[2].d
779         / stdIn:4.2-13.20: X.main[2].a
780         | stdIn:4.2-13.20: X.main[2].b
781    GOTO-\ stdIn:4.2-13.20: X.main[2].c
782    CALL   stdIn:4.2-13.20: X.main[2].g
783    GOTO   stdIn:4.2-13.20: X.main[2].f
784    GOTO-( stdIn:4.2-13.20: X.main[2].e
785    GOTO   stdIn:4.2-13.20: X.main[2].d
786         / stdIn:4.2-13.20: X.main[2].a
787         | stdIn:4.2-13.20: X.main[2].b
788    GOTO-\ stdIn:4.2-13.20: X.main[2].c
789    CALL   stdIn:4.2-13.20: X.main[2].g
790    GOTO   stdIn:4.2-13.20: X.main[2].f
791    CALL   stdIn:2.15-17.4: X.main[2]
792    -
793    
794    (Note that because of a FLINt bug the above code currently does not
795    compile without BTrace turned on.)
796    
797    Here is another example, using my modified Tiger compiler:
798    
799    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
800    - SMLofNJ.Internals.BTrace.mode (SOME true);
801    [autoloading]
802    [autoloading done]
803    val it = false : bool
804    - CM.make "sources.cm";
805    [autoloading]
806    ...
807    [autoloading done]
808    [scanning sources.cm]
809    [parsing (sources.cm):parse.sml]
810    [creating directory CM/SKEL ...]
811    [parsing (sources.cm):tiger.lex.sml]
812    ...
813    [wrote CM/sparc-unix/semant.sml]
814    [compiling (sources.cm):main.sml]
815    [wrote CM/sparc-unix/main.sml]
816    [New bindings added.]
817    val it = true : bool
818    - Main.compile ("../testcases/merge.tig", "foo.out");
819    *** BACK-TRACE ***
820    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
821    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
822    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
823    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
824    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
825    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
826    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
827         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
828    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
829    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
830    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
831    CALL   main.sml:1.18-118.4: Main.compile[2]
832    -
833    
834    ----------------------------------------------------------------------
835    Name: Matthias Blumen
836    Date: 2000/06/21 18:00:00 JST
837    Tag: blume-20000621-manual
838    Description:
839    
840    CM manual update: Path environments documented.
841    
842    ----------------------------------------------------------------------
843    Name: Matthias Blume
844    Date: 2000/06/19 13:40:00
845    Tag: blume-20000619-manual
846    Description:
847    
848    CM manual and system/README update.  This only covers the fact that
849    there are no more implicit anchors.  (Path environments and the "bind"
850    option to "cm" have yet to be documented.)
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume
854    Date: 2000/06/19 11:05:00 JST
855    Tag: blume-20000619-chdir-bugfix
856    Description:
857    
858    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
859    
860    ----------------------------------------------------------------------
861    Name: Matthias Blume
862    Date: 2000/06/18 22:00:10 JST
863    Tag: blume-20000618-implicit-anchors-really-gone
864    Description:
865    
866    I updates the previous HISTORY entry where I forgot to mention that
867    implicit anchors are no longer with us.
868    
869    The current update also gets rid of the (now useless) controller
870    CM.Control.implicit_anchors.
871    
872    ----------------------------------------------------------------------
873    Name: Matthias Blume
874    Date: 2000/06/16 17:30:00 JST
875    Tag: blume-20000616-anchorenv
876    Description:
877    
878    This patch implements the long anticipated (just kidding :) "anchor
879    environment" mechanism.  In the course of doing this, I also
880    re-implemented CM's internal "SrcPath" module from scratch.  The new
881    one should be more robust in certain boundary cases.  In any case, it
882    is a lot cleaner than its predecessor (IMHO).
883    
884    This time, although there is yet another boot file format change, I
885    kept the unpickler backward-compatible.  As a result, no new bootfiles
886    are necessary and bootstrapping is straightforward.  (You cannot read
887    new bootfiles into an old system, but the other way around is no
888    problem.)
889    
890    Visible changes:
891    
892    ** 0. Implicit path anchors (without the leading $-symbol) are no
893    longer recognized at all. This means that such path names are not
894    illegal either.  For example, the name basis.cm simply refers to a
895    local file called "basis.cm" (i.e, the name is an ordinary path
896    relative to .cm-files directory).  Or, to put it differently, only
897    names that start with $ are anchored paths.
898    
899    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
900    vanished.
901    
902    John (Reppy) had critizised this as soon as I originally proposed and
903    implemented it, but at that time I did not really deeply believe
904    him. :) Now I came full-circle because I need the $<singlearc> syntax
905    in another place where it cannot be seen as an abbreviation for
906    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
907    seems to mean (i.e., it "expands" into the corresponding anchor
908    value).
909    
910    However, when paths are used as members in CM description files, it
911    continues to be true that there must be at least another arc after the
912    anchor.  This is now enforced separately during semantic analysis
913    (i.e., from a lexical/syntactical point of view, the notation is ok.)
914    
915    ** 2. The "cm" class now accepts an option "bind".  The option's value
916    is a sub-option list of precisely two items -- one labeled "anchor"
917    and the other one labeled "value".  As you might expect, "anchor" is
918    used to specify an anchor name to be bound, and "value" specifies what
919    the anchor is being bound to.
920    
921    The value must be a directory name and can be given in either standard
922    syntax (including the possibility that it is itself an anchored path)
923    or native syntax.
924    
925    Examples:
926    
927       foo.cm (bind:(anchor:bar value:$mystuff/bar))
928       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
929    
930    and so on.
931    
932    The meaning of this is that the .cm-file will be processed with an
933    augmented anchor environment where the given anchor(s) is/are bound to
934    the given values(s).
935    
936    The rationale for having this feature is this: Suppose you are trying
937    to use two different (already stable) libraries a.cm and b.cm (that
938    you perhaps didn't write yourself).  Further, suppose each of these
939    two libraries internally uses its own auxiliary library $aux/lib.cm.
940    Normally you would now have a problem because the anchor "lib" can not
941    be bound to more than one value globally.  Therefore, the project that
942    uses both a.cm and b.cm must locally redirect the anchor to some other
943    place:
944    
945       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
946       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
947    
948    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
949    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
950    
951    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
952    you when you do so at the time of CM.stabilize).  Therefore, you can
953    also use an anchored path as the value:
954    
955      a.cm (bind:(anchor:lib value:$a-lib))
956      b.cm (bind:(anchor:lib value:$b-lib))
957    
958    Now you can globally configure (using the usual CM.Anchor.anchor or
959    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
960    itself is always locally bound, setting it globally is no longer
961    meaningful or necessary (but it does not hurt either).  In fact, "lib"
962    can still be used as a global anchor for separate purposes.  As a
963    matter of fact, one can locally define "lib" in terms of a global
964    "lib":
965    
966      a.cm (bind:(anchor:lib value:$lib/a))
967      b.cm (bind:(anchor:lib value:$lib/b))
968    
969    ** 3: The encoding of path names has changed.  This affects the way
970    path names are shown in CM's progress report and also the internal
971    protocol encoding used for parallel make.
972    
973    The encoding now uses one or more ':'-separated segments.  Each
974    segments corresponds to a file that has been specified relative to the
975    file given by its preceding segment.  The first segment is either
976    relative to the CWD, absolute, or anchored.  Each segment itself is
977    basically a Unix pathname; all segments but the first are relative.
978    
979    Example:
980    
981       $foo/bar/baz.cm:a/b/c.sml
982    
983    This path denotes the file bar/a/b/c.sml relative to the directory
984    denoted by anchor "foo".  Notice that the encoding also includes
985    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
986    paths are resolved relative to the .cm-files directory, so baz.cm must
987    be ignored to get the "real" pathname.
988    
989    To make this fact more obvious, CM puts the names of such "virtual
990    arcs" into parentheses when they appear in progress reports. (No
991    parentheses will appear in the internal protocol encoding.)  Thus,
992    what you really see is:
993    
994      $foo/bar/(baz.cm):a/b/c.sml
995    
996    I find this notation to be much more informative than before.
997    
998    Another new feature of the encoding is that special characters
999    including parentheses, colons, (back)slashes, and white space are
1000    written as \ddd (where ddd is the decimal encoding of the character).
1001    
1002    *** The CM manual still needs to be updated.
1003    
1004    ----------------------------------------------------------------------
1005    Name: Allen Leung
1006    Date: 2000/06/15 00:38:00
1007    Tag: leunga-20000615-x86-peephole
1008    
1009    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1010    
1011    ----------------------------------------------------------------------
1012    Name: Matthias Blume
1013    Date: 2000/06/12 11:40:00
1014    Tag: blume-20000612-parmakefix
1015    Description:
1016    
1017    More cleanup after changing the file naming scheme: This time I
1018    repaired the parallel make mechanism for CMB.make which I broke earlier.
1019    
1020    ----------------------------------------------------------------------
1021    Name: Allen Leung
1022    Date: 2000/06/09 01:25:00
1023    Tag: leunga-20000609-various
1024    
1025    None of these things should affect normal SML/NJ operations
1026    
1027    1. Peephole improvements provided by Fermin (c--)
1028    2. New annotation DEFUSE for adding extra dependence (moby)
1029    3. New X86 LOCK instructions (moby)
1030    4. New machine description language for reservation tables (scheduling)
1031    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1032       trees etc.)
1033    6. I've changed the CM files so that they can work with versions
1034       110.0.6, 110.25 and 110.28
1035    
1036    ----------------------------------------------------------------------
1037    Name: Matthias Blume
1038    Date: 2000/06/09 12:40:00
1039    Tag: blume-20000609-log
1040    Description:
1041    
1042    - Removed all(?) remaining RCS Log entries from sources.
1043    
1044    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1045      anchored paths).
1046    
1047    ----------------------------------------------------------------------
1048    Name: Matthias Blume
1049    Date: 2000/06/07 17:00:00 JST
1050    Tag: blume-20000607-no-implicit-anchors
1051    Description:
1052    
1053    1. This update changes the default setting for
1054    CM.Control.implicit_anchors from true to false.  This means that
1055    implicit anchors are no longer permitted by default.  I also tried to
1056    make sure that nothing else still relies on implicit anchors.
1057    (This is the next step on the schedule towards a CM that does not even
1058    have the notion of implicit anchors anymore.)
1059    
1060    2. More CM manual updates.
1061    
1062    3. I managed to track down and fix the pickling bug I mentioned last
1063    time.  Because of the previously existing workaround, this entails no
1064    immediate practical changes.
1065    
1066    ----------------------------------------------------------------------
1067    Name: Matthias Blume
1068    Date: 2000/06/06 11:15:00 JST
1069    Tag: blume-20000606-lazierpickle
1070    Description:
1071    
1072    !!!! NEW BOOT FILES !!!!
1073    
1074    * The main purpose of this update is to make library pickles lazier in
1075    order to reduce the initial space penalty for autoloading a library.
1076    As a result, it is now possible to have $smlnj/compiler.cm
1077    pre-registered.  This should take care of the many complaints or
1078    inquiries about missing structure Compiler.  This required changes to
1079    CM's internal data structures and small tweaks to some algorithms.
1080    
1081    As a neat additional effect, it is no longer necessary (for the sake
1082    of lean heap image files) to distinguish between a "minimal" CM and a
1083    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1084    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1085    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1086    
1087    To make the life of compiler-hackers easier, "makeml" now also
1088    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1089    words, after you bootstrap a new sml for the first time, you will not
1090    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1091    around you will still have to do it, though.)
1092    
1093    * A second change consists of major updates to the CM manual.  There
1094    are now several appendices with summary information and also a full
1095    specification of the CM description file syntax.
1096    
1097    * In directory src/system I added the script "allcross".  This script
1098    invokes sml and cross-compiles the compiler for all supported
1099    architectures.  (Useful when providing a new set of boot files.)
1100    
1101    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1102    added a small tweak to pickle-util.sml to work around this problem,
1103    but it is not a proper fix yet.  I will investigate further.  (The
1104    effect of the bug was an inflation of library pickle size.)
1105    
1106    * Version number increased to 110.28.1 (to avoid compatibility problems).
1107    
1108    ----------------------------------------------------------------------
1109    Name: Allen Leung
1110    Date: 2000/05/25 17:28 EDT
1111    Tag: leunga-20000525-ra
1112    Description:
1113    
1114      Fixed a bug in freezing phase of the register allocator.
1115    
1116    ----------------------------------------------------------------------
1117    Name: Allen Leung
1118    Date: 2000/05/15 22:53 EDT
1119    Tag: leunga-20000515-alpha-x86-ra
1120    Description:
1121    
1122      1. Alpha
1123    
1124          Slight cleanup.  Removed the instruction SGNXL
1125    
1126      2. X86
1127    
1128          Added the following instructions to the instruction set:
1129    
1130            ROLx, RORx,
1131            BTx, BTSx, BTLx, BTRx,
1132            XCHGx, and variants with the LOCK prefix
1133    
1134      3. Register Allocation
1135    
1136          The module ra-rewrite-with-renaming has been improved.
1137    
1138      These have no effect on SML/NJ.
1139    
1140    ----------------------------------------------------------------------
1141    Name: Matthias Blume
1142    Date: 2000/05/15 16:20:00 JST
1143    Tag: blume-20000515-lightrebuild
1144    Description:
1145    
1146    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1147       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1148       defined.  In effect, the command will not build any cross-compiler
1149       backends and therefore finish more quickly.
1150    
1151       The "fixpt" script also takes a "-light" switch to be able to use
1152       this new facility while compiling for a fixpoint.
1153    
1154    2. I replaced all mentions of anchored paths in group owner specifications
1155       with simple relative paths (usually starting with "..").
1156       The rationale is that a library's internal workings should not be
1157       compromised by the lack of some anchor.  (An anchor is necessary
1158       for someone who wants to refer to the library by an anchored path,
1159       but it should not be necessary to build the same library in the first
1160       place.)
1161    
1162    3. I changed the way CM's tool mechanism determines the shell command
1163       string used for things like ml-yacc etc. so that it does not break
1164       when CM.Control.implicit_anchors is turned off.
1165    
1166    ----------------------------------------------------------------------
1167    Name: Matthias Blume
1168    Date: 2000/05/12 18:20:00 JST
1169    Tag: blume-20000512-ml-build
1170    Description:
1171    
1172    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1173    getting installed properly (by config/install.sh).
1174    
1175    ----------------------------------------------------------------------
1176    Name: Matthias Blume
1177    Date: 2000/05/12 17:30:00 JST
1178    Tag: blume-20000512-anchors
1179    Description:
1180    
1181    !!! NEW BOOT FILES !!!
1182    
1183    This change is in preparation of fading out support for "implicitly
1184    anchored path names".  I went through all sources and used the
1185    explicit (and relatively new) $-notation.  See system/README and the
1186    CM manual for more info on this.
1187    
1188    I also modified the anchoring scheme for some things such as "smlnj",
1189    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1190    anchors are more expressive: anchor name and first arc do not have to
1191    coincide.  This entails the following user-visible change:
1192    
1193    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1194    particular, when you fire up sml with a command-line argument, say,
1195    e.g.:
1196    
1197       sml '$smlnj/cmb.cm'
1198    
1199    At the ML toplevel prompt:
1200    
1201       CM.autoload "$smlnj/cmb.cm";
1202    
1203    There is also a new controller in CM.Control that can be used to turn
1204    off all remaining support for implicit anchors by saying:
1205    
1206        CM.autoload "$smlnj/
1207        #set CM.Control.implicit_anchors false;
1208    
1209    This causes CM to reject implicitly anchored paths.  This is (for the
1210    time being) less permissive than the "final" version where there will
1211    be no more such implicit anchors and relative paths will be just that:
1212    relative.
1213    
1214    The next step (version after next version?) will be to make the
1215    default for CM.Control.implicit_anchors false.  After the dust has
1216    settled, I can then produce the "final" version of this...
1217    
1218    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1219    
1220    ----------------------------------------------------------------------
1221    Name: Matthias Blume
1222    Date: 2000/05/11 16:30:00 JST
1223    Tag: blume-20000511-sources
1224    Description:
1225    
1226    The main change is that I added function CM.sources as a generalized
1227    version of the earlier CM.makedepend.  This entails the following
1228    additional changes:
1229    
1230      - CM.makedepend has been dropped.
1231    
1232      - CM manual has been updated.
1233    
1234      - TOOLS signature and API have been changed.
1235    
1236    ----------------------------------------------------------------------
1237    Name: Allen Leung
1238    Date: 2000/05/10 21:17 EDT
1239    Tag: leunga-20000510-moby-c--ssa
1240    Description:
1241    
1242      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1243    None of these affect SML/NJ.
1244    
1245    1. Register Allocation
1246    
1247        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1248           This module tries to remove local (i.e. basic block level) redundancies
1249           during spilling.
1250    
1251        b. A new framework for performing region based register allocation.
1252           Not yet entirely functional.
1253    
1254    2. X86
1255    
1256       a. DefUse for POP was missing the stack pointer [found by Lal]
1257       b. Reload for CALL was incorrect in X86Spill [found by John]
1258       c. Various fixes in X86Spill so that it can be used correctly for
1259          the new spilling module.
1260    
1261    3. SSA/IR
1262    
1263       a. New module ir/dj-dataflow.sml implements elimination based
1264          data flow analysis.
1265    
1266    4. MLRiscGen
1267    
1268       a. Fix for gc type annotation
1269    
1270    5. MDGen
1271    
1272       Various fixes for machine description -> ml code translation.  For ssa
1273       only.
1274    
1275    ----------------------------------------------------------------------
1276    Name: Allen Leung
1277    Date: 2000/05/08 22:17 EDT
1278    Tag: leunga-20000508-labexp
1279    Description:
1280    
1281      Fermin has found a few assembly problems with constant expressions
1282      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1283      which choke on dumb assemblers.  This is his fix.
1284    
1285    ----------------------------------------------------------------------
1286    Name: Dave MacQueen
1287    Date: 2000/04/09 14:00 EDT
1288    Tag: dbm-20000502-Version_110_28
1289    Description:
1290    
1291    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1292    
1293    2. Updated config/version to 110.28
1294    
1295    3. Updated config/srcarchiveurl
1296    
1297    3. New boot files!
1298       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1299    
1300    ----------------------------------------------------------------------
1301    Name: Matthias Blume
1302    Date: 2000/05/01 19:05:00 JST
1303    Tag: blume-20000501-noweb
1304    Description:
1305    
1306    A new noweb tool has been added.  The existing system is entirely
1307    unaffected by this, but some CM users have asked for renewed noweb
1308    support.  Everything is documented in the CM manual.
1309    
1310    New (plugin) libraries:
1311    
1312       noweb-tool.cm
1313       nw-ext.cm
1314    
1315    ----------------------------------------------------------------------
1316    Name: Dave MacQueen
1317    Date: 2000/04/30 12:40PM EDT
1318    Tag: dbm-20000430-bug_fixes
1319    Description:
1320    
1321    1. Fix for bug 1498
1322       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1323       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1324         added toRealArray function
1325       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1326         added check for tag Obj.RealArray to array printing case in ppObj
1327    
1328    2. Fix for bug 1510
1329       smlnj/src/compiler/Semant/types/typesutil.sml
1330         fixed definition of dummyargs (used by equalTycon) so that
1331         dummy args are distinct types
1332    
1333    ----------------------------------------------------------------------
1334    Name: Matthias Blume
1335    Date: 2000/04/30 01:00:00 JST
1336    Tag: blume-20000430-versions
1337    Description:
1338    
1339    1. CM version numbering added.  This is an implementation of Lal's
1340       proposal for adding version numbers and version checking to .cm
1341       files.  Lal said that his proposal was just that -- a proposal.
1342       For the time being I went ahead and implemented it so that people
1343       can comment on it.  Everything is completely backward-compatible
1344       (except for the stable library format, i.e., new bootfiles!).
1345    
1346       As usual, see the CM manual for details.
1347    
1348    2. An alternative syntax for anchored paths has been implemented.
1349       Dave has recently voiced the same concerns that I had when I did
1350       this, so there should be some support.  My take is that eventually
1351       I will let support for the current syntax (where anchors are
1352       "implicit") fade out in favor of the new, explicit syntax.
1353       In order to be backward-compatible, both old and new syntax are
1354       currently supported.
1355    
1356       Again, see the CM manual for details.
1357    
1358    3. Parallel make is trying to be slightly smarter:  When the master
1359       process finds a "bottleneck", i.e., when there is only one
1360       compilation unit that can be compiled and everybody else is
1361       waiting on it, then it will simply compile it directly instead
1362       of clumsily telling one of the slaves to do it.
1363    
1364    4. Support for "unsharing" added.  This is necessary in order to be
1365       able to have two different versions of the same library running
1366       at the same time (e.g., for trying out a new MLRISC while still
1367       having the old MLRISC linked into the current compiler, etc.)
1368       See the CM manual.
1369    
1370    5. Simple "makedepend" functionality added for generating Makefile
1371       dependency information.  (This is rather crude at the moment.
1372       Expect some changes here in the future.)
1373    
1374    6. ".fun" added as a recognized suffix for ML files. Also documented
1375       explicitly in the manual that the fallback behavior (unknown suffix
1376       -> ML file) is not an official feature!
1377    
1378    7. Small changes to the pickler for stable libraries.
1379    
1380    8. Several internal changes to CM (for cleanup/improvement).
1381    
1382    
1383    !!!! NEW BINFILES !!!!
1384    
1385    ----------------------------------------------------------------------
1386    Name: Matthias Blume
1387    Date: 2000/04/28 17:30:00 JST
1388    Tag: blume-20000428-pathconfig
1389    Description:
1390    
1391    1. I changed config/install.sh to remove duplicate entries from the
1392       lib/pathconfig file at the end.  Moreover, the final version of
1393       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1394       in src/system/installml.
1395    
1396    2. The config/install.sh script now consistently uses relative
1397       pathnames in lib/pathconfig whenever the anchor is in the lib
1398       directory.  (So far this was true for the libraries that come
1399       pre-compiled and bundled as part of the bootfiles but not for
1400       libraries that are compiled by the script itself.)
1401    
1402    ----------------------------------------------------------------------
1403    Name: Matthias Blume
1404    Date: 2000/04/26 13:10:00 JST
1405    Tag: blume-20000426-fun_suffix
1406    Description:
1407    
1408    Added ".fun" as a recognized file name suffix (for ML code).
1409    
1410    ----------------------------------------------------------------------
1411    Name: Allen Leung
1412    Date: 2000/04/25 17:00:00 EST
1413    Tag: leunga-20000425-alpha-ra
1414    Description:
1415    
1416    1. Alpha
1417    
1418        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1419    in C--.
1420    
1421    2. RA
1422    
1423       Added a flag "ra-dump-size" to print out the size of the flowgraph
1424       and the interference graph.
1425    
1426    ----------------------------------------------------------------------
1427    Name: Dave MacQueen
1428    Date: 2000/04/25/
1429    Tag: dbm-20000425-mlyacc_doc_examples
1430    Description:
1431      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1432      Updated all three examples in src/ml-yacc/examples to run
1433      under 110.* using CM.make.
1434    
1435    ----------------------------------------------------------------------
1436    Name: Allen Leung
1437    Date: 2000/04/20 23:04:00 EST
1438    Tag: leunga-20000420-ssa-c---stuff
1439    Description:
1440    
1441      This update synchronizes my repository with Yale's.  Most of these
1442    changes, however, do not affect SML/NJ at all (the RA is an exception).
1443    
1444    1. Register Allocator
1445    
1446       a. An improvement in the interference graph construction:
1447          Given a copy
1448    
1449                s <- t
1450    
1451          no interference edge between s and t is added for this definition of s.
1452    
1453       b. I've added two new spill heuristic modules that Fermin and I developed
1454          (in the new library RA.cm). These are unused in SML/NJ but maybe
1455          useful for others (Moby?)
1456    
1457    2. X86
1458    
1459       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1460    
1461    3. Alpha
1462    
1463       a. Added the BSR instruction and code generation that goes with it [C--]
1464       b. Other fixes too numerous to recount provided by Fermin [C--]
1465    
1466    4. Regmaps
1467    
1468       a. The regmaps are not initialized with the identity physical bindings
1469          at creation time.  This is unneeded.
1470    
1471    5. MLRISC Optimizations
1472    
1473       a. The DJ-Graph module can now compute the iterated dominance frontiers
1474          intersects with liveness incrementally in linear time! Woohoo!
1475          This is now used in my new SSA construction algorithm.
1476    
1477       b. THe branch reorganization module is now smarter about linear chains of
1478          basic blocks.
1479    
1480    
1481    ----------------------------------------------------------------------
1482    Name: Matthias Blume
1483  Date: 2000/04/12 13:52:00 JST  Date: 2000/04/12 13:52:00 JST
1484  Tag: blume_main_v110p27_1  Tag: blume_main_v110p27_1
1485  Description:  Description:
# Line 129  Line 1596 
1596    
1597  3.  Assembly  3.  Assembly
1598    
1599      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
1600      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
1601      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
1602    
# Line 152  Line 1619 
1619    
1620  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
1621  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
1622  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
1623  legal surface-syntax symbol).  legal surface-syntax symbol).
1624    
1625  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 162  Line 1629 
1629  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
1630    
1631  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1632  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
1633  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
1634  more details.  more details.
1635    
# Line 277  Line 1744 
1744     (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
1745     used).     used).
1746    
1747     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1748     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1749     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1750    
1751     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 290  Line 1757 
1757    
1758     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1759    
1760     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1761     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1762    
1763  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 424  Line 1891 
1891    
1892       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1893       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1894       testl is perferable to andl.       testl is preferable to andl.
1895    
1896  7. RA (x86 only)  7. RA (x86 only)
1897    
# Line 606  Line 2073 
2073    
2074  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2075  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2076  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
2077  explanation.  explanation.
2078    
2079  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2080  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2081  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2082  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 653  Line 2120 
2120  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
2121  <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,
2122  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
2123  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2124  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2125  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2126    
# Line 703  Line 2170 
2170    it from that remote directory.    it from that remote directory.
2171    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2172    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2173    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2174    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2175    
2176    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.615  
changed lines
  Added in v.758

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