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 605, Fri Apr 7 14:06:42 2000 UTC revision 757, Thu Dec 21 14:54:56 2000 UTC
# Line 12  Line 12 
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14  ----------------------------------------------------------------------  ----------------------------------------------------------------------
15    Name: Matthias Blume
16    Date: 2000/12/21 23:55:00 JST
17    Tag: blume-20001221-exn
18    Description:
19    
20    Probably most important: CM no longer silently swallows all exceptions
21    in the compiler.
22    Plus: some other minor CM changes.  For example, CM now reports some
23    sizes for generated binfiles (code, data, envpickle, lambdapickle).
24    
25    ----------------------------------------------------------------------
26    Name: Matthias Blume
27    Date: 2000/12/15 00:01:05 JST
28    Tag: blume-20001215-dirtool
29    Description:
30    
31    - "dir" tool added.
32    - improvements and cleanup to Tools structure
33    - documentation updates
34    
35    ----------------------------------------------------------------------
36    Name: Allen Leung
37    Date: Thu Dec 14 03:45:24 EST 2000
38    Description:
39    Tag:  leunga-20001214-int-inf
40    Description:
41    
42       In IntInf, added these standard functions, which are missing from our
43    implementation:
44    
45        andb : int * int -> int
46        xorb : int * int -> int
47        orb  : int * int -> int
48        notb : int -> int
49         <<   : int * word -> int
50        ~>>  : int * word -> int
51    
52       Not tested, I hope they are correct.
53    
54    ----------------------------------------------------------------------
55    Name: Allen Leung
56    Date: Fri Dec  8 19:23:26 EST 2000
57    Description:
58    Tag:  leunga-20001208-nowhere
59    Description:
60    
61      Slight improvements to the 'nowhere' tool to handle OR-patterns,
62    to generate better error messages etc.  Plus a brief manual.
63    
64    ----------------------------------------------------------------------
65    Name: Lal George
66    Date: 2000/12/08 09:54:02 EST 2000
67    Tag: Release_110_31
68    Description:
69    
70    - Version 110.31
71    ----------------------------------------------------------------------
72    Name: Allen Leung
73    Date: Thu Dec  7 22:01:04 EST 2000
74    Tag:  leunga-20001207-cell-monster-hack
75    Description:
76    
77    Major MLRISC internal changes.  Affect all clients.
78    Summary:
79    
80    1.  Type CELLS.cell = int is now replaced by a datatype.
81        As a result, the old regmap is now gone.  Almost all interfaces
82        in MLRISC change as a consequence.
83    
84    2.  A new brand version of machine description tool (v3.0) that generates
85        modules expecting the new interface.  The old version is removed.
86    
87    3.  The RA interface has been further abstracted into two new functors.
88        RISC_RA and X86RA.  These functors have much simpler interfaces.
89        [See also directory MLRISC/demo.]
90    
91    4.  Some other new source->source code generation tools are available:
92    
93        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
94        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
95           I use this tool to generate the peephole optimizers---with the new
96           cell type changes, peephole rules are becoming difficult to write
97           without conditional pattern matching.
98    
99    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
100        cover the entire MLRISC source tree so many things broke.
101    
102    6.  CM files have been moved to the subdirectory MLRISC/cm.
103        They are moved because there are a lot of them and they clutter up the
104        root dir.
105    
106    7.  More detailed documentation to come...
107    
108        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
109        a makeml -rebuild first.  This is because of other other
110        changes that Matthias has made (see below).
111    
112    
113    ----------------------------------------------------------------------
114    Name: Matthias Blume
115    Date: 2000/11/30 23:12:00 JST
116    Tag: blume-20001130-filereorg
117    Description:
118    
119    Some manual updates and some file reorganizations in CM.
120    
121    ----------------------------------------------------------------------
122    Name: Matthias Blume
123    Date: 2000/11/24 17:45:00 JST
124    Tag: blume-20001124-link
125    Description:
126    
127    Drastically improved link traversal code for the case that the dynamic
128    value was already loaded at bootstrap time.  As a result, CM and CMB
129    now both load blazingly fast -- even on a very slow machine.  Also,
130    memory consumption has been further reduced by this.
131    
132    Warning: The format of the PIDMAP file has changed.  THerefore, to
133    bootstrap you have to do this:
134    
135    1. Run CMB.make
136    2. Make a symbolic link for the boot directory:
137         ln -s sml.boot.ARCH-OS xxx
138    3. "Rebuild" the boot directory:
139         ./makeml -boot xxx -rebuild sml ; rm xxx
140    4. Boot normally:
141          ./makeml
142    
143    ----------------------------------------------------------------------
144    Name: Matthias Blume
145    Date: 2000/11/21 21:20:00 JST
146    Tag: blume-20001121-tools
147    Description:
148    
149    Continued hacking on autoloading problem -- with success this time.
150    Also changed tool-plugin mechanism.  See new CM manual.
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume
154    Date: 2000/11/19 14:30:00 JST
155    Tag:  blume-20001119-autoload
156    Description:
157    
158    Some hacking to make autoloading faster.  Success for CMB, no success
159    so far for CM.  There is a reduced structure CM' that autoloads faster.
160    (This is a temporary, non-documented hack to be eliminated again when
161    the general problem is solved.)
162    
163    ----------------------------------------------------------------------
164    Name: Matthias Blume
165    Date: 2000/11/17 14:10:00 JST
166    Tag: blume-20001117-pickle-lib
167    Description:
168    
169    1. Eliminated comp-lib.cm
170    2. Made pickle-lib.cm
171    3. Eliminated all uses of intset.sml (from comp-lib.cm)
172    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
173       equivalent constructs from smlnj-lib.cm (INtHashTable).
174    5. Point 4. also goes for those uses of intmap.* in MLRISC.
175       Duplicated intmap modules thrown out.
176    6. Hunted down all duplicated SCC code and replaced it with
177       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
178    7. Rewrote Feedback module.
179    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
180       should be thrown out and equivalent modules from smlnj-lib.cm
181       should be used (IntRedBlackSet, IntListSet, ...).
182    
183    Confirmed that compiler compiles to fixpoint.
184    
185    ----------------------------------------------------------------------
186    Name: Allen Leung
187    Date: 2000/11/10 18:00:00
188    Tag: leunga-20001110-new-x86-fp
189    
190    A new x86 floating point code generator has been added.
191    By default this is turned off.  To turn this on, do:
192    
193        CM.autoload "$smlnj/compiler.cm";
194        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
195    
196    Changes:
197    
198    1.  Changed FTAN to FPTAN so that the assembly output is correct.
199    2.  Changed the extension callback for FTANGENT to generate:
200    
201              fptan
202              fstp  %st(0)
203        instead of
204              fptan
205              fstpl ftempmem
206    
207    3.  Numerous assembly fixes for x86.
208    
209    5.  Cleaned up the machine code output module x86/x86MC.sml and added
210        support for a whole bunch of instructions and addressing modes:
211    
212          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
213          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
214          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
215          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
216          fxch %st(n)
217          fld %st(n)
218          fst %st(n)
219          fst mem
220          fstp %st(n)
221          fucom %st(n)
222          fucomp %st(n)
223    
224        All these are now generated when the fast fp mode is turned on.
225    
226    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
227    
228    ----------------------------------------------------------------------
229    Name: Matthias Blume
230    Date: 2000/11/09 11:20:00 JST
231    Tag: blume-20001109-scc
232    Description:
233    
234    Eliminated some code duplication:
235    
236    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
237       (Otherwise the functor is useless.)
238    2. Used GraphSCCFn where SCCUtilFun was used previously.
239    3. Got rid of SCCUtilFun (in comp-lib.cm).
240    
241    ----------------------------------------------------------------------
242    Name: Lal George
243    Date: 2000/11/06 09:02:21 EST 2000
244    Tag: Release_110_30
245    Description:
246    
247    - Version 110.30
248    ----------------------------------------------------------------------
249    Name: Matthias Blume
250    Date: 2000/11/04 14:45:00
251    Tag: blume-20001104-mlbuild
252    Description:
253    
254    - Made ml-build faster on startup.
255    - Documentation fixes.
256    
257    ----------------------------------------------------------------------
258    Name: Matthias Blume
259    Date: 2000/11/02 17:00:00 JST
260    Tag: blume-20001102-condcomp
261    Description:
262    
263    - Small tweaks to pickler -- new BOOTFILES!
264    - Version bumped to 110.29.2.
265    - Added conditional compilation facility to init.cmi (see comment there).
266    ----------------------------------------------------------------------
267    Name: Allen Leung
268    Date: 2000/10/23 19:31:00
269    Tag: leunga-20001023-demo-ra
270    
271    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
272    2. Test programs for the graph library updated
273    3. Some new MLRISC demo programs added
274    
275    ----------------------------------------------------------------------
276    Name: Matthias Blume
277    Date: 2000/08/31 22:15:00 JST
278    Tag: blume-20001017-errmsg
279    Description:
280    
281    More error message grief: Where there used to be no messages, there
282    now were some that had bogus error regions.  Fixed.
283    
284    ----------------------------------------------------------------------
285    Name: Matthias Blume
286    Date: 2000/08/31 17:30:00 JST
287    Tag: blume-20001017-v110p29p1
288    Description:
289    
290    I made a version 110.29.1 with new bootfiles.
291    
292    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
293              CM documentation changes and a small bugfix in CM's error reporting.
294    
295    ----------------------------------------------------------------------
296    Name: Lal George
297    Date: 2000/09/27 14:42:35 EDT
298    Tag: george-20000927-nodestatus
299    Description:
300    
301    Changed the type of the nodestatus, so that:
302    
303            SPILLED(~1)             is now SPILLED
304            SPILLED(m) where m>=0   is now MEMREG(m)
305            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
306    
307    ----------------------------------------------------------------------
308    Name: Matthias Blume
309    Date: 2000/09/07 14:45:00 JST
310    Tag: blume-20000907-cmerrmsg
311    Description:
312    
313    Small tweak to CM to avoid getting ML syntax error messages twice.
314    
315    ----------------------------------------------------------------------
316    Name: Matthias Blume
317    Date: 2000/08/31 18:00:00 JST
318    Tag: blume-20000831-cvsbootfiles
319    Description:
320    
321    New URL for boot files (because the 110.29 files on the BL server do
322    now work correctly with my updated install scripts for yacc and lex).
323    
324    ----------------------------------------------------------------------
325    Name: Matthias Blume
326    Date: 2000/08/08 12:33:00 JST
327    Tag: blume-20000808-manual
328    Description:
329    
330    Tiny update to CM manual.
331    
332    ----------------------------------------------------------------------
333    Name: Allen Leung
334    Date: 2000/08/7 19:31:00
335    Tag: leunga-20000807-a-whole-bunch-of-stuff
336    
337      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
338    and Mobdy.
339    
340    1.  x86
341    
342       a.  Fixes to peephole module by John and Dan.
343       b.  Assembly fix to SETcc by Allen.
344       c.  Fix to c-call by John.
345       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
346       e.  Instruction selection optimization to SETcc as suggested by John.
347    
348           For example,
349    
350            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
351    
352           should generate:
353    
354            MOVL a, x
355            SUBL b, x
356            SHRL 31, x
357    
358    2.  IR stuff
359    
360         A bunch of new DJ-graph related algorithms added.  These
361         speed up SSA construction.
362    
363    3.  SSA + Scheduling
364    
365         Added code for SSA and scheduling to the repository
366    
367    ----------------------------------------------------------------------
368    Name: Lal George
369    Date: 2000/07/27 11:53:14 EDT
370    
371    Tag: lal-20000727-linux-ppc
372    Description:
373    
374     Made changes to support Linux PPC.
375     p.s. I have confirmation that the 110.29 boot files work fine.
376    
377    ----------------------------------------------------------------------
378    Name: Matthias Blume
379    Date: 2000/07/27 17:40:00 JST
380    Tag: blume-20000727-scripts
381    Description:
382    
383    !!!! WARNING !!!!
384    You must recompile the runtime system!
385    !!!! WARNING !!!!
386    
387    This is basically another round of script-enhancements:
388    
389    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
390       and undefine CM preprocessor symbols.
391    
392    2. ml-build avoids generating a new heap image if it finds that the
393       existing one is still ok.  (The condition is that no ML file had to
394       be recompiled and all ML files are found to be older that the heap
395       file.)
396    
397       To make this work smoothly, I also hacked the runtime system as
398       well as SMLofNJ.SysInfo to get access to the heap image suffix
399       (.sparc-solaris, ...) that is currently being used.
400    
401       Moreover, the signature of CM.mk_standalone has changed.  See the
402       CM manual.
403    
404    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
405       CM manual for details.)
406    
407    4. More CM manual updates:
408        - all of the above has been documented.
409        - there is now a section describing the (CM-related) command line
410          arguments that are accepted by the "sml" command
411    
412    ----------------------------------------------------------------------
413    Name: Matthias Blume
414    Date: 2000/07/25 16:20:00 JST
415    Tag: blume-20000725-makedepend
416    Description:
417    
418    Added a script called ml-makedepend.  This can be used in makefiles
419    for Unix' make in a way very similar to the "makedepend" command for
420    C.
421    
422    The script internally uses function CM.sources.
423    
424    Synopsis:
425    
426        ml-makedepend [-f makefile] cmfile targetname
427    
428    The default for the makefile is "makefile" (or "Makefile" should
429    "makefile" not exist).
430    
431    ml-makedepend adds a cmfile/targetname-specific section to this
432    makefile (after removing the previous version of this section).  The
433    section contains a single dependency specification with targetname on
434    the LHS (targetname is an arbitrary name), and a list of files derived
435    from the cmfile on the RHS.  Some of the files on the RHS are
436    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
437    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
438    path names.  The makefile writer is responsible for making sure that
439    these variables have correct at the time "make" is invoked.
440    
441    ----------------------------------------------------------------------
442    Name: Matthias Blume
443    Date: 2000/07/22 23:30:00 JST
444    Tag: blume-20000722-urlupdate
445    Description:
446    
447    Changed BOOT and config/srcarchiveurl to point to BL server:
448    
449        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
450    
451    ----------------------------------------------------------------------
452    Name: Matthias Blume
453    Date: 2000/07/18 18:00:00 JST
454    Tag: blume-20000718-Version_110_29
455    Description:
456    
457    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
458    
459    2. Updated config/version to 110.29
460    
461    3. Updated config/srcarchiveurl
462    
463    3. New boot files!
464       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
465    
466    ----------------------------------------------------------------------
467    Name: Matthias Blume
468    Date: 2000/07/11 13:58:00 JST
469    Tag: blume-20000711-doctypo
470    Description:
471    
472    Fixed a few typos in CM manual.
473    
474    ----------------------------------------------------------------------
475    Name: Allen Leung
476    Date: 2000/06/15 00:38:00
477    Tag: leunga-20000704-sparc-x86
478    
479    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
480    2. fix to x86 RET bug [found by Dan Grossman]
481    3. sparc assembly bug fix for ticc instructions [found by Fermin]
482    
483       Affects c-- and moby only
484    
485    ----------------------------------------------------------------------
486    Name: Matthias Blume
487    Date: 2000/07/04 15:26:00
488    Tag: blume-20000704-trigger
489    Description:
490    
491    1. Improvements to CM manual.
492    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
493       of getting a back-trace.  The function, when called, raises an
494       internal exception which explicitly carries the full back-trace history,
495       so it is unaffected by any intervening handle-raise pairs ("trivial"
496       or not).  The interactive loop will print that history once it arrives
497       at top level.
498       Short of having all exceptions implicitly carry the full history, the
499       recommended way of using this facility is:
500         - compile your program with instrumentation "on"
501         - run it, when it raises an exception, look at the history
502         - if the history is "cut off" because of some handler, go and modify
503           your program so that it explicitly calls BTrace.trigger
504         - recompile (still instrumented), and rerun; look at the full history
505    
506    ----------------------------------------------------------------------
507    Name: Matthias Blume
508    Date: 2000/07/03 15:36:00 JST
509    Tag: blume-20000702-manual
510    Description:
511    
512    Small corrections and updates to CM manual.
513    
514    ----------------------------------------------------------------------
515    Name: Matthias Blume
516    Date: 2000/06/29 16:04:00 JST
517    Tag: blume-20000629-yacctool
518    Description:
519    
520    Changes:
521    
522    1. Class "mlyacc" now takes separate arguments to pass options to
523       generated .sml- and .sig-files independently.
524    2. Corresponding CM manual updates.
525    3. BTrace module now also reports call sites.  (However, for loop clusters
526       it only shows from where the cluster was entered.)  There are associated
527       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
528    
529    ----------------------------------------------------------------------
530    Name: Matthias Blume
531    Date: 2000/06/27 16:51:00 JST
532    Tag: blume-20000627-noweb
533    Description:
534    
535    Changes:
536    
537     1. Implemented "subdir" and "witness" options for noweb tool.
538        This caused some slight internal changes in CM's tool implementation.
539     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
540        some remaining issues from earlier path anchor changes.
541     3. Updated CM manual accordingly.
542    
543     4. Changed implementation of back-tracing so that I now consider it
544        ready for prime-time.
545    
546        In particular, you don't have to explicitly trigger the back-trace
547        anymore.  Instead, if you are running BTrace-instrumented code and
548        there is an uncaught exception (regardless of whether or not it was
549        raised in instrumented code), the top-level evalloop will print
550        the back-trace.
551    
552        Features:
553    
554          - Instrumented and uninstrumented code work together seemlessly.
555            (Of course, uninstrumented code is never mentioned in actual
556             back-traces.)
557    
558          - Asymptotic time- and space-complexity of instrumented code is
559            equal to that of uninstrumented code.  (This means that
560            tail-recursion is preserved by the instrumentation phase.)
561    
562          - Modules whose code has been instrumented in different sessions
563            work together without problem.
564    
565          - There is no penalty whatsoever on uninstrumented code.
566    
567          - There is no penalty on "raise" expressions, even in
568            instrumented code.
569    
570        A potential bug (or perhaps it is a feature, too):
571    
572          A back-trace reaches no further than the outermost instrumented
573          non-trivial "raise".  Here, a "trivial" raise is one that is the
574          sole RHS of a "handle" rule.  Thus, back-traces reach trough
575    
576               <exp> handle e => raise e
577    
578          and even
579    
580               <exp> handle Foo => raise Bar
581    
582          and, of course, through
583    
584               <exp> handle Foo => ...
585    
586         if the exception was not Foo.
587    
588         Back-traces always reach right through any un-instrumented code
589         including any of its "handle" expressions, trivial or not.
590    
591       To try this out, do the following:
592    
593         - Erase all existing binfiles for your program.
594           (You may keep binfiles for those modules where you think you
595            definitely don't need back-tracing.)
596         - Turn on back-trace instrumentation:
597              SMLofNJ.Internals.BTrace.mode (SOME true);
598         - Recompile your program.  (I.e., run "CM.make" or "use".)
599         - You may now turn instrumentation off again (if you want):
600              SMLofNJ.Internals.BTrace.mode (SOME false);
601         - Run your program as usual.  If it raises an exception that
602           reaches the interactive toplevel, then a back-trace will
603           automatically be printed.  After that, the toplevel loop
604           will print the exception history as usual.
605    
606    ----------------------------------------------------------------------
607    Name: Matthias Blume
608    Date: 2000/06/26 09:56:46 JST
609    Tag: blume-20000626-setup
610    Description:
611    
612    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
613          ML code before and after compiling a file (e.g., to set compiler
614          flags)
615    
616    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
617              - associated changes to btrace.sml (BTrace instrumentation pass)
618              - cleaner implementation of btimp.sml (BTrace tracing and report
619                module)
620    
621    CM manual: * new path encoding documented
622               * description of setup-parameter to "sml" added
623    
624    The biggest user-visible change to back-tracing is that it is no
625    longer necessary to compile all traced modules within the same
626    session.  (This was a real limitation.)
627    
628    ----------------------------------------------------------------------
629    Name: Matthias Blume
630    Date: 2000/06/24 12:40:00 JST
631    Tag: blume-20000624-startup
632    Description:
633    
634    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
635    bit too often -- to put it mildly. :)
636    
637    ----------------------------------------------------------------------
638    Name: Matthias Blume
639    Date: 2000/06/23 18:20:00 JST
640    Tag: blume-20000623-btrace
641    Description:
642    
643    This updates adds a backtrace facility to aid programmers in debugging
644    their programs.  This involves the following changes:
645    
646    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
647       keeping track of the current call stack.  When programs are compiled
648       in a special mode, the compiler will insert calls to these hooks
649       into the user program.
650       "Hook" means that it is possible for different implementations of
651       back-tracing to register themselves (at different times).
652    
653    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
654       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
655       It can be turned on using this call:
656         SMLofNJ.Internals.BTrace.mode (SOME true);
657       Turning it off again:
658         SMLofNJ.Internals.BTrace.mode (SOME false);
659       Querying the current status:
660         SMLofNJ.Internals.BTrace.mode NONE;
661       Annotated programs are about twice as big as normal ones, and they
662       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
663       where all hooks do nothing).  The slowdown with a plugin that is
664       actually useful (such as the one supplied by default) is even greater,
665       but in the case of the default plugin it is still only an constant
666       factor (amortized).
667    
668    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
669       with a sub-structure BTrace for controlling back-tracing.  In particular,
670       the above-mentioned function "mode" controls whether the annotation
671       phase is invoked by the compiler.  Another important function is
672       "trigger": when called it aborts the current execution and causes
673       the top-level loop to print a full back-trace.
674    
675    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
676       for back-tracing.  It keeps track of the dynamic call stack and in
677       addition to that it keeps a partial history at each "level" of that
678       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
679       d and b (at separate times, dynamically), then the report will show:
680    
681       GOTO   d
682             /c
683       GOTO  \b
684       CALL   a
685    
686       This shows that there was an initial non-tail call of a, then a
687       tail-call to b or c, looping behavior in a cluster of functions that
688       consist of b and c, and then a goto from that cluster (i.e., either from
689       b or from c) to d.
690    
691       Note that (depending on the user program) the amount of information
692       that the back-trace module has to keep track of at each level is bounded
693       by a constant.  Thus, the whole implementation has the same asymptotical
694       complexity as the original program (both in space and in time).
695    
696    5. compiler/TopLevel/interact/evalloop.sml has been modified to
697       handle the special exception SMLofNJ.Internals.BTrace.BTrace
698       which is raised by the "trigger" function mentioned above.
699    
700    Notes on usage:
701    
702    - Annotated code works well together with unannotated code:
703    Unannotated calls simply do not show up at all in the backtrace.
704    
705    - It is not a good idea to let modules that were annotated during
706    different sessions run at the same time.  This is because the compiler
707    chooses small integers to identify individual functions, and there
708    will be clashes if different modules were compiled in separate sessions.
709    (Nothing will crash, and you will even be told about the clashes, but
710    back-trace information will in general not be useful.)
711    
712    - Back-tracing can be confused by callcc and capture.
713    
714    - The only way of getting a back-trace right now is to explicitly
715    invoke the "trigger" function from your user program.  Eventually, we
716    should make every exception carry back-trace information (if
717    available).  But since this creates more overhead at "raise"-time
718    (similar to the current exnHistory overhead), I have not yet
719    implemented this.  (The implementation will be rather easy.)  With
720    exceptions carrying back-trace information, this facility will be even
721    more useful because users don't need to modify their programs...
722    
723    - While it is possible to compile the compiler with back-trace
724    annotations turned on (I did it to get some confidence in
725    correctness), you must make absolutely sure that core.sml and
726    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
727    be compiled with annotation because there is no core access yet, but
728    if you compile btimp.sml with annotation, then the system will go into
729    an infinite recursion and crash.)
730    Since CM currently does not know about BTrace, the only way to turn
731    annotations on and off for different modules of the compiler is to
732    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
733    this is awkward and clumsy.
734    
735    Sample sessions:
736    
737    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
738    - SMLofNJ.Internals.BTrace.mode (SOME true);
739    [autoloading]
740    [autoloading done]
741    val it = false : bool
742    - structure X = struct
743    -     fun main n = let
744    -         fun a (x, 0) = d x
745    -           | a (x, n) = b (x, n - 1)
746    -         and b (x, n) = c (x, n)
747    -         and c (x, n) = a (x, n)
748    -         and d x = e (x, 3)
749    -         and e (x, 0) = f x
750    -           | e (x, n) = e (x, n - 1)
751    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
752    -           | f n = n * g (n - 1)
753    -         and g n = a (n, 3)
754    -     in
755    -         f n
756    -     end
757    - end;
758    structure X : sig val main : int -> int end
759    - X.main 3;
760    *** BACK-TRACE ***
761    GOTO   stdIn:4.2-13.20: X.main[2].f
762    GOTO-( stdIn:4.2-13.20: X.main[2].e
763    GOTO   stdIn:4.2-13.20: X.main[2].d
764         / stdIn:4.2-13.20: X.main[2].a
765         | stdIn:4.2-13.20: X.main[2].b
766    GOTO-\ stdIn:4.2-13.20: X.main[2].c
767    CALL   stdIn:4.2-13.20: X.main[2].g
768    GOTO   stdIn:4.2-13.20: X.main[2].f
769    GOTO-( stdIn:4.2-13.20: X.main[2].e
770    GOTO   stdIn:4.2-13.20: X.main[2].d
771         / stdIn:4.2-13.20: X.main[2].a
772         | stdIn:4.2-13.20: X.main[2].b
773    GOTO-\ stdIn:4.2-13.20: X.main[2].c
774    CALL   stdIn:4.2-13.20: X.main[2].g
775    GOTO   stdIn:4.2-13.20: X.main[2].f
776    GOTO-( stdIn:4.2-13.20: X.main[2].e
777    GOTO   stdIn:4.2-13.20: X.main[2].d
778         / stdIn:4.2-13.20: X.main[2].a
779         | stdIn:4.2-13.20: X.main[2].b
780    GOTO-\ stdIn:4.2-13.20: X.main[2].c
781    CALL   stdIn:4.2-13.20: X.main[2].g
782    GOTO   stdIn:4.2-13.20: X.main[2].f
783    CALL   stdIn:2.15-17.4: X.main[2]
784    -
785    
786    (Note that because of a FLINt bug the above code currently does not
787    compile without BTrace turned on.)
788    
789    Here is another example, using my modified Tiger compiler:
790    
791    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
792    - SMLofNJ.Internals.BTrace.mode (SOME true);
793    [autoloading]
794    [autoloading done]
795    val it = false : bool
796    - CM.make "sources.cm";
797    [autoloading]
798    ...
799    [autoloading done]
800    [scanning sources.cm]
801    [parsing (sources.cm):parse.sml]
802    [creating directory CM/SKEL ...]
803    [parsing (sources.cm):tiger.lex.sml]
804    ...
805    [wrote CM/sparc-unix/semant.sml]
806    [compiling (sources.cm):main.sml]
807    [wrote CM/sparc-unix/main.sml]
808    [New bindings added.]
809    val it = true : bool
810    - Main.compile ("../testcases/merge.tig", "foo.out");
811    *** BACK-TRACE ***
812    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
813    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
814    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
815    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
816    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
817    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
818    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
819         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
820    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
821    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
822    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
823    CALL   main.sml:1.18-118.4: Main.compile[2]
824    -
825    
826    ----------------------------------------------------------------------
827    Name: Matthias Blumen
828    Date: 2000/06/21 18:00:00 JST
829    Tag: blume-20000621-manual
830    Description:
831    
832    CM manual update: Path environments documented.
833    
834    ----------------------------------------------------------------------
835    Name: Matthias Blume
836    Date: 2000/06/19 13:40:00
837    Tag: blume-20000619-manual
838    Description:
839    
840    CM manual and system/README update.  This only covers the fact that
841    there are no more implicit anchors.  (Path environments and the "bind"
842    option to "cm" have yet to be documented.)
843    
844    ----------------------------------------------------------------------
845    Name: Matthias Blume
846    Date: 2000/06/19 11:05:00 JST
847    Tag: blume-20000619-chdir-bugfix
848    Description:
849    
850    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume
854    Date: 2000/06/18 22:00:10 JST
855    Tag: blume-20000618-implicit-anchors-really-gone
856    Description:
857    
858    I updates the previous HISTORY entry where I forgot to mention that
859    implicit anchors are no longer with us.
860    
861    The current update also gets rid of the (now useless) controller
862    CM.Control.implicit_anchors.
863    
864    ----------------------------------------------------------------------
865    Name: Matthias Blume
866    Date: 2000/06/16 17:30:00 JST
867    Tag: blume-20000616-anchorenv
868    Description:
869    
870    This patch implements the long anticipated (just kidding :) "anchor
871    environment" mechanism.  In the course of doing this, I also
872    re-implemented CM's internal "SrcPath" module from scratch.  The new
873    one should be more robust in certain boundary cases.  In any case, it
874    is a lot cleaner than its predecessor (IMHO).
875    
876    This time, although there is yet another boot file format change, I
877    kept the unpickler backward-compatible.  As a result, no new bootfiles
878    are necessary and bootstrapping is straightforward.  (You cannot read
879    new bootfiles into an old system, but the other way around is no
880    problem.)
881    
882    Visible changes:
883    
884    ** 0. Implicit path anchors (without the leading $-symbol) are no
885    longer recognized at all. This means that such path names are not
886    illegal either.  For example, the name basis.cm simply refers to a
887    local file called "basis.cm" (i.e, the name is an ordinary path
888    relative to .cm-files directory).  Or, to put it differently, only
889    names that start with $ are anchored paths.
890    
891    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
892    vanished.
893    
894    John (Reppy) had critizised this as soon as I originally proposed and
895    implemented it, but at that time I did not really deeply believe
896    him. :) Now I came full-circle because I need the $<singlearc> syntax
897    in another place where it cannot be seen as an abbreviation for
898    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
899    seems to mean (i.e., it "expands" into the corresponding anchor
900    value).
901    
902    However, when paths are used as members in CM description files, it
903    continues to be true that there must be at least another arc after the
904    anchor.  This is now enforced separately during semantic analysis
905    (i.e., from a lexical/syntactical point of view, the notation is ok.)
906    
907    ** 2. The "cm" class now accepts an option "bind".  The option's value
908    is a sub-option list of precisely two items -- one labeled "anchor"
909    and the other one labeled "value".  As you might expect, "anchor" is
910    used to specify an anchor name to be bound, and "value" specifies what
911    the anchor is being bound to.
912    
913    The value must be a directory name and can be given in either standard
914    syntax (including the possibility that it is itself an anchored path)
915    or native syntax.
916    
917    Examples:
918    
919       foo.cm (bind:(anchor:bar value:$mystuff/bar))
920       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
921    
922    and so on.
923    
924    The meaning of this is that the .cm-file will be processed with an
925    augmented anchor environment where the given anchor(s) is/are bound to
926    the given values(s).
927    
928    The rationale for having this feature is this: Suppose you are trying
929    to use two different (already stable) libraries a.cm and b.cm (that
930    you perhaps didn't write yourself).  Further, suppose each of these
931    two libraries internally uses its own auxiliary library $aux/lib.cm.
932    Normally you would now have a problem because the anchor "lib" can not
933    be bound to more than one value globally.  Therefore, the project that
934    uses both a.cm and b.cm must locally redirect the anchor to some other
935    place:
936    
937       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
938       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
939    
940    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
941    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
942    
943    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
944    you when you do so at the time of CM.stabilize).  Therefore, you can
945    also use an anchored path as the value:
946    
947      a.cm (bind:(anchor:lib value:$a-lib))
948      b.cm (bind:(anchor:lib value:$b-lib))
949    
950    Now you can globally configure (using the usual CM.Anchor.anchor or
951    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
952    itself is always locally bound, setting it globally is no longer
953    meaningful or necessary (but it does not hurt either).  In fact, "lib"
954    can still be used as a global anchor for separate purposes.  As a
955    matter of fact, one can locally define "lib" in terms of a global
956    "lib":
957    
958      a.cm (bind:(anchor:lib value:$lib/a))
959      b.cm (bind:(anchor:lib value:$lib/b))
960    
961    ** 3: The encoding of path names has changed.  This affects the way
962    path names are shown in CM's progress report and also the internal
963    protocol encoding used for parallel make.
964    
965    The encoding now uses one or more ':'-separated segments.  Each
966    segments corresponds to a file that has been specified relative to the
967    file given by its preceding segment.  The first segment is either
968    relative to the CWD, absolute, or anchored.  Each segment itself is
969    basically a Unix pathname; all segments but the first are relative.
970    
971    Example:
972    
973       $foo/bar/baz.cm:a/b/c.sml
974    
975    This path denotes the file bar/a/b/c.sml relative to the directory
976    denoted by anchor "foo".  Notice that the encoding also includes
977    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
978    paths are resolved relative to the .cm-files directory, so baz.cm must
979    be ignored to get the "real" pathname.
980    
981    To make this fact more obvious, CM puts the names of such "virtual
982    arcs" into parentheses when they appear in progress reports. (No
983    parentheses will appear in the internal protocol encoding.)  Thus,
984    what you really see is:
985    
986      $foo/bar/(baz.cm):a/b/c.sml
987    
988    I find this notation to be much more informative than before.
989    
990    Another new feature of the encoding is that special characters
991    including parentheses, colons, (back)slashes, and white space are
992    written as \ddd (where ddd is the decimal encoding of the character).
993    
994    *** The CM manual still needs to be updated.
995    
996    ----------------------------------------------------------------------
997    Name: Allen Leung
998    Date: 2000/06/15 00:38:00
999    Tag: leunga-20000615-x86-peephole
1000    
1001    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1002    
1003    ----------------------------------------------------------------------
1004    Name: Matthias Blume
1005    Date: 2000/06/12 11:40:00
1006    Tag: blume-20000612-parmakefix
1007    Description:
1008    
1009    More cleanup after changing the file naming scheme: This time I
1010    repaired the parallel make mechanism for CMB.make which I broke earlier.
1011    
1012    ----------------------------------------------------------------------
1013    Name: Allen Leung
1014    Date: 2000/06/09 01:25:00
1015    Tag: leunga-20000609-various
1016    
1017    None of these things should affect normal SML/NJ operations
1018    
1019    1. Peephole improvements provided by Fermin (c--)
1020    2. New annotation DEFUSE for adding extra dependence (moby)
1021    3. New X86 LOCK instructions (moby)
1022    4. New machine description language for reservation tables (scheduling)
1023    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1024       trees etc.)
1025    6. I've changed the CM files so that they can work with versions
1026       110.0.6, 110.25 and 110.28
1027    
1028    ----------------------------------------------------------------------
1029    Name: Matthias Blume
1030    Date: 2000/06/09 12:40:00
1031    Tag: blume-20000609-log
1032    Description:
1033    
1034    - Removed all(?) remaining RCS Log entries from sources.
1035    
1036    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1037      anchored paths).
1038    
1039    ----------------------------------------------------------------------
1040    Name: Matthias Blume
1041    Date: 2000/06/07 17:00:00 JST
1042    Tag: blume-20000607-no-implicit-anchors
1043    Description:
1044    
1045    1. This update changes the default setting for
1046    CM.Control.implicit_anchors from true to false.  This means that
1047    implicit anchors are no longer permitted by default.  I also tried to
1048    make sure that nothing else still relies on implicit anchors.
1049    (This is the next step on the schedule towards a CM that does not even
1050    have the notion of implicit anchors anymore.)
1051    
1052    2. More CM manual updates.
1053    
1054    3. I managed to track down and fix the pickling bug I mentioned last
1055    time.  Because of the previously existing workaround, this entails no
1056    immediate practical changes.
1057    
1058    ----------------------------------------------------------------------
1059    Name: Matthias Blume
1060    Date: 2000/06/06 11:15:00 JST
1061    Tag: blume-20000606-lazierpickle
1062    Description:
1063    
1064    !!!! NEW BOOT FILES !!!!
1065    
1066    * The main purpose of this update is to make library pickles lazier in
1067    order to reduce the initial space penalty for autoloading a library.
1068    As a result, it is now possible to have $smlnj/compiler.cm
1069    pre-registered.  This should take care of the many complaints or
1070    inquiries about missing structure Compiler.  This required changes to
1071    CM's internal data structures and small tweaks to some algorithms.
1072    
1073    As a neat additional effect, it is no longer necessary (for the sake
1074    of lean heap image files) to distinguish between a "minimal" CM and a
1075    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1076    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1077    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1078    
1079    To make the life of compiler-hackers easier, "makeml" now also
1080    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1081    words, after you bootstrap a new sml for the first time, you will not
1082    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1083    around you will still have to do it, though.)
1084    
1085    * A second change consists of major updates to the CM manual.  There
1086    are now several appendices with summary information and also a full
1087    specification of the CM description file syntax.
1088    
1089    * In directory src/system I added the script "allcross".  This script
1090    invokes sml and cross-compiles the compiler for all supported
1091    architectures.  (Useful when providing a new set of boot files.)
1092    
1093    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1094    added a small tweak to pickle-util.sml to work around this problem,
1095    but it is not a proper fix yet.  I will investigate further.  (The
1096    effect of the bug was an inflation of library pickle size.)
1097    
1098    * Version number increased to 110.28.1 (to avoid compatibility problems).
1099    
1100    ----------------------------------------------------------------------
1101    Name: Allen Leung
1102    Date: 2000/05/25 17:28 EDT
1103    Tag: leunga-20000525-ra
1104    Description:
1105    
1106      Fixed a bug in freezing phase of the register allocator.
1107    
1108    ----------------------------------------------------------------------
1109    Name: Allen Leung
1110    Date: 2000/05/15 22:53 EDT
1111    Tag: leunga-20000515-alpha-x86-ra
1112    Description:
1113    
1114      1. Alpha
1115    
1116          Slight cleanup.  Removed the instruction SGNXL
1117    
1118      2. X86
1119    
1120          Added the following instructions to the instruction set:
1121    
1122            ROLx, RORx,
1123            BTx, BTSx, BTLx, BTRx,
1124            XCHGx, and variants with the LOCK prefix
1125    
1126      3. Register Allocation
1127    
1128          The module ra-rewrite-with-renaming has been improved.
1129    
1130      These have no effect on SML/NJ.
1131    
1132    ----------------------------------------------------------------------
1133    Name: Matthias Blume
1134    Date: 2000/05/15 16:20:00 JST
1135    Tag: blume-20000515-lightrebuild
1136    Description:
1137    
1138    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1139       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1140       defined.  In effect, the command will not build any cross-compiler
1141       backends and therefore finish more quickly.
1142    
1143       The "fixpt" script also takes a "-light" switch to be able to use
1144       this new facility while compiling for a fixpoint.
1145    
1146    2. I replaced all mentions of anchored paths in group owner specifications
1147       with simple relative paths (usually starting with "..").
1148       The rationale is that a library's internal workings should not be
1149       compromised by the lack of some anchor.  (An anchor is necessary
1150       for someone who wants to refer to the library by an anchored path,
1151       but it should not be necessary to build the same library in the first
1152       place.)
1153    
1154    3. I changed the way CM's tool mechanism determines the shell command
1155       string used for things like ml-yacc etc. so that it does not break
1156       when CM.Control.implicit_anchors is turned off.
1157    
1158    ----------------------------------------------------------------------
1159    Name: Matthias Blume
1160    Date: 2000/05/12 18:20:00 JST
1161    Tag: blume-20000512-ml-build
1162    Description:
1163    
1164    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1165    getting installed properly (by config/install.sh).
1166    
1167    ----------------------------------------------------------------------
1168    Name: Matthias Blume
1169    Date: 2000/05/12 17:30:00 JST
1170    Tag: blume-20000512-anchors
1171    Description:
1172    
1173    !!! NEW BOOT FILES !!!
1174    
1175    This change is in preparation of fading out support for "implicitly
1176    anchored path names".  I went through all sources and used the
1177    explicit (and relatively new) $-notation.  See system/README and the
1178    CM manual for more info on this.
1179    
1180    I also modified the anchoring scheme for some things such as "smlnj",
1181    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1182    anchors are more expressive: anchor name and first arc do not have to
1183    coincide.  This entails the following user-visible change:
1184    
1185    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1186    particular, when you fire up sml with a command-line argument, say,
1187    e.g.:
1188    
1189       sml '$smlnj/cmb.cm'
1190    
1191    At the ML toplevel prompt:
1192    
1193       CM.autoload "$smlnj/cmb.cm";
1194    
1195    There is also a new controller in CM.Control that can be used to turn
1196    off all remaining support for implicit anchors by saying:
1197    
1198        CM.autoload "$smlnj/
1199        #set CM.Control.implicit_anchors false;
1200    
1201    This causes CM to reject implicitly anchored paths.  This is (for the
1202    time being) less permissive than the "final" version where there will
1203    be no more such implicit anchors and relative paths will be just that:
1204    relative.
1205    
1206    The next step (version after next version?) will be to make the
1207    default for CM.Control.implicit_anchors false.  After the dust has
1208    settled, I can then produce the "final" version of this...
1209    
1210    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1211    
1212    ----------------------------------------------------------------------
1213    Name: Matthias Blume
1214    Date: 2000/05/11 16:30:00 JST
1215    Tag: blume-20000511-sources
1216    Description:
1217    
1218    The main change is that I added function CM.sources as a generalized
1219    version of the earlier CM.makedepend.  This entails the following
1220    additional changes:
1221    
1222      - CM.makedepend has been dropped.
1223    
1224      - CM manual has been updated.
1225    
1226      - TOOLS signature and API have been changed.
1227    
1228    ----------------------------------------------------------------------
1229    Name: Allen Leung
1230    Date: 2000/05/10 21:17 EDT
1231    Tag: leunga-20000510-moby-c--ssa
1232    Description:
1233    
1234      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1235    None of these affect SML/NJ.
1236    
1237    1. Register Allocation
1238    
1239        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1240           This module tries to remove local (i.e. basic block level) redundancies
1241           during spilling.
1242    
1243        b. A new framework for performing region based register allocation.
1244           Not yet entirely functional.
1245    
1246    2. X86
1247    
1248       a. DefUse for POP was missing the stack pointer [found by Lal]
1249       b. Reload for CALL was incorrect in X86Spill [found by John]
1250       c. Various fixes in X86Spill so that it can be used correctly for
1251          the new spilling module.
1252    
1253    3. SSA/IR
1254    
1255       a. New module ir/dj-dataflow.sml implements elimination based
1256          data flow analysis.
1257    
1258    4. MLRiscGen
1259    
1260       a. Fix for gc type annotation
1261    
1262    5. MDGen
1263    
1264       Various fixes for machine description -> ml code translation.  For ssa
1265       only.
1266    
1267    ----------------------------------------------------------------------
1268    Name: Allen Leung
1269    Date: 2000/05/08 22:17 EDT
1270    Tag: leunga-20000508-labexp
1271    Description:
1272    
1273      Fermin has found a few assembly problems with constant expressions
1274      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1275      which choke on dumb assemblers.  This is his fix.
1276    
1277    ----------------------------------------------------------------------
1278    Name: Dave MacQueen
1279    Date: 2000/04/09 14:00 EDT
1280    Tag: dbm-20000502-Version_110_28
1281    Description:
1282    
1283    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1284    
1285    2. Updated config/version to 110.28
1286    
1287    3. Updated config/srcarchiveurl
1288    
1289    3. New boot files!
1290       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1291    
1292    ----------------------------------------------------------------------
1293    Name: Matthias Blume
1294    Date: 2000/05/01 19:05:00 JST
1295    Tag: blume-20000501-noweb
1296    Description:
1297    
1298    A new noweb tool has been added.  The existing system is entirely
1299    unaffected by this, but some CM users have asked for renewed noweb
1300    support.  Everything is documented in the CM manual.
1301    
1302    New (plugin) libraries:
1303    
1304       noweb-tool.cm
1305       nw-ext.cm
1306    
1307    ----------------------------------------------------------------------
1308    Name: Dave MacQueen
1309    Date: 2000/04/30 12:40PM EDT
1310    Tag: dbm-20000430-bug_fixes
1311    Description:
1312    
1313    1. Fix for bug 1498
1314       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1315       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1316         added toRealArray function
1317       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1318         added check for tag Obj.RealArray to array printing case in ppObj
1319    
1320    2. Fix for bug 1510
1321       smlnj/src/compiler/Semant/types/typesutil.sml
1322         fixed definition of dummyargs (used by equalTycon) so that
1323         dummy args are distinct types
1324    
1325    ----------------------------------------------------------------------
1326    Name: Matthias Blume
1327    Date: 2000/04/30 01:00:00 JST
1328    Tag: blume-20000430-versions
1329    Description:
1330    
1331    1. CM version numbering added.  This is an implementation of Lal's
1332       proposal for adding version numbers and version checking to .cm
1333       files.  Lal said that his proposal was just that -- a proposal.
1334       For the time being I went ahead and implemented it so that people
1335       can comment on it.  Everything is completely backward-compatible
1336       (except for the stable library format, i.e., new bootfiles!).
1337    
1338       As usual, see the CM manual for details.
1339    
1340    2. An alternative syntax for anchored paths has been implemented.
1341       Dave has recently voiced the same concerns that I had when I did
1342       this, so there should be some support.  My take is that eventually
1343       I will let support for the current syntax (where anchors are
1344       "implicit") fade out in favor of the new, explicit syntax.
1345       In order to be backward-compatible, both old and new syntax are
1346       currently supported.
1347    
1348       Again, see the CM manual for details.
1349    
1350    3. Parallel make is trying to be slightly smarter:  When the master
1351       process finds a "bottleneck", i.e., when there is only one
1352       compilation unit that can be compiled and everybody else is
1353       waiting on it, then it will simply compile it directly instead
1354       of clumsily telling one of the slaves to do it.
1355    
1356    4. Support for "unsharing" added.  This is necessary in order to be
1357       able to have two different versions of the same library running
1358       at the same time (e.g., for trying out a new MLRISC while still
1359       having the old MLRISC linked into the current compiler, etc.)
1360       See the CM manual.
1361    
1362    5. Simple "makedepend" functionality added for generating Makefile
1363       dependency information.  (This is rather crude at the moment.
1364       Expect some changes here in the future.)
1365    
1366    6. ".fun" added as a recognized suffix for ML files. Also documented
1367       explicitly in the manual that the fallback behavior (unknown suffix
1368       -> ML file) is not an official feature!
1369    
1370    7. Small changes to the pickler for stable libraries.
1371    
1372    8. Several internal changes to CM (for cleanup/improvement).
1373    
1374    
1375    !!!! NEW BINFILES !!!!
1376    
1377    ----------------------------------------------------------------------
1378    Name: Matthias Blume
1379    Date: 2000/04/28 17:30:00 JST
1380    Tag: blume-20000428-pathconfig
1381    Description:
1382    
1383    1. I changed config/install.sh to remove duplicate entries from the
1384       lib/pathconfig file at the end.  Moreover, the final version of
1385       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1386       in src/system/installml.
1387    
1388    2. The config/install.sh script now consistently uses relative
1389       pathnames in lib/pathconfig whenever the anchor is in the lib
1390       directory.  (So far this was true for the libraries that come
1391       pre-compiled and bundled as part of the bootfiles but not for
1392       libraries that are compiled by the script itself.)
1393    
1394    ----------------------------------------------------------------------
1395    Name: Matthias Blume
1396    Date: 2000/04/26 13:10:00 JST
1397    Tag: blume-20000426-fun_suffix
1398    Description:
1399    
1400    Added ".fun" as a recognized file name suffix (for ML code).
1401    
1402    ----------------------------------------------------------------------
1403    Name: Allen Leung
1404    Date: 2000/04/25 17:00:00 EST
1405    Tag: leunga-20000425-alpha-ra
1406    Description:
1407    
1408    1. Alpha
1409    
1410        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1411    in C--.
1412    
1413    2. RA
1414    
1415       Added a flag "ra-dump-size" to print out the size of the flowgraph
1416       and the interference graph.
1417    
1418    ----------------------------------------------------------------------
1419    Name: Dave MacQueen
1420    Date: 2000/04/25/
1421    Tag: dbm-20000425-mlyacc_doc_examples
1422    Description:
1423      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1424      Updated all three examples in src/ml-yacc/examples to run
1425      under 110.* using CM.make.
1426    
1427    ----------------------------------------------------------------------
1428    Name: Allen Leung
1429    Date: 2000/04/20 23:04:00 EST
1430    Tag: leunga-20000420-ssa-c---stuff
1431    Description:
1432    
1433      This update synchronizes my repository with Yale's.  Most of these
1434    changes, however, do not affect SML/NJ at all (the RA is an exception).
1435    
1436    1. Register Allocator
1437    
1438       a. An improvement in the interference graph construction:
1439          Given a copy
1440    
1441                s <- t
1442    
1443          no interference edge between s and t is added for this definition of s.
1444    
1445       b. I've added two new spill heuristic modules that Fermin and I developed
1446          (in the new library RA.cm). These are unused in SML/NJ but maybe
1447          useful for others (Moby?)
1448    
1449    2. X86
1450    
1451       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1452    
1453    3. Alpha
1454    
1455       a. Added the BSR instruction and code generation that goes with it [C--]
1456       b. Other fixes too numerous to recount provided by Fermin [C--]
1457    
1458    4. Regmaps
1459    
1460       a. The regmaps are not initialized with the identity physical bindings
1461          at creation time.  This is unneeded.
1462    
1463    5. MLRISC Optimizations
1464    
1465       a. The DJ-Graph module can now compute the iterated dominance frontiers
1466          intersects with liveness incrementally in linear time! Woohoo!
1467          This is now used in my new SSA construction algorithm.
1468    
1469       b. THe branch reorganization module is now smarter about linear chains of
1470          basic blocks.
1471    
1472    
1473    ----------------------------------------------------------------------
1474    Name: Matthias Blume
1475    Date: 2000/04/12 13:52:00 JST
1476    Tag: blume_main_v110p27_1
1477    Description:
1478    
1479    Changed install.sh script to handle archive files without version number
1480    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1481    name of the boot file archive.
1482    
1483    ----------------------------------------------------------------------
1484    Name: Dave MacQueen
1485    Date: 2000/04/09 14:00 EDT
1486    Tag: dbm-20000410-Version_110_27
1487    Description:
1488    
1489    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1490    
1491    2. Updated src/config/version to 110.27
1492    
1493    3. New boot files!
1494    
1495    ----------------------------------------------------------------------
1496    Name: Allen Leung
1497    Date: 2000/04/09 19:09:00 EST
1498    Tag: leunga-20000409-misc
1499    Description:
1500    
1501    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1502    
1503    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1504    
1505    ----------------------------------------------------------------------
1506  Name: Stefan  Name: Stefan
1507  Date: 2000/04/07 10:00:00 EDT  Date: 2000/04/07 10:00:00 EDT
1508  Tag: monnier-20000406-branch-handling  Tag: monnier-20000406-branch-handling
# Line 97  Line 1588 
1588    
1589  3.  Assembly  3.  Assembly
1590    
1591      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
1592      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
1593      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
1594    
# Line 120  Line 1611 
1611    
1612  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
1613  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
1614  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
1615  legal surface-syntax symbol).  legal surface-syntax symbol).
1616    
1617  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 130  Line 1621 
1621  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
1622    
1623  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1624  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
1625  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
1626  more details.  more details.
1627    
# Line 245  Line 1736 
1736     (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
1737     used).     used).
1738    
1739     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1740     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1741     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1742    
1743     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 258  Line 1749 
1749    
1750     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1751    
1752     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1753     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1754    
1755  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 392  Line 1883 
1883    
1884       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
1885       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
1886       testl is perferable to andl.       testl is preferable to andl.
1887    
1888  7. RA (x86 only)  7. RA (x86 only)
1889    
# Line 574  Line 2065 
2065    
2066  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2067  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2068  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
2069  explanation.  explanation.
2070    
2071  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2072  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2073  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2074  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 621  Line 2112 
2112  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
2113  <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,
2114  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
2115  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2116  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2117  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2118    
# Line 671  Line 2162 
2162    it from that remote directory.    it from that remote directory.
2163    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2164    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2165    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2166    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2167    
2168    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.605  
changed lines
  Added in v.757

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