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 602, Thu Apr 6 05:31:24 2000 UTC revision 781, Tue Jan 30 22:48:06 2001 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: 2001/01/30 17:50:00 EST
17    Tag: blume-20010130-sync
18    Description:
19    
20    This is just a minor update to sync my devel branch with the main brach.
21    The only visible change is the addition of some README files.
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume
25    Date: 2001/01/12 23:30:00 JST
26    Tag: blume-20010112-bootfiles
27    Description:
28    
29    Made a new set of bootfiles that goes with the current state of the
30    repository.
31    
32    ----------------------------------------------------------------------
33    Name: Matthias Blume
34    Date: 2001/01/12 21:20:00 JST
35    Tag: blume-20010112-sync
36    Description:
37    
38    I am just flushing out some minor changes that had accumulated in
39    my private branch in order to sync with the main tree.  (This is
40    mainly because I had CVS trouble when trying to merge _into_ my
41    private branch.)
42    
43    Most people should be completely unaffected by this.
44    
45    ----------------------------------------------------------------------
46    Name: Allen Leung
47    Date: Thu Jan 11 21:03:00 EST 2001
48    Tag: leunga-20010111-labexp=mltree
49    Description:
50    
51    1.  Removed the type LabelExp and replace it by MLTree.
52    2.  Rewritten mltree-simplify with the pattern matcher tool.
53    3.  There were some bugs in alpha code generator which would break
54        64-bit code generation.
55    4.  Redo the tools to generate code with the
56    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
57        are now generated by perl scripts.
58    
59    ----------------------------------------------------------------------
60    Name: Matthias Blume
61    Date: 2001/01/10 21:55:00 JST
62    Tag: blume-20010110-rcc
63    Description:
64    
65    The RCC stuff now seems to work (but only on the x86).
66    This required hacking of the c-calls interface (and -implementation) in
67    MLRISC.
68    
69    Normal compiler users should be unaffected.
70    
71    ----------------------------------------------------------------------
72    Name: Matthias Blume
73    Date: 2001/01/09 01:20:00 JST
74    Tag: blume-20010109-rcc
75    Description:
76    
77    This is a fairly big patch, flushing out a large number of pending
78    changes that I made to my development copy over the last couple of days.
79    
80    Of practical relevance at this moment is a workaround for a pickling
81    bug that Allen ran into the other day.  The cause of the bug itself is
82    still unknown and it might be hard to fix it properly, but the
83    workaround has some merits of its own (namely somewhat reducing pickling
84    overhead for certain libraries).  Therefore, I think this solution should
85    be satisfactory at this time.
86    
87    The rest of the changes (i.e., the vast majority) has to do with my
88    ongoing efforts of providing direct support for C function calls from
89    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
90    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
91    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
92    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
93    handle RCC.
94    
95    The last part (the changes to mlriscGen.sml) are still known to be
96    wrong on the x86 and not implemented on all other architectures.  But
97    the infrastructure is in place. I had to change a few functor
98    signatures in the backend to be able to route the CCalls interface
99    from MLRISC there, and I had to specialize the mltree type (on the
100    x86) to include the necessary extensions. (The extensions themselves
101    were already there and redy to go in MLRISC/x86).
102    
103    Everything should be very happy as soon as someone helps me with
104    mlriscGen.sml...
105    
106    In any case, nothing of this should matter to anyone as long as the
107    new primop is not being used (which is going to be the case unless you
108    find it where I hid it :). The rest of the compiler is completely
109    unaffected.
110    
111    ----------------------------------------------------------------------
112    Name: Matthias Blume
113    Date: 2001/01/05 00:30:00 JST
114    Tag: blume-20010105-primops
115    Description:
116    
117    Added some experimental support for work that I am doing right now.
118    These changes mostly concern added primops, but there is also a new
119    experimental C library in the runtime system (but currently not enabled
120    anywhere except on Linux/X86).
121    
122    In the course of adding primops (and playing with them), I discovered that
123    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
124    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
125    To recover, I made type information in INL_PRIM mandatory and changed
126    prim.sml as well as built-in.sml accordingly.  The InLine structure now
127    has complete, correct type information (i.e., no bottom types).
128    
129    Since all these changes mean that we need new binfiles, I also bumped the
130    version number to 110.32.1.
131    
132    ----------------------------------------------------------------------
133    Name: Matthias Blume
134    Date: 2000/12/30 22:10:00 JST
135    Tag: blume-20001230-various
136    Description:
137    
138    Added proxy libraries for MLRISC and let MLRISC libraries refer
139    to each other using path anchors.  (See CM manual for explanation.)
140    
141    Updated CM documentation.
142    
143    Fixed some bugs in CM.
144    
145    Implemented "proxy" libraries (= syntactic sugar for CM).
146    
147    Added "-quiet" option to makeml and changed runtime system accordingly.
148    
149    Added cleanup handler for exportML to reset timers and compiler stats.
150    
151    ----------------------------------------------------------------------
152    Name: Lal George
153    Date: 2000/12/22 22:22:58 EST 2000
154    Tag: Release_110_32
155    Description:
156    
157            Infinite precision used throughout MLRISC.
158            see MLRISC/mltree/machine-int.sig
159    
160    ----------------------------------------------------------------------
161    Name: Matthias Blume
162    Date: 2000/12/22 23:16:00 JST
163    Tag: blume-20001222-warn
164    Description:
165    
166    Corrected wording and formatting of some CM warning message which I
167    broke in my previous patch.
168    
169    ----------------------------------------------------------------------
170    Name: Matthias Blume
171    Date: 2000/12/22 21:20:00 JST
172    Tag: blume-20001222-anchorenv
173    Description:
174    
175    Fixed CM's handling of anchor environments in connection with CMB.make.
176    
177    ----------------------------------------------------------------------
178    Name: Matthias Blume
179    Date: 2000/12/22 13:15:00 JST
180    Tag: blume-20001222-cleanup
181    Description:
182    
183    Removed src/cm/ffi which does not (and did not) belong here.
184    
185    ----------------------------------------------------------------------
186    Name: Matthias Blume
187    Date: 2000/12/21 23:55:00 JST
188    Tag: blume-20001221-exn
189    Description:
190    
191    Probably most important: CM no longer silently swallows all exceptions
192    in the compiler.
193    Plus: some other minor CM changes.  For example, CM now reports some
194    sizes for generated binfiles (code, data, envpickle, lambdapickle).
195    
196    ----------------------------------------------------------------------
197    Name: Matthias Blume
198    Date: 2000/12/15 00:01:05 JST
199    Tag: blume-20001215-dirtool
200    Description:
201    
202    - "dir" tool added.
203    - improvements and cleanup to Tools structure
204    - documentation updates
205    
206    ----------------------------------------------------------------------
207    Name: Allen Leung
208    Date: Thu Dec 14 03:45:24 EST 2000
209    Description:
210    Tag:  leunga-20001214-int-inf
211    Description:
212    
213       In IntInf, added these standard functions, which are missing from our
214    implementation:
215    
216        andb : int * int -> int
217        xorb : int * int -> int
218        orb  : int * int -> int
219        notb : int -> int
220         <<   : int * word -> int
221        ~>>  : int * word -> int
222    
223       Not tested, I hope they are correct.
224    
225    ----------------------------------------------------------------------
226    Name: Allen Leung
227    Date: Fri Dec  8 19:23:26 EST 2000
228    Description:
229    Tag:  leunga-20001208-nowhere
230    Description:
231    
232      Slight improvements to the 'nowhere' tool to handle OR-patterns,
233    to generate better error messages etc.  Plus a brief manual.
234    
235    ----------------------------------------------------------------------
236    Name: Lal George
237    Date: 2000/12/08 09:54:02 EST 2000
238    Tag: Release_110_31
239    Description:
240    
241    - Version 110.31
242    ----------------------------------------------------------------------
243    Name: Allen Leung
244    Date: Thu Dec  7 22:01:04 EST 2000
245    Tag:  leunga-20001207-cell-monster-hack
246    Description:
247    
248    Major MLRISC internal changes.  Affect all clients.
249    Summary:
250    
251    1.  Type CELLS.cell = int is now replaced by a datatype.
252        As a result, the old regmap is now gone.  Almost all interfaces
253        in MLRISC change as a consequence.
254    
255    2.  A new brand version of machine description tool (v3.0) that generates
256        modules expecting the new interface.  The old version is removed.
257    
258    3.  The RA interface has been further abstracted into two new functors.
259        RISC_RA and X86RA.  These functors have much simpler interfaces.
260        [See also directory MLRISC/demo.]
261    
262    4.  Some other new source->source code generation tools are available:
263    
264        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
265        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
266           I use this tool to generate the peephole optimizers---with the new
267           cell type changes, peephole rules are becoming difficult to write
268           without conditional pattern matching.
269    
270    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
271        cover the entire MLRISC source tree so many things broke.
272    
273    6.  CM files have been moved to the subdirectory MLRISC/cm.
274        They are moved because there are a lot of them and they clutter up the
275        root dir.
276    
277    7.  More detailed documentation to come...
278    
279        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
280        a makeml -rebuild first.  This is because of other other
281        changes that Matthias has made (see below).
282    
283    
284    ----------------------------------------------------------------------
285    Name: Matthias Blume
286    Date: 2000/11/30 23:12:00 JST
287    Tag: blume-20001130-filereorg
288    Description:
289    
290    Some manual updates and some file reorganizations in CM.
291    
292    ----------------------------------------------------------------------
293    Name: Matthias Blume
294    Date: 2000/11/24 17:45:00 JST
295    Tag: blume-20001124-link
296    Description:
297    
298    Drastically improved link traversal code for the case that the dynamic
299    value was already loaded at bootstrap time.  As a result, CM and CMB
300    now both load blazingly fast -- even on a very slow machine.  Also,
301    memory consumption has been further reduced by this.
302    
303    Warning: The format of the PIDMAP file has changed.  THerefore, to
304    bootstrap you have to do this:
305    
306    1. Run CMB.make
307    2. Make a symbolic link for the boot directory:
308         ln -s sml.boot.ARCH-OS xxx
309    3. "Rebuild" the boot directory:
310         ./makeml -boot xxx -rebuild sml ; rm xxx
311    4. Boot normally:
312          ./makeml
313    
314    ----------------------------------------------------------------------
315    Name: Matthias Blume
316    Date: 2000/11/21 21:20:00 JST
317    Tag: blume-20001121-tools
318    Description:
319    
320    Continued hacking on autoloading problem -- with success this time.
321    Also changed tool-plugin mechanism.  See new CM manual.
322    
323    ----------------------------------------------------------------------
324    Name: Matthias Blume
325    Date: 2000/11/19 14:30:00 JST
326    Tag:  blume-20001119-autoload
327    Description:
328    
329    Some hacking to make autoloading faster.  Success for CMB, no success
330    so far for CM.  There is a reduced structure CM' that autoloads faster.
331    (This is a temporary, non-documented hack to be eliminated again when
332    the general problem is solved.)
333    
334    ----------------------------------------------------------------------
335    Name: Matthias Blume
336    Date: 2000/11/17 14:10:00 JST
337    Tag: blume-20001117-pickle-lib
338    Description:
339    
340    1. Eliminated comp-lib.cm
341    2. Made pickle-lib.cm
342    3. Eliminated all uses of intset.sml (from comp-lib.cm)
343    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
344       equivalent constructs from smlnj-lib.cm (INtHashTable).
345    5. Point 4. also goes for those uses of intmap.* in MLRISC.
346       Duplicated intmap modules thrown out.
347    6. Hunted down all duplicated SCC code and replaced it with
348       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
349    7. Rewrote Feedback module.
350    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
351       should be thrown out and equivalent modules from smlnj-lib.cm
352       should be used (IntRedBlackSet, IntListSet, ...).
353    
354    Confirmed that compiler compiles to fixpoint.
355    
356    ----------------------------------------------------------------------
357    Name: Allen Leung
358    Date: 2000/11/10 18:00:00
359    Tag: leunga-20001110-new-x86-fp
360    
361    A new x86 floating point code generator has been added.
362    By default this is turned off.  To turn this on, do:
363    
364        CM.autoload "$smlnj/compiler.cm";
365        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
366    
367    Changes:
368    
369    1.  Changed FTAN to FPTAN so that the assembly output is correct.
370    2.  Changed the extension callback for FTANGENT to generate:
371    
372              fptan
373              fstp  %st(0)
374        instead of
375              fptan
376              fstpl ftempmem
377    
378    3.  Numerous assembly fixes for x86.
379    
380    5.  Cleaned up the machine code output module x86/x86MC.sml and added
381        support for a whole bunch of instructions and addressing modes:
382    
383          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
384          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
385          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
386          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
387          fxch %st(n)
388          fld %st(n)
389          fst %st(n)
390          fst mem
391          fstp %st(n)
392          fucom %st(n)
393          fucomp %st(n)
394    
395        All these are now generated when the fast fp mode is turned on.
396    
397    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
398    
399    ----------------------------------------------------------------------
400    Name: Matthias Blume
401    Date: 2000/11/09 11:20:00 JST
402    Tag: blume-20001109-scc
403    Description:
404    
405    Eliminated some code duplication:
406    
407    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
408       (Otherwise the functor is useless.)
409    2. Used GraphSCCFn where SCCUtilFun was used previously.
410    3. Got rid of SCCUtilFun (in comp-lib.cm).
411    
412    ----------------------------------------------------------------------
413    Name: Lal George
414    Date: 2000/11/06 09:02:21 EST 2000
415    Tag: Release_110_30
416    Description:
417    
418    - Version 110.30
419    ----------------------------------------------------------------------
420    Name: Matthias Blume
421    Date: 2000/11/04 14:45:00
422    Tag: blume-20001104-mlbuild
423    Description:
424    
425    - Made ml-build faster on startup.
426    - Documentation fixes.
427    
428    ----------------------------------------------------------------------
429    Name: Matthias Blume
430    Date: 2000/11/02 17:00:00 JST
431    Tag: blume-20001102-condcomp
432    Description:
433    
434    - Small tweaks to pickler -- new BOOTFILES!
435    - Version bumped to 110.29.2.
436    - Added conditional compilation facility to init.cmi (see comment there).
437    ----------------------------------------------------------------------
438    Name: Allen Leung
439    Date: 2000/10/23 19:31:00
440    Tag: leunga-20001023-demo-ra
441    
442    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
443    2. Test programs for the graph library updated
444    3. Some new MLRISC demo programs added
445    
446    ----------------------------------------------------------------------
447    Name: Matthias Blume
448    Date: 2000/08/31 22:15:00 JST
449    Tag: blume-20001017-errmsg
450    Description:
451    
452    More error message grief: Where there used to be no messages, there
453    now were some that had bogus error regions.  Fixed.
454    
455    ----------------------------------------------------------------------
456    Name: Matthias Blume
457    Date: 2000/08/31 17:30:00 JST
458    Tag: blume-20001017-v110p29p1
459    Description:
460    
461    I made a version 110.29.1 with new bootfiles.
462    
463    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
464              CM documentation changes and a small bugfix in CM's error reporting.
465    
466    ----------------------------------------------------------------------
467    Name: Lal George
468    Date: 2000/09/27 14:42:35 EDT
469    Tag: george-20000927-nodestatus
470    Description:
471    
472    Changed the type of the nodestatus, so that:
473    
474            SPILLED(~1)             is now SPILLED
475            SPILLED(m) where m>=0   is now MEMREG(m)
476            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
477    
478    ----------------------------------------------------------------------
479    Name: Matthias Blume
480    Date: 2000/09/07 14:45:00 JST
481    Tag: blume-20000907-cmerrmsg
482    Description:
483    
484    Small tweak to CM to avoid getting ML syntax error messages twice.
485    
486    ----------------------------------------------------------------------
487    Name: Matthias Blume
488    Date: 2000/08/31 18:00:00 JST
489    Tag: blume-20000831-cvsbootfiles
490    Description:
491    
492    New URL for boot files (because the 110.29 files on the BL server do
493    now work correctly with my updated install scripts for yacc and lex).
494    
495    ----------------------------------------------------------------------
496    Name: Matthias Blume
497    Date: 2000/08/08 12:33:00 JST
498    Tag: blume-20000808-manual
499    Description:
500    
501    Tiny update to CM manual.
502    
503    ----------------------------------------------------------------------
504    Name: Allen Leung
505    Date: 2000/08/7 19:31:00
506    Tag: leunga-20000807-a-whole-bunch-of-stuff
507    
508      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
509    and Mobdy.
510    
511    1.  x86
512    
513       a.  Fixes to peephole module by John and Dan.
514       b.  Assembly fix to SETcc by Allen.
515       c.  Fix to c-call by John.
516       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
517       e.  Instruction selection optimization to SETcc as suggested by John.
518    
519           For example,
520    
521            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
522    
523           should generate:
524    
525            MOVL a, x
526            SUBL b, x
527            SHRL 31, x
528    
529    2.  IR stuff
530    
531         A bunch of new DJ-graph related algorithms added.  These
532         speed up SSA construction.
533    
534    3.  SSA + Scheduling
535    
536         Added code for SSA and scheduling to the repository
537    
538    ----------------------------------------------------------------------
539    Name: Lal George
540    Date: 2000/07/27 11:53:14 EDT
541    
542    Tag: lal-20000727-linux-ppc
543    Description:
544    
545     Made changes to support Linux PPC.
546     p.s. I have confirmation that the 110.29 boot files work fine.
547    
548    ----------------------------------------------------------------------
549    Name: Matthias Blume
550    Date: 2000/07/27 17:40:00 JST
551    Tag: blume-20000727-scripts
552    Description:
553    
554    !!!! WARNING !!!!
555    You must recompile the runtime system!
556    !!!! WARNING !!!!
557    
558    This is basically another round of script-enhancements:
559    
560    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
561       and undefine CM preprocessor symbols.
562    
563    2. ml-build avoids generating a new heap image if it finds that the
564       existing one is still ok.  (The condition is that no ML file had to
565       be recompiled and all ML files are found to be older that the heap
566       file.)
567    
568       To make this work smoothly, I also hacked the runtime system as
569       well as SMLofNJ.SysInfo to get access to the heap image suffix
570       (.sparc-solaris, ...) that is currently being used.
571    
572       Moreover, the signature of CM.mk_standalone has changed.  See the
573       CM manual.
574    
575    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
576       CM manual for details.)
577    
578    4. More CM manual updates:
579        - all of the above has been documented.
580        - there is now a section describing the (CM-related) command line
581          arguments that are accepted by the "sml" command
582    
583    ----------------------------------------------------------------------
584    Name: Matthias Blume
585    Date: 2000/07/25 16:20:00 JST
586    Tag: blume-20000725-makedepend
587    Description:
588    
589    Added a script called ml-makedepend.  This can be used in makefiles
590    for Unix' make in a way very similar to the "makedepend" command for
591    C.
592    
593    The script internally uses function CM.sources.
594    
595    Synopsis:
596    
597        ml-makedepend [-f makefile] cmfile targetname
598    
599    The default for the makefile is "makefile" (or "Makefile" should
600    "makefile" not exist).
601    
602    ml-makedepend adds a cmfile/targetname-specific section to this
603    makefile (after removing the previous version of this section).  The
604    section contains a single dependency specification with targetname on
605    the LHS (targetname is an arbitrary name), and a list of files derived
606    from the cmfile on the RHS.  Some of the files on the RHS are
607    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
608    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
609    path names.  The makefile writer is responsible for making sure that
610    these variables have correct at the time "make" is invoked.
611    
612    ----------------------------------------------------------------------
613    Name: Matthias Blume
614    Date: 2000/07/22 23:30:00 JST
615    Tag: blume-20000722-urlupdate
616    Description:
617    
618    Changed BOOT and config/srcarchiveurl to point to BL server:
619    
620        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
621    
622    ----------------------------------------------------------------------
623    Name: Matthias Blume
624    Date: 2000/07/18 18:00:00 JST
625    Tag: blume-20000718-Version_110_29
626    Description:
627    
628    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
629    
630    2. Updated config/version to 110.29
631    
632    3. Updated config/srcarchiveurl
633    
634    3. New boot files!
635       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
636    
637    ----------------------------------------------------------------------
638    Name: Matthias Blume
639    Date: 2000/07/11 13:58:00 JST
640    Tag: blume-20000711-doctypo
641    Description:
642    
643    Fixed a few typos in CM manual.
644    
645    ----------------------------------------------------------------------
646    Name: Allen Leung
647    Date: 2000/06/15 00:38:00
648    Tag: leunga-20000704-sparc-x86
649    
650    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
651    2. fix to x86 RET bug [found by Dan Grossman]
652    3. sparc assembly bug fix for ticc instructions [found by Fermin]
653    
654       Affects c-- and moby only
655    
656    ----------------------------------------------------------------------
657    Name: Matthias Blume
658    Date: 2000/07/04 15:26:00
659    Tag: blume-20000704-trigger
660    Description:
661    
662    1. Improvements to CM manual.
663    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
664       of getting a back-trace.  The function, when called, raises an
665       internal exception which explicitly carries the full back-trace history,
666       so it is unaffected by any intervening handle-raise pairs ("trivial"
667       or not).  The interactive loop will print that history once it arrives
668       at top level.
669       Short of having all exceptions implicitly carry the full history, the
670       recommended way of using this facility is:
671         - compile your program with instrumentation "on"
672         - run it, when it raises an exception, look at the history
673         - if the history is "cut off" because of some handler, go and modify
674           your program so that it explicitly calls BTrace.trigger
675         - recompile (still instrumented), and rerun; look at the full history
676    
677    ----------------------------------------------------------------------
678    Name: Matthias Blume
679    Date: 2000/07/03 15:36:00 JST
680    Tag: blume-20000702-manual
681    Description:
682    
683    Small corrections and updates to CM manual.
684    
685    ----------------------------------------------------------------------
686    Name: Matthias Blume
687    Date: 2000/06/29 16:04:00 JST
688    Tag: blume-20000629-yacctool
689    Description:
690    
691    Changes:
692    
693    1. Class "mlyacc" now takes separate arguments to pass options to
694       generated .sml- and .sig-files independently.
695    2. Corresponding CM manual updates.
696    3. BTrace module now also reports call sites.  (However, for loop clusters
697       it only shows from where the cluster was entered.)  There are associated
698       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
699    
700    ----------------------------------------------------------------------
701    Name: Matthias Blume
702    Date: 2000/06/27 16:51:00 JST
703    Tag: blume-20000627-noweb
704    Description:
705    
706    Changes:
707    
708     1. Implemented "subdir" and "witness" options for noweb tool.
709        This caused some slight internal changes in CM's tool implementation.
710     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
711        some remaining issues from earlier path anchor changes.
712     3. Updated CM manual accordingly.
713    
714     4. Changed implementation of back-tracing so that I now consider it
715        ready for prime-time.
716    
717        In particular, you don't have to explicitly trigger the back-trace
718        anymore.  Instead, if you are running BTrace-instrumented code and
719        there is an uncaught exception (regardless of whether or not it was
720        raised in instrumented code), the top-level evalloop will print
721        the back-trace.
722    
723        Features:
724    
725          - Instrumented and uninstrumented code work together seemlessly.
726            (Of course, uninstrumented code is never mentioned in actual
727             back-traces.)
728    
729          - Asymptotic time- and space-complexity of instrumented code is
730            equal to that of uninstrumented code.  (This means that
731            tail-recursion is preserved by the instrumentation phase.)
732    
733          - Modules whose code has been instrumented in different sessions
734            work together without problem.
735    
736          - There is no penalty whatsoever on uninstrumented code.
737    
738          - There is no penalty on "raise" expressions, even in
739            instrumented code.
740    
741        A potential bug (or perhaps it is a feature, too):
742    
743          A back-trace reaches no further than the outermost instrumented
744          non-trivial "raise".  Here, a "trivial" raise is one that is the
745          sole RHS of a "handle" rule.  Thus, back-traces reach trough
746    
747               <exp> handle e => raise e
748    
749          and even
750    
751               <exp> handle Foo => raise Bar
752    
753          and, of course, through
754    
755               <exp> handle Foo => ...
756    
757         if the exception was not Foo.
758    
759         Back-traces always reach right through any un-instrumented code
760         including any of its "handle" expressions, trivial or not.
761    
762       To try this out, do the following:
763    
764         - Erase all existing binfiles for your program.
765           (You may keep binfiles for those modules where you think you
766            definitely don't need back-tracing.)
767         - Turn on back-trace instrumentation:
768              SMLofNJ.Internals.BTrace.mode (SOME true);
769         - Recompile your program.  (I.e., run "CM.make" or "use".)
770         - You may now turn instrumentation off again (if you want):
771              SMLofNJ.Internals.BTrace.mode (SOME false);
772         - Run your program as usual.  If it raises an exception that
773           reaches the interactive toplevel, then a back-trace will
774           automatically be printed.  After that, the toplevel loop
775           will print the exception history as usual.
776    
777    ----------------------------------------------------------------------
778    Name: Matthias Blume
779    Date: 2000/06/26 09:56:46 JST
780    Tag: blume-20000626-setup
781    Description:
782    
783    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
784          ML code before and after compiling a file (e.g., to set compiler
785          flags)
786    
787    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
788              - associated changes to btrace.sml (BTrace instrumentation pass)
789              - cleaner implementation of btimp.sml (BTrace tracing and report
790                module)
791    
792    CM manual: * new path encoding documented
793               * description of setup-parameter to "sml" added
794    
795    The biggest user-visible change to back-tracing is that it is no
796    longer necessary to compile all traced modules within the same
797    session.  (This was a real limitation.)
798    
799    ----------------------------------------------------------------------
800    Name: Matthias Blume
801    Date: 2000/06/24 12:40:00 JST
802    Tag: blume-20000624-startup
803    Description:
804    
805    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
806    bit too often -- to put it mildly. :)
807    
808    ----------------------------------------------------------------------
809    Name: Matthias Blume
810    Date: 2000/06/23 18:20:00 JST
811    Tag: blume-20000623-btrace
812    Description:
813    
814    This updates adds a backtrace facility to aid programmers in debugging
815    their programs.  This involves the following changes:
816    
817    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
818       keeping track of the current call stack.  When programs are compiled
819       in a special mode, the compiler will insert calls to these hooks
820       into the user program.
821       "Hook" means that it is possible for different implementations of
822       back-tracing to register themselves (at different times).
823    
824    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
825       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
826       It can be turned on using this call:
827         SMLofNJ.Internals.BTrace.mode (SOME true);
828       Turning it off again:
829         SMLofNJ.Internals.BTrace.mode (SOME false);
830       Querying the current status:
831         SMLofNJ.Internals.BTrace.mode NONE;
832       Annotated programs are about twice as big as normal ones, and they
833       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
834       where all hooks do nothing).  The slowdown with a plugin that is
835       actually useful (such as the one supplied by default) is even greater,
836       but in the case of the default plugin it is still only an constant
837       factor (amortized).
838    
839    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
840       with a sub-structure BTrace for controlling back-tracing.  In particular,
841       the above-mentioned function "mode" controls whether the annotation
842       phase is invoked by the compiler.  Another important function is
843       "trigger": when called it aborts the current execution and causes
844       the top-level loop to print a full back-trace.
845    
846    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
847       for back-tracing.  It keeps track of the dynamic call stack and in
848       addition to that it keeps a partial history at each "level" of that
849       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
850       d and b (at separate times, dynamically), then the report will show:
851    
852       GOTO   d
853             /c
854       GOTO  \b
855       CALL   a
856    
857       This shows that there was an initial non-tail call of a, then a
858       tail-call to b or c, looping behavior in a cluster of functions that
859       consist of b and c, and then a goto from that cluster (i.e., either from
860       b or from c) to d.
861    
862       Note that (depending on the user program) the amount of information
863       that the back-trace module has to keep track of at each level is bounded
864       by a constant.  Thus, the whole implementation has the same asymptotical
865       complexity as the original program (both in space and in time).
866    
867    5. compiler/TopLevel/interact/evalloop.sml has been modified to
868       handle the special exception SMLofNJ.Internals.BTrace.BTrace
869       which is raised by the "trigger" function mentioned above.
870    
871    Notes on usage:
872    
873    - Annotated code works well together with unannotated code:
874    Unannotated calls simply do not show up at all in the backtrace.
875    
876    - It is not a good idea to let modules that were annotated during
877    different sessions run at the same time.  This is because the compiler
878    chooses small integers to identify individual functions, and there
879    will be clashes if different modules were compiled in separate sessions.
880    (Nothing will crash, and you will even be told about the clashes, but
881    back-trace information will in general not be useful.)
882    
883    - Back-tracing can be confused by callcc and capture.
884    
885    - The only way of getting a back-trace right now is to explicitly
886    invoke the "trigger" function from your user program.  Eventually, we
887    should make every exception carry back-trace information (if
888    available).  But since this creates more overhead at "raise"-time
889    (similar to the current exnHistory overhead), I have not yet
890    implemented this.  (The implementation will be rather easy.)  With
891    exceptions carrying back-trace information, this facility will be even
892    more useful because users don't need to modify their programs...
893    
894    - While it is possible to compile the compiler with back-trace
895    annotations turned on (I did it to get some confidence in
896    correctness), you must make absolutely sure that core.sml and
897    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
898    be compiled with annotation because there is no core access yet, but
899    if you compile btimp.sml with annotation, then the system will go into
900    an infinite recursion and crash.)
901    Since CM currently does not know about BTrace, the only way to turn
902    annotations on and off for different modules of the compiler is to
903    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
904    this is awkward and clumsy.
905    
906    Sample sessions:
907    
908    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
909    - SMLofNJ.Internals.BTrace.mode (SOME true);
910    [autoloading]
911    [autoloading done]
912    val it = false : bool
913    - structure X = struct
914    -     fun main n = let
915    -         fun a (x, 0) = d x
916    -           | a (x, n) = b (x, n - 1)
917    -         and b (x, n) = c (x, n)
918    -         and c (x, n) = a (x, n)
919    -         and d x = e (x, 3)
920    -         and e (x, 0) = f x
921    -           | e (x, n) = e (x, n - 1)
922    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
923    -           | f n = n * g (n - 1)
924    -         and g n = a (n, 3)
925    -     in
926    -         f n
927    -     end
928    - end;
929    structure X : sig val main : int -> int end
930    - X.main 3;
931    *** BACK-TRACE ***
932    GOTO   stdIn:4.2-13.20: X.main[2].f
933    GOTO-( stdIn:4.2-13.20: X.main[2].e
934    GOTO   stdIn:4.2-13.20: X.main[2].d
935         / stdIn:4.2-13.20: X.main[2].a
936         | stdIn:4.2-13.20: X.main[2].b
937    GOTO-\ stdIn:4.2-13.20: X.main[2].c
938    CALL   stdIn:4.2-13.20: X.main[2].g
939    GOTO   stdIn:4.2-13.20: X.main[2].f
940    GOTO-( stdIn:4.2-13.20: X.main[2].e
941    GOTO   stdIn:4.2-13.20: X.main[2].d
942         / stdIn:4.2-13.20: X.main[2].a
943         | stdIn:4.2-13.20: X.main[2].b
944    GOTO-\ stdIn:4.2-13.20: X.main[2].c
945    CALL   stdIn:4.2-13.20: X.main[2].g
946    GOTO   stdIn:4.2-13.20: X.main[2].f
947    GOTO-( stdIn:4.2-13.20: X.main[2].e
948    GOTO   stdIn:4.2-13.20: X.main[2].d
949         / stdIn:4.2-13.20: X.main[2].a
950         | stdIn:4.2-13.20: X.main[2].b
951    GOTO-\ stdIn:4.2-13.20: X.main[2].c
952    CALL   stdIn:4.2-13.20: X.main[2].g
953    GOTO   stdIn:4.2-13.20: X.main[2].f
954    CALL   stdIn:2.15-17.4: X.main[2]
955    -
956    
957    (Note that because of a FLINt bug the above code currently does not
958    compile without BTrace turned on.)
959    
960    Here is another example, using my modified Tiger compiler:
961    
962    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
963    - SMLofNJ.Internals.BTrace.mode (SOME true);
964    [autoloading]
965    [autoloading done]
966    val it = false : bool
967    - CM.make "sources.cm";
968    [autoloading]
969    ...
970    [autoloading done]
971    [scanning sources.cm]
972    [parsing (sources.cm):parse.sml]
973    [creating directory CM/SKEL ...]
974    [parsing (sources.cm):tiger.lex.sml]
975    ...
976    [wrote CM/sparc-unix/semant.sml]
977    [compiling (sources.cm):main.sml]
978    [wrote CM/sparc-unix/main.sml]
979    [New bindings added.]
980    val it = true : bool
981    - Main.compile ("../testcases/merge.tig", "foo.out");
982    *** BACK-TRACE ***
983    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
984    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
985    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
986    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
987    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
988    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
989    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
990         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
991    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
992    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
993    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
994    CALL   main.sml:1.18-118.4: Main.compile[2]
995    -
996    
997    ----------------------------------------------------------------------
998    Name: Matthias Blumen
999    Date: 2000/06/21 18:00:00 JST
1000    Tag: blume-20000621-manual
1001    Description:
1002    
1003    CM manual update: Path environments documented.
1004    
1005    ----------------------------------------------------------------------
1006    Name: Matthias Blume
1007    Date: 2000/06/19 13:40:00
1008    Tag: blume-20000619-manual
1009    Description:
1010    
1011    CM manual and system/README update.  This only covers the fact that
1012    there are no more implicit anchors.  (Path environments and the "bind"
1013    option to "cm" have yet to be documented.)
1014    
1015    ----------------------------------------------------------------------
1016    Name: Matthias Blume
1017    Date: 2000/06/19 11:05:00 JST
1018    Tag: blume-20000619-chdir-bugfix
1019    Description:
1020    
1021    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1022    
1023    ----------------------------------------------------------------------
1024    Name: Matthias Blume
1025    Date: 2000/06/18 22:00:10 JST
1026    Tag: blume-20000618-implicit-anchors-really-gone
1027    Description:
1028    
1029    I updates the previous HISTORY entry where I forgot to mention that
1030    implicit anchors are no longer with us.
1031    
1032    The current update also gets rid of the (now useless) controller
1033    CM.Control.implicit_anchors.
1034    
1035    ----------------------------------------------------------------------
1036    Name: Matthias Blume
1037    Date: 2000/06/16 17:30:00 JST
1038    Tag: blume-20000616-anchorenv
1039    Description:
1040    
1041    This patch implements the long anticipated (just kidding :) "anchor
1042    environment" mechanism.  In the course of doing this, I also
1043    re-implemented CM's internal "SrcPath" module from scratch.  The new
1044    one should be more robust in certain boundary cases.  In any case, it
1045    is a lot cleaner than its predecessor (IMHO).
1046    
1047    This time, although there is yet another boot file format change, I
1048    kept the unpickler backward-compatible.  As a result, no new bootfiles
1049    are necessary and bootstrapping is straightforward.  (You cannot read
1050    new bootfiles into an old system, but the other way around is no
1051    problem.)
1052    
1053    Visible changes:
1054    
1055    ** 0. Implicit path anchors (without the leading $-symbol) are no
1056    longer recognized at all. This means that such path names are not
1057    illegal either.  For example, the name basis.cm simply refers to a
1058    local file called "basis.cm" (i.e, the name is an ordinary path
1059    relative to .cm-files directory).  Or, to put it differently, only
1060    names that start with $ are anchored paths.
1061    
1062    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1063    vanished.
1064    
1065    John (Reppy) had critizised this as soon as I originally proposed and
1066    implemented it, but at that time I did not really deeply believe
1067    him. :) Now I came full-circle because I need the $<singlearc> syntax
1068    in another place where it cannot be seen as an abbreviation for
1069    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1070    seems to mean (i.e., it "expands" into the corresponding anchor
1071    value).
1072    
1073    However, when paths are used as members in CM description files, it
1074    continues to be true that there must be at least another arc after the
1075    anchor.  This is now enforced separately during semantic analysis
1076    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1077    
1078    ** 2. The "cm" class now accepts an option "bind".  The option's value
1079    is a sub-option list of precisely two items -- one labeled "anchor"
1080    and the other one labeled "value".  As you might expect, "anchor" is
1081    used to specify an anchor name to be bound, and "value" specifies what
1082    the anchor is being bound to.
1083    
1084    The value must be a directory name and can be given in either standard
1085    syntax (including the possibility that it is itself an anchored path)
1086    or native syntax.
1087    
1088    Examples:
1089    
1090       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1091       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1092    
1093    and so on.
1094    
1095    The meaning of this is that the .cm-file will be processed with an
1096    augmented anchor environment where the given anchor(s) is/are bound to
1097    the given values(s).
1098    
1099    The rationale for having this feature is this: Suppose you are trying
1100    to use two different (already stable) libraries a.cm and b.cm (that
1101    you perhaps didn't write yourself).  Further, suppose each of these
1102    two libraries internally uses its own auxiliary library $aux/lib.cm.
1103    Normally you would now have a problem because the anchor "lib" can not
1104    be bound to more than one value globally.  Therefore, the project that
1105    uses both a.cm and b.cm must locally redirect the anchor to some other
1106    place:
1107    
1108       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1109       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1110    
1111    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1112    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1113    
1114    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1115    you when you do so at the time of CM.stabilize).  Therefore, you can
1116    also use an anchored path as the value:
1117    
1118      a.cm (bind:(anchor:lib value:$a-lib))
1119      b.cm (bind:(anchor:lib value:$b-lib))
1120    
1121    Now you can globally configure (using the usual CM.Anchor.anchor or
1122    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1123    itself is always locally bound, setting it globally is no longer
1124    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1125    can still be used as a global anchor for separate purposes.  As a
1126    matter of fact, one can locally define "lib" in terms of a global
1127    "lib":
1128    
1129      a.cm (bind:(anchor:lib value:$lib/a))
1130      b.cm (bind:(anchor:lib value:$lib/b))
1131    
1132    ** 3: The encoding of path names has changed.  This affects the way
1133    path names are shown in CM's progress report and also the internal
1134    protocol encoding used for parallel make.
1135    
1136    The encoding now uses one or more ':'-separated segments.  Each
1137    segments corresponds to a file that has been specified relative to the
1138    file given by its preceding segment.  The first segment is either
1139    relative to the CWD, absolute, or anchored.  Each segment itself is
1140    basically a Unix pathname; all segments but the first are relative.
1141    
1142    Example:
1143    
1144       $foo/bar/baz.cm:a/b/c.sml
1145    
1146    This path denotes the file bar/a/b/c.sml relative to the directory
1147    denoted by anchor "foo".  Notice that the encoding also includes
1148    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1149    paths are resolved relative to the .cm-files directory, so baz.cm must
1150    be ignored to get the "real" pathname.
1151    
1152    To make this fact more obvious, CM puts the names of such "virtual
1153    arcs" into parentheses when they appear in progress reports. (No
1154    parentheses will appear in the internal protocol encoding.)  Thus,
1155    what you really see is:
1156    
1157      $foo/bar/(baz.cm):a/b/c.sml
1158    
1159    I find this notation to be much more informative than before.
1160    
1161    Another new feature of the encoding is that special characters
1162    including parentheses, colons, (back)slashes, and white space are
1163    written as \ddd (where ddd is the decimal encoding of the character).
1164    
1165    *** The CM manual still needs to be updated.
1166    
1167    ----------------------------------------------------------------------
1168    Name: Allen Leung
1169    Date: 2000/06/15 00:38:00
1170    Tag: leunga-20000615-x86-peephole
1171    
1172    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1173    
1174    ----------------------------------------------------------------------
1175    Name: Matthias Blume
1176    Date: 2000/06/12 11:40:00
1177    Tag: blume-20000612-parmakefix
1178    Description:
1179    
1180    More cleanup after changing the file naming scheme: This time I
1181    repaired the parallel make mechanism for CMB.make which I broke earlier.
1182    
1183    ----------------------------------------------------------------------
1184    Name: Allen Leung
1185    Date: 2000/06/09 01:25:00
1186    Tag: leunga-20000609-various
1187    
1188    None of these things should affect normal SML/NJ operations
1189    
1190    1. Peephole improvements provided by Fermin (c--)
1191    2. New annotation DEFUSE for adding extra dependence (moby)
1192    3. New X86 LOCK instructions (moby)
1193    4. New machine description language for reservation tables (scheduling)
1194    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1195       trees etc.)
1196    6. I've changed the CM files so that they can work with versions
1197       110.0.6, 110.25 and 110.28
1198    
1199    ----------------------------------------------------------------------
1200    Name: Matthias Blume
1201    Date: 2000/06/09 12:40:00
1202    Tag: blume-20000609-log
1203    Description:
1204    
1205    - Removed all(?) remaining RCS Log entries from sources.
1206    
1207    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1208      anchored paths).
1209    
1210    ----------------------------------------------------------------------
1211    Name: Matthias Blume
1212    Date: 2000/06/07 17:00:00 JST
1213    Tag: blume-20000607-no-implicit-anchors
1214    Description:
1215    
1216    1. This update changes the default setting for
1217    CM.Control.implicit_anchors from true to false.  This means that
1218    implicit anchors are no longer permitted by default.  I also tried to
1219    make sure that nothing else still relies on implicit anchors.
1220    (This is the next step on the schedule towards a CM that does not even
1221    have the notion of implicit anchors anymore.)
1222    
1223    2. More CM manual updates.
1224    
1225    3. I managed to track down and fix the pickling bug I mentioned last
1226    time.  Because of the previously existing workaround, this entails no
1227    immediate practical changes.
1228    
1229    ----------------------------------------------------------------------
1230    Name: Matthias Blume
1231    Date: 2000/06/06 11:15:00 JST
1232    Tag: blume-20000606-lazierpickle
1233    Description:
1234    
1235    !!!! NEW BOOT FILES !!!!
1236    
1237    * The main purpose of this update is to make library pickles lazier in
1238    order to reduce the initial space penalty for autoloading a library.
1239    As a result, it is now possible to have $smlnj/compiler.cm
1240    pre-registered.  This should take care of the many complaints or
1241    inquiries about missing structure Compiler.  This required changes to
1242    CM's internal data structures and small tweaks to some algorithms.
1243    
1244    As a neat additional effect, it is no longer necessary (for the sake
1245    of lean heap image files) to distinguish between a "minimal" CM and a
1246    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1247    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1248    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1249    
1250    To make the life of compiler-hackers easier, "makeml" now also
1251    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1252    words, after you bootstrap a new sml for the first time, you will not
1253    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1254    around you will still have to do it, though.)
1255    
1256    * A second change consists of major updates to the CM manual.  There
1257    are now several appendices with summary information and also a full
1258    specification of the CM description file syntax.
1259    
1260    * In directory src/system I added the script "allcross".  This script
1261    invokes sml and cross-compiles the compiler for all supported
1262    architectures.  (Useful when providing a new set of boot files.)
1263    
1264    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1265    added a small tweak to pickle-util.sml to work around this problem,
1266    but it is not a proper fix yet.  I will investigate further.  (The
1267    effect of the bug was an inflation of library pickle size.)
1268    
1269    * Version number increased to 110.28.1 (to avoid compatibility problems).
1270    
1271    ----------------------------------------------------------------------
1272    Name: Allen Leung
1273    Date: 2000/05/25 17:28 EDT
1274    Tag: leunga-20000525-ra
1275    Description:
1276    
1277      Fixed a bug in freezing phase of the register allocator.
1278    
1279    ----------------------------------------------------------------------
1280    Name: Allen Leung
1281    Date: 2000/05/15 22:53 EDT
1282    Tag: leunga-20000515-alpha-x86-ra
1283    Description:
1284    
1285      1. Alpha
1286    
1287          Slight cleanup.  Removed the instruction SGNXL
1288    
1289      2. X86
1290    
1291          Added the following instructions to the instruction set:
1292    
1293            ROLx, RORx,
1294            BTx, BTSx, BTLx, BTRx,
1295            XCHGx, and variants with the LOCK prefix
1296    
1297      3. Register Allocation
1298    
1299          The module ra-rewrite-with-renaming has been improved.
1300    
1301      These have no effect on SML/NJ.
1302    
1303    ----------------------------------------------------------------------
1304    Name: Matthias Blume
1305    Date: 2000/05/15 16:20:00 JST
1306    Tag: blume-20000515-lightrebuild
1307    Description:
1308    
1309    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1310       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1311       defined.  In effect, the command will not build any cross-compiler
1312       backends and therefore finish more quickly.
1313    
1314       The "fixpt" script also takes a "-light" switch to be able to use
1315       this new facility while compiling for a fixpoint.
1316    
1317    2. I replaced all mentions of anchored paths in group owner specifications
1318       with simple relative paths (usually starting with "..").
1319       The rationale is that a library's internal workings should not be
1320       compromised by the lack of some anchor.  (An anchor is necessary
1321       for someone who wants to refer to the library by an anchored path,
1322       but it should not be necessary to build the same library in the first
1323       place.)
1324    
1325    3. I changed the way CM's tool mechanism determines the shell command
1326       string used for things like ml-yacc etc. so that it does not break
1327       when CM.Control.implicit_anchors is turned off.
1328    
1329    ----------------------------------------------------------------------
1330    Name: Matthias Blume
1331    Date: 2000/05/12 18:20:00 JST
1332    Tag: blume-20000512-ml-build
1333    Description:
1334    
1335    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1336    getting installed properly (by config/install.sh).
1337    
1338    ----------------------------------------------------------------------
1339    Name: Matthias Blume
1340    Date: 2000/05/12 17:30:00 JST
1341    Tag: blume-20000512-anchors
1342    Description:
1343    
1344    !!! NEW BOOT FILES !!!
1345    
1346    This change is in preparation of fading out support for "implicitly
1347    anchored path names".  I went through all sources and used the
1348    explicit (and relatively new) $-notation.  See system/README and the
1349    CM manual for more info on this.
1350    
1351    I also modified the anchoring scheme for some things such as "smlnj",
1352    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1353    anchors are more expressive: anchor name and first arc do not have to
1354    coincide.  This entails the following user-visible change:
1355    
1356    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1357    particular, when you fire up sml with a command-line argument, say,
1358    e.g.:
1359    
1360       sml '$smlnj/cmb.cm'
1361    
1362    At the ML toplevel prompt:
1363    
1364       CM.autoload "$smlnj/cmb.cm";
1365    
1366    There is also a new controller in CM.Control that can be used to turn
1367    off all remaining support for implicit anchors by saying:
1368    
1369        CM.autoload "$smlnj/
1370        #set CM.Control.implicit_anchors false;
1371    
1372    This causes CM to reject implicitly anchored paths.  This is (for the
1373    time being) less permissive than the "final" version where there will
1374    be no more such implicit anchors and relative paths will be just that:
1375    relative.
1376    
1377    The next step (version after next version?) will be to make the
1378    default for CM.Control.implicit_anchors false.  After the dust has
1379    settled, I can then produce the "final" version of this...
1380    
1381    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1382    
1383    ----------------------------------------------------------------------
1384    Name: Matthias Blume
1385    Date: 2000/05/11 16:30:00 JST
1386    Tag: blume-20000511-sources
1387    Description:
1388    
1389    The main change is that I added function CM.sources as a generalized
1390    version of the earlier CM.makedepend.  This entails the following
1391    additional changes:
1392    
1393      - CM.makedepend has been dropped.
1394    
1395      - CM manual has been updated.
1396    
1397      - TOOLS signature and API have been changed.
1398    
1399    ----------------------------------------------------------------------
1400    Name: Allen Leung
1401    Date: 2000/05/10 21:17 EDT
1402    Tag: leunga-20000510-moby-c--ssa
1403    Description:
1404    
1405      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1406    None of these affect SML/NJ.
1407    
1408    1. Register Allocation
1409    
1410        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1411           This module tries to remove local (i.e. basic block level) redundancies
1412           during spilling.
1413    
1414        b. A new framework for performing region based register allocation.
1415           Not yet entirely functional.
1416    
1417    2. X86
1418    
1419       a. DefUse for POP was missing the stack pointer [found by Lal]
1420       b. Reload for CALL was incorrect in X86Spill [found by John]
1421       c. Various fixes in X86Spill so that it can be used correctly for
1422          the new spilling module.
1423    
1424    3. SSA/IR
1425    
1426       a. New module ir/dj-dataflow.sml implements elimination based
1427          data flow analysis.
1428    
1429    4. MLRiscGen
1430    
1431       a. Fix for gc type annotation
1432    
1433    5. MDGen
1434    
1435       Various fixes for machine description -> ml code translation.  For ssa
1436       only.
1437    
1438    ----------------------------------------------------------------------
1439    Name: Allen Leung
1440    Date: 2000/05/08 22:17 EDT
1441    Tag: leunga-20000508-labexp
1442    Description:
1443    
1444      Fermin has found a few assembly problems with constant expressions
1445      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1446      which choke on dumb assemblers.  This is his fix.
1447    
1448    ----------------------------------------------------------------------
1449    Name: Dave MacQueen
1450    Date: 2000/04/09 14:00 EDT
1451    Tag: dbm-20000502-Version_110_28
1452    Description:
1453    
1454    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1455    
1456    2. Updated config/version to 110.28
1457    
1458    3. Updated config/srcarchiveurl
1459    
1460    3. New boot files!
1461       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1462    
1463    ----------------------------------------------------------------------
1464    Name: Matthias Blume
1465    Date: 2000/05/01 19:05:00 JST
1466    Tag: blume-20000501-noweb
1467    Description:
1468    
1469    A new noweb tool has been added.  The existing system is entirely
1470    unaffected by this, but some CM users have asked for renewed noweb
1471    support.  Everything is documented in the CM manual.
1472    
1473    New (plugin) libraries:
1474    
1475       noweb-tool.cm
1476       nw-ext.cm
1477    
1478    ----------------------------------------------------------------------
1479    Name: Dave MacQueen
1480    Date: 2000/04/30 12:40PM EDT
1481    Tag: dbm-20000430-bug_fixes
1482    Description:
1483    
1484    1. Fix for bug 1498
1485       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1486       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1487         added toRealArray function
1488       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1489         added check for tag Obj.RealArray to array printing case in ppObj
1490    
1491    2. Fix for bug 1510
1492       smlnj/src/compiler/Semant/types/typesutil.sml
1493         fixed definition of dummyargs (used by equalTycon) so that
1494         dummy args are distinct types
1495    
1496    ----------------------------------------------------------------------
1497    Name: Matthias Blume
1498    Date: 2000/04/30 01:00:00 JST
1499    Tag: blume-20000430-versions
1500    Description:
1501    
1502    1. CM version numbering added.  This is an implementation of Lal's
1503       proposal for adding version numbers and version checking to .cm
1504       files.  Lal said that his proposal was just that -- a proposal.
1505       For the time being I went ahead and implemented it so that people
1506       can comment on it.  Everything is completely backward-compatible
1507       (except for the stable library format, i.e., new bootfiles!).
1508    
1509       As usual, see the CM manual for details.
1510    
1511    2. An alternative syntax for anchored paths has been implemented.
1512       Dave has recently voiced the same concerns that I had when I did
1513       this, so there should be some support.  My take is that eventually
1514       I will let support for the current syntax (where anchors are
1515       "implicit") fade out in favor of the new, explicit syntax.
1516       In order to be backward-compatible, both old and new syntax are
1517       currently supported.
1518    
1519       Again, see the CM manual for details.
1520    
1521    3. Parallel make is trying to be slightly smarter:  When the master
1522       process finds a "bottleneck", i.e., when there is only one
1523       compilation unit that can be compiled and everybody else is
1524       waiting on it, then it will simply compile it directly instead
1525       of clumsily telling one of the slaves to do it.
1526    
1527    4. Support for "unsharing" added.  This is necessary in order to be
1528       able to have two different versions of the same library running
1529       at the same time (e.g., for trying out a new MLRISC while still
1530       having the old MLRISC linked into the current compiler, etc.)
1531       See the CM manual.
1532    
1533    5. Simple "makedepend" functionality added for generating Makefile
1534       dependency information.  (This is rather crude at the moment.
1535       Expect some changes here in the future.)
1536    
1537    6. ".fun" added as a recognized suffix for ML files. Also documented
1538       explicitly in the manual that the fallback behavior (unknown suffix
1539       -> ML file) is not an official feature!
1540    
1541    7. Small changes to the pickler for stable libraries.
1542    
1543    8. Several internal changes to CM (for cleanup/improvement).
1544    
1545    
1546    !!!! NEW BINFILES !!!!
1547    
1548    ----------------------------------------------------------------------
1549    Name: Matthias Blume
1550    Date: 2000/04/28 17:30:00 JST
1551    Tag: blume-20000428-pathconfig
1552    Description:
1553    
1554    1. I changed config/install.sh to remove duplicate entries from the
1555       lib/pathconfig file at the end.  Moreover, the final version of
1556       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1557       in src/system/installml.
1558    
1559    2. The config/install.sh script now consistently uses relative
1560       pathnames in lib/pathconfig whenever the anchor is in the lib
1561       directory.  (So far this was true for the libraries that come
1562       pre-compiled and bundled as part of the bootfiles but not for
1563       libraries that are compiled by the script itself.)
1564    
1565    ----------------------------------------------------------------------
1566    Name: Matthias Blume
1567    Date: 2000/04/26 13:10:00 JST
1568    Tag: blume-20000426-fun_suffix
1569    Description:
1570    
1571    Added ".fun" as a recognized file name suffix (for ML code).
1572    
1573    ----------------------------------------------------------------------
1574    Name: Allen Leung
1575    Date: 2000/04/25 17:00:00 EST
1576    Tag: leunga-20000425-alpha-ra
1577    Description:
1578    
1579    1. Alpha
1580    
1581        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1582    in C--.
1583    
1584    2. RA
1585    
1586       Added a flag "ra-dump-size" to print out the size of the flowgraph
1587       and the interference graph.
1588    
1589    ----------------------------------------------------------------------
1590    Name: Dave MacQueen
1591    Date: 2000/04/25/
1592    Tag: dbm-20000425-mlyacc_doc_examples
1593    Description:
1594      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1595      Updated all three examples in src/ml-yacc/examples to run
1596      under 110.* using CM.make.
1597    
1598    ----------------------------------------------------------------------
1599    Name: Allen Leung
1600    Date: 2000/04/20 23:04:00 EST
1601    Tag: leunga-20000420-ssa-c---stuff
1602    Description:
1603    
1604      This update synchronizes my repository with Yale's.  Most of these
1605    changes, however, do not affect SML/NJ at all (the RA is an exception).
1606    
1607    1. Register Allocator
1608    
1609       a. An improvement in the interference graph construction:
1610          Given a copy
1611    
1612                s <- t
1613    
1614          no interference edge between s and t is added for this definition of s.
1615    
1616       b. I've added two new spill heuristic modules that Fermin and I developed
1617          (in the new library RA.cm). These are unused in SML/NJ but maybe
1618          useful for others (Moby?)
1619    
1620    2. X86
1621    
1622       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1623    
1624    3. Alpha
1625    
1626       a. Added the BSR instruction and code generation that goes with it [C--]
1627       b. Other fixes too numerous to recount provided by Fermin [C--]
1628    
1629    4. Regmaps
1630    
1631       a. The regmaps are not initialized with the identity physical bindings
1632          at creation time.  This is unneeded.
1633    
1634    5. MLRISC Optimizations
1635    
1636       a. The DJ-Graph module can now compute the iterated dominance frontiers
1637          intersects with liveness incrementally in linear time! Woohoo!
1638          This is now used in my new SSA construction algorithm.
1639    
1640       b. THe branch reorganization module is now smarter about linear chains of
1641          basic blocks.
1642    
1643    
1644    ----------------------------------------------------------------------
1645    Name: Matthias Blume
1646    Date: 2000/04/12 13:52:00 JST
1647    Tag: blume_main_v110p27_1
1648    Description:
1649    
1650    Changed install.sh script to handle archive files without version number
1651    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1652    name of the boot file archive.
1653    
1654    ----------------------------------------------------------------------
1655    Name: Dave MacQueen
1656    Date: 2000/04/09 14:00 EDT
1657    Tag: dbm-20000410-Version_110_27
1658    Description:
1659    
1660    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1661    
1662    2. Updated src/config/version to 110.27
1663    
1664    3. New boot files!
1665    
1666    ----------------------------------------------------------------------
1667    Name: Allen Leung
1668    Date: 2000/04/09 19:09:00 EST
1669    Tag: leunga-20000409-misc
1670    Description:
1671    
1672    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1673    
1674    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1675    
1676    ----------------------------------------------------------------------
1677    Name: Stefan
1678    Date: 2000/04/07 10:00:00 EDT
1679    Tag: monnier-20000406-branch-handling
1680    Description:
1681    
1682    Improved handling of branches (mostly those generated from
1683    polymorphic equality), removed switchoff and changed the
1684    default optimization settings (more cpsopt and less flintopt).
1685    
1686    ----------------------------------------------------------------------
1687  Name: Allen Leung  Name: Allen Leung
1688  Date: 2000/04/06 01:30:00 EST  Date: 2000/04/06 01:30:00 EST
1689  Tag: leunga-20000406-peephole-x86-SSA-2  Tag: leunga-20000406-peephole-x86-SSA-2
# Line 87  Line 1759 
1759    
1760  3.  Assembly  3.  Assembly
1761    
1762      When generating assemby, resolve the value of client defined constants,      When generating assembly, resolve the value of client defined constants,
1763      instead of generating symbolic values.  This is controlled by the      instead of generating symbolic values.  This is controlled by the
1764      new flag "asm-resolve-constants", which is default to true.      new flag "asm-resolve-constants", which is default to true.
1765    
# Line 110  Line 1782 
1782    
1783  To this end, I arranged that instead of "structure Core" as "structure  To this end, I arranged that instead of "structure Core" as "structure
1784  _Core" is bound in the pervasive environment.  Core access is done via  _Core" is bound in the pervasive environment.  Core access is done via
1785  _Core (which can never be accidentially rebound because _Core is not a  _Core (which can never be accidentally rebound because _Core is not a
1786  legal surface-syntax symbol).  legal surface-syntax symbol).
1787    
1788  The current solution is much cleaner because the core environment is  The current solution is much cleaner because the core environment is
# Line 120  Line 1792 
1792  with dynamic and symbolic parts of the core environment.  with dynamic and symbolic parts of the core environment.
1793    
1794  Remaining hackery (to bind the "magic" symbol _Core) is localized in the  Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1795  compilation mananger's bootstrap compiler (actually: in the "init group"  compilation manager's bootstrap compiler (actually: in the "init group"
1796  handling).  See the comments in src/system/smlnj/init/init.cmi for  handling).  See the comments in src/system/smlnj/init/init.cmi for
1797  more details.  more details.
1798    
# Line 235  Line 1907 
1907     (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
1908     used).     used).
1909    
1910     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
1911     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
1912     Makefiles or Shell scripts.     Makefiles or Shell scripts.
1913    
1914     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 248  Line 1920 
1920    
1921     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
1922    
1923     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
1924     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
1925    
1926  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 382  Line 2054 
2054    
2055       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2056       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2057       testl is perferable to andl.       testl is preferable to andl.
2058    
2059  7. RA (x86 only)  7. RA (x86 only)
2060    
# Line 564  Line 2236 
2236    
2237  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2238  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2239  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
2240  explanation.  explanation.
2241    
2242  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2243  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2244  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2245  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 611  Line 2283 
2283  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
2284  <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,
2285  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
2286  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2287  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2288  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2289    
# Line 661  Line 2333 
2333    it from that remote directory.    it from that remote directory.
2334    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2335    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2336    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2337    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2338    
2339    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.602  
changed lines
  Added in v.781

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