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/NOTES/HISTORY
ViewVC logotype

Diff of /sml/trunk/NOTES/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 583, Thu Mar 23 21:52:30 2000 UTC revision 808, Wed Apr 18 15:43:09 2001 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume
17    Date: 2001/04/18 12:45:00 EDT
18    Tag: blume-20010418-ccalls
19    Description:
20    
21     - fixed two off-by-4 errors in the x86-specific c-calls implementation
22       (this bug prevented structure arguments containing pointers from being
23        passed correctly)
24     - changed the raw-C-call code in mlriscGen.sml in such a way that
25       structure arguments are represented as a pointer to the beginning
26       of the structure (instead of having a series of synthesized arguments,
27       one for each structure member)
28    
29     - made makeml script's verbosity level configurable via environment
30       variable (MAKEML_VERBOSITY)
31    
32     - eliminteded placeholder implementations for f32l, w16s, i16s, and f32s
33       in rawmem-x86.sml; we are now using the real thing
34    
35    ----------------------------------------------------------------------
36    Name: Matthias Blume
37    Date: 2001/03/22 16:25:00 EST
38    Tag: blume-20010322-bootfiles
39    Description:
40    
41    Created a new set of bootfiles (for your automatic installation convenience).
42    
43    ----------------------------------------------------------------------
44    Name: Matthias Blume
45    Date: 2001/03/22 15:10:00 EST
46    Tag: blume-20010322-rawmem-parcm
47    Description:
48    
49    1. All "raw memory access" primitives for the new FFI are implemented now
50       (at least on the x86).
51    2. Some further cleanup of CM's parallel make mechanism.
52    
53    ----------------------------------------------------------------------
54    Name: Matthias Blume
55    Date: 2001/03/19 17:53:00 EST
56    Tag: blume-20010319-parallel
57    Description:
58    
59    Parallel make (using compile servers) now works again.
60    
61    To this end, CM.stabilize and CMB.make have been modified to work in
62    two passes when compile servers are attached:
63       1. Compile everything, do not perform stabilization; this pass
64          uses compile servers
65       2. Stabilize everything; this pass does not use compile servers
66    If there are no compile servers, the two passes are combined into one
67    (as before).  Splitting the passes increases the inherent parallelism
68    in the dependency graph because the entire graph including all
69    libraries is available at the same time.  This, in turn, improves
70    server utilization.  The downside is that the master process will
71    have to do some extra work after compilation is done (because for
72    technical reasons it must re-read all the binfiles during stabilization).
73    
74    ----------------------------------------------------------------------
75    Name: Matthias Blume
76    Date: 2001/03/16 12:22:00 EST
77    Tag: blume-20010316-bootfiles
78    Description:
79    
80    Created a new set of bootfiles (for your automatic installation convenience).
81    
82    ----------------------------------------------------------------------
83    Name: Matthias Blume
84    Date: 2001/03/16 11:00:00 EST
85    Tag: blume-20010316-MLTREE-fixup
86    Description:
87    
88    This is a minor fixup for an (untagged) earlier commit by Allen.
89    (A file was missing).
90    
91    ----------------------------------------------------------------------
92    Name: Allen Leung
93    Date: Mon Mar  5 18:54:57 EST 2001
94    Tag: leunga-20010305-cut-support
95    
96    1. New support for alternative control-flow in MLTREE.
97       Currently we support
98    
99          FLOW_TO(CALL ...., [k1,...,kn])
100    
101       This is needed for 'cuts to' in C-- and try/handle-like constructs
102       in Moby
103    
104       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
105    
106    2. Register Allocator
107    
108       Changes in interface [from Fermin, John]
109    
110    3. Alpha 8-bit SLL support [Fermin]
111    
112    4. All architectures
113    
114       A new module (ClusterExpandCopies) for expanding parallel copies.
115    
116    ----------------------------------------------------------------------
117    Name: Allen Leung
118    Date: 2001/02/27 23:07:00 EST
119    Tag: leunga-20010227-minor-stuff
120    
121    1. Alpha bug fix for CMOVNE
122    2. Handle mltree COND(..,FCMP ...,...)
123    3. Bug fix in simplifier
124    
125    ----------------------------------------------------------------------
126    Name: Matthias Blume
127    Date: 2001/01/30 17:50:00 EST
128    Tag: blume-20010130-sync
129    Description:
130    
131    This is just a minor update to sync my devel branch with the main brach.
132    The only visible change is the addition of some README files.
133    
134    ----------------------------------------------------------------------
135    Name: Matthias Blume
136    Date: 2001/01/12 23:30:00 JST
137    Tag: blume-20010112-bootfiles
138    Description:
139    
140    Made a new set of bootfiles that goes with the current state of the
141    repository.
142    
143    ----------------------------------------------------------------------
144    Name: Matthias Blume
145    Date: 2001/01/12 21:20:00 JST
146    Tag: blume-20010112-sync
147    Description:
148    
149    I am just flushing out some minor changes that had accumulated in
150    my private branch in order to sync with the main tree.  (This is
151    mainly because I had CVS trouble when trying to merge _into_ my
152    private branch.)
153    
154    Most people should be completely unaffected by this.
155    
156    ----------------------------------------------------------------------
157    Name: Allen Leung
158    Date: Thu Jan 11 21:03:00 EST 2001
159    Tag: leunga-20010111-labexp=mltree
160    Description:
161    
162    1.  Removed the type LabelExp and replace it by MLTree.
163    2.  Rewritten mltree-simplify with the pattern matcher tool.
164    3.  There were some bugs in alpha code generator which would break
165        64-bit code generation.
166    4.  Redo the tools to generate code with the
167    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
168        are now generated by perl scripts.
169    
170    ----------------------------------------------------------------------
171    Name: Matthias Blume
172    Date: 2001/01/10 21:55:00 JST
173    Tag: blume-20010110-rcc
174    Description:
175    
176    The RCC stuff now seems to work (but only on the x86).
177    This required hacking of the c-calls interface (and -implementation) in
178    MLRISC.
179    
180    Normal compiler users should be unaffected.
181    
182    ----------------------------------------------------------------------
183    Name: Matthias Blume
184    Date: 2001/01/09 01:20:00 JST
185    Tag: blume-20010109-rcc
186    Description:
187    
188    This is a fairly big patch, flushing out a large number of pending
189    changes that I made to my development copy over the last couple of days.
190    
191    Of practical relevance at this moment is a workaround for a pickling
192    bug that Allen ran into the other day.  The cause of the bug itself is
193    still unknown and it might be hard to fix it properly, but the
194    workaround has some merits of its own (namely somewhat reducing pickling
195    overhead for certain libraries).  Therefore, I think this solution should
196    be satisfactory at this time.
197    
198    The rest of the changes (i.e., the vast majority) has to do with my
199    ongoing efforts of providing direct support for C function calls from
200    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
201    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
202    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
203    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
204    handle RCC.
205    
206    The last part (the changes to mlriscGen.sml) are still known to be
207    wrong on the x86 and not implemented on all other architectures.  But
208    the infrastructure is in place. I had to change a few functor
209    signatures in the backend to be able to route the CCalls interface
210    from MLRISC there, and I had to specialize the mltree type (on the
211    x86) to include the necessary extensions. (The extensions themselves
212    were already there and redy to go in MLRISC/x86).
213    
214    Everything should be very happy as soon as someone helps me with
215    mlriscGen.sml...
216    
217    In any case, nothing of this should matter to anyone as long as the
218    new primop is not being used (which is going to be the case unless you
219    find it where I hid it :). The rest of the compiler is completely
220    unaffected.
221    
222    ----------------------------------------------------------------------
223    Name: Matthias Blume
224    Date: 2001/01/05 00:30:00 JST
225    Tag: blume-20010105-primops
226    Description:
227    
228    Added some experimental support for work that I am doing right now.
229    These changes mostly concern added primops, but there is also a new
230    experimental C library in the runtime system (but currently not enabled
231    anywhere except on Linux/X86).
232    
233    In the course of adding primops (and playing with them), I discovered that
234    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
235    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
236    To recover, I made type information in INL_PRIM mandatory and changed
237    prim.sml as well as built-in.sml accordingly.  The InLine structure now
238    has complete, correct type information (i.e., no bottom types).
239    
240    Since all these changes mean that we need new binfiles, I also bumped the
241    version number to 110.32.1.
242    
243    ----------------------------------------------------------------------
244    Name: Matthias Blume
245    Date: 2000/12/30 22:10:00 JST
246    Tag: blume-20001230-various
247    Description:
248    
249    Added proxy libraries for MLRISC and let MLRISC libraries refer
250    to each other using path anchors.  (See CM manual for explanation.)
251    
252    Updated CM documentation.
253    
254    Fixed some bugs in CM.
255    
256    Implemented "proxy" libraries (= syntactic sugar for CM).
257    
258    Added "-quiet" option to makeml and changed runtime system accordingly.
259    
260    Added cleanup handler for exportML to reset timers and compiler stats.
261    
262    ----------------------------------------------------------------------
263    Name: Lal George
264    Date: 2000/12/22 22:22:58 EST 2000
265    Tag: Release_110_32
266    Description:
267    
268            Infinite precision used throughout MLRISC.
269            see MLRISC/mltree/machine-int.sig
270    
271    ----------------------------------------------------------------------
272    Name: Matthias Blume
273    Date: 2000/12/22 23:16:00 JST
274    Tag: blume-20001222-warn
275    Description:
276    
277    Corrected wording and formatting of some CM warning message which I
278    broke in my previous patch.
279    
280    ----------------------------------------------------------------------
281    Name: Matthias Blume
282    Date: 2000/12/22 21:20:00 JST
283    Tag: blume-20001222-anchorenv
284    Description:
285    
286    Fixed CM's handling of anchor environments in connection with CMB.make.
287    
288    ----------------------------------------------------------------------
289    Name: Matthias Blume
290    Date: 2000/12/22 13:15:00 JST
291    Tag: blume-20001222-cleanup
292    Description:
293    
294    Removed src/cm/ffi which does not (and did not) belong here.
295    
296    ----------------------------------------------------------------------
297    Name: Matthias Blume
298    Date: 2000/12/21 23:55:00 JST
299    Tag: blume-20001221-exn
300    Description:
301    
302    Probably most important: CM no longer silently swallows all exceptions
303    in the compiler.
304    Plus: some other minor CM changes.  For example, CM now reports some
305    sizes for generated binfiles (code, data, envpickle, lambdapickle).
306    
307    ----------------------------------------------------------------------
308    Name: Matthias Blume
309    Date: 2000/12/15 00:01:05 JST
310    Tag: blume-20001215-dirtool
311    Description:
312    
313    - "dir" tool added.
314    - improvements and cleanup to Tools structure
315    - documentation updates
316    
317    ----------------------------------------------------------------------
318    Name: Allen Leung
319    Date: Thu Dec 14 03:45:24 EST 2000
320    Description:
321    Tag:  leunga-20001214-int-inf
322    Description:
323    
324       In IntInf, added these standard functions, which are missing from our
325    implementation:
326    
327        andb : int * int -> int
328        xorb : int * int -> int
329        orb  : int * int -> int
330        notb : int -> int
331         <<   : int * word -> int
332        ~>>  : int * word -> int
333    
334       Not tested, I hope they are correct.
335    
336    ----------------------------------------------------------------------
337    Name: Allen Leung
338    Date: Fri Dec  8 19:23:26 EST 2000
339    Description:
340    Tag:  leunga-20001208-nowhere
341    Description:
342    
343      Slight improvements to the 'nowhere' tool to handle OR-patterns,
344    to generate better error messages etc.  Plus a brief manual.
345    
346    ----------------------------------------------------------------------
347    Name: Lal George
348    Date: 2000/12/08 09:54:02 EST 2000
349    Tag: Release_110_31
350    Description:
351    
352    - Version 110.31
353    ----------------------------------------------------------------------
354    Name: Allen Leung
355    Date: Thu Dec  7 22:01:04 EST 2000
356    Tag:  leunga-20001207-cell-monster-hack
357    Description:
358    
359    Major MLRISC internal changes.  Affect all clients.
360    Summary:
361    
362    1.  Type CELLS.cell = int is now replaced by a datatype.
363        As a result, the old regmap is now gone.  Almost all interfaces
364        in MLRISC change as a consequence.
365    
366    2.  A new brand version of machine description tool (v3.0) that generates
367        modules expecting the new interface.  The old version is removed.
368    
369    3.  The RA interface has been further abstracted into two new functors.
370        RISC_RA and X86RA.  These functors have much simpler interfaces.
371        [See also directory MLRISC/demo.]
372    
373    4.  Some other new source->source code generation tools are available:
374    
375        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
376        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
377           I use this tool to generate the peephole optimizers---with the new
378           cell type changes, peephole rules are becoming difficult to write
379           without conditional pattern matching.
380    
381    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
382        cover the entire MLRISC source tree so many things broke.
383    
384    6.  CM files have been moved to the subdirectory MLRISC/cm.
385        They are moved because there are a lot of them and they clutter up the
386        root dir.
387    
388    7.  More detailed documentation to come...
389    
390        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
391        a makeml -rebuild first.  This is because of other other
392        changes that Matthias has made (see below).
393    
394    
395    ----------------------------------------------------------------------
396    Name: Matthias Blume
397    Date: 2000/11/30 23:12:00 JST
398    Tag: blume-20001130-filereorg
399    Description:
400    
401    Some manual updates and some file reorganizations in CM.
402    
403    ----------------------------------------------------------------------
404    Name: Matthias Blume
405    Date: 2000/11/24 17:45:00 JST
406    Tag: blume-20001124-link
407    Description:
408    
409    Drastically improved link traversal code for the case that the dynamic
410    value was already loaded at bootstrap time.  As a result, CM and CMB
411    now both load blazingly fast -- even on a very slow machine.  Also,
412    memory consumption has been further reduced by this.
413    
414    Warning: The format of the PIDMAP file has changed.  THerefore, to
415    bootstrap you have to do this:
416    
417    1. Run CMB.make
418    2. Make a symbolic link for the boot directory:
419         ln -s sml.boot.ARCH-OS xxx
420    3. "Rebuild" the boot directory:
421         ./makeml -boot xxx -rebuild sml ; rm xxx
422    4. Boot normally:
423          ./makeml
424    
425    ----------------------------------------------------------------------
426    Name: Matthias Blume
427    Date: 2000/11/21 21:20:00 JST
428    Tag: blume-20001121-tools
429    Description:
430    
431    Continued hacking on autoloading problem -- with success this time.
432    Also changed tool-plugin mechanism.  See new CM manual.
433    
434    ----------------------------------------------------------------------
435    Name: Matthias Blume
436    Date: 2000/11/19 14:30:00 JST
437    Tag:  blume-20001119-autoload
438    Description:
439    
440    Some hacking to make autoloading faster.  Success for CMB, no success
441    so far for CM.  There is a reduced structure CM' that autoloads faster.
442    (This is a temporary, non-documented hack to be eliminated again when
443    the general problem is solved.)
444    
445    ----------------------------------------------------------------------
446    Name: Matthias Blume
447    Date: 2000/11/17 14:10:00 JST
448    Tag: blume-20001117-pickle-lib
449    Description:
450    
451    1. Eliminated comp-lib.cm
452    2. Made pickle-lib.cm
453    3. Eliminated all uses of intset.sml (from comp-lib.cm)
454    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
455       equivalent constructs from smlnj-lib.cm (INtHashTable).
456    5. Point 4. also goes for those uses of intmap.* in MLRISC.
457       Duplicated intmap modules thrown out.
458    6. Hunted down all duplicated SCC code and replaced it with
459       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
460    7. Rewrote Feedback module.
461    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
462       should be thrown out and equivalent modules from smlnj-lib.cm
463       should be used (IntRedBlackSet, IntListSet, ...).
464    
465    Confirmed that compiler compiles to fixpoint.
466    
467    ----------------------------------------------------------------------
468    Name: Allen Leung
469    Date: 2000/11/10 18:00:00
470    Tag: leunga-20001110-new-x86-fp
471    
472    A new x86 floating point code generator has been added.
473    By default this is turned off.  To turn this on, do:
474    
475        CM.autoload "$smlnj/compiler.cm";
476        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
477    
478    Changes:
479    
480    1.  Changed FTAN to FPTAN so that the assembly output is correct.
481    2.  Changed the extension callback for FTANGENT to generate:
482    
483              fptan
484              fstp  %st(0)
485        instead of
486              fptan
487              fstpl ftempmem
488    
489    3.  Numerous assembly fixes for x86.
490    
491    5.  Cleaned up the machine code output module x86/x86MC.sml and added
492        support for a whole bunch of instructions and addressing modes:
493    
494          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
495          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
496          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
497          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
498          fxch %st(n)
499          fld %st(n)
500          fst %st(n)
501          fst mem
502          fstp %st(n)
503          fucom %st(n)
504          fucomp %st(n)
505    
506        All these are now generated when the fast fp mode is turned on.
507    
508    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
509    
510    ----------------------------------------------------------------------
511    Name: Matthias Blume
512    Date: 2000/11/09 11:20:00 JST
513    Tag: blume-20001109-scc
514    Description:
515    
516    Eliminated some code duplication:
517    
518    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
519       (Otherwise the functor is useless.)
520    2. Used GraphSCCFn where SCCUtilFun was used previously.
521    3. Got rid of SCCUtilFun (in comp-lib.cm).
522    
523    ----------------------------------------------------------------------
524    Name: Lal George
525    Date: 2000/11/06 09:02:21 EST 2000
526    Tag: Release_110_30
527    Description:
528    
529    - Version 110.30
530    ----------------------------------------------------------------------
531    Name: Matthias Blume
532    Date: 2000/11/04 14:45:00
533    Tag: blume-20001104-mlbuild
534    Description:
535    
536    - Made ml-build faster on startup.
537    - Documentation fixes.
538    
539    ----------------------------------------------------------------------
540    Name: Matthias Blume
541    Date: 2000/11/02 17:00:00 JST
542    Tag: blume-20001102-condcomp
543    Description:
544    
545    - Small tweaks to pickler -- new BOOTFILES!
546    - Version bumped to 110.29.2.
547    - Added conditional compilation facility to init.cmi (see comment there).
548    ----------------------------------------------------------------------
549    Name: Allen Leung
550    Date: 2000/10/23 19:31:00
551    Tag: leunga-20001023-demo-ra
552    
553    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
554    2. Test programs for the graph library updated
555    3. Some new MLRISC demo programs added
556    
557    ----------------------------------------------------------------------
558    Name: Matthias Blume
559    Date: 2000/08/31 22:15:00 JST
560    Tag: blume-20001017-errmsg
561    Description:
562    
563    More error message grief: Where there used to be no messages, there
564    now were some that had bogus error regions.  Fixed.
565    
566    ----------------------------------------------------------------------
567    Name: Matthias Blume
568    Date: 2000/08/31 17:30:00 JST
569    Tag: blume-20001017-v110p29p1
570    Description:
571    
572    I made a version 110.29.1 with new bootfiles.
573    
574    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
575              CM documentation changes and a small bugfix in CM's error reporting.
576    
577    ----------------------------------------------------------------------
578    Name: Lal George
579    Date: 2000/09/27 14:42:35 EDT
580    Tag: george-20000927-nodestatus
581    Description:
582    
583    Changed the type of the nodestatus, so that:
584    
585            SPILLED(~1)             is now SPILLED
586            SPILLED(m) where m>=0   is now MEMREG(m)
587            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
588    
589    ----------------------------------------------------------------------
590    Name: Matthias Blume
591    Date: 2000/09/07 14:45:00 JST
592    Tag: blume-20000907-cmerrmsg
593    Description:
594    
595    Small tweak to CM to avoid getting ML syntax error messages twice.
596    
597    ----------------------------------------------------------------------
598    Name: Matthias Blume
599    Date: 2000/08/31 18:00:00 JST
600    Tag: blume-20000831-cvsbootfiles
601    Description:
602    
603    New URL for boot files (because the 110.29 files on the BL server do
604    now work correctly with my updated install scripts for yacc and lex).
605    
606    ----------------------------------------------------------------------
607    Name: Matthias Blume
608    Date: 2000/08/08 12:33:00 JST
609    Tag: blume-20000808-manual
610    Description:
611    
612    Tiny update to CM manual.
613    
614    ----------------------------------------------------------------------
615    Name: Allen Leung
616    Date: 2000/08/7 19:31:00
617    Tag: leunga-20000807-a-whole-bunch-of-stuff
618    
619      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
620    and Mobdy.
621    
622    1.  x86
623    
624       a.  Fixes to peephole module by John and Dan.
625       b.  Assembly fix to SETcc by Allen.
626       c.  Fix to c-call by John.
627       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
628       e.  Instruction selection optimization to SETcc as suggested by John.
629    
630           For example,
631    
632            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
633    
634           should generate:
635    
636            MOVL a, x
637            SUBL b, x
638            SHRL 31, x
639    
640    2.  IR stuff
641    
642         A bunch of new DJ-graph related algorithms added.  These
643         speed up SSA construction.
644    
645    3.  SSA + Scheduling
646    
647         Added code for SSA and scheduling to the repository
648    
649    ----------------------------------------------------------------------
650    Name: Lal George
651    Date: 2000/07/27 11:53:14 EDT
652    
653    Tag: lal-20000727-linux-ppc
654    Description:
655    
656     Made changes to support Linux PPC.
657     p.s. I have confirmation that the 110.29 boot files work fine.
658    
659    ----------------------------------------------------------------------
660    Name: Matthias Blume
661    Date: 2000/07/27 17:40:00 JST
662    Tag: blume-20000727-scripts
663    Description:
664    
665    !!!! WARNING !!!!
666    You must recompile the runtime system!
667    !!!! WARNING !!!!
668    
669    This is basically another round of script-enhancements:
670    
671    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
672       and undefine CM preprocessor symbols.
673    
674    2. ml-build avoids generating a new heap image if it finds that the
675       existing one is still ok.  (The condition is that no ML file had to
676       be recompiled and all ML files are found to be older that the heap
677       file.)
678    
679       To make this work smoothly, I also hacked the runtime system as
680       well as SMLofNJ.SysInfo to get access to the heap image suffix
681       (.sparc-solaris, ...) that is currently being used.
682    
683       Moreover, the signature of CM.mk_standalone has changed.  See the
684       CM manual.
685    
686    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
687       CM manual for details.)
688    
689    4. More CM manual updates:
690        - all of the above has been documented.
691        - there is now a section describing the (CM-related) command line
692          arguments that are accepted by the "sml" command
693    
694    ----------------------------------------------------------------------
695    Name: Matthias Blume
696    Date: 2000/07/25 16:20:00 JST
697    Tag: blume-20000725-makedepend
698    Description:
699    
700    Added a script called ml-makedepend.  This can be used in makefiles
701    for Unix' make in a way very similar to the "makedepend" command for
702    C.
703    
704    The script internally uses function CM.sources.
705    
706    Synopsis:
707    
708        ml-makedepend [-f makefile] cmfile targetname
709    
710    The default for the makefile is "makefile" (or "Makefile" should
711    "makefile" not exist).
712    
713    ml-makedepend adds a cmfile/targetname-specific section to this
714    makefile (after removing the previous version of this section).  The
715    section contains a single dependency specification with targetname on
716    the LHS (targetname is an arbitrary name), and a list of files derived
717    from the cmfile on the RHS.  Some of the files on the RHS are
718    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
719    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
720    path names.  The makefile writer is responsible for making sure that
721    these variables have correct at the time "make" is invoked.
722    
723    ----------------------------------------------------------------------
724    Name: Matthias Blume
725    Date: 2000/07/22 23:30:00 JST
726    Tag: blume-20000722-urlupdate
727    Description:
728    
729    Changed BOOT and config/srcarchiveurl to point to BL server:
730    
731        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
732    
733    ----------------------------------------------------------------------
734    Name: Matthias Blume
735    Date: 2000/07/18 18:00:00 JST
736    Tag: blume-20000718-Version_110_29
737    Description:
738    
739    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
740    
741    2. Updated config/version to 110.29
742    
743    3. Updated config/srcarchiveurl
744    
745    3. New boot files!
746       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
747    
748    ----------------------------------------------------------------------
749    Name: Matthias Blume
750    Date: 2000/07/11 13:58:00 JST
751    Tag: blume-20000711-doctypo
752    Description:
753    
754    Fixed a few typos in CM manual.
755    
756    ----------------------------------------------------------------------
757    Name: Allen Leung
758    Date: 2000/06/15 00:38:00
759    Tag: leunga-20000704-sparc-x86
760    
761    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
762    2. fix to x86 RET bug [found by Dan Grossman]
763    3. sparc assembly bug fix for ticc instructions [found by Fermin]
764    
765       Affects c-- and moby only
766    
767    ----------------------------------------------------------------------
768    Name: Matthias Blume
769    Date: 2000/07/04 15:26:00
770    Tag: blume-20000704-trigger
771    Description:
772    
773    1. Improvements to CM manual.
774    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
775       of getting a back-trace.  The function, when called, raises an
776       internal exception which explicitly carries the full back-trace history,
777       so it is unaffected by any intervening handle-raise pairs ("trivial"
778       or not).  The interactive loop will print that history once it arrives
779       at top level.
780       Short of having all exceptions implicitly carry the full history, the
781       recommended way of using this facility is:
782         - compile your program with instrumentation "on"
783         - run it, when it raises an exception, look at the history
784         - if the history is "cut off" because of some handler, go and modify
785           your program so that it explicitly calls BTrace.trigger
786         - recompile (still instrumented), and rerun; look at the full history
787    
788    ----------------------------------------------------------------------
789    Name: Matthias Blume
790    Date: 2000/07/03 15:36:00 JST
791    Tag: blume-20000702-manual
792    Description:
793    
794    Small corrections and updates to CM manual.
795    
796    ----------------------------------------------------------------------
797    Name: Matthias Blume
798    Date: 2000/06/29 16:04:00 JST
799    Tag: blume-20000629-yacctool
800    Description:
801    
802    Changes:
803    
804    1. Class "mlyacc" now takes separate arguments to pass options to
805       generated .sml- and .sig-files independently.
806    2. Corresponding CM manual updates.
807    3. BTrace module now also reports call sites.  (However, for loop clusters
808       it only shows from where the cluster was entered.)  There are associated
809       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
810    
811    ----------------------------------------------------------------------
812    Name: Matthias Blume
813    Date: 2000/06/27 16:51:00 JST
814    Tag: blume-20000627-noweb
815    Description:
816    
817    Changes:
818    
819     1. Implemented "subdir" and "witness" options for noweb tool.
820        This caused some slight internal changes in CM's tool implementation.
821     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
822        some remaining issues from earlier path anchor changes.
823     3. Updated CM manual accordingly.
824    
825     4. Changed implementation of back-tracing so that I now consider it
826        ready for prime-time.
827    
828        In particular, you don't have to explicitly trigger the back-trace
829        anymore.  Instead, if you are running BTrace-instrumented code and
830        there is an uncaught exception (regardless of whether or not it was
831        raised in instrumented code), the top-level evalloop will print
832        the back-trace.
833    
834        Features:
835    
836          - Instrumented and uninstrumented code work together seemlessly.
837            (Of course, uninstrumented code is never mentioned in actual
838             back-traces.)
839    
840          - Asymptotic time- and space-complexity of instrumented code is
841            equal to that of uninstrumented code.  (This means that
842            tail-recursion is preserved by the instrumentation phase.)
843    
844          - Modules whose code has been instrumented in different sessions
845            work together without problem.
846    
847          - There is no penalty whatsoever on uninstrumented code.
848    
849          - There is no penalty on "raise" expressions, even in
850            instrumented code.
851    
852        A potential bug (or perhaps it is a feature, too):
853    
854          A back-trace reaches no further than the outermost instrumented
855          non-trivial "raise".  Here, a "trivial" raise is one that is the
856          sole RHS of a "handle" rule.  Thus, back-traces reach trough
857    
858               <exp> handle e => raise e
859    
860          and even
861    
862               <exp> handle Foo => raise Bar
863    
864          and, of course, through
865    
866               <exp> handle Foo => ...
867    
868         if the exception was not Foo.
869    
870         Back-traces always reach right through any un-instrumented code
871         including any of its "handle" expressions, trivial or not.
872    
873       To try this out, do the following:
874    
875         - Erase all existing binfiles for your program.
876           (You may keep binfiles for those modules where you think you
877            definitely don't need back-tracing.)
878         - Turn on back-trace instrumentation:
879              SMLofNJ.Internals.BTrace.mode (SOME true);
880         - Recompile your program.  (I.e., run "CM.make" or "use".)
881         - You may now turn instrumentation off again (if you want):
882              SMLofNJ.Internals.BTrace.mode (SOME false);
883         - Run your program as usual.  If it raises an exception that
884           reaches the interactive toplevel, then a back-trace will
885           automatically be printed.  After that, the toplevel loop
886           will print the exception history as usual.
887    
888    ----------------------------------------------------------------------
889    Name: Matthias Blume
890    Date: 2000/06/26 09:56:46 JST
891    Tag: blume-20000626-setup
892    Description:
893    
894    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
895          ML code before and after compiling a file (e.g., to set compiler
896          flags)
897    
898    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
899              - associated changes to btrace.sml (BTrace instrumentation pass)
900              - cleaner implementation of btimp.sml (BTrace tracing and report
901                module)
902    
903    CM manual: * new path encoding documented
904               * description of setup-parameter to "sml" added
905    
906    The biggest user-visible change to back-tracing is that it is no
907    longer necessary to compile all traced modules within the same
908    session.  (This was a real limitation.)
909    
910    ----------------------------------------------------------------------
911    Name: Matthias Blume
912    Date: 2000/06/24 12:40:00 JST
913    Tag: blume-20000624-startup
914    Description:
915    
916    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
917    bit too often -- to put it mildly. :)
918    
919    ----------------------------------------------------------------------
920    Name: Matthias Blume
921    Date: 2000/06/23 18:20:00 JST
922    Tag: blume-20000623-btrace
923    Description:
924    
925    This updates adds a backtrace facility to aid programmers in debugging
926    their programs.  This involves the following changes:
927    
928    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
929       keeping track of the current call stack.  When programs are compiled
930       in a special mode, the compiler will insert calls to these hooks
931       into the user program.
932       "Hook" means that it is possible for different implementations of
933       back-tracing to register themselves (at different times).
934    
935    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
936       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
937       It can be turned on using this call:
938         SMLofNJ.Internals.BTrace.mode (SOME true);
939       Turning it off again:
940         SMLofNJ.Internals.BTrace.mode (SOME false);
941       Querying the current status:
942         SMLofNJ.Internals.BTrace.mode NONE;
943       Annotated programs are about twice as big as normal ones, and they
944       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
945       where all hooks do nothing).  The slowdown with a plugin that is
946       actually useful (such as the one supplied by default) is even greater,
947       but in the case of the default plugin it is still only an constant
948       factor (amortized).
949    
950    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
951       with a sub-structure BTrace for controlling back-tracing.  In particular,
952       the above-mentioned function "mode" controls whether the annotation
953       phase is invoked by the compiler.  Another important function is
954       "trigger": when called it aborts the current execution and causes
955       the top-level loop to print a full back-trace.
956    
957    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
958       for back-tracing.  It keeps track of the dynamic call stack and in
959       addition to that it keeps a partial history at each "level" of that
960       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
961       d and b (at separate times, dynamically), then the report will show:
962    
963       GOTO   d
964             /c
965       GOTO  \b
966       CALL   a
967    
968       This shows that there was an initial non-tail call of a, then a
969       tail-call to b or c, looping behavior in a cluster of functions that
970       consist of b and c, and then a goto from that cluster (i.e., either from
971       b or from c) to d.
972    
973       Note that (depending on the user program) the amount of information
974       that the back-trace module has to keep track of at each level is bounded
975       by a constant.  Thus, the whole implementation has the same asymptotical
976       complexity as the original program (both in space and in time).
977    
978    5. compiler/TopLevel/interact/evalloop.sml has been modified to
979       handle the special exception SMLofNJ.Internals.BTrace.BTrace
980       which is raised by the "trigger" function mentioned above.
981    
982    Notes on usage:
983    
984    - Annotated code works well together with unannotated code:
985    Unannotated calls simply do not show up at all in the backtrace.
986    
987    - It is not a good idea to let modules that were annotated during
988    different sessions run at the same time.  This is because the compiler
989    chooses small integers to identify individual functions, and there
990    will be clashes if different modules were compiled in separate sessions.
991    (Nothing will crash, and you will even be told about the clashes, but
992    back-trace information will in general not be useful.)
993    
994    - Back-tracing can be confused by callcc and capture.
995    
996    - The only way of getting a back-trace right now is to explicitly
997    invoke the "trigger" function from your user program.  Eventually, we
998    should make every exception carry back-trace information (if
999    available).  But since this creates more overhead at "raise"-time
1000    (similar to the current exnHistory overhead), I have not yet
1001    implemented this.  (The implementation will be rather easy.)  With
1002    exceptions carrying back-trace information, this facility will be even
1003    more useful because users don't need to modify their programs...
1004    
1005    - While it is possible to compile the compiler with back-trace
1006    annotations turned on (I did it to get some confidence in
1007    correctness), you must make absolutely sure that core.sml and
1008    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1009    be compiled with annotation because there is no core access yet, but
1010    if you compile btimp.sml with annotation, then the system will go into
1011    an infinite recursion and crash.)
1012    Since CM currently does not know about BTrace, the only way to turn
1013    annotations on and off for different modules of the compiler is to
1014    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1015    this is awkward and clumsy.
1016    
1017    Sample sessions:
1018    
1019    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1020    - SMLofNJ.Internals.BTrace.mode (SOME true);
1021    [autoloading]
1022    [autoloading done]
1023    val it = false : bool
1024    - structure X = struct
1025    -     fun main n = let
1026    -         fun a (x, 0) = d x
1027    -           | a (x, n) = b (x, n - 1)
1028    -         and b (x, n) = c (x, n)
1029    -         and c (x, n) = a (x, n)
1030    -         and d x = e (x, 3)
1031    -         and e (x, 0) = f x
1032    -           | e (x, n) = e (x, n - 1)
1033    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1034    -           | f n = n * g (n - 1)
1035    -         and g n = a (n, 3)
1036    -     in
1037    -         f n
1038    -     end
1039    - end;
1040    structure X : sig val main : int -> int end
1041    - X.main 3;
1042    *** BACK-TRACE ***
1043    GOTO   stdIn:4.2-13.20: X.main[2].f
1044    GOTO-( stdIn:4.2-13.20: X.main[2].e
1045    GOTO   stdIn:4.2-13.20: X.main[2].d
1046         / stdIn:4.2-13.20: X.main[2].a
1047         | stdIn:4.2-13.20: X.main[2].b
1048    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1049    CALL   stdIn:4.2-13.20: X.main[2].g
1050    GOTO   stdIn:4.2-13.20: X.main[2].f
1051    GOTO-( stdIn:4.2-13.20: X.main[2].e
1052    GOTO   stdIn:4.2-13.20: X.main[2].d
1053         / stdIn:4.2-13.20: X.main[2].a
1054         | stdIn:4.2-13.20: X.main[2].b
1055    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1056    CALL   stdIn:4.2-13.20: X.main[2].g
1057    GOTO   stdIn:4.2-13.20: X.main[2].f
1058    GOTO-( stdIn:4.2-13.20: X.main[2].e
1059    GOTO   stdIn:4.2-13.20: X.main[2].d
1060         / stdIn:4.2-13.20: X.main[2].a
1061         | stdIn:4.2-13.20: X.main[2].b
1062    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1063    CALL   stdIn:4.2-13.20: X.main[2].g
1064    GOTO   stdIn:4.2-13.20: X.main[2].f
1065    CALL   stdIn:2.15-17.4: X.main[2]
1066    -
1067    
1068    (Note that because of a FLINt bug the above code currently does not
1069    compile without BTrace turned on.)
1070    
1071    Here is another example, using my modified Tiger compiler:
1072    
1073    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1074    - SMLofNJ.Internals.BTrace.mode (SOME true);
1075    [autoloading]
1076    [autoloading done]
1077    val it = false : bool
1078    - CM.make "sources.cm";
1079    [autoloading]
1080    ...
1081    [autoloading done]
1082    [scanning sources.cm]
1083    [parsing (sources.cm):parse.sml]
1084    [creating directory CM/SKEL ...]
1085    [parsing (sources.cm):tiger.lex.sml]
1086    ...
1087    [wrote CM/sparc-unix/semant.sml]
1088    [compiling (sources.cm):main.sml]
1089    [wrote CM/sparc-unix/main.sml]
1090    [New bindings added.]
1091    val it = true : bool
1092    - Main.compile ("../testcases/merge.tig", "foo.out");
1093    *** BACK-TRACE ***
1094    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1095    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1096    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1097    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1098    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1099    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1100    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1101         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1102    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1103    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1104    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1105    CALL   main.sml:1.18-118.4: Main.compile[2]
1106    -
1107    
1108    ----------------------------------------------------------------------
1109    Name: Matthias Blumen
1110    Date: 2000/06/21 18:00:00 JST
1111    Tag: blume-20000621-manual
1112    Description:
1113    
1114    CM manual update: Path environments documented.
1115    
1116    ----------------------------------------------------------------------
1117    Name: Matthias Blume
1118    Date: 2000/06/19 13:40:00
1119    Tag: blume-20000619-manual
1120    Description:
1121    
1122    CM manual and system/README update.  This only covers the fact that
1123    there are no more implicit anchors.  (Path environments and the "bind"
1124    option to "cm" have yet to be documented.)
1125    
1126    ----------------------------------------------------------------------
1127    Name: Matthias Blume
1128    Date: 2000/06/19 11:05:00 JST
1129    Tag: blume-20000619-chdir-bugfix
1130    Description:
1131    
1132    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1133    
1134    ----------------------------------------------------------------------
1135    Name: Matthias Blume
1136    Date: 2000/06/18 22:00:10 JST
1137    Tag: blume-20000618-implicit-anchors-really-gone
1138    Description:
1139    
1140    I updates the previous HISTORY entry where I forgot to mention that
1141    implicit anchors are no longer with us.
1142    
1143    The current update also gets rid of the (now useless) controller
1144    CM.Control.implicit_anchors.
1145    
1146    ----------------------------------------------------------------------
1147    Name: Matthias Blume
1148    Date: 2000/06/16 17:30:00 JST
1149    Tag: blume-20000616-anchorenv
1150    Description:
1151    
1152    This patch implements the long anticipated (just kidding :) "anchor
1153    environment" mechanism.  In the course of doing this, I also
1154    re-implemented CM's internal "SrcPath" module from scratch.  The new
1155    one should be more robust in certain boundary cases.  In any case, it
1156    is a lot cleaner than its predecessor (IMHO).
1157    
1158    This time, although there is yet another boot file format change, I
1159    kept the unpickler backward-compatible.  As a result, no new bootfiles
1160    are necessary and bootstrapping is straightforward.  (You cannot read
1161    new bootfiles into an old system, but the other way around is no
1162    problem.)
1163    
1164    Visible changes:
1165    
1166    ** 0. Implicit path anchors (without the leading $-symbol) are no
1167    longer recognized at all. This means that such path names are not
1168    illegal either.  For example, the name basis.cm simply refers to a
1169    local file called "basis.cm" (i.e, the name is an ordinary path
1170    relative to .cm-files directory).  Or, to put it differently, only
1171    names that start with $ are anchored paths.
1172    
1173    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1174    vanished.
1175    
1176    John (Reppy) had critizised this as soon as I originally proposed and
1177    implemented it, but at that time I did not really deeply believe
1178    him. :) Now I came full-circle because I need the $<singlearc> syntax
1179    in another place where it cannot be seen as an abbreviation for
1180    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1181    seems to mean (i.e., it "expands" into the corresponding anchor
1182    value).
1183    
1184    However, when paths are used as members in CM description files, it
1185    continues to be true that there must be at least another arc after the
1186    anchor.  This is now enforced separately during semantic analysis
1187    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1188    
1189    ** 2. The "cm" class now accepts an option "bind".  The option's value
1190    is a sub-option list of precisely two items -- one labeled "anchor"
1191    and the other one labeled "value".  As you might expect, "anchor" is
1192    used to specify an anchor name to be bound, and "value" specifies what
1193    the anchor is being bound to.
1194    
1195    The value must be a directory name and can be given in either standard
1196    syntax (including the possibility that it is itself an anchored path)
1197    or native syntax.
1198    
1199    Examples:
1200    
1201       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1202       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1203    
1204    and so on.
1205    
1206    The meaning of this is that the .cm-file will be processed with an
1207    augmented anchor environment where the given anchor(s) is/are bound to
1208    the given values(s).
1209    
1210    The rationale for having this feature is this: Suppose you are trying
1211    to use two different (already stable) libraries a.cm and b.cm (that
1212    you perhaps didn't write yourself).  Further, suppose each of these
1213    two libraries internally uses its own auxiliary library $aux/lib.cm.
1214    Normally you would now have a problem because the anchor "lib" can not
1215    be bound to more than one value globally.  Therefore, the project that
1216    uses both a.cm and b.cm must locally redirect the anchor to some other
1217    place:
1218    
1219       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1220       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1221    
1222    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1223    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1224    
1225    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1226    you when you do so at the time of CM.stabilize).  Therefore, you can
1227    also use an anchored path as the value:
1228    
1229      a.cm (bind:(anchor:lib value:$a-lib))
1230      b.cm (bind:(anchor:lib value:$b-lib))
1231    
1232    Now you can globally configure (using the usual CM.Anchor.anchor or
1233    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1234    itself is always locally bound, setting it globally is no longer
1235    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1236    can still be used as a global anchor for separate purposes.  As a
1237    matter of fact, one can locally define "lib" in terms of a global
1238    "lib":
1239    
1240      a.cm (bind:(anchor:lib value:$lib/a))
1241      b.cm (bind:(anchor:lib value:$lib/b))
1242    
1243    ** 3: The encoding of path names has changed.  This affects the way
1244    path names are shown in CM's progress report and also the internal
1245    protocol encoding used for parallel make.
1246    
1247    The encoding now uses one or more ':'-separated segments.  Each
1248    segments corresponds to a file that has been specified relative to the
1249    file given by its preceding segment.  The first segment is either
1250    relative to the CWD, absolute, or anchored.  Each segment itself is
1251    basically a Unix pathname; all segments but the first are relative.
1252    
1253    Example:
1254    
1255       $foo/bar/baz.cm:a/b/c.sml
1256    
1257    This path denotes the file bar/a/b/c.sml relative to the directory
1258    denoted by anchor "foo".  Notice that the encoding also includes
1259    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1260    paths are resolved relative to the .cm-files directory, so baz.cm must
1261    be ignored to get the "real" pathname.
1262    
1263    To make this fact more obvious, CM puts the names of such "virtual
1264    arcs" into parentheses when they appear in progress reports. (No
1265    parentheses will appear in the internal protocol encoding.)  Thus,
1266    what you really see is:
1267    
1268      $foo/bar/(baz.cm):a/b/c.sml
1269    
1270    I find this notation to be much more informative than before.
1271    
1272    Another new feature of the encoding is that special characters
1273    including parentheses, colons, (back)slashes, and white space are
1274    written as \ddd (where ddd is the decimal encoding of the character).
1275    
1276    *** The CM manual still needs to be updated.
1277    
1278    ----------------------------------------------------------------------
1279    Name: Allen Leung
1280    Date: 2000/06/15 00:38:00
1281    Tag: leunga-20000615-x86-peephole
1282    
1283    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1284    
1285    ----------------------------------------------------------------------
1286    Name: Matthias Blume
1287    Date: 2000/06/12 11:40:00
1288    Tag: blume-20000612-parmakefix
1289    Description:
1290    
1291    More cleanup after changing the file naming scheme: This time I
1292    repaired the parallel make mechanism for CMB.make which I broke earlier.
1293    
1294    ----------------------------------------------------------------------
1295    Name: Allen Leung
1296    Date: 2000/06/09 01:25:00
1297    Tag: leunga-20000609-various
1298    
1299    None of these things should affect normal SML/NJ operations
1300    
1301    1. Peephole improvements provided by Fermin (c--)
1302    2. New annotation DEFUSE for adding extra dependence (moby)
1303    3. New X86 LOCK instructions (moby)
1304    4. New machine description language for reservation tables (scheduling)
1305    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1306       trees etc.)
1307    6. I've changed the CM files so that they can work with versions
1308       110.0.6, 110.25 and 110.28
1309    
1310    ----------------------------------------------------------------------
1311    Name: Matthias Blume
1312    Date: 2000/06/09 12:40:00
1313    Tag: blume-20000609-log
1314    Description:
1315    
1316    - Removed all(?) remaining RCS Log entries from sources.
1317    
1318    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1319      anchored paths).
1320    
1321    ----------------------------------------------------------------------
1322    Name: Matthias Blume
1323    Date: 2000/06/07 17:00:00 JST
1324    Tag: blume-20000607-no-implicit-anchors
1325    Description:
1326    
1327    1. This update changes the default setting for
1328    CM.Control.implicit_anchors from true to false.  This means that
1329    implicit anchors are no longer permitted by default.  I also tried to
1330    make sure that nothing else still relies on implicit anchors.
1331    (This is the next step on the schedule towards a CM that does not even
1332    have the notion of implicit anchors anymore.)
1333    
1334    2. More CM manual updates.
1335    
1336    3. I managed to track down and fix the pickling bug I mentioned last
1337    time.  Because of the previously existing workaround, this entails no
1338    immediate practical changes.
1339    
1340    ----------------------------------------------------------------------
1341    Name: Matthias Blume
1342    Date: 2000/06/06 11:15:00 JST
1343    Tag: blume-20000606-lazierpickle
1344    Description:
1345    
1346    !!!! NEW BOOT FILES !!!!
1347    
1348    * The main purpose of this update is to make library pickles lazier in
1349    order to reduce the initial space penalty for autoloading a library.
1350    As a result, it is now possible to have $smlnj/compiler.cm
1351    pre-registered.  This should take care of the many complaints or
1352    inquiries about missing structure Compiler.  This required changes to
1353    CM's internal data structures and small tweaks to some algorithms.
1354    
1355    As a neat additional effect, it is no longer necessary (for the sake
1356    of lean heap image files) to distinguish between a "minimal" CM and a
1357    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1358    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1359    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1360    
1361    To make the life of compiler-hackers easier, "makeml" now also
1362    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1363    words, after you bootstrap a new sml for the first time, you will not
1364    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1365    around you will still have to do it, though.)
1366    
1367    * A second change consists of major updates to the CM manual.  There
1368    are now several appendices with summary information and also a full
1369    specification of the CM description file syntax.
1370    
1371    * In directory src/system I added the script "allcross".  This script
1372    invokes sml and cross-compiles the compiler for all supported
1373    architectures.  (Useful when providing a new set of boot files.)
1374    
1375    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1376    added a small tweak to pickle-util.sml to work around this problem,
1377    but it is not a proper fix yet.  I will investigate further.  (The
1378    effect of the bug was an inflation of library pickle size.)
1379    
1380    * Version number increased to 110.28.1 (to avoid compatibility problems).
1381    
1382    ----------------------------------------------------------------------
1383    Name: Allen Leung
1384    Date: 2000/05/25 17:28 EDT
1385    Tag: leunga-20000525-ra
1386    Description:
1387    
1388      Fixed a bug in freezing phase of the register allocator.
1389    
1390    ----------------------------------------------------------------------
1391    Name: Allen Leung
1392    Date: 2000/05/15 22:53 EDT
1393    Tag: leunga-20000515-alpha-x86-ra
1394    Description:
1395    
1396      1. Alpha
1397    
1398          Slight cleanup.  Removed the instruction SGNXL
1399    
1400      2. X86
1401    
1402          Added the following instructions to the instruction set:
1403    
1404            ROLx, RORx,
1405            BTx, BTSx, BTLx, BTRx,
1406            XCHGx, and variants with the LOCK prefix
1407    
1408      3. Register Allocation
1409    
1410          The module ra-rewrite-with-renaming has been improved.
1411    
1412      These have no effect on SML/NJ.
1413    
1414    ----------------------------------------------------------------------
1415    Name: Matthias Blume
1416    Date: 2000/05/15 16:20:00 JST
1417    Tag: blume-20000515-lightrebuild
1418    Description:
1419    
1420    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1421       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1422       defined.  In effect, the command will not build any cross-compiler
1423       backends and therefore finish more quickly.
1424    
1425       The "fixpt" script also takes a "-light" switch to be able to use
1426       this new facility while compiling for a fixpoint.
1427    
1428    2. I replaced all mentions of anchored paths in group owner specifications
1429       with simple relative paths (usually starting with "..").
1430       The rationale is that a library's internal workings should not be
1431       compromised by the lack of some anchor.  (An anchor is necessary
1432       for someone who wants to refer to the library by an anchored path,
1433       but it should not be necessary to build the same library in the first
1434       place.)
1435    
1436    3. I changed the way CM's tool mechanism determines the shell command
1437       string used for things like ml-yacc etc. so that it does not break
1438       when CM.Control.implicit_anchors is turned off.
1439    
1440    ----------------------------------------------------------------------
1441    Name: Matthias Blume
1442    Date: 2000/05/12 18:20:00 JST
1443    Tag: blume-20000512-ml-build
1444    Description:
1445    
1446    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1447    getting installed properly (by config/install.sh).
1448    
1449    ----------------------------------------------------------------------
1450    Name: Matthias Blume
1451    Date: 2000/05/12 17:30:00 JST
1452    Tag: blume-20000512-anchors
1453    Description:
1454    
1455    !!! NEW BOOT FILES !!!
1456    
1457    This change is in preparation of fading out support for "implicitly
1458    anchored path names".  I went through all sources and used the
1459    explicit (and relatively new) $-notation.  See system/README and the
1460    CM manual for more info on this.
1461    
1462    I also modified the anchoring scheme for some things such as "smlnj",
1463    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1464    anchors are more expressive: anchor name and first arc do not have to
1465    coincide.  This entails the following user-visible change:
1466    
1467    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1468    particular, when you fire up sml with a command-line argument, say,
1469    e.g.:
1470    
1471       sml '$smlnj/cmb.cm'
1472    
1473    At the ML toplevel prompt:
1474    
1475       CM.autoload "$smlnj/cmb.cm";
1476    
1477    There is also a new controller in CM.Control that can be used to turn
1478    off all remaining support for implicit anchors by saying:
1479    
1480        CM.autoload "$smlnj/
1481        #set CM.Control.implicit_anchors false;
1482    
1483    This causes CM to reject implicitly anchored paths.  This is (for the
1484    time being) less permissive than the "final" version where there will
1485    be no more such implicit anchors and relative paths will be just that:
1486    relative.
1487    
1488    The next step (version after next version?) will be to make the
1489    default for CM.Control.implicit_anchors false.  After the dust has
1490    settled, I can then produce the "final" version of this...
1491    
1492    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1493    
1494    ----------------------------------------------------------------------
1495    Name: Matthias Blume
1496    Date: 2000/05/11 16:30:00 JST
1497    Tag: blume-20000511-sources
1498    Description:
1499    
1500    The main change is that I added function CM.sources as a generalized
1501    version of the earlier CM.makedepend.  This entails the following
1502    additional changes:
1503    
1504      - CM.makedepend has been dropped.
1505    
1506      - CM manual has been updated.
1507    
1508      - TOOLS signature and API have been changed.
1509    
1510    ----------------------------------------------------------------------
1511    Name: Allen Leung
1512    Date: 2000/05/10 21:17 EDT
1513    Tag: leunga-20000510-moby-c--ssa
1514    Description:
1515    
1516      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1517    None of these affect SML/NJ.
1518    
1519    1. Register Allocation
1520    
1521        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1522           This module tries to remove local (i.e. basic block level) redundancies
1523           during spilling.
1524    
1525        b. A new framework for performing region based register allocation.
1526           Not yet entirely functional.
1527    
1528    2. X86
1529    
1530       a. DefUse for POP was missing the stack pointer [found by Lal]
1531       b. Reload for CALL was incorrect in X86Spill [found by John]
1532       c. Various fixes in X86Spill so that it can be used correctly for
1533          the new spilling module.
1534    
1535    3. SSA/IR
1536    
1537       a. New module ir/dj-dataflow.sml implements elimination based
1538          data flow analysis.
1539    
1540    4. MLRiscGen
1541    
1542       a. Fix for gc type annotation
1543    
1544    5. MDGen
1545    
1546       Various fixes for machine description -> ml code translation.  For ssa
1547       only.
1548    
1549    ----------------------------------------------------------------------
1550    Name: Allen Leung
1551    Date: 2000/05/08 22:17 EDT
1552    Tag: leunga-20000508-labexp
1553    Description:
1554    
1555      Fermin has found a few assembly problems with constant expressions
1556      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1557      which choke on dumb assemblers.  This is his fix.
1558    
1559    ----------------------------------------------------------------------
1560    Name: Dave MacQueen
1561    Date: 2000/04/09 14:00 EDT
1562    Tag: dbm-20000502-Version_110_28
1563    Description:
1564    
1565    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1566    
1567    2. Updated config/version to 110.28
1568    
1569    3. Updated config/srcarchiveurl
1570    
1571    3. New boot files!
1572       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1573    
1574    ----------------------------------------------------------------------
1575    Name: Matthias Blume
1576    Date: 2000/05/01 19:05:00 JST
1577    Tag: blume-20000501-noweb
1578    Description:
1579    
1580    A new noweb tool has been added.  The existing system is entirely
1581    unaffected by this, but some CM users have asked for renewed noweb
1582    support.  Everything is documented in the CM manual.
1583    
1584    New (plugin) libraries:
1585    
1586       noweb-tool.cm
1587       nw-ext.cm
1588    
1589    ----------------------------------------------------------------------
1590    Name: Dave MacQueen
1591    Date: 2000/04/30 12:40PM EDT
1592    Tag: dbm-20000430-bug_fixes
1593    Description:
1594    
1595    1. Fix for bug 1498
1596       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1597       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1598         added toRealArray function
1599       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1600         added check for tag Obj.RealArray to array printing case in ppObj
1601    
1602    2. Fix for bug 1510
1603       smlnj/src/compiler/Semant/types/typesutil.sml
1604         fixed definition of dummyargs (used by equalTycon) so that
1605         dummy args are distinct types
1606    
1607    ----------------------------------------------------------------------
1608    Name: Matthias Blume
1609    Date: 2000/04/30 01:00:00 JST
1610    Tag: blume-20000430-versions
1611    Description:
1612    
1613    1. CM version numbering added.  This is an implementation of Lal's
1614       proposal for adding version numbers and version checking to .cm
1615       files.  Lal said that his proposal was just that -- a proposal.
1616       For the time being I went ahead and implemented it so that people
1617       can comment on it.  Everything is completely backward-compatible
1618       (except for the stable library format, i.e., new bootfiles!).
1619    
1620       As usual, see the CM manual for details.
1621    
1622    2. An alternative syntax for anchored paths has been implemented.
1623       Dave has recently voiced the same concerns that I had when I did
1624       this, so there should be some support.  My take is that eventually
1625       I will let support for the current syntax (where anchors are
1626       "implicit") fade out in favor of the new, explicit syntax.
1627       In order to be backward-compatible, both old and new syntax are
1628       currently supported.
1629    
1630       Again, see the CM manual for details.
1631    
1632    3. Parallel make is trying to be slightly smarter:  When the master
1633       process finds a "bottleneck", i.e., when there is only one
1634       compilation unit that can be compiled and everybody else is
1635       waiting on it, then it will simply compile it directly instead
1636       of clumsily telling one of the slaves to do it.
1637    
1638    4. Support for "unsharing" added.  This is necessary in order to be
1639       able to have two different versions of the same library running
1640       at the same time (e.g., for trying out a new MLRISC while still
1641       having the old MLRISC linked into the current compiler, etc.)
1642       See the CM manual.
1643    
1644    5. Simple "makedepend" functionality added for generating Makefile
1645       dependency information.  (This is rather crude at the moment.
1646       Expect some changes here in the future.)
1647    
1648    6. ".fun" added as a recognized suffix for ML files. Also documented
1649       explicitly in the manual that the fallback behavior (unknown suffix
1650       -> ML file) is not an official feature!
1651    
1652    7. Small changes to the pickler for stable libraries.
1653    
1654    8. Several internal changes to CM (for cleanup/improvement).
1655    
1656    
1657    !!!! NEW BINFILES !!!!
1658    
1659    ----------------------------------------------------------------------
1660    Name: Matthias Blume
1661    Date: 2000/04/28 17:30:00 JST
1662    Tag: blume-20000428-pathconfig
1663    Description:
1664    
1665    1. I changed config/install.sh to remove duplicate entries from the
1666       lib/pathconfig file at the end.  Moreover, the final version of
1667       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1668       in src/system/installml.
1669    
1670    2. The config/install.sh script now consistently uses relative
1671       pathnames in lib/pathconfig whenever the anchor is in the lib
1672       directory.  (So far this was true for the libraries that come
1673       pre-compiled and bundled as part of the bootfiles but not for
1674       libraries that are compiled by the script itself.)
1675    
1676    ----------------------------------------------------------------------
1677    Name: Matthias Blume
1678    Date: 2000/04/26 13:10:00 JST
1679    Tag: blume-20000426-fun_suffix
1680    Description:
1681    
1682    Added ".fun" as a recognized file name suffix (for ML code).
1683    
1684    ----------------------------------------------------------------------
1685    Name: Allen Leung
1686    Date: 2000/04/25 17:00:00 EST
1687    Tag: leunga-20000425-alpha-ra
1688    Description:
1689    
1690    1. Alpha
1691    
1692        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1693    in C--.
1694    
1695    2. RA
1696    
1697       Added a flag "ra-dump-size" to print out the size of the flowgraph
1698       and the interference graph.
1699    
1700    ----------------------------------------------------------------------
1701    Name: Dave MacQueen
1702    Date: 2000/04/25/
1703    Tag: dbm-20000425-mlyacc_doc_examples
1704    Description:
1705      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1706      Updated all three examples in src/ml-yacc/examples to run
1707      under 110.* using CM.make.
1708    
1709    ----------------------------------------------------------------------
1710    Name: Allen Leung
1711    Date: 2000/04/20 23:04:00 EST
1712    Tag: leunga-20000420-ssa-c---stuff
1713    Description:
1714    
1715      This update synchronizes my repository with Yale's.  Most of these
1716    changes, however, do not affect SML/NJ at all (the RA is an exception).
1717    
1718    1. Register Allocator
1719    
1720       a. An improvement in the interference graph construction:
1721          Given a copy
1722    
1723                s <- t
1724    
1725          no interference edge between s and t is added for this definition of s.
1726    
1727       b. I've added two new spill heuristic modules that Fermin and I developed
1728          (in the new library RA.cm). These are unused in SML/NJ but maybe
1729          useful for others (Moby?)
1730    
1731    2. X86
1732    
1733       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1734    
1735    3. Alpha
1736    
1737       a. Added the BSR instruction and code generation that goes with it [C--]
1738       b. Other fixes too numerous to recount provided by Fermin [C--]
1739    
1740    4. Regmaps
1741    
1742       a. The regmaps are not initialized with the identity physical bindings
1743          at creation time.  This is unneeded.
1744    
1745    5. MLRISC Optimizations
1746    
1747       a. The DJ-Graph module can now compute the iterated dominance frontiers
1748          intersects with liveness incrementally in linear time! Woohoo!
1749          This is now used in my new SSA construction algorithm.
1750    
1751       b. THe branch reorganization module is now smarter about linear chains of
1752          basic blocks.
1753    
1754    
1755    ----------------------------------------------------------------------
1756    Name: Matthias Blume
1757    Date: 2000/04/12 13:52:00 JST
1758    Tag: blume_main_v110p27_1
1759    Description:
1760    
1761    Changed install.sh script to handle archive files without version number
1762    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1763    name of the boot file archive.
1764    
1765    ----------------------------------------------------------------------
1766    Name: Dave MacQueen
1767    Date: 2000/04/09 14:00 EDT
1768    Tag: dbm-20000410-Version_110_27
1769    Description:
1770    
1771    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1772    
1773    2. Updated src/config/version to 110.27
1774    
1775    3. New boot files!
1776    
1777    ----------------------------------------------------------------------
1778    Name: Allen Leung
1779    Date: 2000/04/09 19:09:00 EST
1780    Tag: leunga-20000409-misc
1781    Description:
1782    
1783    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
1784    
1785    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
1786    
1787    ----------------------------------------------------------------------
1788    Name: Stefan
1789    Date: 2000/04/07 10:00:00 EDT
1790    Tag: monnier-20000406-branch-handling
1791    Description:
1792    
1793    Improved handling of branches (mostly those generated from
1794    polymorphic equality), removed switchoff and changed the
1795    default optimization settings (more cpsopt and less flintopt).
1796    
1797    ----------------------------------------------------------------------
1798    Name: Allen Leung
1799    Date: 2000/04/06 01:30:00 EST
1800    Tag: leunga-20000406-peephole-x86-SSA-2
1801    Description:
1802    
1803       Forgot a few files.
1804    
1805    ----------------------------------------------------------------------
1806    Name: Allen Leung
1807    Date: 2000/04/06 00:36:00 EST
1808    Tag: leunga-20000406-peephole-x86-SSA
1809    Description:
1810    
1811    1.  New Peephole code
1812    
1813    2.  Minor improvement to X86 instruction selection
1814    
1815    3.  Various fixes to SSA and machine description -> code translator
1816    
1817    ----------------------------------------------------------------------
1818    Name: Matthias Blume
1819    Date: 2000/04/05 12:30:00 JST
1820    Tag: blume_main_v110p26p2_3
1821    Description:
1822    
1823    This update just merges three minor cosmetic updates to CM's sources
1824    to get ready for the 110.27 code freeze on Friday.  No functionality
1825    has changed.
1826    
1827    ----------------------------------------------------------------------
1828    Name: Allen Leung
1829    Date: 2000/04/04 19:39:00 EST
1830    Tag: leunga-20000404-x86-asm
1831    Description:
1832    
1833    1.  Fixed a problem in X86 assembly.
1834    
1835        Things like
1836    
1837           jmp %eax
1838           jmp (%eax)
1839    
1840        should be output as
1841    
1842           jmp *%eax
1843           jmp *(%eax)
1844    
1845    2.  Assembly output
1846    
1847          Added a new flag
1848    
1849              "asm-indent-copies" (default to false)
1850    
1851          When this flag is on, parallel copies will be indented an extra level.
1852    
1853    ----------------------------------------------------------------------
1854    Name: Allen Leung
1855    Date: 2000/04/04 03:18:00 EST
1856    Tag: leunga-20000404-C--Moby
1857    Description:
1858    
1859        All of these fixes are related to C--, Moby, and my own optimization
1860        stuff; so they shouldn't affect SML/NJ.
1861    
1862    1.  X86
1863    
1864        Various fixes related floating point, and extensions.
1865    
1866    2.  Alpha
1867    
1868        Some extra patterns related to loads with signed/zero extension
1869        provided by Fermin.
1870    
1871    3.  Assembly
1872    
1873        When generating assembly, resolve the value of client defined constants,
1874        instead of generating symbolic values.  This is controlled by the
1875        new flag "asm-resolve-constants", which is default to true.
1876    
1877    4.  Machine Descriptions
1878    
1879        a. The precedence parser was slightly broken when parsing infixr symbols.
1880        b. The type generalizing code had the bound variables reversed, resulting
1881           in a problem during arity raising.
1882        c. Various fixes in machine descriptions.
1883    
1884    ----------------------------------------------------------------------
1885    Name: Matthias Blume
1886    Date: 2000/04/03 16:05:00 JST
1887    Tag: blume_main_v110p26p2_2
1888    Description:
1889    
1890    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
1891    now done via the ordinary static environment that is context to each
1892    compilation unit.
1893    
1894    To this end, I arranged that instead of "structure Core" as "structure
1895    _Core" is bound in the pervasive environment.  Core access is done via
1896    _Core (which can never be accidentally rebound because _Core is not a
1897    legal surface-syntax symbol).
1898    
1899    The current solution is much cleaner because the core environment is
1900    now simply part of the pervasive environment which is part of every
1901    compilation unit's context anyway.  In particular, this eliminates all
1902    special-case handling that was necessary until now in order to deal
1903    with dynamic and symbolic parts of the core environment.
1904    
1905    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
1906    compilation manager's bootstrap compiler (actually: in the "init group"
1907    handling).  See the comments in src/system/smlnj/init/init.cmi for
1908    more details.
1909    
1910    I also tried to track down all mentions of "Core" (as string argument
1911    to Symbol.strSymbol) in the compiler and replaced them with a
1912    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
1913    name appears in one place only.
1914    
1915    Binfile and bootfile format have not changed, but the switchover from
1916    the old "init.cmi" to the new one is a bit tricky, so I supplied new
1917    bootfiles anyway.
1918    
1919    ----------------------------------------------------------------------
1920    Name: Allen Leung
1921    Date: 2000/04/02 21:17:00 EST
1922    Tag: leunga-20000402-mltree
1923    Description:
1924    
1925       1. Renamed the constructor CALL in MLTREE by popular demand.
1926       2. Added a bunch of files from my repository.  These are currently
1927          used by other non-SMLNJ backends.
1928    
1929    ----------------------------------------------------------------------
1930    Name: Allen Leung
1931    Date: 2000/03/31 21:15:00 EST
1932    Tag: leunga-20000331-aliasing
1933    Description:
1934    
1935    This update contains a rewritten (and hopefully more correct) module
1936    for extracting aliasing information from CPS.
1937    
1938       To turn on this feature:
1939    
1940            Compiler.Control.CG.memDisambiguate := true
1941    
1942       To pretty print the region information with assembly
1943    
1944           Compiler.Control.MLRISC.getFlag "asm-show-region" := true;
1945    
1946       To control how many levels of aliasing information are printed, use:
1947    
1948           Compiler.Control.MLRISC.getInt "points-to-show-level" := n
1949    
1950       The default of n is 3.
1951    
1952    ----------------------------------------------------------------------
1953    Name: David MacQueen
1954    Date: 2000/03/31 11:15:00 EST
1955    Tag: dbm-20000331-runtime_fix
1956    Description:
1957    
1958    This update contains:
1959    
1960    1. runtime/c-lib/c-libraries.c
1961       includes added in revision 1.2 caused compilation errors on hppa-hpux
1962    
1963    2. fix for bug 1556
1964       system/Basis/Implementation/NJ/internal-signals.sml
1965    
1966    ----------------------------------------------------------------------
1967    Name: Matthias Blume
1968    Date: 2000/03/31 18:00:00 JST
1969    Tag: blume_main_v110p26p2_1
1970    Description:
1971    
1972    This update contains:
1973    
1974    1. A small change to CM's handling of stable libraries:
1975       CM now maintains one "global" modmap that is used for all stable
1976       libraries.  The use of such a global modmap maximizes sharing and
1977       minimizes the need for re-traversing parts of environments during
1978       modmap construction.  (However, this has minor impact since modmap
1979       construction seems to account for just one percent or less of total
1980       compile time.)
1981    
1982    2. I added a "genmap" phase to the statistics.  This is where I got the
1983       "one percent" number (see above).
1984    
1985    3. CM's new tool parameter mechanism just became _even_ better. :)
1986       - The parser understands named parameters and recursive options.
1987       - The "make" and "shell" tools use these new features.
1988         (This makes it a lot easier to cascade these tools.)
1989       - There is a small syntax change: named parameters use a
1990    
1991           <name> : ( <option> ... )            or
1992           <name> : <string>
1993    
1994         syntax.  Previously, named parameters were implemented in an
1995         ad-hoc fashion by each tool individually (by parsing strings)
1996         and had the form
1997    
1998           <name>=<string>
1999    
2000       See the CM manual for a full description of these issues.
2001    
2002    ----------------------------------------------------------------------
2003    Name: Matthias Blume
2004    Date: 2000/03/30 18:00:00 JST
2005    Tag: blume_main_v110p26p2_0
2006    Description:
2007    
2008    !!!!! WARNING !!!!!!
2009    !!  New binfiles  !!
2010    !!!!!!!!!!!!!!!!!!!!
2011    
2012    This update contains:
2013    
2014    1. Moderate changes to CM:
2015    
2016       - Changes to CM's tools mechanism.  In particular, it is now possible
2017       to have tools that accept additional "command line" parameters
2018       (specified in the .cm file at each instance where the tool's class is
2019       used).
2020    
2021       This was done to accommodate the new "make" and "shell" tools which
2022       facilitate fairly seamless hookup to portions of code managed using
2023       Makefiles or Shell scripts.
2024    
2025       There are no classes "shared" or "private" anymore.  Instead, the
2026       sharing annotation is now a parameter to the "sml" class.
2027    
2028       There is a bit of generic machinery for implementing one's own
2029       tools that accept command-line parameters.  However, I am not yet fully
2030       satisfied with that part, so expect changes here in the future.
2031    
2032       All existing tools are described in the CM manual.
2033    
2034       - Slightly better error handling.  (CM now suppresses many followup
2035       error messages that tended to be more annoying than helpful.)
2036    
2037    2. Major changes to the compiler's static environment data structures.
2038    
2039       - no CMStaticEnv anymore.
2040            - no CMEnv, no "BareEnvironment" (actually, _only_ BareEnvironment,
2041              but it is called Environment), no conversions between different
2042              kinds of static environments
2043    
2044       - There is still a notion of a "modmap", but such modmaps are generated
2045         on demand at the time when they are needed.  This sounds slow, but I
2046         sped up the code that generates modmaps enough for this not to lead to
2047         a slowdown of the compiler (at least I didn't detect any).
2048    
2049       - To facilitate rapid modmap generation, static environments now
2050         contain an (optional) "modtree" structure.  Modtree annotations are
2051         constructed by the unpickler during unpickling.  (This means that
2052         the elaborator does not have to worry about modtrees at all.)
2053         Modtrees have the advantage that they are compositional in the same
2054         way as the environment data structure itself is compositional.
2055         As a result, modtrees never hang on to parts of an environment that
2056         has already been rendered "stale" by filtering or rebinding.
2057    
2058       - I went through many, many trials and errors before arriving at the
2059         current solution.  (The initial idea of "linkpaths" did not work.)
2060         But the result of all this is that I have touched a lot of files that
2061         depend on the "modules" and "types" data structures (most of the
2062         elaborator). There were a lot of changes during my "linkpath" trials
2063         that could have been reverted to their original state but weren't.
2064         Please, don't be too harsh on me for messing with this code a bit more
2065         than what was strictly necessary...  (I _did_ resist the tempation
2066         of doing any "global reformatting" to avoid an untimely death at
2067         Dave's hands. :)
2068    
2069       - One positive aspect of the previous point:  At least I made sure that
2070         all files that I touched now compile without warnings (other than
2071         "polyEqual").
2072    
2073       - compiler now tends to run "leaner" (i.e., ties up less memory in
2074         redundant modmaps)
2075    
2076    ----------------------------------------------------------------------
2077    Name: Allen Leung
2078    Date: 2000/03/29 18:00:00
2079    Tag: leunga-20000327-mlriscGen_hppa_alpha_x86
2080    Boot files (optional): ftp://react-ilp.cs.nyu.edu/leunga/110.26.1-sml.boot.x86-unix-20000330.tar.gz
2081    Description:
2082    
2083       This update contains *MAJOR* changes to the way code is generated from CPS
2084    in the module mlriscGen, and in various backend modules.
2085    
2086    CHANGES
2087    =======
2088    
2089    1. MLRiscGen: forward propagation fix.
2090    
2091       There was a bug in forward propagation introduced at about the same time
2092       as the MLRISC x86 backend, which prohibits coalescing to be
2093       performed effectively in loops.
2094    
2095       Effect: speed up of loops in RISC architectures.
2096               By itself, this actually slowed down certain benchmarks on the x86.
2097    
2098    2. MLRiscGen:  forward propagating addresses from consing.
2099    
2100       I've changed the way consing code is generated.  Basically I separated
2101       out the initialization part:
2102    
2103            store tag,   offset(allocptr)
2104            store elem1, offset+4(allocptr)
2105            store elem2, offset+8(allocptr)
2106            ...
2107            store elemn, offset+4n(allocptr)
2108    
2109       and the address computation part:
2110    
2111            celladdr <- offset+4+alloctpr
2112    
2113       and move the address computation part
2114    
2115       Effect:  register pressure is generally lower as a result.  This
2116                makes compilation of certain expressions much faster, such as
2117                long lists with non-trivial elements.
2118    
2119                 [(0,0), (0,0), .... (0,0)]
2120    
2121    3. MLRiscGen: base pointer elimination.
2122    
2123        As part of the linkage mechanism, we generate the sequence:
2124    
2125         L:  ...  <- start of the code fragment
2126    
2127         L1:
2128             base pointer <- linkreg - L1 + L
2129    
2130         The base pointer was then used for computing relocatable addresses
2131       in the code fragment.  Frequently (such as in lots of continuations)
2132       this is not needed.  We now eliminate this sequence whenever possible.
2133    
2134         For compile time efficiency, I'm using a very stupid local heuristic.
2135       But in general, this should be done as a control flow analysis.
2136    
2137       Effect:  Smaller code size.  Speed up of most programs.
2138    
2139    4. Hppa back end
2140    
2141         Long jumps in span dependence resolution used to depend on the existence
2142      of the base pointer.
2143    
2144         A jump to a long label L was expanded into the following sequence:
2145    
2146          LDIL %hi(L-8192), %r29
2147          LDO  %lo(L-8192)(%r29), %r29
2148          ADD  %r29, baseptr, %r29
2149          BV,n %r0(%r29)
2150    
2151         In the presence of change (3) above, this will not work.  I've changed
2152       it so that the following sequence of instructions are generated, which
2153       doesn't mention the base pointer at all:
2154    
2155             BL,n  L', %r29           /* branch and link, L' + 4 -> %r29 */
2156        L':  ADDIL L-(L'+4), %r29     /* Compute address of L */
2157             BV,n  %r0(%r29)          /* Jump */
2158    
2159    5. Alpha back end
2160    
2161          New alpha instructions LDB/LDW have been added, as per Fermin's
2162       suggestions.   This is unrelated to all other changes.
2163    
2164    6. X86 back end
2165    
2166         I've changed andl to testl in the floating point test sequence
2167         whenever appropriate.  The Intel optimization guide states that
2168         testl is preferable to andl.
2169    
2170    7. RA (x86 only)
2171    
2172         I've improved the spill propagation algorithm, using an approximation
2173       of maximal weighted independent sets.   This seems to be necessary to
2174       alleviate the negative effect in light of the slow down in (1).
2175    
2176         I'll write down the algorithm one of these days.
2177    
2178    8. MLRiscGen: frequencies
2179    
2180         I've added an annotation that states that all call gc blocks have zero
2181       execution frequencies.  This improves register allocation on the x86.
2182    
2183    BENCHMARKS
2184    ==========
2185    
2186       I've only perform the comparison on 110.25.
2187    
2188       The platforms are:
2189    
2190        HPPA  A four processor HP machine (E9000) with 5G of memory.
2191        X86   A 300Hhz Pentium II with 128M of memory, and
2192        SPARC An Ultra sparc 2 with 512M of memory.
2193    
2194       I used the following parameters for the SML benchmarks:
2195    
2196                 @SMLalloc
2197         HPPA    256k
2198         SPARC   512k
2199         X86     256k
2200    
2201    COMPILATION TIME
2202    ----------------
2203       Here are the numbers comparing the compilation times of the compilers.
2204       I've only compared 110.25 compiling the new sources versus
2205       a fixpoint version of the new compiler compiling the same.
2206    
2207                     110.25                                  New
2208               Total  Time in RA  Spill+Reload   Total  Time In RA Spill+Reload
2209         HPPA   627s    116s        2684+3584     599s    95s       1003+1879
2210         SPARC  892s    173s        2891+3870     708s    116s      1004+1880
2211         X86    999s    315s       94006+130691   987s    296s    108877+141957
2212    
2213                   110.25         New
2214                Code Size      Code Size
2215         HPPA   8596736         8561421
2216         SPARC  8974299         8785143
2217         X86    9029180         8716783
2218    
2219       So in summary, things are at least as good as before.   Dramatic
2220       reduction in compilation is obtained on the Sparc; I can't explain it,
2221       but it is reproducible.  Perhaps someone should try to reproduce this
2222       on their own machines.
2223    
2224    SML BENCHMARKS
2225    --------------
2226    
2227        On the average, all benchmarks perform at least as well as before.
2228    
2229          HPPA         Compilation Time     Spill+Reload      Run Time
2230                     110.25  New            110.25    New   110.25  New
2231    
2232          barnesHut  3.158  3.015  4.75%    1+1       0+0   2.980  2.922   2.00%
2233              boyer  6.152  5.708  7.77%    0+0       0+0   0.218  0.213   2.34%
2234       count-graphs  1.168  1.120  4.32%    0+0       0+0  22.705 23.073  -1.60%
2235                fft  0.877  0.792 10.74%    1+3       1+3   0.602  0.587   2.56%
2236        knuthBendix  3.180  2.857 11.32%    0+0       0+0   0.675  0.662   2.02%
2237             lexgen  6.190  5.290 17.01%    0+0       0+0   0.913  0.788  15.86%
2238               life  0.803  0.703 14.22%   25+25      0+0   0.153  0.140   9.52%
2239              logic  2.048  2.007  2.08%    6+6       1+1   4.133  4.008   3.12%
2240         mandelbrot  0.077  0.080 -4.17%    0+0       0+0   0.765  0.712   7.49%
2241             mlyacc 22.932 20.937  9.53%  154+181    32+57  0.468  0.430   8.91%
2242            nucleic  5.183  5.060  2.44%    2+2       0+0   0.125  0.120   4.17%
2243      ratio-regions  3.357  3.142  6.84%    0+0       0+0  116.225 113.173 2.70%
2244                ray  1.283  1.290 -0.52%    0+0       0+0   2.887  2.855   1.11%
2245             simple  6.307  6.032  4.56%   28+30      5+7   3.705  3.658   1.28%
2246                tsp  0.888  0.862  3.09%    0+0       0+0   7.040  6.893   2.13%
2247               vliw 24.378 23.455  3.94%  106+127    25+45  2.758  2.707   1.91%
2248      --------------------------------------------------------------------------
2249       Average                     6.12%                                   4.09%
2250    
2251          SPARC        Compilation Time     Spill+Reload      Run Time
2252                     110.25  New            110.25    New   110.25  New
2253    
2254          barnesHut  3.778  3.592  5.20%    2+2       0+0   3.648  3.453    5.65%
2255              boyer  6.632  6.110  8.54%    0+0       0+0   0.258  0.242    6.90%
2256       count-graphs  1.435  1.325  8.30%    0+0       0+0  33.672 34.737   -3.07%
2257                fft  0.980  0.940  4.26%    3+9       2+6   0.838  0.827    1.41%
2258        knuthBendix  3.590  3.138 14.39%    0+0       0+0   0.962  0.967   -0.52%
2259             lexgen  6.593  6.072  8.59%    1+1       0+0   1.077  1.078   -0.15%
2260               life  0.972  0.868 11.90%   26+26      0+0   0.143  0.140    2.38%
2261              logic  2.525  2.387  5.80%    7+7       1+1   5.625  5.158    9.05%
2262         mandelbrot  0.090  0.093 -3.57%    0+0       0+0   0.855  0.728   17.39%
2263             mlyacc 26.732 23.827 12.19%  162+189    32+57  0.550  0.560   -1.79%
2264            nucleic  6.233  6.197  0.59%    3+3       0+0   0.163  0.173   -5.77%
2265      ratio-regions  3.780  3.507  7.79%    0+0       0+0 133.993 131.035   2.26%
2266                ray  1.595  1.550  2.90%    1+1       0+0   3.440  3.418    0.63%
2267             simple  6.972  6.487  7.48%   29+32      5+7   3.523  3.525   -0.05%
2268                tsp  1.115  1.063  4.86%    0+0       0+0   7.393  7.265    1.77%
2269               vliw 27.765 24.818 11.87%  110+135    25+45  2.265  2.135    6.09%
2270      ----------------------------------------------------------------------------
2271       Average                     6.94%                                    2.64%
2272    
2273          X86          Compilation Time     Spill+Reload      Run Time
2274                     110.25  New            110.25    New   110.25  New
2275    
2276          barnesHut  5.530  5.420  2.03%  593+893   597+915   3.532  3.440   2.66%
2277              boyer  8.768  7.747 13.19%  493+199   301+289   0.327  0.297  10.11%
2278       count-graphs  2.040  2.010  1.49%  298+394   315+457  26.578 28.660  -7.26%
2279                fft  1.327  1.302  1.92%  112+209   115+210   1.055  0.962   9.71%
2280        knuthBendix  5.218  5.475 -4.69%  451+598   510+650   0.928  0.932  -0.36%
2281             lexgen  9.970  9.623  3.60% 1014+841  1157+885   0.947  0.928   1.97%
2282               life  1.183  1.183  0.00%  162+182   145+148   0.127  0.103  22.58%
2283              logic  3.285  3.512 -6.45%  514+684   591+836   5.682  5.577   1.88%
2284         mandelbrot  0.147  0.143  2.33%   38+41     33+54    0.703  0.690   1.93%
2285             mlyacc 35.457 32.763  8.22% 3496+4564 3611+4860  0.552  0.550   0.30%
2286            nucleic  7.100  6.888  3.07%  239+168   201+158   0.175  0.173   0.96%
2287      ratio-regions  6.388  6.843 -6.65% 1182+257   981+300  120.142 120.345 -0.17%
2288                ray  2.332  2.338 -0.29%  346+398   402+494   3.593  3.540   1.51%
2289             simple  9.912  9.903  0.08% 1475+941  1579+1168  3.057  3.178  -3.83%
2290                tsp  1.623  1.532  5.98%  266+200   250+211   8.045  7.878   2.12%
2291               vliw 33.947 35.470 -4.29% 2629+2774 2877+3171  2.072  1.890   9.61%
2292      ----------------------------------------------------------------------------
2293       Average                     1.22%                                     3.36%
2294    
2295  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2296  Name: Allen Leung  Name: Allen Leung
2297  Date: 2000/03/23 16:25:00  Date: 2000/03/23 16:25:00
# Line 66  Line 2347 
2347    
2348  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2349  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2350  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
2351  explanation.  explanation.
2352    
2353  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2354  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2355  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2356  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 113  Line 2394 
2394  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
2395  <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,
2396  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
2397  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2398  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2399  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2400    
# Line 163  Line 2444 
2444    it from that remote directory.    it from that remote directory.
2445    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2446    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2447    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2448    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2449    
2450    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.583  
changed lines
  Added in v.808

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