Home My Page Projects Code Snippets Project Openings SML/NJ
Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

SCM Repository

[smlnj] Diff of /sml/trunk/HISTORY
ViewVC logotype

Diff of /sml/trunk/HISTORY

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

revision 591, Mon Apr 3 01:19:20 2000 UTC revision 834, Wed May 23 18:30:31 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: 2001/05/23 14:30:00 EDT
18    Tag: Release_110_33
19    Description:
20    
21    New version number (110.33).  New bootfiles.
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume
25    Date: 2001/05/22 18:06:00 EDT
26    Tag: blume-20010522-targets
27    Description:
28    
29    Made install.sh use file config/targets.customized if it exists, falling
30    back to config/targets if it doesn't.  This way one can have a customized
31    version of the targets file without touching the "real thing", thus
32    eliminating the constant fear of accidentally checking something bogus
33    back into the CVS repository...  (File config/targets.customized must
34    not be added to the repository!)
35    
36    ----------------------------------------------------------------------
37    Name: Matthias Blume
38    Date: 2001/05/22 16:30:00 EDT
39    Tag: blume-20010522-minitut
40    Description:
41    
42    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
43       struct returns.
44    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
45       incomplete, preliminary documentation for NLFFI.
46    
47    ----------------------------------------------------------------------
48    Name: Matthias Blume
49    Date: 2001/05/14 11:30:00 EDT
50    Tag: blume-20010514-script
51    Description:
52    
53    Some bugs in install script fixed.
54    
55    In addition to that I also made a slight change to the NLFFI API:
56    Functors generated by ml-nlffigen now take the dynamic library as a
57    straight functor argument, not as a suspended one.  (The original
58    functor code used to force the suspension right away anyway, so there
59    was nothing gained by this complication of the interface.)
60    
61    ----------------------------------------------------------------------
62    Name: Matthias Blume
63    Date: 2001/05/11 14:35:00 EDT
64    Tag: blume-20010511-ml-nlffi
65    Description:
66    
67    I finally took the plunge and added my new FFI code to the main
68    repository.  For x86-linux it is now ready for prime-time.
69    
70    There are two new subdirectories of "src":
71    
72      - ml-nlffi-lib:
73           The utility library for programs using the FFI interface.
74           Here is the implementation of $/c.cm and its associated low-level
75           partners $/c-int.cm and $/memory.cm.
76      - ml-nlffigen:
77           A stand-alone program for generating ML glue code from C source
78           code.
79    
80    Building ml-nlffigen requires $/ckit-lib.cm.
81    
82    The config/install.sh script has been updates to do the Right Thing
83    (hopefully).
84    
85    Notice that the source tree for the C-Kit will not be put under "src"
86    but directly under the installation root directory.  (This is the
87    structure that currently exists on the CVS server when you check out
88    module "sml".)  Fortunately, config/install.sh knows about this oddity.
89    
90    Bugs: No documentation yet.
91    
92    ----------------------------------------------------------------------
93    Name: Matthias Blume
94    Date: 2001/05/09 16:35:00 EDT
95    Tag: blume-20010509-cpscontract
96    Description:
97    
98    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
99    wrapper/unwrapper elimination did not decrement usage counts and some
100    dead variables got overlooked by the dead-up logic.)
101    
102    ----------------------------------------------------------------------
103    Name: Lal George
104    Date: 2001/05/08  17:26:09 EDT
105    Tag: george-20010508-omit-frameptr
106    Description:
107    
108    Changes to implement the omit-frame-pointer optimization to support
109    raw C calls. For now, there is only support on the Intel x86, but
110    other architectures will follow as more experience is gained with this.
111    
112    
113    ----------------------------------------------------------------------
114    Name: Matthias Blume
115    Date: 2001/05/07 14:40:00 EDT
116    Tag: blume-20010507-proxies
117    Description:
118    
119    I made into "proxy libraries" all libraries that qualify for such a
120    change.  (A qualifying library is a library that has another library or
121    groups as its sole member and repeats that member's export list
122    verbatim.  A proxy library avoids this repetition by omitting its export
123    list, effectively inheriting the list that its (only) member exports.
124    See the CM manual for more explanation.)
125    The main effect is that explicit export lists for these libraries
126    do not have to be kepts in sync, making maintenance a bit easier.
127    
128    I also added copyright notices to many .cm-files.
129    
130    Last but not least, I made a new set of bootfiles.
131    
132    ----------------------------------------------------------------------
133    Name: Matthias Blume
134    Date: 2001/05/04 17:00:00 EDT
135    Tag: blume-20010504-cm-lsplit
136    Description:
137    
138    0. John merged pending changes to $/smlnj-lib.cm
139    
140    1. Allen's previous change accidentally backed out of one of Lal's
141       earlier changes.  I undid this mistake (re-introducing Lal's change).
142    
143    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
144       within the compiler where applicable.  There is some code simplification
145       because of that.
146    
147    3. The "split" phase (in FLINT) is now part of the default list of phases.
148       Compiler.Control.LambdaSplitting.* can be used to globally control the
149       lambda-splitting (cross-module-inlining) engine.  In addition to that,
150       it can now also be controlled on a per-source basis: CM has been taught
151       a new tool parameter applicable to ML source files.
152    
153       - To turn lambda-splitting off completely:
154            local open Compiler.Control.LambdaSplitting in
155                val _ = set Off
156            end
157       - To make "no lambda-splitting" the global default (but allow per-source
158         overriding); this is the initial setting:
159            local open Compiler.Control.LambdaSplitting in
160                val _ = set (Default NONE)
161            end
162       - To make "lambda-splitting with aggressiveness a" the global default
163         (and allow per-source overriding):
164            local open Compiler.Control.LambdaSplitting in
165                val _ = set (Default (SOME a))
166            end
167    
168       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
169         write (in the respective .cm-file):
170            a.sml (lambdasplitting:off)
171       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
172            a.sml (lambdasplitting:on)
173       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
174         <a> is a decimal non-negative integer):
175            a.sml (lambdasplitting:<a>)
176       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
177            a.sml (lambdasplitting:infinity)
178       - To use the global default for a.sml:
179            a.sml (lambdasplitting:default)
180         or simply
181            a.sml
182    
183    ----------------------------------------------------------------------
184    Name: Allen Leung
185    Date: 2001/05/04 01:57:00 EDT
186    Tag: leunga-20010504-sync
187    Description:
188    
189      MLRISC features.
190    
191      1. Fix to CMPXCHG instructions.
192      2. Changed RA interface to allow annotations in callbacks.
193      3. Added a new method to the stream interface to allow annotations updates.
194    
195    ----------------------------------------------------------------------
196    Name: Matthias Blume
197    Date: 2001/05/01 11:45:00 EDT
198    Tag: blume-20010501-pcedittmp
199    Description:
200    
201    Changed install.sh to use the current working directory instead of
202    /usr/tmp for a temporary file (pcedittmp).  The previous choice
203    of /usr/tmp caused trouble with MacOS X because of file premission
204    problems.
205    
206    ----------------------------------------------------------------------
207    Name: Matthias Blume
208    Date: 2001/04/20 11:10:00 EDT
209    Tag: blume-20010420-inMLflag
210    Description:
211    
212     - added vp_limitPtrMask to vproc-state.h
213       (for use by the raw-C-calls mechanism to implement proper interrupt
214        handling)
215     - made the ML compiler aware of various data-structure offsets so it
216       can generate code for accessing the vp_inML flag and vp_limitPtrMask
217     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
218       raw C-calls
219    
220    ----------------------------------------------------------------------
221    Name: Lal George
222    Date: 2001/04/20 09:15:28 EDT
223    Tag: george-20010420-macosX
224    Description:
225    
226     - Changes to port to Mac OS X; Darwin.
227    
228     - In the process I found that sqrt was broken on the PPC, because the
229       fsqrt instruction is not implemented.
230    
231    ----------------------------------------------------------------------
232    Name: Matthias Blume
233    Date: 2001/04/18 12:45:00 EDT
234    Tag: blume-20010418-ccalls
235    Description:
236    
237     - fixed two off-by-4 errors in the x86-specific c-calls implementation
238       (this bug prevented structure arguments containing pointers from being
239        passed correctly)
240     - changed the raw-C-call code in mlriscGen.sml in such a way that
241       structure arguments are represented as a pointer to the beginning
242       of the structure (instead of having a series of synthesized arguments,
243       one for each structure member)
244    
245     - made makeml script's verbosity level configurable via environment
246       variable (MAKEML_VERBOSITY)
247    
248     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
249       in rawmem-x86.sml; we are now using the real thing
250    
251    ----------------------------------------------------------------------
252    Name: Matthias Blume
253    Date: 2001/03/22 16:25:00 EST
254    Tag: blume-20010322-bootfiles
255    Description:
256    
257    Created a new set of bootfiles (for your automatic installation convenience).
258    
259    ----------------------------------------------------------------------
260    Name: Matthias Blume
261    Date: 2001/03/22 15:10:00 EST
262    Tag: blume-20010322-rawmem-parcm
263    Description:
264    
265    1. All "raw memory access" primitives for the new FFI are implemented now
266       (at least on the x86).
267    2. Some further cleanup of CM's parallel make mechanism.
268    
269    ----------------------------------------------------------------------
270    Name: Matthias Blume
271    Date: 2001/03/19 17:53:00 EST
272    Tag: blume-20010319-parallel
273    Description:
274    
275    Parallel make (using compile servers) now works again.
276    
277    To this end, CM.stabilize and CMB.make have been modified to work in
278    two passes when compile servers are attached:
279       1. Compile everything, do not perform stabilization; this pass
280          uses compile servers
281       2. Stabilize everything; this pass does not use compile servers
282    If there are no compile servers, the two passes are combined into one
283    (as before).  Splitting the passes increases the inherent parallelism
284    in the dependency graph because the entire graph including all
285    libraries is available at the same time.  This, in turn, improves
286    server utilization.  The downside is that the master process will
287    have to do some extra work after compilation is done (because for
288    technical reasons it must re-read all the binfiles during stabilization).
289    
290    ----------------------------------------------------------------------
291    Name: Matthias Blume
292    Date: 2001/03/16 12:22:00 EST
293    Tag: blume-20010316-bootfiles
294    Description:
295    
296    Created a new set of bootfiles (for your automatic installation convenience).
297    
298    ----------------------------------------------------------------------
299    Name: Matthias Blume
300    Date: 2001/03/16 11:00:00 EST
301    Tag: blume-20010316-MLTREE-fixup
302    Description:
303    
304    This is a minor fixup for an (untagged) earlier commit by Allen.
305    (A file was missing).
306    
307    ----------------------------------------------------------------------
308    Name: Allen Leung
309    Date: Mon Mar  5 18:54:57 EST 2001
310    Tag: leunga-20010305-cut-support
311    
312    1. New support for alternative control-flow in MLTREE.
313       Currently we support
314    
315          FLOW_TO(CALL ...., [k1,...,kn])
316    
317       This is needed for 'cuts to' in C-- and try/handle-like constructs
318       in Moby
319    
320       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
321    
322    2. Register Allocator
323    
324       Changes in interface [from Fermin, John]
325    
326    3. Alpha 8-bit SLL support [Fermin]
327    
328    4. All architectures
329    
330       A new module (ClusterExpandCopies) for expanding parallel copies.
331    
332    ----------------------------------------------------------------------
333    Name: Allen Leung
334    Date: 2001/02/27 23:07:00 EST
335    Tag: leunga-20010227-minor-stuff
336    
337    1. Alpha bug fix for CMOVNE
338    2. Handle mltree COND(..,FCMP ...,...)
339    3. Bug fix in simplifier
340    
341    ----------------------------------------------------------------------
342    Name: Matthias Blume
343    Date: 2001/01/30 17:50:00 EST
344    Tag: blume-20010130-sync
345    Description:
346    
347    This is just a minor update to sync my devel branch with the main brach.
348    The only visible change is the addition of some README files.
349    
350    ----------------------------------------------------------------------
351    Name: Matthias Blume
352    Date: 2001/01/12 23:30:00 JST
353    Tag: blume-20010112-bootfiles
354    Description:
355    
356    Made a new set of bootfiles that goes with the current state of the
357    repository.
358    
359    ----------------------------------------------------------------------
360    Name: Matthias Blume
361    Date: 2001/01/12 21:20:00 JST
362    Tag: blume-20010112-sync
363    Description:
364    
365    I am just flushing out some minor changes that had accumulated in
366    my private branch in order to sync with the main tree.  (This is
367    mainly because I had CVS trouble when trying to merge _into_ my
368    private branch.)
369    
370    Most people should be completely unaffected by this.
371    
372    ----------------------------------------------------------------------
373    Name: Allen Leung
374    Date: Thu Jan 11 21:03:00 EST 2001
375    Tag: leunga-20010111-labexp=mltree
376    Description:
377    
378    1.  Removed the type LabelExp and replace it by MLTree.
379    2.  Rewritten mltree-simplify with the pattern matcher tool.
380    3.  There were some bugs in alpha code generator which would break
381        64-bit code generation.
382    4.  Redo the tools to generate code with the
383    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
384        are now generated by perl scripts.
385    
386    ----------------------------------------------------------------------
387    Name: Matthias Blume
388    Date: 2001/01/10 21:55:00 JST
389    Tag: blume-20010110-rcc
390    Description:
391    
392    The RCC stuff now seems to work (but only on the x86).
393    This required hacking of the c-calls interface (and -implementation) in
394    MLRISC.
395    
396    Normal compiler users should be unaffected.
397    
398    ----------------------------------------------------------------------
399    Name: Matthias Blume
400    Date: 2001/01/09 01:20:00 JST
401    Tag: blume-20010109-rcc
402    Description:
403    
404    This is a fairly big patch, flushing out a large number of pending
405    changes that I made to my development copy over the last couple of days.
406    
407    Of practical relevance at this moment is a workaround for a pickling
408    bug that Allen ran into the other day.  The cause of the bug itself is
409    still unknown and it might be hard to fix it properly, but the
410    workaround has some merits of its own (namely somewhat reducing pickling
411    overhead for certain libraries).  Therefore, I think this solution should
412    be satisfactory at this time.
413    
414    The rest of the changes (i.e., the vast majority) has to do with my
415    ongoing efforts of providing direct support for C function calls from
416    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
417    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
418    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
419    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
420    handle RCC.
421    
422    The last part (the changes to mlriscGen.sml) are still known to be
423    wrong on the x86 and not implemented on all other architectures.  But
424    the infrastructure is in place. I had to change a few functor
425    signatures in the backend to be able to route the CCalls interface
426    from MLRISC there, and I had to specialize the mltree type (on the
427    x86) to include the necessary extensions. (The extensions themselves
428    were already there and redy to go in MLRISC/x86).
429    
430    Everything should be very happy as soon as someone helps me with
431    mlriscGen.sml...
432    
433    In any case, nothing of this should matter to anyone as long as the
434    new primop is not being used (which is going to be the case unless you
435    find it where I hid it :). The rest of the compiler is completely
436    unaffected.
437    
438    ----------------------------------------------------------------------
439    Name: Matthias Blume
440    Date: 2001/01/05 00:30:00 JST
441    Tag: blume-20010105-primops
442    Description:
443    
444    Added some experimental support for work that I am doing right now.
445    These changes mostly concern added primops, but there is also a new
446    experimental C library in the runtime system (but currently not enabled
447    anywhere except on Linux/X86).
448    
449    In the course of adding primops (and playing with them), I discovered that
450    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
451    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
452    To recover, I made type information in INL_PRIM mandatory and changed
453    prim.sml as well as built-in.sml accordingly.  The InLine structure now
454    has complete, correct type information (i.e., no bottom types).
455    
456    Since all these changes mean that we need new binfiles, I also bumped the
457    version number to 110.32.1.
458    
459    ----------------------------------------------------------------------
460    Name: Matthias Blume
461    Date: 2000/12/30 22:10:00 JST
462    Tag: blume-20001230-various
463    Description:
464    
465    Added proxy libraries for MLRISC and let MLRISC libraries refer
466    to each other using path anchors.  (See CM manual for explanation.)
467    
468    Updated CM documentation.
469    
470    Fixed some bugs in CM.
471    
472    Implemented "proxy" libraries (= syntactic sugar for CM).
473    
474    Added "-quiet" option to makeml and changed runtime system accordingly.
475    
476    Added cleanup handler for exportML to reset timers and compiler stats.
477    
478    ----------------------------------------------------------------------
479    Name: Lal George
480    Date: 2000/12/22 22:22:58 EST 2000
481    Tag: Release_110_32
482    Description:
483    
484            Infinite precision used throughout MLRISC.
485            see MLRISC/mltree/machine-int.sig
486    
487    ----------------------------------------------------------------------
488    Name: Matthias Blume
489    Date: 2000/12/22 23:16:00 JST
490    Tag: blume-20001222-warn
491    Description:
492    
493    Corrected wording and formatting of some CM warning message which I
494    broke in my previous patch.
495    
496    ----------------------------------------------------------------------
497    Name: Matthias Blume
498    Date: 2000/12/22 21:20:00 JST
499    Tag: blume-20001222-anchorenv
500    Description:
501    
502    Fixed CM's handling of anchor environments in connection with CMB.make.
503    
504    ----------------------------------------------------------------------
505    Name: Matthias Blume
506    Date: 2000/12/22 13:15:00 JST
507    Tag: blume-20001222-cleanup
508    Description:
509    
510    Removed src/cm/ffi which does not (and did not) belong here.
511    
512    ----------------------------------------------------------------------
513    Name: Matthias Blume
514    Date: 2000/12/21 23:55:00 JST
515    Tag: blume-20001221-exn
516    Description:
517    
518    Probably most important: CM no longer silently swallows all exceptions
519    in the compiler.
520    Plus: some other minor CM changes.  For example, CM now reports some
521    sizes for generated binfiles (code, data, envpickle, lambdapickle).
522    
523    ----------------------------------------------------------------------
524    Name: Matthias Blume
525    Date: 2000/12/15 00:01:05 JST
526    Tag: blume-20001215-dirtool
527    Description:
528    
529    - "dir" tool added.
530    - improvements and cleanup to Tools structure
531    - documentation updates
532    
533    ----------------------------------------------------------------------
534    Name: Allen Leung
535    Date: Thu Dec 14 03:45:24 EST 2000
536    Description:
537    Tag:  leunga-20001214-int-inf
538    Description:
539    
540       In IntInf, added these standard functions, which are missing from our
541    implementation:
542    
543        andb : int * int -> int
544        xorb : int * int -> int
545        orb  : int * int -> int
546        notb : int -> int
547         <<   : int * word -> int
548        ~>>  : int * word -> int
549    
550       Not tested, I hope they are correct.
551    
552    ----------------------------------------------------------------------
553    Name: Allen Leung
554    Date: Fri Dec  8 19:23:26 EST 2000
555    Description:
556    Tag:  leunga-20001208-nowhere
557    Description:
558    
559      Slight improvements to the 'nowhere' tool to handle OR-patterns,
560    to generate better error messages etc.  Plus a brief manual.
561    
562    ----------------------------------------------------------------------
563    Name: Lal George
564    Date: 2000/12/08 09:54:02 EST 2000
565    Tag: Release_110_31
566    Description:
567    
568    - Version 110.31
569    ----------------------------------------------------------------------
570    Name: Allen Leung
571    Date: Thu Dec  7 22:01:04 EST 2000
572    Tag:  leunga-20001207-cell-monster-hack
573    Description:
574    
575    Major MLRISC internal changes.  Affect all clients.
576    Summary:
577    
578    1.  Type CELLS.cell = int is now replaced by a datatype.
579        As a result, the old regmap is now gone.  Almost all interfaces
580        in MLRISC change as a consequence.
581    
582    2.  A new brand version of machine description tool (v3.0) that generates
583        modules expecting the new interface.  The old version is removed.
584    
585    3.  The RA interface has been further abstracted into two new functors.
586        RISC_RA and X86RA.  These functors have much simpler interfaces.
587        [See also directory MLRISC/demo.]
588    
589    4.  Some other new source->source code generation tools are available:
590    
591        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
592        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
593           I use this tool to generate the peephole optimizers---with the new
594           cell type changes, peephole rules are becoming difficult to write
595           without conditional pattern matching.
596    
597    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
598        cover the entire MLRISC source tree so many things broke.
599    
600    6.  CM files have been moved to the subdirectory MLRISC/cm.
601        They are moved because there are a lot of them and they clutter up the
602        root dir.
603    
604    7.  More detailed documentation to come...
605    
606        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
607        a makeml -rebuild first.  This is because of other other
608        changes that Matthias has made (see below).
609    
610    
611    ----------------------------------------------------------------------
612    Name: Matthias Blume
613    Date: 2000/11/30 23:12:00 JST
614    Tag: blume-20001130-filereorg
615    Description:
616    
617    Some manual updates and some file reorganizations in CM.
618    
619    ----------------------------------------------------------------------
620    Name: Matthias Blume
621    Date: 2000/11/24 17:45:00 JST
622    Tag: blume-20001124-link
623    Description:
624    
625    Drastically improved link traversal code for the case that the dynamic
626    value was already loaded at bootstrap time.  As a result, CM and CMB
627    now both load blazingly fast -- even on a very slow machine.  Also,
628    memory consumption has been further reduced by this.
629    
630    Warning: The format of the PIDMAP file has changed.  THerefore, to
631    bootstrap you have to do this:
632    
633    1. Run CMB.make
634    2. Make a symbolic link for the boot directory:
635         ln -s sml.boot.ARCH-OS xxx
636    3. "Rebuild" the boot directory:
637         ./makeml -boot xxx -rebuild sml ; rm xxx
638    4. Boot normally:
639          ./makeml
640    
641    ----------------------------------------------------------------------
642    Name: Matthias Blume
643    Date: 2000/11/21 21:20:00 JST
644    Tag: blume-20001121-tools
645    Description:
646    
647    Continued hacking on autoloading problem -- with success this time.
648    Also changed tool-plugin mechanism.  See new CM manual.
649    
650    ----------------------------------------------------------------------
651    Name: Matthias Blume
652    Date: 2000/11/19 14:30:00 JST
653    Tag:  blume-20001119-autoload
654    Description:
655    
656    Some hacking to make autoloading faster.  Success for CMB, no success
657    so far for CM.  There is a reduced structure CM' that autoloads faster.
658    (This is a temporary, non-documented hack to be eliminated again when
659    the general problem is solved.)
660    
661    ----------------------------------------------------------------------
662    Name: Matthias Blume
663    Date: 2000/11/17 14:10:00 JST
664    Tag: blume-20001117-pickle-lib
665    Description:
666    
667    1. Eliminated comp-lib.cm
668    2. Made pickle-lib.cm
669    3. Eliminated all uses of intset.sml (from comp-lib.cm)
670    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
671       equivalent constructs from smlnj-lib.cm (INtHashTable).
672    5. Point 4. also goes for those uses of intmap.* in MLRISC.
673       Duplicated intmap modules thrown out.
674    6. Hunted down all duplicated SCC code and replaced it with
675       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
676    7. Rewrote Feedback module.
677    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
678       should be thrown out and equivalent modules from smlnj-lib.cm
679       should be used (IntRedBlackSet, IntListSet, ...).
680    
681    Confirmed that compiler compiles to fixpoint.
682    
683    ----------------------------------------------------------------------
684    Name: Allen Leung
685    Date: 2000/11/10 18:00:00
686    Tag: leunga-20001110-new-x86-fp
687    
688    A new x86 floating point code generator has been added.
689    By default this is turned off.  To turn this on, do:
690    
691        CM.autoload "$smlnj/compiler.cm";
692        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
693    
694    Changes:
695    
696    1.  Changed FTAN to FPTAN so that the assembly output is correct.
697    2.  Changed the extension callback for FTANGENT to generate:
698    
699              fptan
700              fstp  %st(0)
701        instead of
702              fptan
703              fstpl ftempmem
704    
705    3.  Numerous assembly fixes for x86.
706    
707    5.  Cleaned up the machine code output module x86/x86MC.sml and added
708        support for a whole bunch of instructions and addressing modes:
709    
710          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
711          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
712          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
713          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
714          fxch %st(n)
715          fld %st(n)
716          fst %st(n)
717          fst mem
718          fstp %st(n)
719          fucom %st(n)
720          fucomp %st(n)
721    
722        All these are now generated when the fast fp mode is turned on.
723    
724    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
725    
726    ----------------------------------------------------------------------
727    Name: Matthias Blume
728    Date: 2000/11/09 11:20:00 JST
729    Tag: blume-20001109-scc
730    Description:
731    
732    Eliminated some code duplication:
733    
734    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
735       (Otherwise the functor is useless.)
736    2. Used GraphSCCFn where SCCUtilFun was used previously.
737    3. Got rid of SCCUtilFun (in comp-lib.cm).
738    
739    ----------------------------------------------------------------------
740    Name: Lal George
741    Date: 2000/11/06 09:02:21 EST 2000
742    Tag: Release_110_30
743    Description:
744    
745    - Version 110.30
746    ----------------------------------------------------------------------
747    Name: Matthias Blume
748    Date: 2000/11/04 14:45:00
749    Tag: blume-20001104-mlbuild
750    Description:
751    
752    - Made ml-build faster on startup.
753    - Documentation fixes.
754    
755    ----------------------------------------------------------------------
756    Name: Matthias Blume
757    Date: 2000/11/02 17:00:00 JST
758    Tag: blume-20001102-condcomp
759    Description:
760    
761    - Small tweaks to pickler -- new BOOTFILES!
762    - Version bumped to 110.29.2.
763    - Added conditional compilation facility to init.cmi (see comment there).
764    ----------------------------------------------------------------------
765    Name: Allen Leung
766    Date: 2000/10/23 19:31:00
767    Tag: leunga-20001023-demo-ra
768    
769    1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
770    2. Test programs for the graph library updated
771    3. Some new MLRISC demo programs added
772    
773    ----------------------------------------------------------------------
774    Name: Matthias Blume
775    Date: 2000/08/31 22:15:00 JST
776    Tag: blume-20001017-errmsg
777    Description:
778    
779    More error message grief: Where there used to be no messages, there
780    now were some that had bogus error regions.  Fixed.
781    
782    ----------------------------------------------------------------------
783    Name: Matthias Blume
784    Date: 2000/08/31 17:30:00 JST
785    Tag: blume-20001017-v110p29p1
786    Description:
787    
788    I made a version 110.29.1 with new bootfiles.
789    
790    Changes:  Modified pickler/unpickler for faster and leaner unpickling.
791              CM documentation changes and a small bugfix in CM's error reporting.
792    
793    ----------------------------------------------------------------------
794    Name: Lal George
795    Date: 2000/09/27 14:42:35 EDT
796    Tag: george-20000927-nodestatus
797    Description:
798    
799    Changed the type of the nodestatus, so that:
800    
801            SPILLED(~1)             is now SPILLED
802            SPILLED(m) where m>=0   is now MEMREG(m)
803            SPILLED(s) where s<~1   is now SPILL_LOC(~s)
804    
805    ----------------------------------------------------------------------
806    Name: Matthias Blume
807    Date: 2000/09/07 14:45:00 JST
808    Tag: blume-20000907-cmerrmsg
809    Description:
810    
811    Small tweak to CM to avoid getting ML syntax error messages twice.
812    
813    ----------------------------------------------------------------------
814    Name: Matthias Blume
815    Date: 2000/08/31 18:00:00 JST
816    Tag: blume-20000831-cvsbootfiles
817    Description:
818    
819    New URL for boot files (because the 110.29 files on the BL server do
820    now work correctly with my updated install scripts for yacc and lex).
821    
822    ----------------------------------------------------------------------
823    Name: Matthias Blume
824    Date: 2000/08/08 12:33:00 JST
825    Tag: blume-20000808-manual
826    Description:
827    
828    Tiny update to CM manual.
829    
830    ----------------------------------------------------------------------
831    Name: Allen Leung
832    Date: 2000/08/7 19:31:00
833    Tag: leunga-20000807-a-whole-bunch-of-stuff
834    
835      Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
836    and Mobdy.
837    
838    1.  x86
839    
840       a.  Fixes to peephole module by John and Dan.
841       b.  Assembly fix to SETcc by Allen.
842       c.  Fix to c-call by John.
843       d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
844       e.  Instruction selection optimization to SETcc as suggested by John.
845    
846           For example,
847    
848            MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
849    
850           should generate:
851    
852            MOVL a, x
853            SUBL b, x
854            SHRL 31, x
855    
856    2.  IR stuff
857    
858         A bunch of new DJ-graph related algorithms added.  These
859         speed up SSA construction.
860    
861    3.  SSA + Scheduling
862    
863         Added code for SSA and scheduling to the repository
864    
865    ----------------------------------------------------------------------
866    Name: Lal George
867    Date: 2000/07/27 11:53:14 EDT
868    
869    Tag: lal-20000727-linux-ppc
870    Description:
871    
872     Made changes to support Linux PPC.
873     p.s. I have confirmation that the 110.29 boot files work fine.
874    
875    ----------------------------------------------------------------------
876    Name: Matthias Blume
877    Date: 2000/07/27 17:40:00 JST
878    Tag: blume-20000727-scripts
879    Description:
880    
881    !!!! WARNING !!!!
882    You must recompile the runtime system!
883    !!!! WARNING !!!!
884    
885    This is basically another round of script-enhancements:
886    
887    1. sml, ml-build, and ml-makedepend accept options -D and -U to define
888       and undefine CM preprocessor symbols.
889    
890    2. ml-build avoids generating a new heap image if it finds that the
891       existing one is still ok.  (The condition is that no ML file had to
892       be recompiled and all ML files are found to be older that the heap
893       file.)
894    
895       To make this work smoothly, I also hacked the runtime system as
896       well as SMLofNJ.SysInfo to get access to the heap image suffix
897       (.sparc-solaris, ...) that is currently being used.
898    
899       Moreover, the signature of CM.mk_standalone has changed.  See the
900       CM manual.
901    
902    3. ml-makedepend accepts additional options -n, -a, and -o.  (See the
903       CM manual for details.)
904    
905    4. More CM manual updates:
906        - all of the above has been documented.
907        - there is now a section describing the (CM-related) command line
908          arguments that are accepted by the "sml" command
909    
910    ----------------------------------------------------------------------
911    Name: Matthias Blume
912    Date: 2000/07/25 16:20:00 JST
913    Tag: blume-20000725-makedepend
914    Description:
915    
916    Added a script called ml-makedepend.  This can be used in makefiles
917    for Unix' make in a way very similar to the "makedepend" command for
918    C.
919    
920    The script internally uses function CM.sources.
921    
922    Synopsis:
923    
924        ml-makedepend [-f makefile] cmfile targetname
925    
926    The default for the makefile is "makefile" (or "Makefile" should
927    "makefile" not exist).
928    
929    ml-makedepend adds a cmfile/targetname-specific section to this
930    makefile (after removing the previous version of this section).  The
931    section contains a single dependency specification with targetname on
932    the LHS (targetname is an arbitrary name), and a list of files derived
933    from the cmfile on the RHS.  Some of the files on the RHS are
934    ARCH/OPSYS-specific.  Therefore, ml-makedepend inserts references to
935    "make" variables $(ARCH) and $(OPSYS) in place of the corresponding
936    path names.  The makefile writer is responsible for making sure that
937    these variables have correct at the time "make" is invoked.
938    
939    ----------------------------------------------------------------------
940    Name: Matthias Blume
941    Date: 2000/07/22 23:30:00 JST
942    Tag: blume-20000722-urlupdate
943    Description:
944    
945    Changed BOOT and config/srcarchiveurl to point to BL server:
946    
947        ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.29/
948    
949    ----------------------------------------------------------------------
950    Name: Matthias Blume
951    Date: 2000/07/18 18:00:00 JST
952    Tag: blume-20000718-Version_110_29
953    Description:
954    
955    1. Updated src/compiler/TopLevel/main/version.sml to version 110.29
956    
957    2. Updated config/version to 110.29
958    
959    3. Updated config/srcarchiveurl
960    
961    3. New boot files!
962       ftp://ftp.cs.princeton.edu/pub/people/blume/sml/110.29-autofetch
963    
964    ----------------------------------------------------------------------
965    Name: Matthias Blume
966    Date: 2000/07/11 13:58:00 JST
967    Tag: blume-20000711-doctypo
968    Description:
969    
970    Fixed a few typos in CM manual.
971    
972    ----------------------------------------------------------------------
973    Name: Allen Leung
974    Date: 2000/06/15 00:38:00
975    Tag: leunga-20000704-sparc-x86
976    
977    1. x86 peephole improvement sp += k; sp -= k => nop  [from John]
978    2. fix to x86 RET bug [found by Dan Grossman]
979    3. sparc assembly bug fix for ticc instructions [found by Fermin]
980    
981       Affects c-- and moby only
982    
983    ----------------------------------------------------------------------
984    Name: Matthias Blume
985    Date: 2000/07/04 15:26:00
986    Tag: blume-20000704-trigger
987    Description:
988    
989    1. Improvements to CM manual.
990    2. SMLofNJ.Internals.BTrace.trigger reinstated as an alternative way
991       of getting a back-trace.  The function, when called, raises an
992       internal exception which explicitly carries the full back-trace history,
993       so it is unaffected by any intervening handle-raise pairs ("trivial"
994       or not).  The interactive loop will print that history once it arrives
995       at top level.
996       Short of having all exceptions implicitly carry the full history, the
997       recommended way of using this facility is:
998         - compile your program with instrumentation "on"
999         - run it, when it raises an exception, look at the history
1000         - if the history is "cut off" because of some handler, go and modify
1001           your program so that it explicitly calls BTrace.trigger
1002         - recompile (still instrumented), and rerun; look at the full history
1003    
1004    ----------------------------------------------------------------------
1005    Name: Matthias Blume
1006    Date: 2000/07/03 15:36:00 JST
1007    Tag: blume-20000702-manual
1008    Description:
1009    
1010    Small corrections and updates to CM manual.
1011    
1012    ----------------------------------------------------------------------
1013    Name: Matthias Blume
1014    Date: 2000/06/29 16:04:00 JST
1015    Tag: blume-20000629-yacctool
1016    Description:
1017    
1018    Changes:
1019    
1020    1. Class "mlyacc" now takes separate arguments to pass options to
1021       generated .sml- and .sig-files independently.
1022    2. Corresponding CM manual updates.
1023    3. BTrace module now also reports call sites.  (However, for loop clusters
1024       it only shows from where the cluster was entered.)  There are associated
1025       modifications to core.sml, internals.{sig,sml}, btrace.sml, and btimp.sml.
1026    
1027    ----------------------------------------------------------------------
1028    Name: Matthias Blume
1029    Date: 2000/06/27 16:51:00 JST
1030    Tag: blume-20000627-noweb
1031    Description:
1032    
1033    Changes:
1034    
1035     1. Implemented "subdir" and "witness" options for noweb tool.
1036        This caused some slight internal changes in CM's tool implementation.
1037     2. Fixed bug in "tool plugin" mechanism.  This is essentially cleaning
1038        some remaining issues from earlier path anchor changes.
1039     3. Updated CM manual accordingly.
1040    
1041     4. Changed implementation of back-tracing so that I now consider it
1042        ready for prime-time.
1043    
1044        In particular, you don't have to explicitly trigger the back-trace
1045        anymore.  Instead, if you are running BTrace-instrumented code and
1046        there is an uncaught exception (regardless of whether or not it was
1047        raised in instrumented code), the top-level evalloop will print
1048        the back-trace.
1049    
1050        Features:
1051    
1052          - Instrumented and uninstrumented code work together seemlessly.
1053            (Of course, uninstrumented code is never mentioned in actual
1054             back-traces.)
1055    
1056          - Asymptotic time- and space-complexity of instrumented code is
1057            equal to that of uninstrumented code.  (This means that
1058            tail-recursion is preserved by the instrumentation phase.)
1059    
1060          - Modules whose code has been instrumented in different sessions
1061            work together without problem.
1062    
1063          - There is no penalty whatsoever on uninstrumented code.
1064    
1065          - There is no penalty on "raise" expressions, even in
1066            instrumented code.
1067    
1068        A potential bug (or perhaps it is a feature, too):
1069    
1070          A back-trace reaches no further than the outermost instrumented
1071          non-trivial "raise".  Here, a "trivial" raise is one that is the
1072          sole RHS of a "handle" rule.  Thus, back-traces reach trough
1073    
1074               <exp> handle e => raise e
1075    
1076          and even
1077    
1078               <exp> handle Foo => raise Bar
1079    
1080          and, of course, through
1081    
1082               <exp> handle Foo => ...
1083    
1084         if the exception was not Foo.
1085    
1086         Back-traces always reach right through any un-instrumented code
1087         including any of its "handle" expressions, trivial or not.
1088    
1089       To try this out, do the following:
1090    
1091         - Erase all existing binfiles for your program.
1092           (You may keep binfiles for those modules where you think you
1093            definitely don't need back-tracing.)
1094         - Turn on back-trace instrumentation:
1095              SMLofNJ.Internals.BTrace.mode (SOME true);
1096         - Recompile your program.  (I.e., run "CM.make" or "use".)
1097         - You may now turn instrumentation off again (if you want):
1098              SMLofNJ.Internals.BTrace.mode (SOME false);
1099         - Run your program as usual.  If it raises an exception that
1100           reaches the interactive toplevel, then a back-trace will
1101           automatically be printed.  After that, the toplevel loop
1102           will print the exception history as usual.
1103    
1104    ----------------------------------------------------------------------
1105    Name: Matthias Blume
1106    Date: 2000/06/26 09:56:46 JST
1107    Tag: blume-20000626-setup
1108    Description:
1109    
1110    CM: - setup-parameter to "sml" added; this can be used to run arbitrary
1111          ML code before and after compiling a file (e.g., to set compiler
1112          flags)
1113    
1114    Compiler: - improved btrace API (in core.sml, internals.{sig,sml})
1115              - associated changes to btrace.sml (BTrace instrumentation pass)
1116              - cleaner implementation of btimp.sml (BTrace tracing and report
1117                module)
1118    
1119    CM manual: * new path encoding documented
1120               * description of setup-parameter to "sml" added
1121    
1122    The biggest user-visible change to back-tracing is that it is no
1123    longer necessary to compile all traced modules within the same
1124    session.  (This was a real limitation.)
1125    
1126    ----------------------------------------------------------------------
1127    Name: Matthias Blume
1128    Date: 2000/06/24 12:40:00 JST
1129    Tag: blume-20000624-startup
1130    Description:
1131    
1132    Fixes startup slowdown problem.  (I was calling SrcPath.sync a _tad_
1133    bit too often -- to put it mildly. :)
1134    
1135    ----------------------------------------------------------------------
1136    Name: Matthias Blume
1137    Date: 2000/06/23 18:20:00 JST
1138    Tag: blume-20000623-btrace
1139    Description:
1140    
1141    This updates adds a backtrace facility to aid programmers in debugging
1142    their programs.  This involves the following changes:
1143    
1144    1. Module system/smlnj/init/core.sml (structure _Core) now has hooks for
1145       keeping track of the current call stack.  When programs are compiled
1146       in a special mode, the compiler will insert calls to these hooks
1147       into the user program.
1148       "Hook" means that it is possible for different implementations of
1149       back-tracing to register themselves (at different times).
1150    
1151    2. compiler/MiscUtil/profile/btrace.sml implements the annotation phase
1152       as an Absyn.dec->Absyn.dec rewrite.  Normally this phase is turned off.
1153       It can be turned on using this call:
1154         SMLofNJ.Internals.BTrace.mode (SOME true);
1155       Turning it off again:
1156         SMLofNJ.Internals.BTrace.mode (SOME false);
1157       Querying the current status:
1158         SMLofNJ.Internals.BTrace.mode NONE;
1159       Annotated programs are about twice as big as normal ones, and they
1160       run a factor of 2 to 4 slower with a dummy back-trace plugin (one
1161       where all hooks do nothing).  The slowdown with a plugin that is
1162       actually useful (such as the one supplied by default) is even greater,
1163       but in the case of the default plugin it is still only an constant
1164       factor (amortized).
1165    
1166    3. system/Basis/Implementation/NJ/internals.{sig,sml} have been augmented
1167       with a sub-structure BTrace for controlling back-tracing.  In particular,
1168       the above-mentioned function "mode" controls whether the annotation
1169       phase is invoked by the compiler.  Another important function is
1170       "trigger": when called it aborts the current execution and causes
1171       the top-level loop to print a full back-trace.
1172    
1173    4. compiler/MiscUtil/profile/btimp.sml is the current default plugin
1174       for back-tracing.  It keeps track of the dynamic call stack and in
1175       addition to that it keeps a partial history at each "level" of that
1176       stack.  For example, if a tail-calls b, b tail-calls c, and c tail-calls
1177       d and b (at separate times, dynamically), then the report will show:
1178    
1179       GOTO   d
1180             /c
1181       GOTO  \b
1182       CALL   a
1183    
1184       This shows that there was an initial non-tail call of a, then a
1185       tail-call to b or c, looping behavior in a cluster of functions that
1186       consist of b and c, and then a goto from that cluster (i.e., either from
1187       b or from c) to d.
1188    
1189       Note that (depending on the user program) the amount of information
1190       that the back-trace module has to keep track of at each level is bounded
1191       by a constant.  Thus, the whole implementation has the same asymptotical
1192       complexity as the original program (both in space and in time).
1193    
1194    5. compiler/TopLevel/interact/evalloop.sml has been modified to
1195       handle the special exception SMLofNJ.Internals.BTrace.BTrace
1196       which is raised by the "trigger" function mentioned above.
1197    
1198    Notes on usage:
1199    
1200    - Annotated code works well together with unannotated code:
1201    Unannotated calls simply do not show up at all in the backtrace.
1202    
1203    - It is not a good idea to let modules that were annotated during
1204    different sessions run at the same time.  This is because the compiler
1205    chooses small integers to identify individual functions, and there
1206    will be clashes if different modules were compiled in separate sessions.
1207    (Nothing will crash, and you will even be told about the clashes, but
1208    back-trace information will in general not be useful.)
1209    
1210    - Back-tracing can be confused by callcc and capture.
1211    
1212    - The only way of getting a back-trace right now is to explicitly
1213    invoke the "trigger" function from your user program.  Eventually, we
1214    should make every exception carry back-trace information (if
1215    available).  But since this creates more overhead at "raise"-time
1216    (similar to the current exnHistory overhead), I have not yet
1217    implemented this.  (The implementation will be rather easy.)  With
1218    exceptions carrying back-trace information, this facility will be even
1219    more useful because users don't need to modify their programs...
1220    
1221    - While it is possible to compile the compiler with back-trace
1222    annotations turned on (I did it to get some confidence in
1223    correctness), you must make absolutely sure that core.sml and
1224    btimp.sml are compiled WITHOUT annotation!  (core.sml cannot actually
1225    be compiled with annotation because there is no core access yet, but
1226    if you compile btimp.sml with annotation, then the system will go into
1227    an infinite recursion and crash.)
1228    Since CM currently does not know about BTrace, the only way to turn
1229    annotations on and off for different modules of the compiler is to
1230    interrupt CMB.make, change the settings, and re-invoke it.  Of course,
1231    this is awkward and clumsy.
1232    
1233    Sample sessions:
1234    
1235    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1236    - SMLofNJ.Internals.BTrace.mode (SOME true);
1237    [autoloading]
1238    [autoloading done]
1239    val it = false : bool
1240    - structure X = struct
1241    -     fun main n = let
1242    -         fun a (x, 0) = d x
1243    -           | a (x, n) = b (x, n - 1)
1244    -         and b (x, n) = c (x, n)
1245    -         and c (x, n) = a (x, n)
1246    -         and d x = e (x, 3)
1247    -         and e (x, 0) = f x
1248    -           | e (x, n) = e (x, n - 1)
1249    -         and f 0 = SMLofNJ.Internals.BTrace.trigger ()
1250    -           | f n = n * g (n - 1)
1251    -         and g n = a (n, 3)
1252    -     in
1253    -         f n
1254    -     end
1255    - end;
1256    structure X : sig val main : int -> int end
1257    - X.main 3;
1258    *** BACK-TRACE ***
1259    GOTO   stdIn:4.2-13.20: X.main[2].f
1260    GOTO-( stdIn:4.2-13.20: X.main[2].e
1261    GOTO   stdIn:4.2-13.20: X.main[2].d
1262         / stdIn:4.2-13.20: X.main[2].a
1263         | stdIn:4.2-13.20: X.main[2].b
1264    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1265    CALL   stdIn:4.2-13.20: X.main[2].g
1266    GOTO   stdIn:4.2-13.20: X.main[2].f
1267    GOTO-( stdIn:4.2-13.20: X.main[2].e
1268    GOTO   stdIn:4.2-13.20: X.main[2].d
1269         / stdIn:4.2-13.20: X.main[2].a
1270         | stdIn:4.2-13.20: X.main[2].b
1271    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1272    CALL   stdIn:4.2-13.20: X.main[2].g
1273    GOTO   stdIn:4.2-13.20: X.main[2].f
1274    GOTO-( stdIn:4.2-13.20: X.main[2].e
1275    GOTO   stdIn:4.2-13.20: X.main[2].d
1276         / stdIn:4.2-13.20: X.main[2].a
1277         | stdIn:4.2-13.20: X.main[2].b
1278    GOTO-\ stdIn:4.2-13.20: X.main[2].c
1279    CALL   stdIn:4.2-13.20: X.main[2].g
1280    GOTO   stdIn:4.2-13.20: X.main[2].f
1281    CALL   stdIn:2.15-17.4: X.main[2]
1282    -
1283    
1284    (Note that because of a FLINt bug the above code currently does not
1285    compile without BTrace turned on.)
1286    
1287    Here is another example, using my modified Tiger compiler:
1288    
1289    Standard ML of New Jersey v110.28.1 [FLINT v1.5], June 5, 2000
1290    - SMLofNJ.Internals.BTrace.mode (SOME true);
1291    [autoloading]
1292    [autoloading done]
1293    val it = false : bool
1294    - CM.make "sources.cm";
1295    [autoloading]
1296    ...
1297    [autoloading done]
1298    [scanning sources.cm]
1299    [parsing (sources.cm):parse.sml]
1300    [creating directory CM/SKEL ...]
1301    [parsing (sources.cm):tiger.lex.sml]
1302    ...
1303    [wrote CM/sparc-unix/semant.sml]
1304    [compiling (sources.cm):main.sml]
1305    [wrote CM/sparc-unix/main.sml]
1306    [New bindings added.]
1307    val it = true : bool
1308    - Main.compile ("../testcases/merge.tig", "foo.out");
1309    *** BACK-TRACE ***
1310    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trvar
1311    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1312    CALL   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1313    GOTO   lib/semant.sml:289.3-295.22: SemantFun[2].transExp.trexp.check[2]
1314    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1315    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1316    CALL   lib/semant.sml:488.3-505.6: SemantFun[2].transDec.trdec[2].transBody[2]
1317         / lib/semant.sml:411.65-543.8: SemantFun[2].transDec
1318    CALL-\ lib/semant.sml:413.2-540.9: SemantFun[2].transDec.trdec[2]
1319    CALL   lib/semant.sml:99.2-396.21: SemantFun[2].transExp.trexp
1320    CALL   lib/semant.sml:8.52-558.4: SemantFun[2].transProg[2]
1321    CALL   main.sml:1.18-118.4: Main.compile[2]
1322    -
1323    
1324    ----------------------------------------------------------------------
1325    Name: Matthias Blumen
1326    Date: 2000/06/21 18:00:00 JST
1327    Tag: blume-20000621-manual
1328    Description:
1329    
1330    CM manual update: Path environments documented.
1331    
1332    ----------------------------------------------------------------------
1333    Name: Matthias Blume
1334    Date: 2000/06/19 13:40:00
1335    Tag: blume-20000619-manual
1336    Description:
1337    
1338    CM manual and system/README update.  This only covers the fact that
1339    there are no more implicit anchors.  (Path environments and the "bind"
1340    option to "cm" have yet to be documented.)
1341    
1342    ----------------------------------------------------------------------
1343    Name: Matthias Blume
1344    Date: 2000/06/19 11:05:00 JST
1345    Tag: blume-20000619-chdir-bugfix
1346    Description:
1347    
1348    Fixed a bug in new SrcPath module that sometimes led to a bad chDir call.
1349    
1350    ----------------------------------------------------------------------
1351    Name: Matthias Blume
1352    Date: 2000/06/18 22:00:10 JST
1353    Tag: blume-20000618-implicit-anchors-really-gone
1354    Description:
1355    
1356    I updates the previous HISTORY entry where I forgot to mention that
1357    implicit anchors are no longer with us.
1358    
1359    The current update also gets rid of the (now useless) controller
1360    CM.Control.implicit_anchors.
1361    
1362    ----------------------------------------------------------------------
1363    Name: Matthias Blume
1364    Date: 2000/06/16 17:30:00 JST
1365    Tag: blume-20000616-anchorenv
1366    Description:
1367    
1368    This patch implements the long anticipated (just kidding :) "anchor
1369    environment" mechanism.  In the course of doing this, I also
1370    re-implemented CM's internal "SrcPath" module from scratch.  The new
1371    one should be more robust in certain boundary cases.  In any case, it
1372    is a lot cleaner than its predecessor (IMHO).
1373    
1374    This time, although there is yet another boot file format change, I
1375    kept the unpickler backward-compatible.  As a result, no new bootfiles
1376    are necessary and bootstrapping is straightforward.  (You cannot read
1377    new bootfiles into an old system, but the other way around is no
1378    problem.)
1379    
1380    Visible changes:
1381    
1382    ** 0. Implicit path anchors (without the leading $-symbol) are no
1383    longer recognized at all. This means that such path names are not
1384    illegal either.  For example, the name basis.cm simply refers to a
1385    local file called "basis.cm" (i.e, the name is an ordinary path
1386    relative to .cm-files directory).  Or, to put it differently, only
1387    names that start with $ are anchored paths.
1388    
1389    ** 1. The $<singlearc> abbreviation for $/<singlearc> has finally
1390    vanished.
1391    
1392    John (Reppy) had critizised this as soon as I originally proposed and
1393    implemented it, but at that time I did not really deeply believe
1394    him. :) Now I came full-circle because I need the $<singlearc> syntax
1395    in another place where it cannot be seen as an abbreviation for
1396    $/<singlearc>.  To avoid the confusion, $<singlearc> now means what it
1397    seems to mean (i.e., it "expands" into the corresponding anchor
1398    value).
1399    
1400    However, when paths are used as members in CM description files, it
1401    continues to be true that there must be at least another arc after the
1402    anchor.  This is now enforced separately during semantic analysis
1403    (i.e., from a lexical/syntactical point of view, the notation is ok.)
1404    
1405    ** 2. The "cm" class now accepts an option "bind".  The option's value
1406    is a sub-option list of precisely two items -- one labeled "anchor"
1407    and the other one labeled "value".  As you might expect, "anchor" is
1408    used to specify an anchor name to be bound, and "value" specifies what
1409    the anchor is being bound to.
1410    
1411    The value must be a directory name and can be given in either standard
1412    syntax (including the possibility that it is itself an anchored path)
1413    or native syntax.
1414    
1415    Examples:
1416    
1417       foo.cm (bind:(anchor:bar value:$mystuff/bar))
1418       lib.cm (bind:(anchor:a value:"H:\\x\\y\\z"))  (* only works under windows *)
1419    
1420    and so on.
1421    
1422    The meaning of this is that the .cm-file will be processed with an
1423    augmented anchor environment where the given anchor(s) is/are bound to
1424    the given values(s).
1425    
1426    The rationale for having this feature is this: Suppose you are trying
1427    to use two different (already stable) libraries a.cm and b.cm (that
1428    you perhaps didn't write yourself).  Further, suppose each of these
1429    two libraries internally uses its own auxiliary library $aux/lib.cm.
1430    Normally you would now have a problem because the anchor "lib" can not
1431    be bound to more than one value globally.  Therefore, the project that
1432    uses both a.cm and b.cm must locally redirect the anchor to some other
1433    place:
1434    
1435       a.cm (bind:(anchor:lib value:/usr/lib/smlnj/a-stuff))
1436       b.cm (bind:(anchor:lib value:/usr/lib/smlnj/b-stuff))
1437    
1438    This hard-wires $lib/aux.cm to /usr/lib/smlnj/a-stuff/aux.cm or
1439    /usr/lib/smlnj/b-stuff/aux.cm, respectively.
1440    
1441    Hard-wiring path names is a bit inflexible (and CM will verbosely warn
1442    you when you do so at the time of CM.stabilize).  Therefore, you can
1443    also use an anchored path as the value:
1444    
1445      a.cm (bind:(anchor:lib value:$a-lib))
1446      b.cm (bind:(anchor:lib value:$b-lib))
1447    
1448    Now you can globally configure (using the usual CM.Anchor.anchor or
1449    pathconfig machinery) bindings for "a-lib" and "b-lib".  Since "lib"
1450    itself is always locally bound, setting it globally is no longer
1451    meaningful or necessary (but it does not hurt either).  In fact, "lib"
1452    can still be used as a global anchor for separate purposes.  As a
1453    matter of fact, one can locally define "lib" in terms of a global
1454    "lib":
1455    
1456      a.cm (bind:(anchor:lib value:$lib/a))
1457      b.cm (bind:(anchor:lib value:$lib/b))
1458    
1459    ** 3: The encoding of path names has changed.  This affects the way
1460    path names are shown in CM's progress report and also the internal
1461    protocol encoding used for parallel make.
1462    
1463    The encoding now uses one or more ':'-separated segments.  Each
1464    segments corresponds to a file that has been specified relative to the
1465    file given by its preceding segment.  The first segment is either
1466    relative to the CWD, absolute, or anchored.  Each segment itself is
1467    basically a Unix pathname; all segments but the first are relative.
1468    
1469    Example:
1470    
1471       $foo/bar/baz.cm:a/b/c.sml
1472    
1473    This path denotes the file bar/a/b/c.sml relative to the directory
1474    denoted by anchor "foo".  Notice that the encoding also includes
1475    baz.cm which is the .cm-file that listed a/b/c.sml.  As usual, such
1476    paths are resolved relative to the .cm-files directory, so baz.cm must
1477    be ignored to get the "real" pathname.
1478    
1479    To make this fact more obvious, CM puts the names of such "virtual
1480    arcs" into parentheses when they appear in progress reports. (No
1481    parentheses will appear in the internal protocol encoding.)  Thus,
1482    what you really see is:
1483    
1484      $foo/bar/(baz.cm):a/b/c.sml
1485    
1486    I find this notation to be much more informative than before.
1487    
1488    Another new feature of the encoding is that special characters
1489    including parentheses, colons, (back)slashes, and white space are
1490    written as \ddd (where ddd is the decimal encoding of the character).
1491    
1492    *** The CM manual still needs to be updated.
1493    
1494    ----------------------------------------------------------------------
1495    Name: Allen Leung
1496    Date: 2000/06/15 00:38:00
1497    Tag: leunga-20000615-x86-peephole
1498    
1499    x86 Peephole fix by Fermin.  Affects c-- and moby only.
1500    
1501    ----------------------------------------------------------------------
1502    Name: Matthias Blume
1503    Date: 2000/06/12 11:40:00
1504    Tag: blume-20000612-parmakefix
1505    Description:
1506    
1507    More cleanup after changing the file naming scheme: This time I
1508    repaired the parallel make mechanism for CMB.make which I broke earlier.
1509    
1510    ----------------------------------------------------------------------
1511    Name: Allen Leung
1512    Date: 2000/06/09 01:25:00
1513    Tag: leunga-20000609-various
1514    
1515    None of these things should affect normal SML/NJ operations
1516    
1517    1. Peephole improvements provided by Fermin (c--)
1518    2. New annotation DEFUSE for adding extra dependence (moby)
1519    3. New X86 LOCK instructions (moby)
1520    4. New machine description language for reservation tables (scheduling)
1521    5. Fixes to various optimization/analysis modules (branch chaining, dominator
1522       trees etc.)
1523    6. I've changed the CM files so that they can work with versions
1524       110.0.6, 110.25 and 110.28
1525    
1526    ----------------------------------------------------------------------
1527    Name: Matthias Blume
1528    Date: 2000/06/09 12:40:00
1529    Tag: blume-20000609-log
1530    Description:
1531    
1532    - Removed all(?) remaining RCS Log entries from sources.
1533    
1534    - Fixed bug in ml-yacc and ml-lex sources (use explicit anchors for
1535      anchored paths).
1536    
1537    ----------------------------------------------------------------------
1538    Name: Matthias Blume
1539    Date: 2000/06/07 17:00:00 JST
1540    Tag: blume-20000607-no-implicit-anchors
1541    Description:
1542    
1543    1. This update changes the default setting for
1544    CM.Control.implicit_anchors from true to false.  This means that
1545    implicit anchors are no longer permitted by default.  I also tried to
1546    make sure that nothing else still relies on implicit anchors.
1547    (This is the next step on the schedule towards a CM that does not even
1548    have the notion of implicit anchors anymore.)
1549    
1550    2. More CM manual updates.
1551    
1552    3. I managed to track down and fix the pickling bug I mentioned last
1553    time.  Because of the previously existing workaround, this entails no
1554    immediate practical changes.
1555    
1556    ----------------------------------------------------------------------
1557    Name: Matthias Blume
1558    Date: 2000/06/06 11:15:00 JST
1559    Tag: blume-20000606-lazierpickle
1560    Description:
1561    
1562    !!!! NEW BOOT FILES !!!!
1563    
1564    * The main purpose of this update is to make library pickles lazier in
1565    order to reduce the initial space penalty for autoloading a library.
1566    As a result, it is now possible to have $smlnj/compiler.cm
1567    pre-registered.  This should take care of the many complaints or
1568    inquiries about missing structure Compiler.  This required changes to
1569    CM's internal data structures and small tweaks to some algorithms.
1570    
1571    As a neat additional effect, it is no longer necessary (for the sake
1572    of lean heap image files) to distinguish between a "minimal" CM and a
1573    "full" CM.  Now, there is only one CM (i.e., the "full" version:
1574    $smlnj/cm.cm aka $smlnj/cm/full.cm), and it is always available at the
1575    interactive top level. ($smlnj/cm/minimal.cm is gone.)
1576    
1577    To make the life of compiler-hackers easier, "makeml" now also
1578    pre-registers $smlnj/cmb.cm (aka $smlnj/cmb/current.cm).  In other
1579    words, after you bootstrap a new sml for the first time, you will not
1580    have to autoload $smlnj/cmb.cm again afterwards.  (The first time
1581    around you will still have to do it, though.)
1582    
1583    * A second change consists of major updates to the CM manual.  There
1584    are now several appendices with summary information and also a full
1585    specification of the CM description file syntax.
1586    
1587    * In directory src/system I added the script "allcross".  This script
1588    invokes sml and cross-compiles the compiler for all supported
1589    architectures.  (Useful when providing a new set of boot files.)
1590    
1591    * There seems to be a latent bug in my "lazy pickles" mechanism.  I
1592    added a small tweak to pickle-util.sml to work around this problem,
1593    but it is not a proper fix yet.  I will investigate further.  (The
1594    effect of the bug was an inflation of library pickle size.)
1595    
1596    * Version number increased to 110.28.1 (to avoid compatibility problems).
1597    
1598    ----------------------------------------------------------------------
1599    Name: Allen Leung
1600    Date: 2000/05/25 17:28 EDT
1601    Tag: leunga-20000525-ra
1602    Description:
1603    
1604      Fixed a bug in freezing phase of the register allocator.
1605    
1606    ----------------------------------------------------------------------
1607    Name: Allen Leung
1608    Date: 2000/05/15 22:53 EDT
1609    Tag: leunga-20000515-alpha-x86-ra
1610    Description:
1611    
1612      1. Alpha
1613    
1614          Slight cleanup.  Removed the instruction SGNXL
1615    
1616      2. X86
1617    
1618          Added the following instructions to the instruction set:
1619    
1620            ROLx, RORx,
1621            BTx, BTSx, BTLx, BTRx,
1622            XCHGx, and variants with the LOCK prefix
1623    
1624      3. Register Allocation
1625    
1626          The module ra-rewrite-with-renaming has been improved.
1627    
1628      These have no effect on SML/NJ.
1629    
1630    ----------------------------------------------------------------------
1631    Name: Matthias Blume
1632    Date: 2000/05/15 16:20:00 JST
1633    Tag: blume-20000515-lightrebuild
1634    Description:
1635    
1636    1. I added an alternative to "-rebuild" to "makeml".  The difference is
1637       that prior to calling CMB.make' the CM-variable "LIGHT" will be
1638       defined.  In effect, the command will not build any cross-compiler
1639       backends and therefore finish more quickly.
1640    
1641       The "fixpt" script also takes a "-light" switch to be able to use
1642       this new facility while compiling for a fixpoint.
1643    
1644    2. I replaced all mentions of anchored paths in group owner specifications
1645       with simple relative paths (usually starting with "..").
1646       The rationale is that a library's internal workings should not be
1647       compromised by the lack of some anchor.  (An anchor is necessary
1648       for someone who wants to refer to the library by an anchored path,
1649       but it should not be necessary to build the same library in the first
1650       place.)
1651    
1652    3. I changed the way CM's tool mechanism determines the shell command
1653       string used for things like ml-yacc etc. so that it does not break
1654       when CM.Control.implicit_anchors is turned off.
1655    
1656    ----------------------------------------------------------------------
1657    Name: Matthias Blume
1658    Date: 2000/05/12 18:20:00 JST
1659    Tag: blume-20000512-ml-build
1660    Description:
1661    
1662    Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from
1663    getting installed properly (by config/install.sh).
1664    
1665    ----------------------------------------------------------------------
1666    Name: Matthias Blume
1667    Date: 2000/05/12 17:30:00 JST
1668    Tag: blume-20000512-anchors
1669    Description:
1670    
1671    !!! NEW BOOT FILES !!!
1672    
1673    This change is in preparation of fading out support for "implicitly
1674    anchored path names".  I went through all sources and used the
1675    explicit (and relatively new) $-notation.  See system/README and the
1676    CM manual for more info on this.
1677    
1678    I also modified the anchoring scheme for some things such as "smlnj",
1679    "MLRISC", "cm", etc. to take advantage of the fact that explicit
1680    anchors are more expressive: anchor name and first arc do not have to
1681    coincide.  This entails the following user-visible change:
1682    
1683    You have to write $smlnj/foo/bar instead of smlnj/foo/bar.  In
1684    particular, when you fire up sml with a command-line argument, say,
1685    e.g.:
1686    
1687       sml '$smlnj/cmb.cm'
1688    
1689    At the ML toplevel prompt:
1690    
1691       CM.autoload "$smlnj/cmb.cm";
1692    
1693    There is also a new controller in CM.Control that can be used to turn
1694    off all remaining support for implicit anchors by saying:
1695    
1696        CM.autoload "$smlnj/
1697        #set CM.Control.implicit_anchors false;
1698    
1699    This causes CM to reject implicitly anchored paths.  This is (for the
1700    time being) less permissive than the "final" version where there will
1701    be no more such implicit anchors and relative paths will be just that:
1702    relative.
1703    
1704    The next step (version after next version?) will be to make the
1705    default for CM.Control.implicit_anchors false.  After the dust has
1706    settled, I can then produce the "final" version of this...
1707    
1708    Note: Since bootstrapping is a bit tricky, I provided new boot files.
1709    
1710    ----------------------------------------------------------------------
1711    Name: Matthias Blume
1712    Date: 2000/05/11 16:30:00 JST
1713    Tag: blume-20000511-sources
1714    Description:
1715    
1716    The main change is that I added function CM.sources as a generalized
1717    version of the earlier CM.makedepend.  This entails the following
1718    additional changes:
1719    
1720      - CM.makedepend has been dropped.
1721    
1722      - CM manual has been updated.
1723    
1724      - TOOLS signature and API have been changed.
1725    
1726    ----------------------------------------------------------------------
1727    Name: Allen Leung
1728    Date: 2000/05/10 21:17 EDT
1729    Tag: leunga-20000510-moby-c--ssa
1730    Description:
1731    
1732      Various bug fixes and new features for C--, Moby and MLRISC optimizations.
1733    None of these affect SML/NJ.
1734    
1735    1. Register Allocation
1736    
1737        a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented.
1738           This module tries to remove local (i.e. basic block level) redundancies
1739           during spilling.
1740    
1741        b. A new framework for performing region based register allocation.
1742           Not yet entirely functional.
1743    
1744    2. X86
1745    
1746       a. DefUse for POP was missing the stack pointer [found by Lal]
1747       b. Reload for CALL was incorrect in X86Spill [found by John]
1748       c. Various fixes in X86Spill so that it can be used correctly for
1749          the new spilling module.
1750    
1751    3. SSA/IR
1752    
1753       a. New module ir/dj-dataflow.sml implements elimination based
1754          data flow analysis.
1755    
1756    4. MLRiscGen
1757    
1758       a. Fix for gc type annotation
1759    
1760    5. MDGen
1761    
1762       Various fixes for machine description -> ml code translation.  For ssa
1763       only.
1764    
1765    ----------------------------------------------------------------------
1766    Name: Allen Leung
1767    Date: 2000/05/08 22:17 EDT
1768    Tag: leunga-20000508-labexp
1769    Description:
1770    
1771      Fermin has found a few assembly problems with constant expressions
1772      generated in LabelExp.  Mostly, the problems involve extra parentheses,
1773      which choke on dumb assemblers.  This is his fix.
1774    
1775    ----------------------------------------------------------------------
1776    Name: Dave MacQueen
1777    Date: 2000/04/09 14:00 EDT
1778    Tag: dbm-20000502-Version_110_28
1779    Description:
1780    
1781    1. Updated src/compiler/TopLevel/main/version.sml to version 110.28
1782    
1783    2. Updated config/version to 110.28
1784    
1785    3. Updated config/srcarchiveurl
1786    
1787    3. New boot files!
1788       ftp://ftp.research.bell-labs.com/dist/smlnj/working/110.28/
1789    
1790    ----------------------------------------------------------------------
1791    Name: Matthias Blume
1792    Date: 2000/05/01 19:05:00 JST
1793    Tag: blume-20000501-noweb
1794    Description:
1795    
1796    A new noweb tool has been added.  The existing system is entirely
1797    unaffected by this, but some CM users have asked for renewed noweb
1798    support.  Everything is documented in the CM manual.
1799    
1800    New (plugin) libraries:
1801    
1802       noweb-tool.cm
1803       nw-ext.cm
1804    
1805    ----------------------------------------------------------------------
1806    Name: Dave MacQueen
1807    Date: 2000/04/30 12:40PM EDT
1808    Tag: dbm-20000430-bug_fixes
1809    Description:
1810    
1811    1. Fix for bug 1498
1812       smlnj/src/system/Basis/Implementation/Unsafe/object.sig
1813       smlnj/src/system/Basis/Implementation/Unsafe/object.sml
1814         added toRealArray function
1815       smlnj/src/compiler/MiscUtil/print/ppobj.sml
1816         added check for tag Obj.RealArray to array printing case in ppObj
1817    
1818    2. Fix for bug 1510
1819       smlnj/src/compiler/Semant/types/typesutil.sml
1820         fixed definition of dummyargs (used by equalTycon) so that
1821         dummy args are distinct types
1822    
1823    ----------------------------------------------------------------------
1824    Name: Matthias Blume
1825    Date: 2000/04/30 01:00:00 JST
1826    Tag: blume-20000430-versions
1827    Description:
1828    
1829    1. CM version numbering added.  This is an implementation of Lal's
1830       proposal for adding version numbers and version checking to .cm
1831       files.  Lal said that his proposal was just that -- a proposal.
1832       For the time being I went ahead and implemented it so that people
1833       can comment on it.  Everything is completely backward-compatible
1834       (except for the stable library format, i.e., new bootfiles!).
1835    
1836       As usual, see the CM manual for details.
1837    
1838    2. An alternative syntax for anchored paths has been implemented.
1839       Dave has recently voiced the same concerns that I had when I did
1840       this, so there should be some support.  My take is that eventually
1841       I will let support for the current syntax (where anchors are
1842       "implicit") fade out in favor of the new, explicit syntax.
1843       In order to be backward-compatible, both old and new syntax are
1844       currently supported.
1845    
1846       Again, see the CM manual for details.
1847    
1848    3. Parallel make is trying to be slightly smarter:  When the master
1849       process finds a "bottleneck", i.e., when there is only one
1850       compilation unit that can be compiled and everybody else is
1851       waiting on it, then it will simply compile it directly instead
1852       of clumsily telling one of the slaves to do it.
1853    
1854    4. Support for "unsharing" added.  This is necessary in order to be
1855       able to have two different versions of the same library running
1856       at the same time (e.g., for trying out a new MLRISC while still
1857       having the old MLRISC linked into the current compiler, etc.)
1858       See the CM manual.
1859    
1860    5. Simple "makedepend" functionality added for generating Makefile
1861       dependency information.  (This is rather crude at the moment.
1862       Expect some changes here in the future.)
1863    
1864    6. ".fun" added as a recognized suffix for ML files. Also documented
1865       explicitly in the manual that the fallback behavior (unknown suffix
1866       -> ML file) is not an official feature!
1867    
1868    7. Small changes to the pickler for stable libraries.
1869    
1870    8. Several internal changes to CM (for cleanup/improvement).
1871    
1872    
1873    !!!! NEW BINFILES !!!!
1874    
1875    ----------------------------------------------------------------------
1876    Name: Matthias Blume
1877    Date: 2000/04/28 17:30:00 JST
1878    Tag: blume-20000428-pathconfig
1879    Description:
1880    
1881    1. I changed config/install.sh to remove duplicate entries from the
1882       lib/pathconfig file at the end.  Moreover, the final version of
1883       lib/pathconfig is sorted alphabetically.  The same (sorting) is done
1884       in src/system/installml.
1885    
1886    2. The config/install.sh script now consistently uses relative
1887       pathnames in lib/pathconfig whenever the anchor is in the lib
1888       directory.  (So far this was true for the libraries that come
1889       pre-compiled and bundled as part of the bootfiles but not for
1890       libraries that are compiled by the script itself.)
1891    
1892    ----------------------------------------------------------------------
1893    Name: Matthias Blume
1894    Date: 2000/04/26 13:10:00 JST
1895    Tag: blume-20000426-fun_suffix
1896    Description:
1897    
1898    Added ".fun" as a recognized file name suffix (for ML code).
1899    
1900    ----------------------------------------------------------------------
1901    Name: Allen Leung
1902    Date: 2000/04/25 17:00:00 EST
1903    Tag: leunga-20000425-alpha-ra
1904    Description:
1905    
1906    1. Alpha
1907    
1908        PSEUDOARITH was missing in AlphaRewrite.  This causes an endless loop
1909    in C--.
1910    
1911    2. RA
1912    
1913       Added a flag "ra-dump-size" to print out the size of the flowgraph
1914       and the interference graph.
1915    
1916    ----------------------------------------------------------------------
1917    Name: Dave MacQueen
1918    Date: 2000/04/25/
1919    Tag: dbm-20000425-mlyacc_doc_examples
1920    Description:
1921      Updated mlyacc.tex sections 5 and 7 for SML '97 and CM.
1922      Updated all three examples in src/ml-yacc/examples to run
1923      under 110.* using CM.make.
1924    
1925    ----------------------------------------------------------------------
1926    Name: Allen Leung
1927    Date: 2000/04/20 23:04:00 EST
1928    Tag: leunga-20000420-ssa-c---stuff
1929    Description:
1930    
1931      This update synchronizes my repository with Yale's.  Most of these
1932    changes, however, do not affect SML/NJ at all (the RA is an exception).
1933    
1934    1. Register Allocator
1935    
1936       a. An improvement in the interference graph construction:
1937          Given a copy
1938    
1939                s <- t
1940    
1941          no interference edge between s and t is added for this definition of s.
1942    
1943       b. I've added two new spill heuristic modules that Fermin and I developed
1944          (in the new library RA.cm). These are unused in SML/NJ but maybe
1945          useful for others (Moby?)
1946    
1947    2. X86
1948    
1949       a. Various fixes in the backend provided by Fermin [C--] and Lal.
1950    
1951    3. Alpha
1952    
1953       a. Added the BSR instruction and code generation that goes with it [C--]
1954       b. Other fixes too numerous to recount provided by Fermin [C--]
1955    
1956    4. Regmaps
1957    
1958       a. The regmaps are not initialized with the identity physical bindings
1959          at creation time.  This is unneeded.
1960    
1961    5. MLRISC Optimizations
1962    
1963       a. The DJ-Graph module can now compute the iterated dominance frontiers
1964          intersects with liveness incrementally in linear time! Woohoo!
1965          This is now used in my new SSA construction algorithm.
1966    
1967       b. THe branch reorganization module is now smarter about linear chains of
1968          basic blocks.
1969    
1970    
1971    ----------------------------------------------------------------------
1972    Name: Matthias Blume
1973    Date: 2000/04/12 13:52:00 JST
1974    Tag: blume_main_v110p27_1
1975    Description:
1976    
1977    Changed install.sh script to handle archive files without version number
1978    and to use "boot.<arch>-<os>" instead of "sml.boot.<arch>-<os>" for the
1979    name of the boot file archive.
1980    
1981    ----------------------------------------------------------------------
1982    Name: Dave MacQueen
1983    Date: 2000/04/09 14:00 EDT
1984    Tag: dbm-20000410-Version_110_27
1985    Description:
1986    
1987    1. Updated src/compiler/TopLevel/main/version.sml to version 110.27
1988    
1989    2. Updated src/config/version to 110.27
1990    
1991    3. New boot files!
1992    
1993    ----------------------------------------------------------------------
1994    Name: Allen Leung
1995    Date: 2000/04/09 19:09:00 EST
1996    Tag: leunga-20000409-misc
1997    Description:
1998    
1999    1.  Yet another fix for x86 assembly for idivl, imull, mull and friends.
2000    
2001    2.  Miscellaneous improvements to MLRISC (unused in sml/nj)
2002    
2003    ----------------------------------------------------------------------
2004    Name: Stefan
2005    Date: 2000/04/07 10:00:00 EDT
2006    Tag: monnier-20000406-branch-handling
2007    Description:
2008    
2009    Improved handling of branches (mostly those generated from
2010    polymorphic equality), removed switchoff and changed the
2011    default optimization settings (more cpsopt and less flintopt).
2012    
2013    ----------------------------------------------------------------------
2014    Name: Allen Leung
2015    Date: 2000/04/06 01:30:00 EST
2016    Tag: leunga-20000406-peephole-x86-SSA-2
2017    Description:
2018    
2019       Forgot a few files.
2020    
2021    ----------------------------------------------------------------------
2022    Name: Allen Leung
2023    Date: 2000/04/06 00:36:00 EST
2024    Tag: leunga-20000406-peephole-x86-SSA
2025    Description:
2026    
2027    1.  New Peephole code
2028    
2029    2.  Minor improvement to X86 instruction selection
2030    
2031    3.  Various fixes to SSA and machine description -> code translator
2032    
2033    ----------------------------------------------------------------------
2034    Name: Matthias Blume
2035    Date: 2000/04/05 12:30:00 JST
2036    Tag: blume_main_v110p26p2_3
2037    Description:
2038    
2039    This update just merges three minor cosmetic updates to CM's sources
2040    to get ready for the 110.27 code freeze on Friday.  No functionality
2041    has changed.
2042    
2043    ----------------------------------------------------------------------
2044    Name: Allen Leung
2045    Date: 2000/04/04 19:39:00 EST
2046    Tag: leunga-20000404-x86-asm
2047    Description:
2048    
2049    1.  Fixed a problem in X86 assembly.
2050    
2051        Things like
2052    
2053           jmp %eax
2054           jmp (%eax)
2055    
2056        should be output as
2057    
2058           jmp *%eax
2059           jmp *(%eax)
2060    
2061    2.  Assembly output
2062    
2063          Added a new flag
2064    
2065              "asm-indent-copies" (default to false)
2066    
2067          When this flag is on, parallel copies will be indented an extra level.
2068    
2069    ----------------------------------------------------------------------
2070    Name: Allen Leung
2071    Date: 2000/04/04 03:18:00 EST
2072    Tag: leunga-20000404-C--Moby
2073    Description:
2074    
2075        All of these fixes are related to C--, Moby, and my own optimization
2076        stuff; so they shouldn't affect SML/NJ.
2077    
2078    1.  X86
2079    
2080        Various fixes related floating point, and extensions.
2081    
2082    2.  Alpha
2083    
2084        Some extra patterns related to loads with signed/zero extension
2085        provided by Fermin.
2086    
2087    3.  Assembly
2088    
2089        When generating assembly, resolve the value of client defined constants,
2090        instead of generating symbolic values.  This is controlled by the
2091        new flag "asm-resolve-constants", which is default to true.
2092    
2093    4.  Machine Descriptions
2094    
2095        a. The precedence parser was slightly broken when parsing infixr symbols.
2096        b. The type generalizing code had the bound variables reversed, resulting
2097           in a problem during arity raising.
2098        c. Various fixes in machine descriptions.
2099    
2100    ----------------------------------------------------------------------
2101    Name: Matthias Blume
2102    Date: 2000/04/03 16:05:00 JST
2103    Tag: blume_main_v110p26p2_2
2104    Description:
2105    
2106    I eliminated coreEnv from compInfo.  Access to the "Core" structure is
2107    now done via the ordinary static environment that is context to each
2108    compilation unit.
2109    
2110    To this end, I arranged that instead of "structure Core" as "structure
2111    _Core" is bound in the pervasive environment.  Core access is done via
2112    _Core (which can never be accidentally rebound because _Core is not a
2113    legal surface-syntax symbol).
2114    
2115    The current solution is much cleaner because the core environment is
2116    now simply part of the pervasive environment which is part of every
2117    compilation unit's context anyway.  In particular, this eliminates all
2118    special-case handling that was necessary until now in order to deal
2119    with dynamic and symbolic parts of the core environment.
2120    
2121    Remaining hackery (to bind the "magic" symbol _Core) is localized in the
2122    compilation manager's bootstrap compiler (actually: in the "init group"
2123    handling).  See the comments in src/system/smlnj/init/init.cmi for
2124    more details.
2125    
2126    I also tried to track down all mentions of "Core" (as string argument
2127    to Symbol.strSymbol) in the compiler and replaced them with a
2128    reference to the new CoreSym.coreSym.  Seems cleaner since the actual
2129    name appears in one place only.
2130    
2131    Binfile and bootfile format have not changed, but the switchover from
2132    the old "init.cmi" to the new one is a bit tricky, so I supplied new
2133    bootfiles anyway.
2134    
2135  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2136  Name: Allen Leung  Name: Allen Leung
2137  Date: 2000/04/02 21:17:00 EST  Date: 2000/04/02 21:17:00 EST
# Line 113  Line 2234 
2234     (specified in the .cm file at each instance where the tool's class is     (specified in the .cm file at each instance where the tool's class is
2235     used).     used).
2236    
2237     This was done to accomodate the new "make" and "shell" tools which     This was done to accommodate the new "make" and "shell" tools which
2238     facilitate fairly seemless hookup to portions of code managed using     facilitate fairly seamless hookup to portions of code managed using
2239     Makefiles or Shell scripts.     Makefiles or Shell scripts.
2240    
2241     There are no classes "shared" or "private" anymore.  Instead, the     There are no classes "shared" or "private" anymore.  Instead, the
# Line 126  Line 2247 
2247    
2248     All existing tools are described in the CM manual.     All existing tools are described in the CM manual.
2249    
2250     - Slightly better error handling.  (CM now surpresses many followup     - Slightly better error handling.  (CM now suppresses many followup
2251     error messages that tended to be more annoying than helpful.)     error messages that tended to be more annoying than helpful.)
2252    
2253  2. Major changes to the compiler's static environment data structures.  2. Major changes to the compiler's static environment data structures.
# Line 260  Line 2381 
2381    
2382       I've changed andl to testl in the floating point test sequence       I've changed andl to testl in the floating point test sequence
2383       whenever appropriate.  The Intel optimization guide states that       whenever appropriate.  The Intel optimization guide states that
2384       testl is perferable to andl.       testl is preferable to andl.
2385    
2386  7. RA (x86 only)  7. RA (x86 only)
2387    
# Line 442  Line 2563 
2563    
2564  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an  1. Tools.registerStdShellCmdTool (from smlnj/cm/tool.cm) takes an
2565  additional argument called "template" which is an optional string that  additional argument called "template" which is an optional string that
2566  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
2567  explanation.  explanation.
2568    
2569  2. A special-purpose tool can be "regisitered" by simply dropping the  2. A special-purpose tool can be "registered" by simply dropping the
2570  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same  corresponding <...>-tool.cm (and/or <...>-ext.cm) into the same
2571  directory where the .cm file lives that uses this tool.  (The  directory where the .cm file lives that uses this tool.  (The
2572  behavior/misfeature until now was to look for the tool description  behavior/misfeature until now was to look for the tool description
# Line 489  Line 2610 
2610  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
2611  <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,
2612  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
2613  perhaps got accidentially killed).  In this case, fixpt will quickly  perhaps got accidentally killed).  In this case, fixpt will quickly
2614  move through what exists before continuing where it left off earlier,  move through what exists before continuing where it left off earlier,
2615  and, thus, saves a lot of time.  and, thus, saves a lot of time.
2616    
# Line 539  Line 2660 
2660    it from that remote directory.    it from that remote directory.
2661    This should simplify installation further:  For machines that have    This should simplify installation further:  For machines that have
2662    access to the internet, just fetch <version>-config.tgz, unpack it,    access to the internet, just fetch <version>-config.tgz, unpack it,
2663    edit config/targets, and go (run config/install.sh).  The scipt will    edit config/targets, and go (run config/install.sh).  The script will
2664    fetch everything else that it might need all by itself.    fetch everything else that it might need all by itself.
2665    
2666    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.591  
changed lines
  Added in v.834

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