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 590, Sat Apr 1 02:24:08 2000 UTC revision 1758, Tue Feb 15 23:52:33 2005 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2005/02/15 17:35:00 CST
18    Tag: blume-20050215-tools
19    Description:
20    
21    Created new "tools" directory under "src" and moved "TraceDebugProf"
22    there.
23    
24    ----------------------------------------------------------------------
25    Name: Matthias Blume (blume (at) tti - c (dot) org)
26    Date: 2005/02/10 17:55:00 CST
27    Tag: blume-20050210-longlong
28    Description:
29    
30    Implemented "long long" arguments and results for NLFFI.  (Only the
31    PPC/MacOS implementation is complete, the other backends still need to
32    be updated.)
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume (blume (at) tti - c (dot) org)
36    Date: 2005/01/24 17:40:00 CST
37    Tag: blume-20050124-mlyacc
38    Description:
39    
40    Minor cleanup in ML-Yacc rule printing mechanism.  This should fix a
41    problem with certain "as" patterns which previously got rendered
42    using incorrect syntax.
43    
44    ----------------------------------------------------------------------
45    Name: Matthias Blume (blume (at) tti - c (dot) org)
46    Date: 2005/01/18 12:00:00 CST
47    Tag: blume-20050118-profile
48    Description:
49    
50    Made time profiling code (interrupt handler) in runtime system aware
51    of new array representation.
52    
53    ----------------------------------------------------------------------
54    Name: Matthias Blume (blume (at) tti - c (dot) org)
55    Date: 2005/01/14 18:00:00 CST
56    Tag: blume-20050114-heap2exec
57    Description:
58    
59    Implemented new (but still experimental) heap2exec facility.  This is
60    tested under Mac OS X and should work under Linux (will test shortly).
61    It will probably also work on the Sparc (will test some time later).
62    
63      - removed old "HACKED_STANDALONE" hack from runtime
64    
65    To be able to test this, uncomment the request for "heap2asm" in
66    config/targets prior to installation.  (Notice that this is different
67    from "heap2exec" mentioned below.  Not a typo.)
68    
69    To perform an actual test, run the command
70    
71       $ bin/heap2exec heapfile execfile
72    
73    (You can put heap2exec on your shell's path.)
74    
75    For example, run
76    
77      $ bin/heap2exec bin/.heap/ml-yacc.ppc-darwin mly
78    
79    This will create a standalone executable called "mly" which you can
80    then invoke directly as a command.
81    
82    ----------------------------------------------------------------------
83    Name: Matthias Blume (blume (at) tti - c (dot) org)
84    Date: 2005/01/07 11:44:00 CST
85    Tag: blume-20050107-mlstring
86    Description:
87    
88    fixed off-by-one error in ML_STRING macro (globals.c)
89    
90    ----------------------------------------------------------------------
91    Name: Matthias Blume (blume (at) tti - c (dot) org)
92    Date: 2004/12/23 18:00:00 CST
93    Tag: blume-20041223-santa
94    Description:
95    
96    Made ml-build script "smarter" (but only very little).
97    
98    ----------------------------------------------------------------------
99    Name: Matthias Blume (blume (at) tti - c (dot) org)
100    Date: 2004/12/21 15:05:00 CST
101    Tag: blume-20041221-longlong
102    Description:
103    
104    * Implemented access to signed and unsigned long long data in NLFFI.
105    (The parameter-passing part of the picture has not complete. But
106    data structure access seems to work.)
107    
108    * Fixed CM's incorrect assumption that the PPC is little-endian.
109      (On the Mac, it is big-endian.  And that's currently our only
110      PPC platform.)
111    
112    ----------------------------------------------------------------------
113    Name: Matthias Blume (blume (at) tti - c (dot) org)
114    Date: 2004/12/21 12:50:00 CST
115    Tag: blume-20041221-memory
116    Description:
117    
118    Some cleanup in the $c/memory.cm library: separated some concerns by
119    moving allocation code and memory access code each into their own
120    files.
121    
122    ----------------------------------------------------------------------
123    Name: Allen Leung (leunga (at) reservoir (dot) com)
124    Date: 2004/12/17 16:12:00 EST
125    Tag: leunga-20041217-cygwin-smlnj-home
126    Description:
127    
128        The Unix I/O library of SML/NJ on cygwin does not understand
129    Windows style pathname, so problems arise when SMLNJ_HOME is set
130    to a Windows style pathname.  _run-sml now converts SMLNJ_HOME
131    to a POSIX pathname on cygwin.
132    
133    ----------------------------------------------------------------------
134    Name: Matthias Blume (blume (at) tti - c (dot) org)
135    Date: 2004/12/16 13:35:00 CST
136    Tag: Release_110_52
137    Description:
138    
139    Last-minute changes incorporated into 110.52.  Release tag moved.
140    
141    The changes:
142       - HashString.hashString' -> HashString.hashSubstring
143       - bug fix in UnivariateStats
144    
145    ----------------------------------------------------------------------
146    Name: Matthias Blume (blume (at) tti - c (dot) org)
147    Date: 2004/12/15 23:40:00 CST
148    Tag: blume-20041215-hashSubstring
149    Description:
150    
151    - HashString.hashString' -> HashString.hashSubstring
152    - corresponding changes in atom.sml
153    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
154      added some comments
155    
156    ----------------------------------------------------------------------
157    Name: Matthias Blume (blume (at) tti - c (dot) org)
158    Date: 2004/12/15 15:30:00 CST
159    Tag: (Release_110_52)
160    Description:
161    
162    New working version (110.52).  NEW BOOTFILES!
163    
164    ----------------------------------------------------------------------
165    Name: Matthias Blume (blume (at) tti - c (dot) org)
166    Date: 2004/12/15 12:45:00 CST
167    Tag: blume-20041215-spaces
168    Description:
169    
170    More on the space problem (this time for Win32).
171    
172    ----------------------------------------------------------------------
173    Name: Matthias Blume (blume (at) tti - c (dot) org)
174    Date: 2004/12/14 17:30:00 CST
175    Tag: blume-20041214-spaces
176    Description:
177    
178    Hacked some of the scripts (in particular: the installer) to cope with
179    spaces in filenames a bit better.  But beware: the current "solution"
180    is likely still full of bugs and inherently incomplete.  (We need to
181    do away with those shell scripts for a comprehensive solution.)
182    
183    ----------------------------------------------------------------------
184    Name: Matthias Blume (blume (at) tti - c (dot) org)
185    Date: 2004/12/13 14:45:00 CST
186    Tag: blume-20041213-ml-makedepend
187    Description:
188    
189    Fixed bug in code for ml-makedepend.
190    
191    ----------------------------------------------------------------------
192    Name: Matthias Blume (blume (at) tti - c (dot) org)
193    Date: 2004/12/09 16:30:00 CST
194    Tag: blume-20041209-statistics
195    Description:
196    
197    Added two simple but potentially useful statistics modules to SML/NJ Library.
198    (See CHANGES file there.)
199    
200    ----------------------------------------------------------------------
201    Name: Matthias Blume (blume (at) tti - c (dot) org)
202    Date: 2004/12/01 16:50:00 CST
203    Tag: blume-20041201-atom
204    Description:
205    
206    smlnj-lib:
207    
208      Added function HashString.hashString' for substrings.
209      Hand-inlined CharVector.foldl into HashString (for speed).
210      Modified implementation of structure Atom to avoid extracting
211      strings from substrings unless necessary.
212      (Also see CHANGES file for smlnj-lib.)
213    
214    ----------------------------------------------------------------------
215    Name: Matthias Blume (blume (at) tti - c (dot) org)
216    Date: 2004/11/24 22:15:00 CST
217    Tag: blume-20041124-cml
218    Description:
219    
220    Made sure CML compiles when Position = Int64.
221    
222    ----------------------------------------------------------------------
223    Name: Matthias Blume (blume (at) tti - c (dot) org)
224    Date: 2004/11/24 14:30:00 CST
225    Tag: blume-20041124-position
226    Description:
227    
228    The compiler can now be compiled in a mode that makes structure Position
229    equal to Int64.  The default, however, is unchanged (Position = Int31)
230    for the time being.
231    
232    To enable 64-bit positions, use the following procedure:
233    
234    1. Start sml
235    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
236    3. Type
237          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
238    4. Run CMB.make() as usual.
239    
240    This is barely tested.  The only test so far was a little SML program
241    counting the number of characters in an 8-gigabyte file by
242    reading it character-by-character.  That test was successful.
243    
244    In support of 64-bit positions, a number of new functions have been
245    added to the runtime system.
246    
247    ----------------------------------------------------------------------
248    Name: Matthias Blume (blume (at) tti - c (dot) org)
249    Date: 2004/11/23 14:45:00 CST
250    Tag: blume-20041123-useFile
251    Description:
252    
253    Fixed a problem with unhelpful error messages related to problems with
254    .cm- or .sml files that appear as part of the sml command line.
255    
256    ----------------------------------------------------------------------
257    Name: Matthias Blume (blume (at) tti - c (dot) org)
258    Date: 2004/11/18 15:40:00 CST
259    Tag: Release_110_51
260    Description:
261    
262    New working version (110.51).  NEW BOOTFILES!
263    
264    ----------------------------------------------------------------------
265    Name: Matthias Blume (blume (at) tti - c (dot) org)
266    Date: 2004/11/18 15:35:00 CST
267    Tag: <none>
268    Description:
269    
270    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
271    Joemann).
272    
273    ----------------------------------------------------------------------
274    Name: Allen Leung (leunga (at) reservoir (dot) com)
275    Date: 2004/11/17 16:05:21 EST 2004
276    Tag: leunga-20041117-mlrisc-live-kill
277    Description:
278    
279       Added support for MLTree constructs LIVE and KILL
280    to all the architectures.
281    
282    ----------------------------------------------------------------------
283    Name: Matthias Blume (blume (at) tti - c (dot) org)
284    Date: 2004/11/13 00:20:00 CST
285    Tag: blume-20041113-versiontool
286    Description:
287    
288    - Stripped down the versiontool: It now only handles the version number.
289      The date string is generated at bootstrap time (during makeml).
290    
291    - In a previous commit, fixed a minor issue with how polyequal is being
292      translated.  In particular, the code now "looks through" abstractions.
293      This results in slightly fewer polyEqual warnings and hopefully slightly
294      more efficient code.  Important examples for where this matters are
295      the new int64 and word64 types.
296    
297    ----------------------------------------------------------------------
298    Name: Matthias Blume (blume (at) tti - c (dot) org)
299    Date: 2004/11/12 00:30:00 CST
300    Tag: blume-20041112-int64
301    Description:
302    
303    Structure Int64 fully hooked in.  (The implementation is not very
304    efficient, though.)
305    
306    ----------------------------------------------------------------------
307    Name: Matthias Blume (blume (at) tti - c (dot) org)
308    Date: 2004/11/11 17:30:00 CST
309    Tag: blume-20041111-more64
310    Description:
311    
312    All the pieces of Word64 are now there, with the exception of the
313    conversions from and to LargeWord. (Eventually these need to be identities,
314    but for the time being they don't even make sense because LargeWord is
315    32-bit wide.)
316    
317    Also started to add similar support for Int64, but major pieces of that
318    are still missing.
319    
320    ----------------------------------------------------------------------
321    Name: Matthias Blume (blume (at) tti - c (dot) org)
322    Date: 2004/11/11 00:15:00 CST
323    Tag: blume-20041111-word64
324    Description:
325    
326    Structure Word64 is now (almost) complete, word literals and patterns
327    seem to work.  There are a few odd pieces missing.  In particular,
328    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
329    at the moment.
330    
331    Making Word64 official would mean that LargeWord becomes Word64.  But
332    this requires extreme care because most word-word conversions have to
333    go through LargeWord, so making a mistake means loss of efficiency or
334    worse.  Eventually there will be a solution similar to (but actually
335    simpler than) what I did with IntInf.
336    
337    ----------------------------------------------------------------------
338    Name: Matthias Blume (blume (at) tti - c (dot) org)
339    Date: 2004/11/10 18:12:00 CST
340    Tag: blume-20041110-64bit
341    Description:
342    
343    More 64-bit hacking (but still not even half-way there yet).
344    Also, some assorted improvements to the handling of 8-bit words.
345    
346    ----------------------------------------------------------------------
347    Name: Matthias Blume (blume (at) tti - c (dot) org)
348    Date: 2004/11/09 17:50:00 CST
349    Tag: <none>
350    Description:
351    
352    Started adding some infrastructure for supporting 64-bit int- and
353    word-types.  (Still in its very early stages.)
354    
355    ----------------------------------------------------------------------
356    Name: Matthias Blume (blume (at) tti - c (dot) org)
357    Date: 2004/10/28 10:45:00 CDT
358    Tag: Release_110_50
359    Description:
360    
361    New working version (110.50).  NEW BOOTFILES!
362    
363                 =====================
364    
365    Also:
366    
367      - Changend config/srcarchiveurl from a file just
368        containing the URL string into a file containing
369        shell script code.  The code has access to the $VERSION variable.
370      - Made corresponding changes to config/install.sh and config/unpack.
371      - Default contents of config/srcarchiveurl uses $VERSION and
372        normally does not have to be edited to reflect a version change.
373    
374      (As a result, a version change can be done by just editing
375       config/version, the rest is now automatic.)
376    
377    ----------------------------------------------------------------------
378    Name: Matthias Blume (blume (at) tti - c (dot) org)
379    Date: 2004/10/27 17:50:00 CDT
380    Tag: blume-20041027-btrace-msg
381    Description:
382    
383    BackTrace.monitor now also reports the source of the exception that
384    triggered the trace.
385    
386    ----------------------------------------------------------------------
387    Name: Matthias Blume (blume (at) tti - c (dot) org)
388    Date: 2004/10/27 17:20:00 CDT
389    Tag: blume-20041027-x86-c-calls
390    Description:
391    
392    This is the HISTORY entry for two earlier commits, both concerning
393    the x86 c-calls code in MLRISC:
394    
395      - added a missing LOAD in the code that deals with struct arguments
396      - made sure the caller does not add the wrong number of bytes to the
397        stack pointer after a call of a function returning a struct
398        (the callee already pops the implicit argument which points to
399         the space reserved for the result)
400    
401    ----------------------------------------------------------------------
402    Name: Allen Leung (leunga (at) reservoir (dot) com)
403    Date: 2004/10/24 14:00:00 EST
404    Tag: leunga-20041024-x86-gas-fucomip
405    Description:
406    
407       John discovered a bug in the syntax of fucomip.
408       The opcodes FU?COMIP? have been changed to
409    
410           fu?comip? %st(i), %st
411    
412    ----------------------------------------------------------------------
413    Name: Matthias Blume (blume (at) tti - c (dot) org)
414    Date: 2004/10/20 15:06:00 CDT
415    Tag: blume-20041020-standalone-backtrace
416    Description:
417    
418    Added a mechanism for getting back-trace information from standalone
419    programs.  Here is how it works:
420    
421       1. The part of the program from which you want to get backtrace
422          information (usually the whole program) should be wrapped with
423          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
424          main program could be modified from something like
425    
426             fun main (pgm, args) = ...
427    
428         to
429    
430            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
431    
432       2. To be able to access BackTrace.monitor, you have to add
433          library $smlnj-tdp/plugins.cm to the .cm file that contains your
434          main function.
435    
436       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
437          CM might have created in the past for your project).
438    
439       4. Build the system using this command line:
440    
441             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
442                      myprog.cm MyProg.main myprog
443    
444          instead of the usual
445    
446             ml-build myprog.cm MyProg.main myprog
447    
448    I changed a library name:
449    
450       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
451    
452    New libraries:
453    
454       $smlnj-tdp/back-trace.cm
455          -- when loaded causes the back-trace plugin to be installed
456       $smlnj-tdp/coverage.cm
457          -- when loaded causes the coverage plugin to be installed
458    
459    ----------------------------------------------------------------------
460    Name: Matthias Blume (blume (at) tti - c (dot) org)
461    Date: 2004/10/18 16:45:00 CDT
462    Tag: blume-20041018-groupowner
463    Description:
464    
465    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
466    
467    Eliminated group owner specs from .cm files throughout the source tree.
468    
469    ----------------------------------------------------------------------
470    Name: Matthias Blume (blume (at) tti - c (dot) org)
471    Date: 2004/10/15 15:45:00 CDT
472    Tag: blume-20041015-coverage
473    Description:
474    
475    * Test coverage tool added!
476    * Further reorganization of tracing-, debugging-, and profiling support:
477    
478        - moved original BTImp -- now called BackTrace -- into a separate
479          library called $/trace-debug-profile.cm
480        - eliminated all mentions of BTrace from SMLofNJ.Internals
481        - only the instrumentation mechanism is now left in the compiler proper
482        - BackTrace module is a plugin which is NOT plugged in by default
483        - Coverage module is another such plugin
484    
485    To get the benefits of any of these plugin modules, the code in
486    question must be compiled with tdp instrumentation turned on.  This
487    can be done by setting SMLofNJ.Internals.TDP.mode to true.
488    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
489    
490    Plugins are selected at link time.  (Pre-compiled instrumented code
491    can be re-loaded with different plugins in effect.)  When an
492    instrumented module is linked, whatever plugins are at that time
493    enabled will come into effect for that module.
494    
495    To enable the back-trace plugin, load library $/trace-debug-profile.cm
496    and invoke BackTrace.install() (e.g., from the interactive prompt).
497    To enable the coverage plugin, load the same library and invoke
498    Coverage.install().
499    
500    Back-traces are generated automatically on uncaught exceptions and
501    when the code in question explicitly invokes BackTrace.trigger().
502    
503    Coverage (and execution frequency-) information must be queried
504    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
505    
506    ----------------------------------------------------------------------
507    Name: Matthias Blume (blume (at) tti - c (dot) org)
508    Date: 2004/10/14 17:40:00 CDT
509    Tag: blume-20041014-tdp-core
510    Description:
511    
512    Snapshot of a significant overhaul of how the trace/debug/profile support
513    is hooked into the system (specifically: Core and SMLofNJ.Internals).
514    
515    ----------------------------------------------------------------------
516    Name: Matthias Blume (blume (at) tti - c (dot) org)
517    Date: 2004/10/13 16:34:00 CDT
518    Tag: blume-20041013-tdp
519    Description:
520    
521    Some rationalization of names:
522    
523       structure BTrace -> structure TDPInstrument
524       etc.
525    
526    This is is preparation of using the original back-trace
527    instrumentation for other purposes.  "TDP" stands for
528    Trace/Debug/Profile.
529    
530    The control flag controlling whether instrumentation is on or off is now
531    registered under a different name, so instead of running sml as
532    
533     sml -Cinstrument.btrace-mode=true
534    
535    one has to say
536    
537     sml -Ctdp.instrument=true
538    
539    ----------------------------------------------------------------------
540    Name: Matthias Blume (blume (at) tti - c (dot) org)
541    Date: 2004/10/11 16:37:00 CDT
542    Tag: blume-20041011-regions
543    Description:
544    
545    Made some minor modifications to elabcore.sml to have source regions
546    be propagated more tightly -- resulting in better (i.e., smaller)
547    regions being reported in error- and debug messages.
548    
549    ----------------------------------------------------------------------
550    Name: Matthias Blume (blume (at) tti - c (dot) org)
551    Date: 2004/10/08 22:50:00 CDT
552    Tag: blume-20041008-cmkw
553    Description:
554    
555    Fixed handling of keywords in .cm files: After seeing "is" the lexer
556    treats subsequent occurrences of "group", "library", "source", "is",
557    "*", and "-" as ordinary identifiers rather than keywords.
558    
559    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
560    is supposed to accept a tool argument called "source", but this did
561    not work because of the clash with the keyword.
562    
563    ----------------------------------------------------------------------
564    Name: Matthias Blume (blume (at) tti - c (dot) org)
565    Date: 2004/10/07 16:00:00 CDT
566    Tag: blume-20041007-cleanup
567    Description:
568    
569    Assorted cleanup work:
570    
571      - got rid of intstrmap in favor of using the library's
572        hash table implementation
573      - threw out most of the pathnames stuff, as it was not used anyway
574      - simplified tokentable implementation
575      - fixed some minor spelling errors
576    
577    ----------------------------------------------------------------------
578    Name: Matthias Blume (blume (at) tti - c (dot) org)
579    Date: 2004/10/06 15:15:15 CDT
580    Tag: blume-20041006-handler
581    Description:
582    
583    Cleaned up the absyn to reflect the invariant that HANDLE always
584    carries a FNexp as part of the type definition.  This eliminates some
585    superfluous sanity checks at runtime down the road.
586    
587    Some minor cleanup of the btrace code.
588    
589    ----------------------------------------------------------------------
590    Name: Matthias Blume (blume (at) tti - c (dot) org)
591    Date: 2004/10/01 10:20:30 CDT
592    Tag: blume-20041001-slave
593    Description:
594    
595    Added hack to make slave mode work in the presence of the version
596    tool.  (Still, since the master does two passes over the code for
597    CMB.make, the release number gets bumped twice when slaves are
598    attached.  I don't know if this is worth fixing...)
599    
600    ----------------------------------------------------------------------
601    Name: Matthias Blume (blume (at) tti - c (dot) org)
602    Date: 2004/09/30 10:55:00 CDT
603    Tag: blume-20040930-version
604    Description:
605    
606    * Moved the "version" magic into its own little library under
607      src/system/smlnj/internal.  This avoids expensive reconstruction of
608      a stable src/compiler/core.cm.
609    
610    * At the same time, structure CompilerVersion is now known as
611      structure SMLNJVersion.
612    
613    * Arranged for the version tool to NOT kick in when rebuilding the system
614      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
615      Also, loading the versiontool does not work when rebuilding the system
616      because CM is not properly initialized at that time.
617    
618    ----------------------------------------------------------------------
619    Name: Matthias Blume (blume (at) tti - c (dot) org)
620    Date: 2004/09/29 14:00:00 CDT
621    Tag: blume-20040929-autoversion
622    Description:
623    
624    Implemented some CM magic to have
625        file src/compiler/TopLevel/main/version.sml
626    generated automagically.
627    The version is taken from two files: config/version and config/release.
628    The first is expected to contain a two-part version number such as 110.49.
629    The second should contain a single number, but it may be missing.
630    
631    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
632    time the version tool is loaded (which is the first time you say CMB.make),
633    then the tool will increment the value stored in config/release every
634    time CMB.make is invoked.
635    
636    The binfile format is now insensitive to anything beyond the first
637    two components of a version number, so bumping the release does not render
638    binfiles incompatible.  Auto-bumping can be used to keep track of versions
639    during development without invalidating existing binfiles.
640    
641    In any case, every CMB.make updates the date information in version.sml.
642    (This is the date that is printed in the banner.)
643    
644    ----------------------------------------------------------------------
645    Name: Matthias Blume (blume (at) tti - c (dot) org)
646    Date: 2004/09/28 10:53:00 CDT
647    Tag: blume-20040928-controls
648    Description:
649    
650    Some cleanup of the controls code.
651    
652    ----------------------------------------------------------------------
653    Name: Matthias Blume (blume (at) tti - c (dot) org)
654    Date: 2004/09/27 22:08:00 CDT
655    Tag: blume-20040927-controls
656    Description:
657    
658    Added two pieces of functionality to the Controls interface:
659    
660    1.   val save'restore: 'a control -> unit -> unit
661    
662       grabs the current value of the control in stage 1 and restores it
663       in stage 2.
664    
665    2.   val set' : 'a control * 'a -> unit -> unit
666    
667       stores the given value into the control in stage 2 (i.e., delayed)
668       but does all error checking in stage 1.
669       (This is for string controls that need to do parse their argument --
670        something that might fail.  In some cases, notably in CM, one
671        already knows the intended argument but wants to delay the actual
672        assignment until a time when error recovery would be more difficult.)
673    
674    Changed the handling of controls in tool arguments to classes "sml" and
675    "lazysml":
676       - use Controls.save'restore as a more robust way of restoring the
677         old value (in particular: without having to re-parse the string)
678       - use controls to handle the "overload" keyword in the init group
679         (I believe this change actually fixes a long-standing obscure bug.)
680    
681    ----------------------------------------------------------------------
682    Name: Matthias Blume (blume (at) tti - c (dot) org)
683    Date: 2004/09/27 17:00:00 CDT
684    Tag: blume-20040927-lazysml
685    Description:
686    
687    Added a new tool class called "lazysml" to CM's tool chest.  The only
688    difference to "sml" is that compilation is done with Control.lazysml
689    set to true.  A source of class "lazysml" is automatically recognized
690    by a file name suffix of ".lml".
691    
692    In addition to the above feature, the original class "sml" now also
693    supports a tool argument "lazy" which has the same effect.  As a
694    result, the following three lines are equivalent:
695    
696        foo.sml : lazysml
697        foo.sml : sml (lazy)
698        foo.sml (lazy)
699    
700    The setting goes into effect both during parsing and during
701    compilation.  The original setting is restored right after parsing and
702    after compilation, respectively.
703    
704    In addition to all the above, there is also a general mechanism to set
705    ANY of the "controls" that are available at the command line via
706    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
707    apply as well. (In fact, "lazy" is implemented as a special case of
708    the general mechanism.)
709    
710    The .cm file syntax uses a new keyword tool argument called "with".
711    There are several ways of indicating the desired settings:
712    
713        foo.sml (with:parser.quotations=true)
714        foo.sml (with:(name:parser.quotations value:true))
715        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
716        foo.sml (with:(name1=value1 name2=value2 ...))
717        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
718    
719        etc.
720    
721    Another possible abbreviation is to leave out the =v or value:v part
722    if the name refers to a boolean control (in which case the value is
723    taken to be true).  Thus, one could get lazy sml also by saying:
724    
725        foo.sml (with:parser.lazy-keyword=true)
726        foo.sml (with:parser.lazy-keyword)
727        foo.sml (with:(name:parser.lazy-keyword value:true))
728        foo.sml (with:(name:parser.lazy-keyword))
729    
730    ----------------------------------------------------------------------
731    Name: Matthias Blume (blume (at) tti - c (dot) org)
732    Date: 2004/09/24 16:22:00 CDT
733    Tag: blume-20040924-ppc-long-branch
734    Description:
735    
736    Turned message about "emiting long form of branch" off by default.
737    Added a control flag to turn it back on when desired.
738    
739    ----------------------------------------------------------------------
740    Name: Matthias Blume (blume (at) tti - c (dot) org)
741    Date: 2004/09/24 16:05:00 CDT
742    Tag: blume-20040924-rounding
743    Description:
744    
745    Applied patch for setting rounding modes under Mac OS X.  Thanks to
746    Melissa O'Neill for providing the code!
747    
748    ----------------------------------------------------------------------
749    Name: Matthias Blume (blume (at) tti - c (dot) org)
750    Date: 2004/09/23 17:30:00 CDT
751    Tag: blume-20040923-envvars
752    Description:
753    
754    1. Changed definition of type ControlRegistry.registry_tree to
755       include control_info (i.e., the name of the controlling environment
756       variable).
757    
758    2. Added command-line flags -e and -E to print the names of environment
759       variables that can be used to control internal settings.
760       (This uses the new API mentioned in 1.)
761    
762    ----------------------------------------------------------------------
763    Name: Matthias Blume (blume (at) tti - c (dot) org)
764    Date: 2004/09/13 16:50:00 CDT
765    Tag: Release_110_49
766    Description:
767    
768    New working version (110.49).  NEW BOOTFILES!
769    
770    ----------------------------------------------------------------------
771    Name: Matthias Blume
772    Date: 2004/09/13 16:20:00 CDT
773    Tag: blume-20040913-config-mlrisc
774    Description:
775    
776    Put target "mlrisc" back into the default list.
777    (There is no harm in having it, and some users have expressed their
778    wish to have "mlrisc" included by default.)
779    
780    ----------------------------------------------------------------------
781    Name: John Reppy
782    Date: 2004/09/13
783    Tag: jhr-20040913-signals
784    Description:
785    
786    Fixed the signal masking code to properly nest mask/unmask operations
787    on a per-signal basis.
788    
789    ----------------------------------------------------------------------
790    Name: Matthias Blume (blume (at) tti - c (dot) org)
791    Date: 2004/09/08 13:20:00 CDT
792    Tag: blume-20040908-heap-magic
793    Description:
794    
795    Bumped the heap macig number to 0x09082004 to account for the changed
796    layout of the ML frame under MacOS X.
797    
798    ----------------------------------------------------------------------
799    Name: Allen Leung (leunga (at) reservoir (dot) com)
800    Date: 2004/09/03 11:26:00 EST
801    Tag: leunga-20040903-cygwin-install
802    Description:
803    
804       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
805    The Cygwin runtime is turned on by setting the environment
806    variable SMLNJ_CYGWIN_RUNTIME to 1.
807    
808    ----------------------------------------------------------------------
809    Name: Matthias Blume (blume (at) tti - c (dot) org)
810    Date: 2004/08/31 17:15:00 CDT
811    Tag: blume-20040831-core
812    Description:
813    
814    Added some exports to src/compiler/core.cm upon request by J. Joemann.
815    
816    ----------------------------------------------------------------------
817    Name: Matthias Blume (blume (at) tti - c (dot) org)
818    Date: 2004/08/30 17:55:00 CDT
819    Tag: blume-20040830-installer
820    Description:
821    
822    Upon request by Johannes Joemann:
823    
824    - improved ML code of installer to fall back to coping when renaming
825      fails (i.e., when source and target are on different file systems);
826      the code compiles but has yet to be tested in anger
827    - removed mlrisc from list of default targets (config/targets)
828    
829    ----------------------------------------------------------------------
830    Name: Matthias Blume (blume (at) tti - c (dot) org)
831    Date: 2004/08/27 17:20:00 CDT
832    Tag: blume-20040827-ptreql
833    Description:
834    
835    Added ptreql primop to structure InlineT (upon request from Larry
836    Paulson).
837    
838    ----------------------------------------------------------------------
839    Name: Allen Leung (leunga (at) reservoir (dot) com)
840    Date: 2004/08/15 21:21:00 EST
841    Tag: leunga-110_48-udgraph
842    Description:
843    
844    Another bug fix from Carl Hauser:
845    
846    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
847    > 48c48
848    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
849    > e::L)
850    > ---
851    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
852    > Without this, any deletion of an edge in an undirected graph does severe
853    > violence to the graph.
854    
855    ----------------------------------------------------------------------
856    Name: Allen Leung (leunga (at) reservoir (dot) com)
857    Date: 2004/08/10 23:35:00 EST
858    Tag: leunga-110_48-ppc
859    Description:
860    
861       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
862    
863    ----------------------------------------------------------------------
864    Name: Matthias Blume (blume (at) tti - c (dot) org)
865    Date: 2004/08/10 12:00:00 CDT
866    Tag: Release_110_48
867    Description:
868    
869    New working version (110.48).  NEW BOOTFILES!
870    
871    ----------------------------------------------------------------------
872    Name: Allen Leung (leunga (at) reservoir (dot) com)
873    Date: 2004/08/09 12:21:00 EST
874    Tag: leunga-110_47-dijsktra
875    Description:
876    
877    Bug fix from Carl Hauser:
878    
879    single_source_shortest_paths in dijkstra.sml was observed to get wrong
880    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
881    
882    The problem is that following the expression A.update(dist,s,Num.zero)
883    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
884    
885    ----------------------------------------------------------------------
886    Name: Matthias Blume (blume (at) tti - c (dot) org)
887    Date: 2004/08/06 18:10:00 CDT
888    Tag: blume-20040806-cmdline
889    Description:
890    
891    Fiddled with handling of command-line options:
892    
893      * sml now quits after processing the command line
894        if -H, -S, -h<n>, or -s<n> appears as the last
895        command-line argument
896      * a new option -q terminates the session when encountered on
897        the command line; subsequent arguments will be ignored
898      * bug fixes: short (erroneous) arguments are no longer ignored
899        completely
900    
901    ----------------------------------------------------------------------
902    Name: Allen Leung (leunga (at) reservoir (dot) com)
903    Date: 2004/08/04 18:17:00 EST
904    Tag: leunga-110_47-ppc-ibm-asm
905    Description:
906    
907      - Added minimal IBM assembly syntax support for PowerPC.
908    
909      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
910        been altered in the runtime.  We need an automatic way to keep the file
911        in sync.
912    
913    ----------------------------------------------------------------------
914    Name: Matthias Blume (blume (at) tti - c (dot) org)
915    Date: 2004/08/04 14:00:00 CDT
916    Tag: Release_110_47
917    Description:
918    
919    New working version (110.47).  NEW BOOTFILES!
920    
921    ----------------------------------------------------------------------
922    Name: Matthias Blume (blume (at) tti - c (dot) org)
923    Date: 2004/08/03 14:25:00 CDT
924    Tag: blume-20040803-callingconv
925    Description:
926    
927    Added low-level support for choosing C calling conventions by
928    twiddling the type of rawccall.  (See
929    src/compiler/Semant/types/cproto.sml for details.)
930    
931    ----------------------------------------------------------------------
932    Name: Matthias Blume (blume (at) tti - c (dot) org)
933    Date: 2004/08/02 15:55:00 CDT
934    Tag: blume-20040802-backout
935    Description:
936    
937    Backed out of change to win32-filesys.c.  The earlier patch to
938    get_file_time caused CM to produce files with the wrong time stamp.
939    
940    ----------------------------------------------------------------------
941    Name: Matthias Blume (blume (at) tti - c (dot) org)
942    Date: 2004/08/02 14:45:00 CDT
943    Tag: blume-20040802-nlffi-win32
944    Description:
945    
946    Added NLFFI support for Win32, adapted from a patch provided by David
947    Hansel.  This is currently completely untested.  Also, the issue
948    concerning stdcall vs. ccall is still unresolved.
949    
950    ----------------------------------------------------------------------
951    Name: Matthias Blume (blume (at) tti - c (dot) org)
952    Date: 2004/07/30 17:55:00 CDT
953    Tag: blume-20040730-various
954    Description:
955    
956    Gearing up towards 110.47...
957    
958    - various minor bugfixes to ml-nlffigen
959    - a beginning of a manual for nlffi
960    
961    - eliminated 'export name=value' in config/install.sh as this does
962      not work with certain versions of /bin/sh
963      (Thanks to David King at Motorola for catching this.)
964    
965    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
966      - added a test for tm==NULL to gmtime.c and localtime.c
967      - applied patch for incorrect GetFileTime under win32
968      - toSeconds -> toMilliseconds in Win32/win32-process.sml
969    
970    ----------------------------------------------------------------------
971    Name: Matthias Blume (blume (at) tti - c (dot) org)
972    Date: 2004/07/21 18:20:00 CDT
973    Tag: blume-20040721-nlffigen
974    Description:
975    
976    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
977      for a typedef to an incomplete type, but leave out the "typ" member.
978      (This is just for consistency.)
979    - Started to produce what is supposed to become better (i.e., comprehensive)
980      documentation of what ml-nlffigen does and produces.
981    
982    ----------------------------------------------------------------------
983    Name: Matthias Blume (blume (at) tti - c (dot) org)
984    Date: 2004/07/14 16:25:00 CDT
985    Tag: blume-20040714-union
986    Description:
987    
988    Added C_UNION to c-calls/c-types.sml and updated the machinery
989    (ml-nlffigen, cproto.sml) that conveys C function interface
990    information to the code generator.
991    
992    However, the actual architecture-specific implementation of function
993    arguments and results that are C unions is still not implemented.
994    
995    ----------------------------------------------------------------------
996    Name: Allen Leung (leunga (at) reservoir (dot) com)
997    Date: 2004/07/14 14:38:00 EST
998    Tag: leunga-110_46_1-ppc-lwzu
999    Description:
1000    
1001        Added these instructions to the PowerPC architecture:
1002        LBZU(X), LHZU(X), LWZU(X),
1003        STWU(X), STFDU, STFSU
1004    
1005        etc...
1006    
1007        Note: I haven't added their instruction encoding into the description.
1008    
1009    ----------------------------------------------------------------------
1010    Name: Allen Leung (leunga (at) reservoir (dot) com)
1011    Date: 2004/07/13 15:04:00 EST
1012    Tag: leunga-110_46_1-ppc-lwarx
1013    Description:
1014    
1015       Added the two instructions LWARX and STWCX to the PowerPC
1016    instruction set.
1017    
1018       A (untested) rewrite of loop-structure.sml.  The old version
1019    is completely broken.
1020    
1021    ----------------------------------------------------------------------
1022    Name: Matthias Blume (blume (at) tti - c (dot) org)
1023    Date: 2004/07/13 13:50:00 CDT
1024    Tag: blume-20040713-nlffi
1025    Description:
1026    
1027    - use paramAlloc to report c-calls with too many arguments
1028      (for PPC version where parameter area is pre-allocated)
1029    - added ccall_maxargspace to machspec (to implement the above)
1030    - made "make" commend in CM's "make" tool configurable
1031    - added option (default: on) for passing the name of the SML/NJ's "bin"
1032      directory to "make"; the call looks like this:
1033    
1034         make <options> SMLNJ_BINDIR=<dir> <target>
1035    
1036      This can be used by the Makefile to, e.g., pick the "right" version
1037      of ml-nlffigen.
1038    - minor code tweaks
1039    
1040    ----------------------------------------------------------------------
1041    Name: Matthias Blume (blume (at) tti - c (dot) org)
1042    Date: 2004/07/12 22:50:00 CDT
1043    Tag: blume-110_46_1-macosx-nlffi
1044    Description:
1045    
1046    NLFFI under Mac OS X now working (sort of).  This is largely untested,
1047    though.
1048    
1049    Note:  1. You have to make a new, clean build of the runtime system.
1050           2. There are new BOOTFILES, you have to use them!
1051              (Doing the bootstrap process yourself would be *very* painful!
1052               If you absolutely have to do it, build the system under
1053               a different architecture and then cross-compile.)
1054    
1055    Version bumped to 110.46.1 to account for runtime data format changes.
1056    
1057    ----------------------------------------------------------------------
1058    Name: Matthias Blume (blume (at) tti - c (dot) org)
1059    Date: 2004/06/18 14:30:00 CDT
1060    Tag: blume-20040618-unix
1061    Description:
1062    
1063    Changed the implementation of structure Unix so that the same stream
1064    is returned every time one of the {text,bin}{In,Out}streamOf functions
1065    is invoked on the same proc.  This is not what the spec currently
1066    says -- although IMO it arguably should.  (See discussion below.)
1067    
1068    ----------------------------------------------------------------------
1069    Name: Matthias Blume (blume (at) tti - c (dot) org)
1070    Date: 2004/06/17 18:15:00 CDT
1071    Tag: Release_110_46
1072    Description:
1073    
1074    New working version (110.46).  NEW BOOTFILES!
1075    
1076    ----------------------------------------------------------------------
1077    Name: Matthias Blume (blume (at) tti - c (dot) org)
1078    Date: 2004/06/17 17:20:00 CDT
1079    Tag: blume-20040617-timer-unix
1080    Description:
1081    
1082    Changed the interface of structures Timer and Unix to match the most
1083    recent Basis spec.
1084    
1085    In the case of Unix there still seems to be an open/weird issue:
1086    
1087       The {text,bin}{In,Out}streamOf functions are supposed to create
1088       fresh streams whenever they are called -- as opposed to have them
1089       return the same stream every time.  This design is supposed to
1090       prevent space leaks caused by proc values hanging on to streams.
1091    
1092       The reap function, on the other hand, is supposed to close the
1093       streams.  This cannot be done without having a handle on the
1094       stream in proc after all...
1095    
1096    I took the liberty to implement the following stopgap solution:
1097    
1098      The proc value hangs on to the most recently created stream(s).
1099      Reap closes those.  If either or both of the two streams hadn't
1100      been created at all yet, then reap will close the corresponding
1101      file descriptors directly.
1102    
1103    PS: I don't understand the original space leak argument anymore.  If
1104    a proc hangs on to the imperative stream, then I/O operations on those
1105    will advance the state of the cached stream and avoid the space leak.
1106    
1107    ----------------------------------------------------------------------
1108    Name: Matthias Blume (blume (at) tti - c (dot) org)
1109    Date: 2004/05/28 16:45:00 CDT
1110    Tag: blume-20040528-basis
1111    Description:
1112    
1113    Added signature PACK_REAL and exported functor PrimIO.
1114    
1115    ----------------------------------------------------------------------
1116    Name: Matthias Blume (blume (at) tti - c (dot) org)
1117    Date: 2004/05/25 16:00:00 CDT
1118    Tag: blume-20040525-group-owner
1119    Description:
1120    
1121    CM now ignores (but still accepts) the "owner" information in group
1122    descriptions.  The owner of a group is its next enclosing
1123    library. Each group must have a unique owner.  (There is a virtual
1124    "toplevel" library that owns groups which are not nested within a real
1125    library.)  Previously, each group had to explicitly declare its owner,
1126    and CM would check that such a declaration is correct.  The new scheme
1127    is to have CM check that for each group there is precisely one owning
1128    library.
1129    
1130    The advantage of the new scheme is that the programmer no longer needs
1131    to maintain the somewhat annoying owner information.  The downside is
1132    that CM cannot enforce the ownership rule across multiple runs of
1133    CM.make.  Fortunately, enclosing the same group in two different
1134    libraries A and B which are not part of the same program does not
1135    cause real problems.
1136    
1137    ----------------------------------------------------------------------
1138    Name: Matthias Blume (blume (at) tti - c (dot) org)
1139    Date: 2004/05/20 16:00:00 CDT
1140    Tag: blume-20040520-win32
1141    Description:
1142    
1143    Made the win32 version work again.  (Strangely, a misplaced comma had
1144    slipped into win32-process.c which prevented the runtime from being
1145    compiled correctly.)
1146    
1147    Also, included a minor addition to ml-build.bat analogous to what was
1148    done in blume-20040519-ml-build.
1149    
1150    ----------------------------------------------------------------------
1151    Name: Matthias Blume (blume (at) tti - c (dot) org)
1152    Date: 2004/05/19 22:10:00 CDT
1153    Tag: blume-20040519-ml-build
1154    Description:
1155    
1156    Arranged for ml-build to clean up after itself a little bit better.
1157    The script generates a temporary SML source file and compiles it using
1158    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1159    gets rid of those at the end, so they don't accumulate under .cm.
1160    
1161    This required a minor change to install.sh because the name of the
1162    metadata directory (default: .cm) is actually configurable at
1163    installation time.
1164    
1165    ----------------------------------------------------------------------
1166    Name: Matthias Blume (blume (at) tti - c (dot) org)
1167    Date: 2004/05/18 15:50:00 CDT
1168    Tag: blume-20040518-mkreader
1169    Description:
1170    
1171    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1172    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1173    
1174    ----------------------------------------------------------------------
1175    Name: Matthias Blume (blume (at) tti - c (dot) org)
1176    Date: 2004/05/11 14:35:00 CDT
1177    Tag: blume-20040511-win32sock
1178    Description:
1179    
1180    Added previously missing support for many socket-related functions
1181    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1182    for the voluminous patch!
1183    
1184    (I have not tested this patch under win32 yet.)
1185    
1186    Here is David's e-mail:
1187    
1188    Hi,
1189    
1190    Attached to this email you find a diff against sml/nj 110.45
1191    that will enable socket support under Windows.
1192    
1193    To apply the patch (using unix or cygwin)
1194    1) gunzip runtime.diff.gz
1195    2) "cd" into "src/runtime" in the source tree of a fresh
1196       110.45 installation.
1197    3) patch -p 1 < [your/path/to]runtime.diff
1198    
1199    The code compiles fine but has NOT yet been extensively tested.
1200    I only ran a few tests for basic socket client functionality
1201    (which worked fine).  Especially the functions that use ioctl
1202    are not tested at all and might not work (see below).
1203    
1204    I implemented this since we want to move to a newer version of sml/nj
1205    but need socket support in order to use it.  This is the first time I
1206    even had a look at the sml/nj source,  so please review my changes
1207    before making this part of the distribution!  Here are a few issues
1208    that I think might be better for someone to solve who is more
1209    familiar with the sml/nj source (and socket programming):
1210    
1211    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1212      exception since I could not figure out what the windows equivalent
1213      of these functions is
1214    
1215    - In sockets-osdep.h there are a some #include statements that are
1216      only used in a few files that include sockets-osdep.h
1217    
1218    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1219      does not process its return value since I don't know how to
1220      report an error upwards.
1221    
1222    - It would probably be good to have a call to WSACleanup() when
1223      the library is unloaded (if there is such a possibility).
1224      Otherwise I think Windows will take care of this automatically
1225      when the process finishes.
1226    
1227    - I used ioctlsocket() as a replacement for ioctl() but I have
1228      no idea if that is actually the proper replacement on Windows.
1229    
1230    - All these issues are marked in the code by "FIXME" comments.
1231    
1232    We use sml/nj extensively in our products and are quite happy
1233    with it.  I hope this contribution will help you.
1234    
1235    Keep up the good work!
1236    
1237    David
1238    
1239    ----------------------------------------------------------------------
1240    Name: Matthias Blume (blume (at) tti - c (dot) org)
1241    Date: 2004/05/11 14:20:00 CDT
1242    Tag: blume-20040511-installml
1243    Description:
1244    
1245    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1246    the patch.)
1247    
1248    ----------------------------------------------------------------------
1249    Name: Matthias Blume (blume (at) tti - c (dot) org)
1250    Date: 2004/05/11 14:05:00 CDT
1251    Tag: blume-20040511-nlffi-netbsd
1252    Description:
1253    
1254    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1255    the patch.)
1256    
1257    ----------------------------------------------------------------------
1258    Name: Matthias Blume (blume (at) tti - c (dot) org)
1259    Date: 2004/05/11 12:05:00 CDT
1260    Tag: blume-20040511-exports
1261    Description:
1262    
1263    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1264    export lists in compiler-related .cm-files.
1265    
1266    ----------------------------------------------------------------------
1267    Name: Matthias Blume (blume (at) tti - c (dot) org)
1268    Date: 2004/05/11 11:35:00 CDT
1269    Tag: blume-20040511-allsource
1270    Description:
1271    
1272    The installer now honors the "src-smlnj" target again, although its meaning
1273    has changed from "all sources required for the compiler" to "all sources
1274    the installer knows about".  In other words, if you enable "src-smlnj"
1275    in the "targets" file, then the installer will pull in sources for
1276    everything.  (Notice that this refers to source code only.  Compiled
1277    code is still only installed for modules that were requested explicitly
1278    or which are required for other modules that were requested explicitly.)
1279    
1280    ----------------------------------------------------------------------
1281    Name: Matthias Blume (blume (at) tti - c (dot) org)
1282    Date: 2004/04/23 17:40:00 CDT
1283    Tag: blume-20040423-ieee-scan
1284    Description:
1285    
1286    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1287    in the exponent calculation we get INF or ZERO (depending on the mantissa
1288    and the sign of the exponent).
1289    
1290    ----------------------------------------------------------------------
1291    Name: Matthias Blume (blume (at) tti - c (dot) org)
1292    Date: 2004/04/23 10:40:00 CDT
1293    Tag: blume-20040423-ml-build
1294    Description:
1295    
1296    The ml-build script now terminates with a non-0 status when something
1297    goes wrong.
1298    
1299    ----------------------------------------------------------------------
1300    Name: Matthias Blume (blume (at) tti - c (dot) org)
1301    Date: 2004/04/22 16:35:00 CDT
1302    Tag: blume-20040422-Option
1303    Description:
1304    
1305    Made exception Option to be the same as exception Option.Option
1306    (as it should be).
1307    
1308    ----------------------------------------------------------------------
1309    Name: Allen Leung (leunga (at) reservoir (dot) com)
1310    Date: 2004/03/19 14:40:00 EST
1311    Tag: leunga-20040319-cygwin-nlffi
1312    Description:
1313    
1314    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1315    of SML/NJ.  The problem is that
1316    
1317       lib = dlopen(NULL, ...)
1318       f   = dlsym(lib, "malloc");
1319    
1320    does not work on Windows unless we explicitly export symbols
1321    such as 'malloc' during linking.  We fixed this by explicitly
1322    exporting the required symbols with the magic gcc incantation:
1323    
1324        -Wl,--export-all cygwin.def
1325    
1326    where cygwin.def is a file containing all the symbols that
1327    we wish to export.
1328    
1329    I suspect this is a Windows problem and we'll have to
1330    do the same (somehow with windows compilers) when
1331    we build the native win32 version with the system
1332    calls LoadLibrary/GetProcAddress.
1333    
1334    ----------------------------------------------------------------------
1335    Name: Matthias Blume (blume (at) tti - c (dot) org)
1336    Date: 2004/03/04 16:35:00 CST
1337    Tag: blume-20040304-intinf-fmt
1338    Description:
1339    
1340    Fixed problem with IntInf.fmt (sign would show up on the right instead
1341    of on the left for BIN, OCT, and HEX).
1342    
1343    ----------------------------------------------------------------------
1344    Name: Matthias Blume (blume (at) tti - c (dot) org)
1345    Date: 2004/03/04 11:25:00 CST
1346    Tag: blume-20040304-symlinks
1347    Description:
1348    
1349    Fixed problem with installer script (unix only) where bin/ml-yacc and
1350    friends pointed (via symlinks) to absolute locations instead of just
1351    .run-sml.  This was reported by Vesa A Norrman.
1352    
1353    ----------------------------------------------------------------------
1354    Name: Matthias Blume (blume (at) tti - c (dot) org)
1355    Date: 2004/02/13 14:50:00 CST
1356    Tag: Release_110_45
1357    Description:
1358    
1359    New working version (110.45). New bootfiles.
1360    
1361    ----------------------------------------------------------------------
1362    Name: Matthias Blume (blume (at) tti - c (dot) org)
1363    Date: 2004/01/26 15:15:15 CST
1364    Tag: blume-20040126-toplevel
1365    Description:
1366    
1367    Improved handling of exceptions at the interactive toplevel.
1368    
1369    ----------------------------------------------------------------------
1370    Name: Matthias Blume (blume (at) tti - c (dot) org)
1371    Date: 2004/01/26 11:25:00
1372    Tag: blume-20040126-app
1373    Description:
1374    
1375    Type of top-level "app" corrected.
1376    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1377    
1378    ----------------------------------------------------------------------
1379    Name: Matthias Blume (blume (at) tti - c (dot) org)
1380    Date: 2003/11/18 17:10 CST
1381    Tag: blume-20031118-basis-fiddle
1382    Description:
1383    
1384    - changed Timer interface to what might become the spec
1385    - POSIX_FLAGS -> BIT_FLAGS according to spec
1386    - some other minor discrepancies wrt. spec eliminated
1387    
1388    ----------------------------------------------------------------------
1389    Name: Matthias Blume (blume (at) tti - c (dot) org)
1390    Date: 2003/11/06 12:00:00 CST
1391    Tag: Release_110_44
1392    Description:
1393    
1394    New working version (110.44). New bootfiles.
1395    
1396    ----------------------------------------------------------------------
1397    Name: Matthias Blume (blume (at) tti - c (dot) org)
1398    Date: 2003/11/04 11:50:00 CST
1399    Tag: blume-20031104-move-libraries
1400    Description:
1401    
1402    Eliminated the "dont_move_libraries" directive in config/targets.
1403    (The mechanism was broken and could not be fixed easily.  Moreover,
1404    there does not seem to be any reason not to move all libraries into
1405    lib during installation.  I originally implemented this directive as a
1406    backward-compatibility feature when I first introduced the new CM.
1407    Now that things have been stable for a long time and going back to the
1408    old CM is not an option, there is no reason to keep it around.)
1409    
1410    ----------------------------------------------------------------------
1411    Name: Matthias Blume (blume (at) tti - c (dot) org)
1412    Date: 2003/11/03 16:00:00 CST
1413    Tag: blume-20031103-installdir
1414    Description:
1415    
1416    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1417    Richards for pointing out the problem and providing the solution.)
1418    
1419    ----------------------------------------------------------------------
1420    Name: Matthias Blume (blume (at) tti - c (dot) org)
1421    Date: 2003/10/01 17:05:00 CDT
1422    Tag: blume-20031001-lal-mlrisc
1423    Description:
1424    
1425    MLRISC bug fix from Lal.
1426    
1427    ----------------------------------------------------------------------
1428    Name: Matthias Blume (blume (at) tti - c (dot) org)
1429    Date: 2003/09/30 16:10:00 CDT
1430    Tag: blume-20030930-primio-bat
1431    Description:
1432    
1433    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1434    2. Improved .bat files (for Win32 port) to make things work under Win95.
1435       (thanks to Aaron S. Hawley for this one)
1436    
1437    ----------------------------------------------------------------------
1438    Name: Matthias Blume (blume (at) tti - c (dot) org)
1439    Date: 2003/09/26 16:05:00 CDT
1440    Tag: blume-20030926-wrappriv
1441    Description:
1442    
1443    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1444    
1445    ----------------------------------------------------------------------
1446    Name: Matthias Blume (blume (at) tti - c (dot) org)
1447    Date: 2003/09/26 15:00:00 CDT
1448    Tag: blume-20030926-110_43_3
1449    Description:
1450    
1451    - additional cleanup
1452    - version number bump, NEW BOOTFILES
1453    
1454    ----------------------------------------------------------------------
1455    Name: Matthias Blume (blume (at) tti - c (dot) org)
1456    Date: 2003/09/26 12:00:00 CDT
1457    Tag: blume-20030926-ppautoload
1458    Description:
1459    
1460    I modified the read-eval-print loop so that the autoloader gets
1461    invoked whenever the prettyprinter tries to look up a symbol that
1462    is not currently defined in the toplevel environment but which
1463    appears in CM's autoload registry.  As a result, we see far fewer of
1464    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1465    
1466    In addition to this I tried to clean up some pieces of the Basis
1467    implementation (e.g., Socket, Word8Array) in order to prevent other
1468    instances of these ?.Foo.Bar.xxx names from being printed.
1469    
1470    The mechanism that picks names for types still needs some work, though.
1471    (Right now it seems that if there is a type A.t which is defined to
1472    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1473    "?.B.u" although the perhaps more sensible solution would be to use
1474    "A.t" in this case.  In other words, the prettyprinter should follow
1475    a chain of DEFtycs not farther than there are corresponding toplevel
1476    names in the current environment.)
1477    
1478    ----------------------------------------------------------------------
1479    Name: Matthias Blume (blume (at) tti - c (dot) org)
1480    Date: 2003/09/24 16:31:00 CDT
1481    Tag: blume-20030924-installer
1482    Description:
1483    
1484    Another installer tweak:  All the ML code for the installer is now
1485    compiled during CMB.make and put into a little library called
1486    $smlnj/installer.cm.  The installation then simply invokes
1487    
1488       sml -m $smlnj/installer.cm
1489    
1490    and everything happens automagically.
1491    
1492    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1493    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1494            and BINDIR.
1495    
1496    The new scheme guarantees that the ML code responsible for the installation
1497    is in sync with the APIs of the main system.  Also, the installer is
1498    somewhat faster because the installer script is precompiled.
1499    
1500    ----------------------------------------------------------------------
1501    Name: Matthias Blume (blume (at) tti - c (dot) org)
1502    Date: 2003/09/24 15:35:00 CDT
1503    Tag: blume-20030924-synsock
1504    Description:
1505    
1506    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1507    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1508    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1509    implementations of non-blocking operations from CML's Socket
1510    structure.
1511    
1512    ----------------------------------------------------------------------
1513    Name: Matthias Blume (blume (at) tti - c (dot) org)
1514    Date: 2003/09/24 15:10:05 CDT
1515    Tag: blume-20030924-sockets
1516    Description:
1517    
1518    1. Fixed SOCKET API and implementation to match Basis spec.
1519       This required changing the internal representation of sockets to one
1520       that remembers (for each socket file descriptor) whether it is currently
1521       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1522       call is made only if the state actually needs to change).
1523    
1524    2. OS-specific details of sockets were moved into separate files, thus
1525       making it possible to unify the bulk of the socket implementations
1526       between Unix and Win32.
1527    
1528    3. CML's socket API changed accordingly.
1529       (Note that we need to remove non-blocking functions from this API
1530       since they are redundant in the case of CML!)
1531    
1532    4. CML's socket implementation now makes use of non-blocking functions
1533       provided by Basis, thus removing all OS-dependent code from this part
1534       of CML.
1535    
1536    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1537    
1538    ----------------------------------------------------------------------
1539    Name: Matthias Blume (blume (at) tti - c (dot) org)
1540    Date: 2003/09/22 12:10:00 CDT
1541    Tag: blume-20030922-110_43_2
1542    Description:
1543    
1544    Made a new interim version and bootfiles for developer's bootstrapping
1545    convenience.
1546    
1547       110.43.2 -- NEW BOOTFILES
1548    
1549    ----------------------------------------------------------------------
1550    Name: Matthias Blume (blume (at) tti - c (dot) org)
1551    Date: 2003/09/19 15:55:00 CDT
1552    Tag: blume-20030919-cmdir
1553    Description:
1554    
1555    1. new-install.sh -> install.sh
1556    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1557    3. tweaked installer so that another name instead of .cm can be chosen
1558       at install time (by setting the CM_DIR_ARC environment variable
1559       during installation); once installation is complete, the name is
1560       fixed
1561    
1562    ----------------------------------------------------------------------
1563    Name: Matthias Blume (blume (at) tti - c (dot) org)
1564    Date: 2003/09/18 16:00:00 CDT
1565    Tag: blume-20030918-110_43_1
1566    Description:
1567    
1568    Made a new interim version and bootfiles for developer's bootstrapping
1569    convenience.
1570    
1571       110.43.1 -- NEW BOOTFILES
1572    
1573    ----------------------------------------------------------------------
1574    Name: Matthias Blume (blume (at) tti - c (dot) org)
1575    Date: 2003/09/18 15:20:00 CDT
1576    Tag: blume-20030918-misc
1577    Description:
1578    
1579    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1580       this seems to be controversial at the moment) and used those in
1581       Posix.ProcEnv.times.
1582    
1583    2. Added Time.{from,to}Nanoseconds to Time.
1584    
1585    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1586       For example, fromLargeInt never needs to look at more than 3 "big
1587       digits" to get its 53 bits of precision.
1588    
1589    ----------------------------------------------------------------------
1590    Name: Matthias Blume (blume (at) tti - c (dot) org)
1591    Date: 2003/09/17 16:30:00 CDT
1592    Tag: blume-20030917-real32-slices
1593    Description:
1594    
1595    Added an entry to the primitive environment
1596    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1597    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1598    
1599    Removed some of the "magic" constants in real64.sml and replaced them
1600    with code that generates these values from their corresponding
1601    integer counterparts.
1602    
1603    Made all(?) the slice-related changes to the Basis and made everything
1604    compile again...
1605    
1606    ----------------------------------------------------------------------
1607    Name: Matthias Blume (blume (at) tti - c (dot) org)
1608    Date: 2003/09/15 17:45:00 CDT
1609    Tag: blume-20030915-rbase
1610    Description:
1611    
1612    Fixed bug in Real.fromLargeInt.
1613    
1614    ----------------------------------------------------------------------
1615    Name: Matthias Blume (blume (at) tti - c (dot) org)
1616    Date: 2003/09/13 18:11:00 CDT
1617    Tag: blume-20030913-libinstall
1618    Description:
1619    
1620    Minor bugfix in config/libinstall (set anchor with path to
1621    standalone tool after installing it, otherwise libraries that
1622    need ml-lex or ml-yacc won't compile the first time the installer
1623    runs).
1624    
1625    ----------------------------------------------------------------------
1626    Name: Matthias Blume (blume (at) tti - c (dot) org)
1627    Date: 2003/09/12 11:45:00 CDT
1628    Tag: blume-20030912-various
1629    Description:
1630    
1631     - fixed bug in Real.toLargeInt
1632     - fixed bug in Posix.ProcEnv.times
1633     - changed inputLine functions to return an option
1634     - minor installer improvements / bugfixes
1635     - changed default @SMLalloc parameter for x86/celeron to 64k
1636    
1637    ----------------------------------------------------------------------
1638    Name: Matthias Blume (blume (at) tti - c (dot) org)
1639    Date: 2003/09/09 22:00:00 CDT
1640    Tag: Release_110_43
1641    Description:
1642    
1643        New working release 110.43.  New bootfiles.
1644    
1645    ----------------------------------------------------------------------
1646    Name: Matthias Blume (blume (at) tti - c (dot) org)
1647    Date: 2003/09/09 19:20:00 CDT
1648    Tag: blume-20030909-installer
1649    Description:
1650    
1651    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1652    Modified config/install.bat to take advantage of it.  Also modified
1653    config/install.sh (and called it config/new-install.sh) to take advantage
1654    of it on Unix systems.  (The SML code is (supposed to be) platform-
1655    independent.)
1656    
1657    The installer can now install everything under Win32
1658    as well as under *nix as long as it compiles.
1659    
1660    Other changes:
1661    
1662      - made CML compile again under Win32
1663      - made eXene compile under Win32 (by providing a fake structure UnixSock
1664        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1665      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1666        same as type int;  under Win32 it isn't
1667      - fixed some slice-related problems in the win32-specific parts of CML
1668      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1669        (under Win32, the volume name is case-insensitive, and the
1670        OS.Path code compares volume names for equality)
1671    
1672    ----------------------------------------------------------------------
1673    Name: Matthias Blume (blume (at) tti - c (dot) org)
1674    Date: 2003/09/08 11:55:00 CDT
1675    Tag: blume-20030908-fullpath
1676    Description:
1677    
1678    Made Win32 version of OS.FileSys.fullPath return current directory
1679    when given an empty string.  This is what the spec says, and incidentally,
1680    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1681    cases when presented with the name of a non-existing .cm file.)
1682    
1683    ----------------------------------------------------------------------
1684    Name: Matthias Blume (blume (at) tti - c (dot) org)
1685    Date: 2003/09/04 16:30:00 CDT
1686    Tag: blume-20030905-slices-etc
1687    Description:
1688    
1689      1. Changed interface to vectors and arrays in Basis to match
1690         (draft) Basis spec.
1691      2. Added signatures and implementations of slices according to
1692         Basis spec.
1693      3. Edited source code throughout the system to make it compile again
1694         under 1. and 2.  (In some cases code had to be added to have it
1695         match the new signatures.)
1696      4. MLRISC should be backward-compatible: the copies of the originals
1697         of files that needed to change under 3. were retained, the .cm files
1698         check the compiler version number and use old versions when
1699         appropriate.
1700      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1701         dirstream -> string option (in accordance with Basis spec).
1702      6. When generating code that counts lines, ml-lex used function
1703         CharVector.foldli, taking advantage of its old interface.
1704         This has been replaced with the corresponding code from
1705         CharVectorSlice.  (html-lex must be re-lexed!)
1706      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1707         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1708      8. Removed temporary additions (fromInternal, toInternal) from the
1709         (now obsolete) IntInf in smlnj-lib/Util.
1710      9. Cleaned up structure Byte.
1711     10. Added localOffset, scan, and fromString to Date (according to spec).
1712         Cleaned/corrected implementation of Date.
1713         (Still need to check for correctness; implement better canonicalizeDate.)
1714     11. Added "scan" to signature IEEE_REAL.
1715     12. Some improvements to IntInf [in particular: efficiency-hack for
1716         mod and rem when second operand is 2 (for parity checks).]
1717     13. Changed representation of type Time.time, using a single IntInf.int
1718         value counting microseconds.  This considerably simplified the
1719         implementation of structure Time.  We now support negative time
1720         values; scan and fromString handle signs.
1721     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1722         ArraySlice).
1723    
1724    ----------------------------------------------------------------------
1725    Name: Matthias Blume (blume (at) tti - c (dot) org)
1726    Date: 2003/08/28 17:00:00 CDT
1727    Tag: blume-20030828-intinf
1728    Description:
1729    
1730    This is a major update which comes with a version number bump
1731    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1732    and an implementation of IntInf in the Basis.
1733    
1734    There are a fairly large number of related changes and updates throughout
1735    the system:
1736    
1737    Basis:
1738            - Implemented IntInf.
1739            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1740            - Added some missing Real64 operations, most notably Real.toLargeInt.
1741            - Added FixedInt as a synonym for Int32.
1742    
1743    compiler:
1744            * Added support for a built-in intinf type.
1745              - literals
1746              - pattern matching
1747              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1748              - overloading on literals and operations
1749    
1750            This required adding a primitive type intinf, some additional
1751            primops, and implementations for several non-trivial intinf
1752            operations in Core.  (The intinf type is completely abstract
1753            to the compiler; all operations get delegated back to the Core.)
1754    
1755            * Intinf equality is handled by polyequal.  However, the compiler
1756            does not print its usual warning in this case (since polyequal
1757            is the right thing to do there).
1758    
1759            * Improved the organization of structure InlineT.
1760    
1761            * A word about conversion primops:
1762            If conversions involving intinf do not cancel out during
1763            CPS contract, then the compiler must insert calls to Core functions.
1764            Since all core access must be resolved already during the FLINT
1765            translate phase, it would be too late a the time of CPS contract
1766            to add new Core calls.  For this reason, conversion primops
1767            for intinf carry two arguments:  1. the numeric argument that
1768            they are supposed to convert, and 2. the Core function that
1769            can help with this conversion if necessary.  If CPS contract
1770            eliminates a primop, then the associated Core function becomes
1771            dead and goes away.  Intinf conversion primops that do not get
1772            eliminated by CPS contract get rewritten into calls of their
1773            core functions by a separate, new phase.
1774    
1775    interactive system:
1776            - Control.Print.intinfDepth controls max length of intinf constants
1777            being printed. (Analogous to Control.Print.stringDepth.)
1778            - Cleanup in printutil and pputil: got rid of unused stuff and
1779            duplicates; replaced some of the code with code that makes better
1780            use of library functionality.
1781    
1782    CM:
1783            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1784            are no longer silent.
1785    
1786    CKIT:
1787            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1788            in favor of LargeInt -- which is now the same as IntInf.
1789            CKIT-knowledgable people should check whether this is what's
1790            intended and otherwise change things back to using Int32 or
1791            FixedInt.
1792    
1793    Throughout the code:
1794            Started using IntInf.int literals and built-in operations
1795            (e.g., comparison with 0) where this seems appropriate.
1796    
1797    
1798    ----------------------------------------------------------------------
1799    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1800    Date: 2003/08/13 11:36:00 CDT
1801    Tag: dbm-20030813-mcz-merge1
1802    Description:
1803    
1804    Merging changes from the mcz-branch development branch into trunk.
1805    These changes involve replacement of the emulated old prettyprinter
1806    interface with direct use of the SML/NJ Lib PP library, and fixing
1807    of a couple of bugs (895, 1186) relating to error messages.  A new
1808    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1809    has been added.
1810    
1811    ----------------------------------------------------------------------
1812    Name: Matthias Blume (blume (at) tti - c (dot) org)
1813    Date: 2003/08/11 15:45:00 CDT
1814    Tag: blume-20030811-windows
1815    Description:
1816    
1817    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1818    
1819        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1820    
1821    This patch restores SML/NJ's ability to run under win32.  There are a
1822    number of changes, including fixes for several bugs that had gone
1823    unnoticed until now:
1824    
1825      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1826        trying to run the win32 version from within cygwin.)
1827    
1828      - There are a number of simple .bat scripts that substitute for their
1829        corresponding Unix shell-scripts.  (See below.)
1830    
1831      - The internals of ml-build have been modified slightly.  The main
1832        difference is that instead of calling ".link-sml" (or link-sml.bat)
1833        using OS.Process.system, the ML process delegates this task back
1834        to the script.  Otherwise problems arise in mixed environments such
1835        as Cygwin where scripts look and work like Unix scripts, but
1836        where OS.Process.system cannot run them.
1837    
1838      - In CM, the srcpath pickler used native pathname syntax -- which
1839        is incorrect in the case of cross-compilation.  The new pickle format
1840        is independent of platform-specific naming conventions.
1841    
1842      - Path configuration files (such as lib/pathconfig) can now choose
1843        between native and standard syntax.  Placing a line of the form
1844    
1845            standard!
1846    
1847        into the file causes all subsequent paths to be interpreted using
1848        CM standard pathname syntax (= Unix conventions); a line
1849    
1850            native!
1851    
1852        switches back to native style.  This was needed so that
1853        path config files can be written portably, see src/system/pathconfig.
1854    
1855      - Runtime system:
1856    
1857          - win32-filesys.c:  get_file_time and set_file_time now
1858            access modification time, not creation time.
1859    
1860          - I/O code made aware of new array representation.
1861    
1862          - Bug fixes in X86.prim.masm.
1863    
1864       - src/system/makeml made aware of win32. (For use under cygwin
1865            and other Unix-environments for windows.)
1866    
1867       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1868         which caused BinIO.inputAll to fail consistently.
1869    
1870    .bat scripts:
1871    
1872       Windows .bat scripts assume that SMLNJ_HOME is defined.
1873    
1874       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1875         applications (sml, ml-yacc, ml-lex).
1876       - ml-build.bat: analogous to ml-build.
1877       - config\install.bat:  Analogous to config/install.sh.  This requires
1878         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1879         (nmake etc. must be on the path, and vcvars32 must have been run.)
1880         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1881         and the bootfile hierarchy must have been unpacked under
1882         sml.boot.x86-win32.
1883         The script is very primitive and does a poor job at error checking.
1884         It only installs the base system, ml-lex, and ml-yacc.  No other
1885         libraries are being installed (i.e., you get only those that
1886         are part of the compiler.)
1887       - link-sml.bat:  analogous to .link-sml, but not currently used
1888    
1889    Unrelated bug fixes:
1890    
1891      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1892        types.
1893      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1894        otherwise.)
1895    
1896    ----------------------------------------------------------------------
1897    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1898    Date: 2003/06/17
1899    Tag: macqueen-20030617-bug895
1900    Description:
1901    
1902    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1903    Tag will be used for new development branch (mcz-branch) for
1904    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1905    
1906    ----------------------------------------------------------------------
1907    Name: Matthias Blume (blume (at) tti - c (dot) org)
1908    Date: 2003/05/27 16:55:00 CDT
1909    Tag: blume-20030527-polyeq
1910    Description:
1911    
1912    Tried to eliminated most cases of polymorphic equality.
1913    
1914    ----------------------------------------------------------------------
1915    Name: Matthias Blume (blume (at) tti - c (dot) org)
1916    Date: 2003/05/21 17:45:00 CDT
1917    Tag: blume-20030517-complete
1918    Description:
1919    
1920    Two changes:
1921    
1922      1. Added a flag for controlling whether non-exhaustive bindings will
1923         be treated as errors (default is false).
1924      2. Cleaned up the *entire* source tree so that CMB.make goes through
1925         without a single non-exhaustive match- or bind warning.
1926    
1927    ----------------------------------------------------------------------
1928    Name: Matthias Blume (blume (at) tti - c (dot) org)
1929    Date: 2003/05/17 10:20:00 CDT
1930    Tag: blume-20030517-absyn
1931    Description:
1932    
1933    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1934    
1935       This mainly affects the quality of error messages.  However, some
1936       of the code is now more straightforward than before. (Treatment of
1937       the above four constructs in translate.sml is much simpler than
1938       the "macro-expansion" that was going on before.  Plus, the mach-
1939       compiler no longer gets invoked just to be able to compile an
1940       if-expression.)
1941    
1942    2. The ErrorMsg.Error exception is now caught and absorbed by the
1943       interactive loop.
1944    
1945    ----------------------------------------------------------------------
1946    Name: Allen Leung
1947    Date: 2003/05/16 13:05:00 CDT
1948    Tag: leunga-20030516-cygwin-runtime
1949    Description:
1950    
1951        Ported the runtime system to cygwin, which uses the unix
1952    x86-unix bin files.  Missing/buggy features:
1953    
1954       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1955         the Cygwin library.
1956       o Ctrl-C handling may be flaky.
1957       o Windows system calls and Windows I/O are not supported.
1958    
1959       A new set of binfiles is located at:
1960    
1961         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1962    
1963       This is only needed for bootstrapping the cygwin version of smlnj.
1964       Other x86 versions can use the existing binfiles.
1965    
1966    ----------------------------------------------------------------------
1967    Name: Matthias Blume
1968    Date: 2003/04/08 15:42:00 CDT
1969    Tag: blume-20030408-listpair
1970    Description:
1971    
1972    1. Added a target 'mlrisc' to installer.
1973    
1974    2. Added missing elements to structure ListPair.
1975    
1976    ----------------------------------------------------------------------
1977    Name: Allen Leung
1978    Date: 2003/01/07 10:40:00 EST
1979    Tag: leunga-20030107-int-rem
1980    Description:
1981    
1982    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1983    The arguments to the SUBL instruction were swapped.
1984    
1985    ----------------------------------------------------------------------
1986    Name: Matthias Blume
1987    Date: 2002/12/12 16:25:00 EST
1988    Tag: blume-20021212-risc-ra
1989    Description:
1990    
1991    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1992    sent the RA into an infinite loop when floating point registers get
1993    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1994    110.37 and 110.38.)
1995    There was another set of potential problems related to the handling of
1996    MLRISC annotations (but those did not yet cause real problems, apparently).
1997    
1998    ----------------------------------------------------------------------
1999    Name: Matthias Blume
2000    Date: 2002/12/06 22:40:00 EST
2001    Tag: blume-20021206-cm-fileid
2002    Description:
2003    
2004    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
2005    This fixes the problem of CM getting confused by files that suddenly
2006    change their identity (e.g., by getting unlinked and recreated by some
2007    text editor such as vi).  There might be a better/cheaper/cleaner way
2008    of doing this, but for now this will have to do.
2009    
2010    ----------------------------------------------------------------------
2011    Name: Matthias Blume
2012    Date: 2002/10/28 09:50:00 EST
2013    Tag: blume-20021028-typecheck
2014    Description:
2015    
2016    Exported structure Typecheck from $smlnj/viscomp/core.cm.
2017    
2018    ----------------------------------------------------------------------
2019    Name: Matthias Blume
2020    Date: 2002/10/17 09:10:00 EDT
2021    Tag: Release_110_42
2022    Description:
2023    
2024    In good old tradition, there has been a slight hiccup so that we have
2025    to patch 110.42 after the fact.   The old release tag has been replaced
2026    (see below).
2027    
2028    The change solves a problem with two competing approaches the
2029    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
2030    each other's way.
2031    
2032    This change only affects the runtime system code and the installer script.
2033    (No new bootfiles.)
2034    
2035    ----------------------------------------------------------------------
2036    Name: Matthias Blume
2037    Date: 2002/10/16 12:00:00 EDT
2038    Tag: Release_110_42_removed
2039    Description:
2040    
2041    New working release.  New bootfiles.
2042    
2043    ----------------------------------------------------------------------
2044    Name: Matthias Blume
2045    Date: 2002/10/10 13:10:00 EDT
2046    Tag: blume-20021010-ppc-divs
2047    Description:
2048    
2049    The mltree operator DIVS must be implemented with an overflow check on
2050    the PPC because the hardware indicates divide-by-zero using "overflow" as
2051    well.
2052    
2053    ----------------------------------------------------------------------
2054    Name: Matthias Blume
2055    Date: 2002/07/23 11:45:00
2056    Tag: blume-20020723-smlnj-home
2057    Description:
2058    
2059    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
2060    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
2061    CM_PATHCONFIG is also set), the path configuration file is assumed
2062    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
2063    the entire tree to some other place and everything will "just work".
2064    
2065    (Companion commands such as ml-build and ml-makedepend also sense this
2066    variable.)
2067    
2068    ----------------------------------------------------------------------
2069    Name: Matthias Blume
2070    Date: 2002/07/12 21:19:00 EDT
2071    Tag: blume-20020712-liveness
2072    Description:
2073    
2074    Exported two useful "step" functions from liveness module (MLRISC).
2075    
2076    ----------------------------------------------------------------------
2077    Name: Matthias Blume
2078    Date: 2002/07/05 16:00 EDT
2079    Tag: Release_110_41
2080    Description:
2081    
2082    New working release.  New bootfiles.
2083    
2084    ----------------------------------------------------------------------
2085    Name: Matthias Blume
2086    Date: 2002/07/05 10:25:00 EDT
2087    Tag: blume-20020705-btimp
2088    Description:
2089    
2090    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
2091    other clients can set up backtracing support.
2092    
2093    ----------------------------------------------------------------------
2094    Name: Matthias Blume
2095    Date: 2002/06/25 17:23:00 EDT
2096    Tag: blume-20020625-fpmax
2097    Description:
2098    
2099    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
2100    case.  (The sense of the isNaN test was reversed -- which made min and
2101    max always return their first argument.)
2102    
2103    ----------------------------------------------------------------------
2104    Name: Matthias Blume
2105    Date: 2002/06/11
2106    Tag: blume-20020611-unixpath
2107    Description:
2108    
2109    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
2110    
2111    ----------------------------------------------------------------------
2112    Name: Matthias Blume
2113    Date: 2002/06/10 16:35:00 EDT
2114    Tag: blume-20020610-ieeereal
2115    Description:
2116    
2117    I back-ported my implementation of IEEEReal.fromString from the
2118    idlbasis-devel branch so that we can test it.
2119    
2120    Another small change is that ppDec tries to give more information
2121    than just "<sig>" in the case of functors.  However, this code is
2122    broken in some mysterious way if the functor's body's signature
2123    has not been declared by ascription but gets inferred from the
2124    implementation.  This needs fixing...
2125    
2126    ----------------------------------------------------------------------
2127    Name: Matthias Blume
2128    Date: 2002/05/31
2129    Tag: blume-20020531-btrace-mode
2130    Description:
2131    
2132    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2133    the wayside when I switched over to using Controls everywhere.)
2134    
2135    ----------------------------------------------------------------------
2136    Name: Lal George
2137    Date: 2002/05/23 12:21:40 EDT
2138    Tag: george-20020523-visual-labels
2139    Description:
2140    
2141    Labels are now displayed in the graphical output to make
2142    the fall-through and target blocks obvious.
2143    
2144    ----------------------------------------------------------------------
2145    Name: Matthias Blume
2146    Date: 2002/05/22 11:03:00 EDT
2147    Tag: blume-20020522-shrink
2148    Description:
2149    
2150    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2151    situation that comes up when allocating huge arrays.
2152    
2153    ----------------------------------------------------------------------
2154    Name: Matthias Blume
2155    Date: 2002/05/21 16:00:00 EDT
2156    Tag: Release_110_40
2157    Description:
2158    
2159    New working release (110.40).  New bootfiles.
2160    
2161    [Also: John Reppy fixed GC bug 1131.]
2162    
2163    ----------------------------------------------------------------------
2164    Name: Matthias Blume
2165    Date: 2002/05/21 12:35:00 EDT
2166    Tag: blume-20020521-cmdoc
2167    Description:
2168    
2169    CM documentation update.
2170    
2171    ----------------------------------------------------------------------
2172    Name: Matthias Blume
2173    Date: 2002/05/21 10:55:00 EDT
2174    Tag: blume-20020521-misc
2175    Description:
2176    
2177    - John tweaked runtime to be silent on heap export (except when
2178      GC messages are on).
2179    - I added a few more things (cross-compiling versions of CMB) to
2180      config/preloads (as suggestions).
2181    
2182    ----------------------------------------------------------------------
2183    Name: Matthias Blume
2184    Date: 2002/05/20 22:25:00 EDT
2185    Tag: blume-20020520-controls
2186    Description:
2187    
2188    - Added ControlUtil structure to control-lib.cm.
2189    - Use it throughout.
2190    - Used Controls facility to define MLRISC controls (as opposed to
2191      registering MLRISC control ref cells with Controls after the fact)
2192    - Fixed messed-up controls priorities.
2193    
2194    * Removed again all the stuff from config/preloads that one wouldn't
2195      be able to preload at the time the initial heap image is built.
2196      (Many libraries, e.g., CML, do not exist yet at this time.  The
2197       only libraries that can be preloaded via config/preloads are those
2198       that come bundled with the bootfiles.)
2199    
2200    ----------------------------------------------------------------------
2201    Name: Matthias Blume
2202    Date: 2002/05/20 10:59:00 EDT
2203    Tag: blume-20020520-preloads
2204    Description:
2205    
2206    Added a lot of commented-out suggestions for things to be included
2207    in config/preloads.
2208    
2209    ----------------------------------------------------------------------
2210    Name: Allen Leung
2211    Date: 2002/05/18 14:20:00 EDT
2212    Tag: leunga-20020518-mdl
2213    Description:
2214    
2215    o Made the mdl tool stuff compile and run again.
2216    o I've disabled all the stuff that depends on RTL specifications; they
2217      are all badly broken anyway.
2218    
2219    ----------------------------------------------------------------------
2220    Name: Matthias Blume
2221    Date: 2002/05/17 16:49:00 EDT
2222    Tag: blume-20020517-controls
2223    Description:
2224    
2225    0. John Reppy made several modifications to the SML/NJ library.
2226       In particular, there is a shiny new controls-lib.cm.
2227    
2228    1. Pushed new controls interface through compiler so that everything
2229       compiles again.
2230    
2231    2. Added FormatComb and FORMAT_COMB to the CML version of the
2232       SML/NJ library (so that CML compiles again).
2233    
2234    3. Modified init scripts because XXX_DEFAULT environment variables
2235       are no longer with us.  (Boot-time initialization is now done
2236       using the same environment variables that are also used for
2237       startup-time initialization of controls.)
2238    
2239    ----------------------------------------------------------------------
2240    Name: Lal George
2241    Date: 2002/05/15 09:20:10 EDT
2242    Tag: george-20020515-pseudo-op-decls
2243    Description:
2244    
2245            All pseudo-ops emitted before the first segment declaration
2246    such as TEXT, DATA, and BSS directives are assumed to be global
2247    declarations and are emitted first in the assembly file. This is
2248    useful in a number of situations where one has pseudo-ops that are not
2249    specific to any segment, and also works around the constraint that one
2250    cannot have client pseudo-ops in the TEXT segment.
2251    
2252            Because no segment is associated with these declarations it is
2253    an error to allocate any space or objects before the first segment
2254    directive and an exception will be raised. However, we cannot make
2255    this check for client pseudo-ops.
2256    
2257            These top level declarations are a field in the CFG graph_info.
2258    In theory you can continue to add to this field after the CFG has been
2259    built -- provided you know what you are doing;-)
2260    
2261    ----------------------------------------------------------------------
2262    Name: Matthias Blume
2263    Date: 2002/05/13 16:40:00 EDT
2264    Tag: blume-20020513-pp-etc
2265    Description:
2266    
2267    A few minor bugfixes:
2268    
2269      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2270        (Bogus printouts for redefined bindings still occur.  Compiler
2271        bug should no longer occur now.  We need to redo the prettyprinter
2272        from scratch.)
2273    
2274      - CM pathname printer now also adds escape sequences for ( and )
2275    
2276      - commend and docu fixes for ml-nlffi
2277    
2278    ----------------------------------------------------------------------
2279    Name: Matthias Blume
2280    Date: 2002/05/10 16:40:00 EDT
2281    Tag: blume-20020510-erg-textio
2282    Description:
2283    
2284    Applied the following bugfix provided by Emden Gansner:
2285    
2286        Output is corrupted when outputSubstr is used rather than output.
2287    
2288        The problem occurs when a substring
2289    
2290            ss = (s, dataStart, dataLen)
2291    
2292        where dataStart > 0, fills a stream buffer with avail bytes left.
2293        avail bytes of s, starting at index dataStart, are copied into the
2294        buffer, the buffer is flushed, and then the remaining dataLen-avail
2295        bytes of ss are copied into the beginning of the buffer. Instead of
2296        starting this copy at index dataStart+avail in s, the current code
2297        starts the copy at index avail.
2298    
2299        Fix:
2300        In text-io-fn.sml, change line 695 from
2301             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2302        to
2303             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2304    
2305    ----------------------------------------------------------------------
2306    Name: Matthias Blume
2307    Date: 2002/04/12 13:55:00 EDT
2308    Tag: blume-20020412-assyntax
2309    Description:
2310    
2311    1. Grabbed newer assyntax.h from the XFree86 project.
2312    2. Fiddled with how to compile X86.prim.asm without warnings.
2313    3. (Very) Minor cleanup in CM.
2314    
2315    ----------------------------------------------------------------------
2316    Name: Matthias Blume
2317    Date: 2002/04/01 (no joke!) 17:07:00 EST
2318    Tag: blume-20020401-x86div
2319    Description:
2320    
2321    Added full support for div/mod/rem/quot on the x86, using the machine
2322    instruction's two results (without clumsily recomputing the remainder)
2323    directly where appropriate.
2324    
2325    Some more extensive power-of-two support was added to the x86 instruction
2326    selector (avoiding expensive divs, mods, and muls where they can be
2327    replaced with cheaper shifts and masks).  However, this sort of thing
2328    ought to be done earlier, e.g., within the CPS optimizer so that
2329    all architectures benefit from it.
2330    
2331    The compiler compiles to a fixed point, but changes might be somewhat
2332    fragile nevertheless.  Please, report any strange things that you might
2333    see wrt. div/mod/quot/rem...
2334    
2335    ----------------------------------------------------------------------
2336    Name: Matthias Blume
2337    Date: 2002/03/29 17:22:00
2338    Tag: blume-20020329-div
2339    Description:
2340    
2341    Fixed my broken div/mod logic.  Unfortunately, this means that the
2342    inline code for div/mod now has one more comparison than before.
2343    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2344    The problem was with quotient = 0, because that alone does not tell
2345    us which way the rounding went.  One then has to look at whether
2346    remainder and divisor have the same sign...  :(
2347    
2348    Anyway, I replaced the bootfiles with fresh ones...
2349    
2350    ----------------------------------------------------------------------
2351    Name: Matthias Blume
2352    Date: 2002/03/29 14:10:00 EST
2353    Tag: blume-20020329-inlprims
2354    Description:
2355    
2356    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2357    
2358    Primops have changed. This means that the bin/boot-file formats have
2359    changed as well.
2360    
2361    To make sure that there is no confusion, I made a new version.
2362    
2363    
2364    CHANGES:
2365    
2366    * removed REMT from mltree (remainder should never overflow).
2367    
2368    * added primops to deal with divisions of all flavors to the frontend
2369    
2370    * handled these primops all the way through so they map to their respective
2371      MLRISC support
2372    
2373    * used these primops in the implementation of Int, Int32, Word, Word32
2374    
2375    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2376    
2377    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2378    
2379    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2380      floating point
2381    
2382    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2383      and Real (but Real.abs maps to a separate floating-point-only primop)
2384    
2385    
2386    TODO items:
2387    
2388    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2389      instructions because the machine instruction encoder cannot handle
2390      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2391      REM in the first place, and REM is supported by the x86 machine coder.)
2392    
2393    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2394      selection phase.  (The two can be streamlined because the hardware
2395      delivers both quotient and remainder at the same time anyway.)
2396    
2397    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2398      (Currently the behavior is inconsistent both across architectures and
2399      wrt. the draft Basis spec.)
2400    
2401    * Word8 should eventually be handled natively, too.
2402    
2403    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2404      as if there currently is no execution path that could trigger it in
2405      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2406      hold for things like multiplication and division.)
2407    
2408    ----------------------------------------------------------------------
2409    Name: Matthias Blume
2410    Date: 2002/03/27 16:27:00 EST
2411    Tag: blume-20020327-mlrisc-divisions
2412    Description:
2413    
2414    Added support for all four division operations (ML's div, mod, quot,
2415    and rem) to MLRISC.  In the course of doing so, I also rationalized
2416    the naming (no more annoying switch-around of DIV and QUOT), by
2417    parameterizing the operation by div_rounding_mode (which can be either
2418    DIV_TO_ZERO or DIV_TO_NEGINF).
2419    
2420    The generic MLTreeGen functor takes care of compiling all four
2421    operations down to only round-to-zero div.
2422    
2423    Missing pieces:
2424    
2425      * Doing something smarter than relying on MLTreeGen on architectures
2426        like, e.g., the x86 where hardware division delivers both quotient and
2427        remainder at the same time.  With this, the implementation of the
2428        round-to-neginf operations could be further streamlined.
2429    
2430      * Remove inlining support for div/mod/rem from the frontend and replace it
2431        with primops that get carried through to the backend.  Do this for all
2432        int and word types.
2433    
2434    ----------------------------------------------------------------------
2435    Name: Matthias Blume
2436    Date: 2002/03/25 17:25:00 EST
2437    Tag: blume-20020325-divmod
2438    Description:
2439    
2440    I improved (hopefully without breaking them) the implementation of Int.div,
2441    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2442    advantage of the following observations:
2443    
2444      Let  q = x quot y      r = x rem y
2445           d = x div  y      m = x mod y
2446    
2447    where "quot" is the round-to-zero version of integer division that
2448    hardware usually provides.  Then we have:
2449    
2450         r = x - q * y        where neither the * nor the - will overflow
2451         d = if q >= 0 orelse x = q * y then q else q - 1
2452                              where neither the * nor the - will overflow
2453         m = if q >= 0 orelse r = 0 then r else r + y
2454                              where the + will not overflow
2455    
2456    This results in substantial simplification of the generated code.
2457    The following table shows the number of CFG nodes and edges generated
2458    for
2459            fun f (x, y) = x OPER y
2460            (* with OPER \in div, mod, quot, rem *)
2461    
2462    
2463        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2464        --------------------------------------------------------
2465         div |         24 |         39 |         12 |         16
2466         mod |         41 |         71 |         12 |         16
2467        quot |          8 |         10 |          8 |         10
2468         rem |         10 |         14 |          8 |         10
2469    
2470    
2471    ----------------------------------------------------------------------
2472    Name: Matthias Blume
2473    Date: 2002/03/25 22:06:00 EST
2474    Tag: blume-20020325-cprotobug
2475    Description:
2476    
2477    Fixed a bug in cproto (c prototype decoder).
2478    
2479    ----------------------------------------------------------------------
2480    Name: Matthias Blume
2481    Date: 2002/03/25 16:00:00 EST
2482    Tag: blume-20020325-raw-primops
2483    Description:
2484    
2485    I did some cleanup to Allen's new primop code and
2486    replaced yesterday's bootfiles with new ones.
2487    (But they are stored in the same place.)
2488    
2489    ----------------------------------------------------------------------
2490    Name: Matthias Blume
2491    Date: 2002/03/24 22:40:00 EST
2492    Tag: blume-20020324-bootfiles
2493    Description:
2494    
2495    Made the bootfiles that Allen asked for.
2496    
2497    ----------------------------------------------------------------------
2498    Name: Allen Leung
2499    Date: 2002/03/23 15:50:00 EST
2500    Tag: leunga-20020323-flint-cps-rcc-primops
2501    Description:
2502    
2503      1. Changes to FLINT primops:
2504    
2505        (* make a call to a C-function;
2506         * The primop carries C function prototype information and specifies
2507         * which of its (ML-) arguments are floating point. C prototype
2508         * information is for use by the backend, ML information is for
2509         * use by the CPS converter. *)
2510      | RAW_CCALL of { c_proto: CTypes.c_proto,
2511                       ml_args: ccall_type list,
2512                       ml_res_opt: ccall_type option,
2513                       reentrant : bool
2514                     } option
2515       (* Allocate uninitialized storage on the heap.
2516        * The record is meant to hold short-lived C objects, i.e., they
2517        * are not ML pointers.  With the tag, the representation is
2518        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2519        *)
2520      | RAW_RECORD of {tag:bool,sz:int}
2521      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2522    
2523      2.  These CPS primops are now overloaded:
2524    
2525           rawload of {kind:numkind}
2526           rawstore of {kind:numkind}
2527    
2528          The one argument form is:
2529    
2530             rawload {kind} address
2531    
2532          The two argument form is:
2533    
2534             rawload {kind} [ml object, byte-offset]
2535    
2536      3. RAW_CCALL/RCC now takes two extra arguments:
2537    
2538         a. The first is whether the C call is reentrant, i.e., whether
2539            ML state should be saved and restored.
2540         b. The second argument is a string argument specifying the name of
2541            library and the C function.
2542    
2543         These things are currently not handled in the code generator, yet.
2544    
2545      4. In CProto,
2546    
2547         An encoding type of "bool" means "ml object" and is mapped into
2548         C prototype of PTR.  Note that "bool" is different than "string",
2549         even though "string" is also mapped into PTR, because "bool"
2550         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2551    
2552      5. Pickler/unpicker
2553    
2554         Changed to handle RAW_RECORD and newest RAW_CCALL
2555    
2556      6. MLRiscGen,
2557    
2558         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2559         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2560            in the file CodeGen/cpscompile/cps-c-calls.sml
2561    
2562      7. Added the conditional move operator
2563    
2564             condmove of branch
2565    
2566         to cps.  Generation of this is still buggy so it is currently
2567         disabled.
2568    
2569    ----------------------------------------------------------------------
2570    Name: Lal George
2571    Date: 2002/03/22 14:18:25 EST
2572    Tag: george-20020322-cps-branch-prob
2573    Description:
2574    
2575    Implemented the Ball-Larus branch prediction-heuristics, and
2576    incorporated graphical viewers for control flow graphs.
2577    
2578    Ball-Larus Heuristics:
2579    ---------------------
2580    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2581    
2582    By design it uses the Dempster-Shafer theory for combining
2583    probabilities.  For example, in the function:
2584    
2585        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2586    
2587    the ball-larus heuristics predicts that the n=0 is unlikely
2588    (OH-heuristic), and the 'then' branch is unlikely because of the
2589    RH-heuristic -- giving the 'then' branch an even lower combined
2590    probability using the Dempster-Shafer theory.
2591    
2592    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2593    probability of the 'then' branch because of the loop in the else
2594    branch.
2595    
2596    
2597    Graphical Viewing:
2598    ------------------
2599    I merely plugged in Allen's graphical viewers into the compiler. The
2600    additional code is not much. At the top level, saying:
2601    
2602            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2603    
2604    will display the graphical view of the control flow graph just before
2605    back-patching.  daVinci must be in your path for this to work. If
2606    daVinci is not available, then the default viewer can be changed
2607    using:
2608    
2609            Control.MLRISC.getString "viewer"
2610    
2611    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2612    course, these viewers must be in your path.
2613    
2614    The above will display the compilation unit at the level of clusters,
2615    many of which are small, boring, and un-interesting. Also setting:
2616    
2617            Control.MLRISC.getInt "cfg-graphical-view_size"
2618    
2619    will display clusters that are larger than the value set by the above.
2620    
2621    
2622    ----------------------------------------------------------------------
2623    Name: Matthias Blume
2624    Date: 2002/03/21 22:20:00 EST
2625    Tag: blume-20020321-kmp-bugfix
2626    Description:
2627    
2628    Changed the interface to the KMP routine in PreString and fixed
2629    a minor bug in one place where it was used.
2630    
2631    ----------------------------------------------------------------------
2632    Name: Allen Leung
2633    Date: 2002/03/21 20:30:00 EST
2634    Tag: leunga-20020321-cfg
2635    Description:
2636    
2637      Fixed a potential problem in cfg edge splitting.
2638    
2639    ----------------------------------------------------------------------
2640    Name: Allen Leung
2641    Date: 2002/03/21 17:15:00 EST
2642    Tag: leunga-20020321-x86-fp-cfg
2643    Description:
2644    
2645      1. Recoded the buggy parts of x86-fp.
2646    
2647         a. All the block reordering code has been removed.
2648            We now depend on the block placement phases to do this work.
2649    
2650         b. Critical edge splitting code has been simplified and moved into the
2651            CFG modules, as where they belong.
2652    
2653         Both of these were quite buggy and complex.  The code is now much, much
2654         simpler.
2655    
2656      2. X86 backend.
2657    
2658         a. Added instructions for 64-bit support.  Instruction selection for
2659            64-bit has not been committed, however, since that
2660            requires changes to MLTREE which haven't been approved by
2661            Lal and John.
2662    
2663         b. Added support for FUCOMI and FUCOMIP when generating code for
2664            PentiumPro and above.  We only generate these instructions in
2665            the fast-fp mode.
2666    
2667         c. Added cases for JP and JNP in X86FreqProps.
2668    
2669      3. CFG
2670    
2671         CFG now has a bunch of methods for edge splitting and merging.
2672    
2673      4. Machine description.
2674    
2675         John's simplification of MLTREE_BASIS.fcond broke a few machine
2676         description things:
2677    
2678         rtl-build.{sig,sml} and hppa.mdl fixed.
2679    
2680         NOTE: the machine description stuff in the repository is still broken.
2681               Again, I can't put my fixes in because that involves
2682               changes to MLTREE.
2683    
2684    ----------------------------------------------------------------------
2685    Name: Matthias Blume
2686    Date: 2002/03/20 15:55:00 EST
2687    Tag: blume-20020320-kmp
2688    Description:
2689    
2690    Implemented Knuth-Morris-Pratt string matching in PreString and used
2691    it for String.isSubstring, Substring.isSubstring, and
2692    Substring.position.
2693    
2694    (Might need some stress-testing.  Simple examples worked fine.)
2695    
2696    ----------------------------------------------------------------------
2697    Name: Matthias Blume
2698    Date: 2002/03/19 16:37:00 EST
2699    Tag: blume-20020319-witnesses
2700    Description:
2701    
2702    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2703    
2704    This implements a generic mechanism for changing constness qualifiers
2705    anywhere within big C types without resorting to outright "casts".
2706    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2707    modify the constness at the outermost level.)
2708    The implementation of "convert" is based on the idea of "witness"
2709    values -- values that are not used by the operation but whose types
2710    "testify" to their applicability.  On the implementation side, "convert"
2711    is simply a projection (returning its second curried argument).  With
2712    cross-module inlining, it should not result in any machine code being
2713    generated.
2714    
2715    ----------------------------------------------------------------------
2716    Name: Matthias Blume
2717    Date: 2002/03/15 16:40:00 EST
2718    Tag: blume-20020315-basis
2719    Description:
2720    
2721    Provided (preliminary?) implementations for
2722    
2723      {String,Substring}.{concatWith,isSuffix,isSubstring}
2724    
2725    and
2726    
2727      Substring.full
2728    
2729    Those are in the Basis spec but they were missing in SML/NJ.
2730    
2731    ----------------------------------------------------------------------
2732    Name: Matthias Blume
2733    Date: 2002/03/14 21:30:00 EST
2734    Tag: blume-20020314-controls
2735    Description:
2736    
2737    Controls:
2738    ---------
2739    
2740    1. Factored out the recently-added Controls : CONTROLS stuff and put
2741       it into its own library $/controls-lib.cm.  The source tree for
2742       this is under src/smlnj-lib/Controls.
2743    
2744    2. Changed the names of types and functions in this interface, so they
2745       make a bit more "sense":
2746    
2747          module -> registry
2748          'a registry -> 'a group
2749    
2750    3. The interface now deals in ref cells only.  The getter/setter interface
2751       is (mostly) gone.
2752    
2753    4. Added a function that lets one register an already-existing ref cell.
2754    
2755    5. Made the corresponding modifications to the rest of the code so that
2756       everything compiles again.
2757    
2758    6. Changed the implementation of Controls.MLRISC back to something closer
2759       to the original.  In particular, this module (and therefore MLRISC)
2760       does not depend on Controls.  There now is some link-time code in
2761       int-sys.sml that registers the MLRISC controls with the Controls
2762       module.
2763    
2764    CM:
2765    ---
2766    
2767      * One can now specify the lambda-split aggressiveness in init.cmi.
2768    
2769    ----------------------------------------------------------------------
2770    Name: Allen Leung
2771    Date: 2002/03/13 17:30:00 EST
2772    Tag: leunga-20020313-x86-fp-unary
2773    Description:
2774    
2775    Bug fix for:
2776    
2777    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2778    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2779    > - fun f(x,(y,z)) = Real.~ y;
2780    > [autoloading]
2781    > [autoloading done]
2782    >       fchsl   (%eax), 184(%esp)
2783    > Error: MLRisc bug: X86MCEmitter.emitInstr
2784    >
2785    > uncaught exception Error
2786    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2787    
2788    The problem was that the code generator did not generate any fp registers
2789    in this case, and the ra didn't know that it needed to run the X86FP phase to
2790    translate the pseudo fp instruction.   This only happened with unary fp
2791    operators in certain situations.
2792    
2793    ----------------------------------------------------------------------
2794    Name: Matthias Blume
2795    Date: 2002/03/13 14:00:00 EST
2796    Tag: blume-20020313-overload-etc
2797    Description:
2798    
2799    1. Added _overload as a synonym for overload for backward compatibility.
2800       (Control.overloadKW must be true for either version to be accepted.)
2801    
2802    2. Fixed bug in install script that caused more things to be installed
2803       than what was requested in config/targets.
2804    
2805    3. Made CM aware of the (_)overload construct so that autoloading
2806       works.
2807    
2808    ----------------------------------------------------------------------
2809    Name: Matthias Blume
2810    Date: 2002/03/12 22:03:00 EST
2811    Tag: blume-20020312-url
2812    Description:
2813    
2814    Forgot to update BOOT and srcarchiveurl.
2815    
2816    ----------------------------------------------------------------------
2817    Name: Matthias Blume
2818    Date: 2002/03/12 17:30:00 EST
2819    Tag: blume-20020312-version110392
2820    Description:
2821    
2822    Yet another version number bump (because of small changes to the
2823    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2824    
2825    Changes:
2826    
2827      The new pid generation scheme described a few weeks ago was overly
2828      complicated.  I implemented a new mechanism that is simpler and
2829      provides a bit more "stability":  Once CM has seen a compilation
2830      unit, it keeps its identity constant (as long as you do not delete
2831      those crucial CM/GUID/* files).  This means that when you change
2832      an interface, compile, then go back to the old interface, and
2833      compile again, you arrive at the original pid.
2834    
2835      There now also is a mechanism that instructs CM to use the plain
2836      environment hash as a module's pid (effectively making its GUID
2837      the empty string).  For this, "noguid" must be specified as an
2838      option to the .sml file in question within its .cm file.
2839      This is most useful for code that is being generated by tools such
2840      as ml-nlffigen (because during development programmers tend to
2841      erase the tool's entire output directory tree including CM's cached
2842      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2843      revert to the old, broken behavior of SML/NJ, but in specific cases
2844      where there is no danger of interface confusion, its use is ok
2845      (I think).
2846    
2847      ml-nlffigen by default generates "noguid" annotations.  They can be
2848      turned off by specifying -guid in its command line.
2849    
2850    ----------------------------------------------------------------------
2851    Name: Lal George
2852    Date: 2002/03/12 12 14:42:36 EST
2853    Tag: george-20020312-frequency-computation
2854    Description:
2855    
2856    Integrated jump chaining and static block frequency into the
2857    compiler. More details and numbers later.
2858    
2859    ----------------------------------------------------------------------
2860    Name: Lal George
2861    Date: 2002/03/11 11 22:38:53 EST
2862    Tag: george-20020311-jump-chain-elim
2863    Description:
2864    
2865    Tested the jump chain elimination on all architectures (except the
2866    hppa).  This is on by default right now and is profitable for the
2867    alpha and x86, however, it may not be profitable for the sparc and ppc
2868    when compiling the compiler.
2869    
2870    The gc test will typically jump to a label at the end of the cluster,
2871    where there is another jump to an external cluster containing the actual
2872    code to invoke gc. This is to allow factoring of common gc invocation
2873    sequences. That is to say, we generate:
2874    
2875            f:
2876               testgc
2877               ja   L1      % jump if above to L1
2878    
2879            L1:
2880               jmp L2
2881    
2882    
2883    After jump chain elimination the 'ja L1' instructions is converted to
2884    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2885    up being implemented in their long form (if L2 is far away) using:
2886    
2887            jbe     L3      % jump if below or equal to L3
2888            jmp     L2
2889         L3:
2890            ...
2891    
2892    
2893    For large compilation units L2  may be far away.
2894    
2895    
2896    ----------------------------------------------------------------------
2897    Name: Matthias Blume
2898    Date: 2002/03/11 13:30:00 EST
2899    Tag: blume-20020311-mltreeeval
2900    Description:
2901    
2902    A functor parameter was missing.
2903    
2904    ----------------------------------------------------------------------
2905    Name: Allen Leung
2906    Date: 2002/03/11 10:30:00 EST
2907    Tag: leunga-20020311-runtime-string0
2908    Description:
2909    
2910       The representation of the empty string now points to a
2911    legal null terminated C string instead of unit.  It is now possible
2912    to convert an ML string into C string with InlineT.CharVector.getData.
2913    This compiles into one single machine instruction.
2914    
2915    ----------------------------------------------------------------------
2916    Name: Allen Leung
2917    Date: 2002/03/10 23:55:00 EST
2918    Tag: leunga-20020310-x86-call
2919    Description:
2920    
2921       Added machine generation for CALL instruction (relative displacement mode)
2922    
2923    ----------------------------------------------------------------------
2924    Name: Matthias Blume
2925    Date: 2002/03/08 16:05:00
2926    Tag: blume-20020308-entrypoints
2927    Description:
2928    
2929    Version number bumped to 110.39.1.  NEW BOOTFILES!
2930    
2931    Entrypoints: non-zero offset into a code object where execution should begin.
2932    
2933    - Added the notion of an entrypoint to CodeObj.
2934    - Added reading/writing of entrypoint info to Binfile.
2935    - Made runtime system bootloader aware of entrypoints.
2936    - Use the address of the label of the first function given to mlriscGen
2937      as the entrypoint.  This address is currently always 0, but it will
2938      not be 0 once we turn on block placement.
2939    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2940      with entry points) from mlriscGen.
2941    
2942    ----------------------------------------------------------------------
2943    Name: Allen Leung
2944    Date: 2002/03/07 20:45:00 EST
2945    Tag: leunga-20020307-x86-cmov
2946    Description:
2947    
2948       Bug fixes for CMOVcc on x86.
2949    
2950       1. Added machine code generation for CMOVcc
2951       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2952       3. CMOVcc cannot have an immediate operand as argument.
2953    
2954    ----------------------------------------------------------------------
2955    Name: Matthias Blume
2956    Date: 2002/03/07 16:15:00 EST
2957    Tag: blume-20020307-controls
2958    Description:
2959    
2960    This is a very large but mostly boring patch which makes (almost)
2961    every tuneable compiler knob (i.e., pretty much everything under
2962    Control.* plus a few other things) configurable via both the command
2963    line and environment variables in the style CM did its configuration
2964    until now.
2965    
2966    Try starting sml with '-h' (or, if you are brave, '-H')
2967    
2968    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2969    implements the underlying generic mechanism.
2970    
2971    The interface to some of the existing such facilities has changed somewhat.
2972    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2973    (The getFoo interface is still there for backward-compatibility, but its
2974    use is deprecated.)
2975    
2976    The ml-build script passes -Cxxx=yyy command-line arguments through so
2977    that one can now twiddle the compiler settings when using this "batch"
2978    compiler.
2979    
2980    TODO items:
2981    
2982    We should go through and throw out all controls that are no longer
2983    connected to anything.  Moreover, we should go through and provide
2984    meaningful (and correct!) documentation strings for those controls
2985    that still are connected.
2986    
2987    Currently, multiple calls to Controls.new are accepted (only the first
2988    has any effect).  Eventually we should make sure that every control
2989    is being made (via Controls.new) exactly once.  Future access can then
2990    be done using Controls.acc.
2991    
2992    Finally, it would probably be a good idea to use the getter-setter
2993    interface to controls rather than ref cells.  For the time being, both
2994    styles are provided by the Controls module, but getter-setter pairs are
2995    better if thread-safety is of any concern because they can be wrapped.
2996    
2997    *****************************************
2998    
2999    One bug fix: The function blockPlacement in three of the MLRISC
3000    backpatch files used to be hard-wired to one of two possibilities at
3001    link time (according to the value of the placementFlag).  But (I
3002    think) it should rather sense the flag every time.
3003    
3004    *****************************************
3005    
3006    Other assorted changes (by other people who did not supply a HISTORY entry):
3007    
3008    1. the cross-module inliner now works much better (Monnier)
3009    2. representation of weights, frequencies, and probabilities in MLRISC
3010       changed in preparation of using those for weighted block placement
3011       (Reppy, George)
3012    
3013    ----------------------------------------------------------------------
3014    Name: Lal George
3015    Date: 2002/03/07 14:44:24 EST 2002
3016    Tag: george-20020307-weighted-block-placement
3017    
3018    Tested the weighted block placement optimization on all architectures
3019    (except the hppa) using AMPL to generate the block and edge frequencies.
3020    Changes were required in the machine properties to correctly
3021    categorize trap instructions. There is an MLRISC flag
3022    "weighted-block-placement" that can be used to enable weighted block
3023    placement, but this will be ineffective without block/edge
3024    frequencies (coming soon).
3025    
3026    
3027    ----------------------------------------------------------------------
3028    Name: Lal George
3029    Date: 2002/03/05 17:24:48 EST
3030    Tag: george-20020305-linkage-cluster
3031    
3032    In order to support the block placement optimization, a new cluster
3033    is generated as the very first cluster (called the linkage cluster).
3034    It contains a single jump to the 'real' entry point for the compilation
3035    unit. Block placement has no effect on the linkage cluster itself, but
3036    all the other clusters  have full freedom in the manner in which they
3037    reorder blocks or functions.
3038    
3039    On the x86 the typical linkage code that is generated is:
3040       ----------------------
3041            .align 2
3042       L0:
3043            addl    $L1-L0, 72(%esp)
3044            jmp     L1
3045    
3046    
3047            .align  2
3048       L1:
3049       ----------------------
3050    
3051    72(%esp) is the memory location for the stdlink register. This
3052    must contain the address of the CPS function being called. In the
3053    above example, it contains the address of  L0; before
3054    calling L1 (the real entry point for the compilation unit), it
3055    must contain the address for L1, and hence
3056    
3057            addl $L1-L0, 72(%esp)
3058    
3059    I have tested this on all architectures except the hppa.The increase
3060    in code size is of course negligible
3061    
3062    ----------------------------------------------------------------------
3063    Name: Allen Leung
3064    Date: 2002/03/03 13:20:00 EST
3065    Tag: leunga-20020303-mlrisc-tools
3066    
3067      Added #[ ... ] expressions to mlrisc tools
3068    
3069    ----------------------------------------------------------------------
3070    Name: Matthias Blume
3071    Date: 2002/02/27 12:29:00 EST
3072    Tag: blume-20020227-cdebug
3073    Description:
3074    
3075    - made types in structure C and C_Debug to be equal
3076    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
3077    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
3078    
3079    ----------------------------------------------------------------------
3080    Name: Matthias Blume
3081    Date: 2002/02/26 12:00:00 EST
3082    Tag: blume-20020226-ffi
3083    Description:
3084    
3085    1. Fixed a minor bug in CM's "noweb" tool:
3086       If numbering is turned off, then truly don't number (i.e., do not
3087       supply the -L option to noweb).  The previous behavior was to supply
3088       -L'' -- which caused noweb to use the "default" line numbering scheme.
3089       Thanks to Chris Richards for pointing this out (and supplying the fix).
3090    
3091    2. Once again, I reworked some aspects of the FFI:
3092    
3093       A. The incomplete/complete type business:
3094    
3095       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
3096         gone!
3097       - ML types representing an incomplete type are now *equal* to
3098         ML types representing their corresponding complete types (just like
3099         in C).  This is still safe because ml-nlffigen will not generate
3100         RTTI for incomplete types, nor will it generate functions that
3101         require access to such RTTI.   But when ML code generated from both
3102         incomplete and complete versions of the C type meet, the ML types
3103         are trivially interoperable.
3104    
3105         NOTE:  These changes restore the full generality of the translation
3106         (which was previously lost when I eliminated functorization)!
3107    
3108       B. Enum types:
3109    
3110       - Structure C now has a type constructor "enum" that is similar to
3111         how the "su" constructor works.  However, "enum" is not a phantom
3112         type because each "T enum" has values (and is isomorphic to
3113         MLRep.Signed.int).
3114       - There are generic access operations for enum objects (using
3115         MLRep.Signed.int).
3116       - ml-nlffigen will generate a structure E_foo for each "enum foo".
3117         * The structure contains the definition of type "mlrep" (the ML-side
3118         representation type of the enum).  Normally, mlrep is the same
3119         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3120         then mlrep will be defined as a datatype -- thus facilitating
3121         pattern matching on mlrep values.
3122         ("-ec" will be suppressed if there are duplicate values in an
3123          enumeration.)
3124         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3125         will be generated for each C enum constant xxx.
3126         * Conversion functions m2i and i2m convert between mlrep and
3127         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
3128         * Coversion functions c and ml convert between mlrep and "tag enum".
3129         * Access functions (get/set) fetch and store mlrep values.
3130       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3131         enumerations are merged into one single enumeration represented by
3132         structure E_'.
3133    
3134    ----------------------------------------------------------------------
3135    Name: Allen Leung
3136    Date: 2002/02/25 04:45:00 EST
3137    Tag: leunga-20020225-cps-spill
3138    
3139    This is a new implementation of the CPS spill phase.
3140    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3141    In case of problems, replace it with the old file spill.sml
3142    
3143    The current compiler runs into some serious performance problems when
3144    constructing a large record.  This can happen when we try to compile a
3145    structure with many items.  Even a very simple structure like the following
3146    makes the compiler slow down.
3147    
3148        structure Foo = struct
3149           val x_1 = 0w1 : Word32.int
3150           val x_2 = 0w2 : Word32.int
3151           val x_3 = 0w3 : Word32.int
3152           ...
3153           val x_N = 0wN : Word32.int
3154        end
3155    
3156    The following table shows the compile time, from N=1000 to N=4000,
3157    with the old compiler:
3158    
3159    N
3160    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3161           MLRISC ra                               0.06u  0.00s  0.05g
3162              (spills = 0 reloads = 0)
3163           TOTAL                                   0.63u  0.07s  0.21g
3164    
3165    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3166           MLRISC ra                               5.68u  0.59s  3.93g
3167              (spills = 0 reloads = 0)
3168           TOTAL                                   14.71u  0.99s  4.81g
3169    
3170    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3171           MLRISC ra                               5.54u  0.65s  3.91g
3172              (spills = 543 reloads = 1082)
3173           TOTAL                                   65.40u  3.13s  6.00g
3174    
3175    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3176           MLRISC ra                               0.80u  0.10s  0.55g
3177              (spills = 42 reloads = 84)
3178           TOTAL                                   129.42u  5.10s  4.13g
3179    
3180    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3181           MLRISC ra                               2.69u  0.27s  1.38g
3182              (spills = 62 reloads = 124)
3183           TOTAL                                   682.48u  19.61s  13.99g
3184    
3185    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3186           MLRISC ra                               4.96u  0.27s  2.72g
3187              (spills = 85 reloads = 170)
3188           TOTAL                                   2375.26u  57.21s  48.00g
3189    
3190    As you can see the old cps spill module suffers from some serious
3191    performance problem.  But since I cannot decipher the old code fully,
3192    instead of patching the problems up, I'm reimplementing it
3193    with a different algorithm.  The new code is more modular,
3194    smaller when compiled, and substantially faster
3195    (O(n log n) time and O(n) space).  Timing of the new spill module:
3196    
3197    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3198          MLRISC ra                               0.25u  0.02s  0.15g
3199             (spills=1 reloads=3)
3200          TOTAL                                   7.74u  0.34s  1.62g
3201    
3202    Implementation details:
3203    
3204    As far as I can tell, the purpose of the CPS spill module is to make sure the
3205    number of live variables at any program point (the bandwidth)
3206    does not exceed a certain limit, which is determined by the
3207    size of the spill area.
3208    
3209    When the bandwidth is too large, we decrease the register pressure by
3210    packing live variables into spill records.  How we achieve this is
3211    completely different than what we did in the old code.
3212    
3213    First, there is something about the MLRiscGen code generator
3214    that we should be aware of:
3215    
3216    o MLRiscGen performs code motion!
3217    
3218       In particular, it will move floating point computations and
3219       address computations involving only the heap pointer to
3220       their use sites (if there is only a single use).
3221       What this means is that if we have a CPS record construction
3222       statement
3223    
3224           RECORD(k,vl,w,e)
3225    
3226       we should never count the new record address w as live if w
3227       has only one use (which is often the case).
3228    
3229       We should do something similar to floating point, but the transformation
3230       there is much more complex, so I won't deal with that.
3231    
3232    Secondly, there are now two new cps primops at our disposal:
3233    
3234     1. rawrecord of record_kind option
3235        This pure operator allocates some uninitialized storage from the heap.
3236        There are two forms:
3237    
3238         rawrecord NONE [INT n]  allocates a tagless record of length n
3239         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3240                                     and initializes the tag.
3241    
3242     2. rawupdate of cty
3243          rawupdate cty (v,i,x)
3244          Assigns to x to the ith component of record v.
3245          The storelist is not updated.
3246    
3247    We use these new primops for both spilling and increment record construction.
3248    
3249     1. Spilling.
3250    
3251        This is implemented with a linear scan algorithm (but generalized
3252        to trees).  The algorithm will create a single spill record at the
3253        beginning of the cps function and use rawupdate to spill to it,
3254        and SELECT or SELp to reload from it.  So both spills and reloads
3255        are fine-grain operations.  In contrast, in the old algorithm
3256        "spills" have to be bundled together in records.
3257    
3258        Ideally, we should sink the spill record construction to where
3259        it is needed.  We can even split the spill record into multiple ones
3260        at the places where they are needed.  But CPS is not a good
3261        representation for global code motion, so I'll keep it simple and
3262        am not attempting this.
3263    
3264     2. Incremental record construction (aka record splitting).
3265    
3266        Long records with many component values which are simulatenously live
3267        (recall that single use record addresses are not considered to
3268         be live) are constructed with rawrecord and rawupdate.
3269        We allocate space on the heap with rawrecord first, then gradually
3270        fill it in with rawupdate.  This is the technique suggested to me
3271        by Matthias.
3272    
3273        Some restrictions on when this is applicable:
3274        1. It is not a VECTOR record.  The code generator currently does not handle
3275           this case. VECTOR record uses double indirection like arrays.
3276        2. All the record component values are defined in the same "basic block"
3277           as the record constructor.  This is to prevent speculative
3278           record construction.
3279    
3280    ----------------------------------------------------------------------
3281    Name: Allen Leung
3282    Date: 2002/02/22 01:02:00 EST
3283    Tag: leunga-20020222-mlrisc-tools
3284    
3285    Minor bug fixes in the parser and rewriter
3286    
3287    ----------------------------------------------------------------------
3288    Name: Allen Leung
3289    Date: 2002/02/21 20:20:00 EST
3290    Tag: leunga-20020221-peephole
3291    
3292    Regenerated the peephole files.  Some contained typos in the specification
3293    and some didn't compile because of pretty printing bugs in the old version
3294    of 'nowhere'.
3295    
3296    ----------------------------------------------------------------------
3297    Name: Allen Leung
3298    Date: 2002/02/19 20:20:00 EST
3299    Tag: leunga-20020219-mlrisc-tools
3300    Description:
3301    
3302       Minor bug fixes to the mlrisc-tools library:
3303    
3304       1.  Fixed up parsing colon suffixed keywords
3305       2.  Added the ability to shut the error messages up
3306       3.  Reimplemented the pretty printer and fixed up/improved
3307           the pretty printing of handle and -> types.
3308       4.  Fixed up generation of literal symbols in the nowhere tool.
3309       5.  Added some SML keywords to to sml.sty
3310    
3311    ----------------------------------------------------------------------
3312    Name: Matthias Blume
3313    Date: 2002/02/19 16:20:00 EST
3314    Tag: blume-20020219-cmffi
3315    Description:
3316    
3317    A wild mix of changes, some minor, some major:
3318    
3319    * All C FFI-related libraries are now anchored under $c:
3320        $/c.cm      --> $c/c.cm
3321        $/c-int.cm  --> $c/internals/c-int.cm
3322        $/memory.cm --> $c/memory/memory.cm
3323    
3324    * "make" tool (in CM) now treats its argument pathname slightly
3325      differently:
3326        1. If the native expansion is an absolute name, then before invoking
3327           the "make" command on it, CM will apply OS.Path.mkRelative
3328           (with relativeTo = OS.FileSys.getDir()) to it.
3329        2. The argument will be passed through to subsequent phases of CM
3330           processing without "going native".  In particular, if the argument
3331           was an anchored path, then "make" will not lose track of that anchor.
3332    
3333    * Compiler backends now "know" their respective C calling conventions
3334      instead of having to be told about it by ml-nlffigen.  This relieves
3335      ml-nlffigen from one of its burdens.
3336    
3337    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3338    
3339    * Export C_DEBUG and C_Debug from $c/c.cm.
3340    
3341    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3342      subtyping relationship between incomplete pointers and their complete
3343      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3344      with the convention of instantiating 'o with ('t, 'c) obj whenever
3345      the pointer target type is complete.  In the incomplete case, 'o
3346      will be instantiated with some "'c iobj" -- a type obtained by
3347      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3348    
3349      Operations that work on both incomplete and complete pointer types are
3350      typed as taking an 'o ptr while operations that require the target to
3351      be known are typed as taking some ('t, 'c) obj ptr.
3352    
3353      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3354      where void is an eqtype without any values.  This makes it possible
3355      to work on voidptr values using functions meant to operate on light
3356      incomplete pointers.
3357    
3358    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3359      been vastly simplified.
3360    
3361    ----------------------------------------------------------------------
3362    Name: Matthias Blume
3363    Date: 2002/02/19 10:48:00 EST
3364    Tag: blume-20020219-pqfix
3365    Description:
3366    
3367    Applied Chris Okasaki's bug fix for priority queues.
3368    
3369    ----------------------------------------------------------------------
3370    Name: Matthias Blume
3371    Date: 2002/02/15 17:05:00
3372    Tag: Release_110_39
3373    Description:
3374    
3375    Last-minute retagging is becoming a tradition... :-(
3376    
3377    This is the working release 110.39.
3378    
3379    ----------------------------------------------------------------------
3380    Name: Matthias Blume
3381    Date: 2002/02/15 16:00:00 EST
3382    Tag: Release_110_39-orig
3383    Description:
3384    
3385    Working release 110.39.  New bootfiles.
3386    
3387    (Update: There was a small bug in the installer so it wouldn't work
3388    with all shells.  So I retagged. -Matthias)
3389    
3390    ----------------------------------------------------------------------
3391    Name: Matthias Blume
3392    Date: 2002/02/15 14:17:00 EST
3393    Tag: blume-20020215-showbindings
3394    Description:
3395    
3396    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3397    the latter can be useful for exploring what bindings are available at
3398    the interactive prompt.  (The first function returns only the list
3399    of symbols that are really bound, the second prints those but also the
3400    ones that CM's autoloading mechanism knows about.)
3401    
3402    ----------------------------------------------------------------------
3403    Name: Matthias Blume
3404    Date: 2002/02/15 12:08:00 EST
3405    Tag: blume-20020215-iptrs
3406    Description:
3407    
3408    Two improvements to ml-nlffigen:
3409    
3410      1. Write files only if they do not exist or if their current contents
3411         do not coincide with what's being written.  (That is, avoid messing
3412         with the time stamps unless absolutely necessary.)
3413    
3414      2. Implement a "repository" mechanism for generated files related
3415         to "incomplete pointer types".   See the README file for details.
3416    
3417    ----------------------------------------------------------------------
3418    Name: Matthias Blume
3419    Date: 2002/02/14 11:50:00 EST
3420    Tag: blume-20020214-quote
3421    Description:
3422    
3423    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3424    because of the new and improved tag generation scheme.  (Thanks to Allen
3425    Leung for pointing it out.)
3426    
3427    ----------------------------------------------------------------------
3428    Name: Lal George
3429    Date: 2002/02/14 09:55:27 EST 2002
3430    Tag: george-20020214-isabelle-bug
3431    Description:
3432    
3433    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3434    of Isabelle on the x86.
3435    
3436    From Allen:
3437    -----------
3438     I've found the problem:
3439    
3440         in ra-core.sml, I use the counter "blocked" to keep track of the
3441         true number of elements in the freeze queue.  When the counter goes
3442         to zero, I skip examining the queue.  But I've messed up the
3443         bookkeeping in combine():
3444    
3445             else ();
3446             case !ucol of
3447               PSEUDO => (if !cntv > 0 then
3448                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3449                                        ^^^^^^^^^^^^^^^^^^^^^^^
3450                      moveu := mergeMoveList(!movev, !moveu)
3451                     )
3452                  else ();
3453    
3454         combine() is called to coalesce two nodes u and v.
3455         I think I was thinking that if the move counts of u and v are both
3456         greater than zero then after they are coalesced then one node is
3457         removed from the freeze queue.  Apparently I was thinking that
3458         both u and v are of low degree, but that's clearly not necessarily true.
3459    
3460    
3461    02/12/2002:
3462        Here's the patch.  HOL now compiles.
3463    
3464        I don't know how this impact on performance (compile
3465        time or runtime).  This bug caused the RA (especially on the x86)
3466        to go thru the potential spill phase when there are still nodes on the
3467        freeze queue.
3468    
3469    
3470    
3471    
3472    ----------------------------------------------------------------------
3473    Name: Matthias Blume
3474    Date: 2002/02/13 22:40:00 EST
3475    Tag: blume-20020213-fptr-rtti
3476    Description:
3477    
3478    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3479    updates.
3480    
3481    ----------------------------------------------------------------------
3482    Name: Matthias Blume
3483    Date: 2002/02/13 16:41:00 EST
3484    Tag: blume-20020213-cmlpq
3485    Description:
3486    
3487    Added new priority queue export symbols (which have just been added to
3488    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3489    not compile and the installer would choke.)
3490    
3491    ----------------------------------------------------------------------
3492    Name: Matthias Blume
3493    Date: 2002/02/13 16:15:00 EST
3494    Tag: blume-20020213-various
3495    Description:
3496    
3497    1. More tweaks to ml-nlffigen:
3498    
3499       - better internal datastructures (resulting in slight speedup)
3500       - "-match" option requires exact match
3501       - "localized" gensym counters (untagged structs/unions nested within
3502         other structs/unions or within typedefs get a fresh counter; their
3503         tag will be prefixed by a concatenation of their parents' tags)
3504       - bug fixes (related to calculation of transitive closure of types
3505         to be included in the output)
3506    
3507    2. Minor Basis updates:
3508    
3509       - added implementations for List.collate and Option.app
3510    
3511    ----------------------------------------------------------------------
3512    Name: Matthias Blume
3513    Date: 2002/02/11 15:55:00 EST
3514    Tag: blume-20020211-gensym
3515    Description:
3516    
3517    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3518    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3519    names (ML structure names that correspond to unnamed C structs, unions,
3520    and enums), so that separate runs of ml-nlffigen do not clash.
3521    
3522    ----------------------------------------------------------------------
3523    Name: Matthias Blume
3524    Date: 2002/02/11 12:05:00 EST
3525    Tag: blume-20020211-gensml
3526    Description:
3527    
3528    A quick fix for a problem with GenSML (in the pgraph-util library):
3529    Make generation of toplevel "local" optional.  (Strictly speaking,
3530    signature definitions within "local" are not legal SML.)
3531    
3532    Other than that: updates to INSTALL and cm/TODO.
3533    
3534    ----------------------------------------------------------------------
3535    Name: Matthias Blume
3536    Date: 2002/02/08 15:00:00 EST
3537    Tag: blume-20020208-uniquepid
3538    Description:
3539    
3540    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3541    
3542    1. The installer (config/install.sh) has gotten smarter:
3543    
3544         - Configuration options are a bit easier to specify now
3545           (in config/targets).
3546         - Bug in recognizing .tar.bz2 files fixed.
3547         - Installer automatically resolves dependencies between
3548           configuration options (e.g., if you ask for eXene, you will
3549           also get cml -- regardless whether you asked for it or not).
3550         - Installer can run in "quieter mode" by setting the environment
3551           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3552           "completely silent", though.
3553         - Build HashCons library as part of smlnj-lib.
3554    
3555    2. A new scheme for assigning persistent identifiers to compilation
3556       units (and, by extension, to types etc.) has been put into place.
3557       This fixes a long-standing bug where types and even dynamic values
3558       can get internally confused, thereby compromising type safety
3559       (abstraction) and dynamic correctness.  See
3560    
3561         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3562    
3563       for an example of how things could go wrong until now.
3564    
3565       The downside of the new scheme is that pids are not quite as
3566       persistent as they used to be: CM will generate a fresh pid
3567       for every compilation unit that it thinks it sees for the first
3568       time.  That means that if you compile starting from a clean, fresh
3569       source tree at two different times, you end up with different
3570       binaries.
3571    
3572       Cutoff recompilation, however, has not been compromised because
3573       CM keeps pid information in special caches between runs.
3574    
3575    ----------------------------------------------------------------------
3576    Name: Lal George
3577    Date: 2002/02/07 15:34:13 EST 2002
3578    Tag: <none>
3579    Description:
3580    
3581    Compilers that generate assembly code may produce  global labels
3582    whose value is resolved at link time. The various peephole optimization
3583    modules did not take this in account.
3584    
3585    TODO. The Labels.addrOf function should really return an option
3586    type so that clients are forced to deal with this issue, rather
3587    than an exception being raised.
3588    
3589    ----------------------------------------------------------------------
3590    Name: Lal George
3591    Date: 2002/02/06 13:55:02 EST
3592    Tag: george-20020206-ra-breakup
3593    Description:
3594    
3595    1. A bug fix from Allen.
3596    
3597        A typo causes extra fstp %st(0)'s to be generated at compensation
3598        edges, which might cause stack underflow traps at runtime.  This
3599        occurs in fft where there are extraneous fstps right before the 'into'
3600        trap instruction (in this case they are harmless since none of the
3601        integers overflow.)
3602    
3603    2. Pulled out various utility modules that were embedded in the modules
3604       of the register allocator. I need these modules for other purposes, but
3605       they are not complete enough to put into a library (just yet).
3606    ----------------------------------------------------------------------
3607    Name: Matthias Blume
3608    Date: 2002/01/31 16:05:00 EST
3609    Tag: blume-20020131-sparc-ccalls
3610    Description:
3611    
3612    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3613       of extra stack space by mistake.  Fixed.
3614    
3615    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3616    
3617    ----------------------------------------------------------------------
3618    Name: Allen Leung
3619    Date: 2002/01/30
3620    Tag: leunga-20020130-nowhere-bug-fix
3621    Description:
3622    
3623       MLRISC bug fixes:
3624       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3625       2. MachineInt.fromString was negating its value.
3626    
3627    ----------------------------------------------------------------------
3628    Name: Matthias Blume
3629    Date: 2002/01/29
3630    Tag: blume-20020129-INSTALL
3631    Description:
3632    
3633    - Added somewhat detailed installation instructions (file INSTALL).
3634    - Fixed curl-detection bug in config/install.sh.
3635    - It is now possible to select the URL getter using the URLGETTER
3636      environment variable:
3637    
3638          not set / "unknown"      --> automatic detection (script tries wget,
3639                                       curl, and lynx)
3640          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3641                                       how to properly invoke them)
3642          other                    --> use $URLGETTER directly, it must take
3643                                       precisely two command-line arguments
3644                                       (source URL and destination file name)
3645    
3646    ----------------------------------------------------------------------
3647    Name: Matthias Blume
3648    Date: 2002/01/28
3649    Tag: blume-20020128-sparc-ccalls
3650    Description:
3651    
3652    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3653    - Make use of the allocParam argument in sparc-c-calls.
3654    
3655    ----------------------------------------------------------------------
3656    Name: Matthias Blume
3657    Date: 2002/01/28
3658    Tag: blume-20020128-allocParam
3659    Description:
3660    
3661    John Reppy:  Changes c-calls API to accept client-callback for
3662    allocating extra stack space.
3663    me: Corresponding changes to mlriscGen (using a dummy argument that
3664        does not change the current behavior).
3665    
3666    ----------------------------------------------------------------------
3667    Name: Matthias Blume
3668    Date: 2002/01/28 12:00:00
3669    Tag: Release_110_38
3670    Description:
3671    
3672    This time for real!!!
3673    
3674    ----------------------------------------------------------------------
3675    Name: Matthias Blume
3676    Date: 2002/01/28 10:56:00 EST
3677    Tag: blume-20020128-retraction
3678    Description:
3679    
3680    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3681       with blume-Release_110_38-retracted.)
3682    
3683    1. Fixed a problem with incorrect rounding modes in real64.sml.
3684       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3685    
3686    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3687       and enums fixed.  The naming of corresponding ML identifiers should
3688       now be consistent again.
3689    
3690    ----------------------------------------------------------------------
3691    Name: Allen Leung
3692    Date: 2002/01/27
3693    Tag: leunga-20020127-nowhere
3694    Description:
3695    
3696       Added a target called nowhere in the configuration scripts.
3697       Enabling this will build the MLRISC 'nowhere' tool (for translating
3698       programs with where-clauses into legal SML code) during installation.
3699    
3700    ----------------------------------------------------------------------
3701    Name: Matthias Blume
3702    Date: 2002/01/25 21:27:00 EST
3703    Tag: blume-Release_110_38-retracted
3704    Description:
3705    
3706    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3707    
3708    README will be added later.
3709    
3710    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3711    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3712    
3713    ----------------------------------------------------------------------
3714    Name: Matthias Blume
3715    Date: 2002/01/25
3716    Tag: blume-20020125-ffi
3717    Description:
3718    
3719    A large number of tweaks and improvements to ml-nlffi-lib and
3720    ml-nlffigen:
3721    
3722       - ML represenation types have been streamlined
3723       - getter and setter functions work with concrete values, not abstract
3724         ones where possible
3725       - ml-nlffigen command line more flexible (see README file there)
3726       - some bugs have been fixed (hopefully)
3727    
3728    ----------------------------------------------------------------------
3729    Name: Lal George
3730    Date: 2002/01/24
3731    Tag: george-20020124-risc-ra-interface
3732    Description:
3733    
3734       There is a dramatic simplification in the interface to the
3735       register allocator for RISC architectures as a result of making
3736       parallel copy instructions explicit.
3737    
3738    ----------------------------------------------------------------------
3739    Name: Matthias Blume
3740    Date: 2002/01/22
3741    Tag: blume-20020122-x86-ccalls
3742    Description:
3743    
3744    Bug fix for c-calls on x86 (having to do with how char- and
3745    short-arguments are being handled).
3746    
3747    ----------------------------------------------------------------------
3748    Name: Matthias Blume
3749    Date: 2002/01/21
3750    Tag: blume-20020121-ff
3751    Description:
3752    
3753    Another day of fiddling with the FFI...
3754    
3755    1. Bug fix/workaround:  CKIT does not complain about negative array
3756       dimensions, so ml-nlffigen has to guard itself against this possibility.
3757       (Otherwise a negative dimension would send it into an infinite loop.)
3758    
3759    2. Some of the abstract types (light objects, light pointers, most "base"
3760       types) in structure C are now eqtypes.
3761    
3762    3. Added constructors and test functions for NULL function pointers.
3763    
3764    ----------------------------------------------------------------------
3765    Name: Matthias Blume
3766    Date: 2002/01/18
3767    Tag: blume-20020118-ready-for-new-release
3768    Description:
3769    
3770    Made config/srcarchiveurl point to a new place.  (Will provide boot
3771    files shortly.)
3772    
3773    Maybe we christen this to be 110.38?
3774    
3775    ----------------------------------------------------------------------
3776    Name: Matthias Blume
3777    Date: 2002/01/18
3778    Tag: blume-20020118-more-ffifiddle
3779    Description:
3780    
3781    Today's FFI fiddling:
3782    
3783      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3784        C.Set.* that use concrete (MLRep.*) arguments and results instead
3785        of abstract ones.
3786    
3787      - Provided word-style bit operations etc. for "int" representation
3788        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3789        and Long.
3790    
3791    ----------------------------------------------------------------------
3792    Name: Matthias Blume
3793    Date: 2002/01/18
3794    Tag: blume-20020118-use-x86-fp
3795    Description:
3796    
3797    Now that x86-fast-fp seems to be working, I turned it back on again
3798    by default.  (Seems to work fine now, even with the FFI.)
3799    
3800    Other than that, I added some documentation about the FFI to
3801    src/ml-nlffigen/README and updated the FFI test examples in
3802    src/ml-nlffi-lib/Tests/*.
3803    
3804    ----------------------------------------------------------------------
3805    Name: Allen Leung
3806    Date: 2002/01/17
3807    Tag: leunga-20020117-x86-fast-fp-call
3808    Description:
3809    
3810       1. Fixed a problem with handling return fp values when x86's fast fp
3811          mode is turned on.
3812    
3813       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3814          of %f32.
3815    
3816       3. Added a constructor INT32lit to the ast of MLRISC tools.
3817    
3818    ----------------------------------------------------------------------
3819    Name: Matthias Blume
3820    Date: 2002/01/16
3821    Tag: blume-20020116-ffifiddle
3822    Description:
3823    
3824    More fiddling with the FFI interface:
3825    
3826     - Make constness 'c instead of rw wherever possible.  This eliminates
3827       the need for certain explicit coercions.  (However, due to ML's
3828       value polymorphism, there will still be many cases where explicit
3829       coercions are necessary.  Phantom types are not the whole answer
3830       to modeling a subtyping relationship in ML.)
3831    
3832     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3833    
3834     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3835       clumsy workaround for varargs problem.)
3836    
3837    ----------------------------------------------------------------------
3838    Name: Lal George
3839    Date: 2002/01/15
3840    Tag: <none>
3841    Description:
3842    
3843    1. Since COPY instructions are no longer native to the architecture,
3844       a generic functor can be used to implement the expandCopies function.
3845    
3846    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3847       TEXT segment.
3848    
3849    ----------------------------------------------------------------------
3850    Name: Matthias Blume
3851    Date: 2002/01/15
3852    Tag: blume-20020115-ffiupdates
3853    Description:
3854    
3855    1. Fix for