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