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