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 1044, Mon Jan 28 21:36:08 2002 UTC revision 1724, Mon Dec 13 20:45:51 2004 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/12/13 14:45:00 CST
18    Tag: blume-20041213-ml-makedepend
19    Description:
20    
21    Fixed bug in code for ml-makedepend.
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume (blume (at) tti - c (dot) org)
25    Date: 2004/12/09 16:30:00 CST
26    Tag: blume-20041209-statistics
27    Description:
28    
29    Added two simple but potentially useful statistics modules to SML/NJ Library.
30    (See CHANGES file there.)
31    
32    ----------------------------------------------------------------------
33    Name: Matthias Blume (blume (at) tti - c (dot) org)
34    Date: 2004/12/01 16:50:00 CST
35    Tag: blume-20041201-atom
36    Description:
37    
38    smlnj-lib:
39    
40      Added function HashString.hashString' for substrings.
41      Hand-inlined CharVector.foldl into HashString (for speed).
42      Modified implementation of structure Atom to avoid extracting
43      strings from substrings unless necessary.
44      (Also see CHANGES file for smlnj-lib.)
45    
46    ----------------------------------------------------------------------
47    Name: Matthias Blume (blume (at) tti - c (dot) org)
48    Date: 2004/11/24 22:15:00 CST
49    Tag: blume-20041124-cml
50    Description:
51    
52    Made sure CML compiles when Position = Int64.
53    
54    ----------------------------------------------------------------------
55    Name: Matthias Blume (blume (at) tti - c (dot) org)
56    Date: 2004/11/24 14:30:00 CST
57    Tag: blume-20041124-position
58    Description:
59    
60    The compiler can now be compiled in a mode that makes structure Position
61    equal to Int64.  The default, however, is unchanged (Position = Int31)
62    for the time being.
63    
64    To enable 64-bit positions, use the following procedure:
65    
66    1. Start sml
67    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
68    3. Type
69          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
70    4. Run CMB.make() as usual.
71    
72    This is barely tested.  The only test so far was a little SML program
73    counting the number of characters in an 8-gigabyte file by
74    reading it character-by-character.  That test was successful.
75    
76    In support of 64-bit positions, a number of new functions have been
77    added to the runtime system.
78    
79    ----------------------------------------------------------------------
80    Name: Matthias Blume (blume (at) tti - c (dot) org)
81    Date: 2004/11/23 14:45:00 CST
82    Tag: blume-20041123-useFile
83    Description:
84    
85    Fixed a problem with unhelpful error messages related to problems with
86    .cm- or .sml files that appear as part of the sml command line.
87    
88    ----------------------------------------------------------------------
89    Name: Matthias Blume (blume (at) tti - c (dot) org)
90    Date: 2004/11/18 15:40:00 CST
91    Tag: Release_110_51
92    Description:
93    
94    New working version (110.51).  NEW BOOTFILES!
95    
96    ----------------------------------------------------------------------
97    Name: Matthias Blume (blume (at) tti - c (dot) org)
98    Date: 2004/11/18 15:35:00 CST
99    Tag: <none>
100    Description:
101    
102    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
103    Joemann).
104    
105    ----------------------------------------------------------------------
106    Name: Allen Leung (leunga (at) reservoir (dot) com)
107    Date: 2004/11/17 16:05:21 EST 2004
108    Tag: leunga-20041117-mlrisc-live-kill
109    Description:
110    
111       Added support for MLTree constructs LIVE and KILL
112    to all the architectures.
113    
114    ----------------------------------------------------------------------
115    Name: Matthias Blume (blume (at) tti - c (dot) org)
116    Date: 2004/11/13 00:20:00 CST
117    Tag: blume-20041113-versiontool
118    Description:
119    
120    - Stripped down the versiontool: It now only handles the version number.
121      The date string is generated at bootstrap time (during makeml).
122    
123    - In a previous commit, fixed a minor issue with how polyequal is being
124      translated.  In particular, the code now "looks through" abstractions.
125      This results in slightly fewer polyEqual warnings and hopefully slightly
126      more efficient code.  Important examples for where this matters are
127      the new int64 and word64 types.
128    
129    ----------------------------------------------------------------------
130    Name: Matthias Blume (blume (at) tti - c (dot) org)
131    Date: 2004/11/12 00:30:00 CST
132    Tag: blume-20041112-int64
133    Description:
134    
135    Structure Int64 fully hooked in.  (The implementation is not very
136    efficient, though.)
137    
138    ----------------------------------------------------------------------
139    Name: Matthias Blume (blume (at) tti - c (dot) org)
140    Date: 2004/11/11 17:30:00 CST
141    Tag: blume-20041111-more64
142    Description:
143    
144    All the pieces of Word64 are now there, with the exception of the
145    conversions from and to LargeWord. (Eventually these need to be identities,
146    but for the time being they don't even make sense because LargeWord is
147    32-bit wide.)
148    
149    Also started to add similar support for Int64, but major pieces of that
150    are still missing.
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume (blume (at) tti - c (dot) org)
154    Date: 2004/11/11 00:15:00 CST
155    Tag: blume-20041111-word64
156    Description:
157    
158    Structure Word64 is now (almost) complete, word literals and patterns
159    seem to work.  There are a few odd pieces missing.  In particular,
160    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
161    at the moment.
162    
163    Making Word64 official would mean that LargeWord becomes Word64.  But
164    this requires extreme care because most word-word conversions have to
165    go through LargeWord, so making a mistake means loss of efficiency or
166    worse.  Eventually there will be a solution similar to (but actually
167    simpler than) what I did with IntInf.
168    
169    ----------------------------------------------------------------------
170    Name: Matthias Blume (blume (at) tti - c (dot) org)
171    Date: 2004/11/10 18:12:00 CST
172    Tag: blume-20041110-64bit
173    Description:
174    
175    More 64-bit hacking (but still not even half-way there yet).
176    Also, some assorted improvements to the handling of 8-bit words.
177    
178    ----------------------------------------------------------------------
179    Name: Matthias Blume (blume (at) tti - c (dot) org)
180    Date: 2004/11/09 17:50:00 CST
181    Tag: <none>
182    Description:
183    
184    Started adding some infrastructure for supporting 64-bit int- and
185    word-types.  (Still in its very early stages.)
186    
187    ----------------------------------------------------------------------
188    Name: Matthias Blume (blume (at) tti - c (dot) org)
189    Date: 2004/10/28 10:45:00 CDT
190    Tag: Release_110_50
191    Description:
192    
193    New working version (110.50).  NEW BOOTFILES!
194    
195                 =====================
196    
197    Also:
198    
199      - Changend config/srcarchiveurl from a file just
200        containing the URL string into a file containing
201        shell script code.  The code has access to the $VERSION variable.
202      - Made corresponding changes to config/install.sh and config/unpack.
203      - Default contents of config/srcarchiveurl uses $VERSION and
204        normally does not have to be edited to reflect a version change.
205    
206      (As a result, a version change can be done by just editing
207       config/version, the rest is now automatic.)
208    
209    ----------------------------------------------------------------------
210    Name: Matthias Blume (blume (at) tti - c (dot) org)
211    Date: 2004/10/27 17:50:00 CDT
212    Tag: blume-20041027-btrace-msg
213    Description:
214    
215    BackTrace.monitor now also reports the source of the exception that
216    triggered the trace.
217    
218    ----------------------------------------------------------------------
219    Name: Matthias Blume (blume (at) tti - c (dot) org)
220    Date: 2004/10/27 17:20:00 CDT
221    Tag: blume-20041027-x86-c-calls
222    Description:
223    
224    This is the HISTORY entry for two earlier commits, both concerning
225    the x86 c-calls code in MLRISC:
226    
227      - added a missing LOAD in the code that deals with struct arguments
228      - made sure the caller does not add the wrong number of bytes to the
229        stack pointer after a call of a function returning a struct
230        (the callee already pops the implicit argument which points to
231         the space reserved for the result)
232    
233    ----------------------------------------------------------------------
234    Name: Allen Leung (leunga (at) reservoir (dot) com)
235    Date: 2004/10/24 14:00:00 EST
236    Tag: leunga-20041024-x86-gas-fucomip
237    Description:
238    
239       John discovered a bug in the syntax of fucomip.
240       The opcodes FU?COMIP? have been changed to
241    
242           fu?comip? %st(i), %st
243    
244    ----------------------------------------------------------------------
245    Name: Matthias Blume (blume (at) tti - c (dot) org)
246    Date: 2004/10/20 15:06:00 CDT
247    Tag: blume-20041020-standalone-backtrace
248    Description:
249    
250    Added a mechanism for getting back-trace information from standalone
251    programs.  Here is how it works:
252    
253       1. The part of the program from which you want to get backtrace
254          information (usually the whole program) should be wrapped with
255          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
256          main program could be modified from something like
257    
258             fun main (pgm, args) = ...
259    
260         to
261    
262            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
263    
264       2. To be able to access BackTrace.monitor, you have to add
265          library $smlnj-tdp/plugins.cm to the .cm file that contains your
266          main function.
267    
268       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
269          CM might have created in the past for your project).
270    
271       4. Build the system using this command line:
272    
273             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
274                      myprog.cm MyProg.main myprog
275    
276          instead of the usual
277    
278             ml-build myprog.cm MyProg.main myprog
279    
280    I changed a library name:
281    
282       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
283    
284    New libraries:
285    
286       $smlnj-tdp/back-trace.cm
287          -- when loaded causes the back-trace plugin to be installed
288       $smlnj-tdp/coverage.cm
289          -- when loaded causes the coverage plugin to be installed
290    
291    ----------------------------------------------------------------------
292    Name: Matthias Blume (blume (at) tti - c (dot) org)
293    Date: 2004/10/18 16:45:00 CDT
294    Tag: blume-20041018-groupowner
295    Description:
296    
297    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
298    
299    Eliminated group owner specs from .cm files throughout the source tree.
300    
301    ----------------------------------------------------------------------
302    Name: Matthias Blume (blume (at) tti - c (dot) org)
303    Date: 2004/10/15 15:45:00 CDT
304    Tag: blume-20041015-coverage
305    Description:
306    
307    * Test coverage tool added!
308    * Further reorganization of tracing-, debugging-, and profiling support:
309    
310        - moved original BTImp -- now called BackTrace -- into a separate
311          library called $/trace-debug-profile.cm
312        - eliminated all mentions of BTrace from SMLofNJ.Internals
313        - only the instrumentation mechanism is now left in the compiler proper
314        - BackTrace module is a plugin which is NOT plugged in by default
315        - Coverage module is another such plugin
316    
317    To get the benefits of any of these plugin modules, the code in
318    question must be compiled with tdp instrumentation turned on.  This
319    can be done by setting SMLofNJ.Internals.TDP.mode to true.
320    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
321    
322    Plugins are selected at link time.  (Pre-compiled instrumented code
323    can be re-loaded with different plugins in effect.)  When an
324    instrumented module is linked, whatever plugins are at that time
325    enabled will come into effect for that module.
326    
327    To enable the back-trace plugin, load library $/trace-debug-profile.cm
328    and invoke BackTrace.install() (e.g., from the interactive prompt).
329    To enable the coverage plugin, load the same library and invoke
330    Coverage.install().
331    
332    Back-traces are generated automatically on uncaught exceptions and
333    when the code in question explicitly invokes BackTrace.trigger().
334    
335    Coverage (and execution frequency-) information must be queried
336    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
337    
338    ----------------------------------------------------------------------
339    Name: Matthias Blume (blume (at) tti - c (dot) org)
340    Date: 2004/10/14 17:40:00 CDT
341    Tag: blume-20041014-tdp-core
342    Description:
343    
344    Snapshot of a significant overhaul of how the trace/debug/profile support
345    is hooked into the system (specifically: Core and SMLofNJ.Internals).
346    
347    ----------------------------------------------------------------------
348    Name: Matthias Blume (blume (at) tti - c (dot) org)
349    Date: 2004/10/13 16:34:00 CDT
350    Tag: blume-20041013-tdp
351    Description:
352    
353    Some rationalization of names:
354    
355       structure BTrace -> structure TDPInstrument
356       etc.
357    
358    This is is preparation of using the original back-trace
359    instrumentation for other purposes.  "TDP" stands for
360    Trace/Debug/Profile.
361    
362    The control flag controlling whether instrumentation is on or off is now
363    registered under a different name, so instead of running sml as
364    
365     sml -Cinstrument.btrace-mode=true
366    
367    one has to say
368    
369     sml -Ctdp.instrument=true
370    
371    ----------------------------------------------------------------------
372    Name: Matthias Blume (blume (at) tti - c (dot) org)
373    Date: 2004/10/11 16:37:00 CDT
374    Tag: blume-20041011-regions
375    Description:
376    
377    Made some minor modifications to elabcore.sml to have source regions
378    be propagated more tightly -- resulting in better (i.e., smaller)
379    regions being reported in error- and debug messages.
380    
381    ----------------------------------------------------------------------
382    Name: Matthias Blume (blume (at) tti - c (dot) org)
383    Date: 2004/10/08 22:50:00 CDT
384    Tag: blume-20041008-cmkw
385    Description:
386    
387    Fixed handling of keywords in .cm files: After seeing "is" the lexer
388    treats subsequent occurrences of "group", "library", "source", "is",
389    "*", and "-" as ordinary identifiers rather than keywords.
390    
391    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
392    is supposed to accept a tool argument called "source", but this did
393    not work because of the clash with the keyword.
394    
395    ----------------------------------------------------------------------
396    Name: Matthias Blume (blume (at) tti - c (dot) org)
397    Date: 2004/10/07 16:00:00 CDT
398    Tag: blume-20041007-cleanup
399    Description:
400    
401    Assorted cleanup work:
402    
403      - got rid of intstrmap in favor of using the library's
404        hash table implementation
405      - threw out most of the pathnames stuff, as it was not used anyway
406      - simplified tokentable implementation
407      - fixed some minor spelling errors
408    
409    ----------------------------------------------------------------------
410    Name: Matthias Blume (blume (at) tti - c (dot) org)
411    Date: 2004/10/06 15:15:15 CDT
412    Tag: blume-20041006-handler
413    Description:
414    
415    Cleaned up the absyn to reflect the invariant that HANDLE always
416    carries a FNexp as part of the type definition.  This eliminates some
417    superfluous sanity checks at runtime down the road.
418    
419    Some minor cleanup of the btrace code.
420    
421    ----------------------------------------------------------------------
422    Name: Matthias Blume (blume (at) tti - c (dot) org)
423    Date: 2004/10/01 10:20:30 CDT
424    Tag: blume-20041001-slave
425    Description:
426    
427    Added hack to make slave mode work in the presence of the version
428    tool.  (Still, since the master does two passes over the code for
429    CMB.make, the release number gets bumped twice when slaves are
430    attached.  I don't know if this is worth fixing...)
431    
432    ----------------------------------------------------------------------
433    Name: Matthias Blume (blume (at) tti - c (dot) org)
434    Date: 2004/09/30 10:55:00 CDT
435    Tag: blume-20040930-version
436    Description:
437    
438    * Moved the "version" magic into its own little library under
439      src/system/smlnj/internal.  This avoids expensive reconstruction of
440      a stable src/compiler/core.cm.
441    
442    * At the same time, structure CompilerVersion is now known as
443      structure SMLNJVersion.
444    
445    * Arranged for the version tool to NOT kick in when rebuilding the system
446      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
447      Also, loading the versiontool does not work when rebuilding the system
448      because CM is not properly initialized at that time.
449    
450    ----------------------------------------------------------------------
451    Name: Matthias Blume (blume (at) tti - c (dot) org)
452    Date: 2004/09/29 14:00:00 CDT
453    Tag: blume-20040929-autoversion
454    Description:
455    
456    Implemented some CM magic to have
457        file src/compiler/TopLevel/main/version.sml
458    generated automagically.
459    The version is taken from two files: config/version and config/release.
460    The first is expected to contain a two-part version number such as 110.49.
461    The second should contain a single number, but it may be missing.
462    
463    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
464    time the version tool is loaded (which is the first time you say CMB.make),
465    then the tool will increment the value stored in config/release every
466    time CMB.make is invoked.
467    
468    The binfile format is now insensitive to anything beyond the first
469    two components of a version number, so bumping the release does not render
470    binfiles incompatible.  Auto-bumping can be used to keep track of versions
471    during development without invalidating existing binfiles.
472    
473    In any case, every CMB.make updates the date information in version.sml.
474    (This is the date that is printed in the banner.)
475    
476    ----------------------------------------------------------------------
477    Name: Matthias Blume (blume (at) tti - c (dot) org)
478    Date: 2004/09/28 10:53:00 CDT
479    Tag: blume-20040928-controls
480    Description:
481    
482    Some cleanup of the controls code.
483    
484    ----------------------------------------------------------------------
485    Name: Matthias Blume (blume (at) tti - c (dot) org)
486    Date: 2004/09/27 22:08:00 CDT
487    Tag: blume-20040927-controls
488    Description:
489    
490    Added two pieces of functionality to the Controls interface:
491    
492    1.   val save'restore: 'a control -> unit -> unit
493    
494       grabs the current value of the control in stage 1 and restores it
495       in stage 2.
496    
497    2.   val set' : 'a control * 'a -> unit -> unit
498    
499       stores the given value into the control in stage 2 (i.e., delayed)
500       but does all error checking in stage 1.
501       (This is for string controls that need to do parse their argument --
502        something that might fail.  In some cases, notably in CM, one
503        already knows the intended argument but wants to delay the actual
504        assignment until a time when error recovery would be more difficult.)
505    
506    Changed the handling of controls in tool arguments to classes "sml" and
507    "lazysml":
508       - use Controls.save'restore as a more robust way of restoring the
509         old value (in particular: without having to re-parse the string)
510       - use controls to handle the "overload" keyword in the init group
511         (I believe this change actually fixes a long-standing obscure bug.)
512    
513    ----------------------------------------------------------------------
514    Name: Matthias Blume (blume (at) tti - c (dot) org)
515    Date: 2004/09/27 17:00:00 CDT
516    Tag: blume-20040927-lazysml
517    Description:
518    
519    Added a new tool class called "lazysml" to CM's tool chest.  The only
520    difference to "sml" is that compilation is done with Control.lazysml
521    set to true.  A source of class "lazysml" is automatically recognized
522    by a file name suffix of ".lml".
523    
524    In addition to the above feature, the original class "sml" now also
525    supports a tool argument "lazy" which has the same effect.  As a
526    result, the following three lines are equivalent:
527    
528        foo.sml : lazysml
529        foo.sml : sml (lazy)
530        foo.sml (lazy)
531    
532    The setting goes into effect both during parsing and during
533    compilation.  The original setting is restored right after parsing and
534    after compilation, respectively.
535    
536    In addition to all the above, there is also a general mechanism to set
537    ANY of the "controls" that are available at the command line via
538    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
539    apply as well. (In fact, "lazy" is implemented as a special case of
540    the general mechanism.)
541    
542    The .cm file syntax uses a new keyword tool argument called "with".
543    There are several ways of indicating the desired settings:
544    
545        foo.sml (with:parser.quotations=true)
546        foo.sml (with:(name:parser.quotations value:true))
547        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
548        foo.sml (with:(name1=value1 name2=value2 ...))
549        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
550    
551        etc.
552    
553    Another possible abbreviation is to leave out the =v or value:v part
554    if the name refers to a boolean control (in which case the value is
555    taken to be true).  Thus, one could get lazy sml also by saying:
556    
557        foo.sml (with:parser.lazy-keyword=true)
558        foo.sml (with:parser.lazy-keyword)
559        foo.sml (with:(name:parser.lazy-keyword value:true))
560        foo.sml (with:(name:parser.lazy-keyword))
561    
562    ----------------------------------------------------------------------
563    Name: Matthias Blume (blume (at) tti - c (dot) org)
564    Date: 2004/09/24 16:22:00 CDT
565    Tag: blume-20040924-ppc-long-branch
566    Description:
567    
568    Turned message about "emiting long form of branch" off by default.
569    Added a control flag to turn it back on when desired.
570    
571    ----------------------------------------------------------------------
572    Name: Matthias Blume (blume (at) tti - c (dot) org)
573    Date: 2004/09/24 16:05:00 CDT
574    Tag: blume-20040924-rounding
575    Description:
576    
577    Applied patch for setting rounding modes under Mac OS X.  Thanks to
578    Melissa O'Neill for providing the code!
579    
580    ----------------------------------------------------------------------
581    Name: Matthias Blume (blume (at) tti - c (dot) org)
582    Date: 2004/09/23 17:30:00 CDT
583    Tag: blume-20040923-envvars
584    Description:
585    
586    1. Changed definition of type ControlRegistry.registry_tree to
587       include control_info (i.e., the name of the controlling environment
588       variable).
589    
590    2. Added command-line flags -e and -E to print the names of environment
591       variables that can be used to control internal settings.
592       (This uses the new API mentioned in 1.)
593    
594    ----------------------------------------------------------------------
595    Name: Matthias Blume (blume (at) tti - c (dot) org)
596    Date: 2004/09/13 16:50:00 CDT
597    Tag: Release_110_49
598    Description:
599    
600    New working version (110.49).  NEW BOOTFILES!
601    
602    ----------------------------------------------------------------------
603    Name: Matthias Blume
604    Date: 2004/09/13 16:20:00 CDT
605    Tag: blume-20040913-config-mlrisc
606    Description:
607    
608    Put target "mlrisc" back into the default list.
609    (There is no harm in having it, and some users have expressed their
610    wish to have "mlrisc" included by default.)
611    
612    ----------------------------------------------------------------------
613    Name: John Reppy
614    Date: 2004/09/13
615    Tag: jhr-20040913-signals
616    Description:
617    
618    Fixed the signal masking code to properly nest mask/unmask operations
619    on a per-signal basis.
620    
621    ----------------------------------------------------------------------
622    Name: Matthias Blume (blume (at) tti - c (dot) org)
623    Date: 2004/09/08 13:20:00 CDT
624    Tag: blume-20040908-heap-magic
625    Description:
626    
627    Bumped the heap macig number to 0x09082004 to account for the changed
628    layout of the ML frame under MacOS X.
629    
630    ----------------------------------------------------------------------
631    Name: Allen Leung (leunga (at) reservoir (dot) com)
632    Date: 2004/09/03 11:26:00 EST
633    Tag: leunga-20040903-cygwin-install
634    Description:
635    
636       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
637    The Cygwin runtime is turned on by setting the environment
638    variable SMLNJ_CYGWIN_RUNTIME to 1.
639    
640    ----------------------------------------------------------------------
641    Name: Matthias Blume (blume (at) tti - c (dot) org)
642    Date: 2004/08/31 17:15:00 CDT
643    Tag: blume-20040831-core
644    Description:
645    
646    Added some exports to src/compiler/core.cm upon request by J. Joemann.
647    
648    ----------------------------------------------------------------------
649    Name: Matthias Blume (blume (at) tti - c (dot) org)
650    Date: 2004/08/30 17:55:00 CDT
651    Tag: blume-20040830-installer
652    Description:
653    
654    Upon request by Johannes Joemann:
655    
656    - improved ML code of installer to fall back to coping when renaming
657      fails (i.e., when source and target are on different file systems);
658      the code compiles but has yet to be tested in anger
659    - removed mlrisc from list of default targets (config/targets)
660    
661    ----------------------------------------------------------------------
662    Name: Matthias Blume (blume (at) tti - c (dot) org)
663    Date: 2004/08/27 17:20:00 CDT
664    Tag: blume-20040827-ptreql
665    Description:
666    
667    Added ptreql primop to structure InlineT (upon request from Larry
668    Paulson).
669    
670    ----------------------------------------------------------------------
671    Name: Allen Leung (leunga (at) reservoir (dot) com)
672    Date: 2004/08/15 21:21:00 EST
673    Tag: leunga-110_48-udgraph
674    Description:
675    
676    Another bug fix from Carl Hauser:
677    
678    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
679    > 48c48
680    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
681    > e::L)
682    > ---
683    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
684    > Without this, any deletion of an edge in an undirected graph does severe
685    > violence to the graph.
686    
687    ----------------------------------------------------------------------
688    Name: Allen Leung (leunga (at) reservoir (dot) com)
689    Date: 2004/08/10 23:35:00 EST
690    Tag: leunga-110_48-ppc
691    Description:
692    
693       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
694    
695    ----------------------------------------------------------------------
696    Name: Matthias Blume (blume (at) tti - c (dot) org)
697    Date: 2004/08/10 12:00:00 CDT
698    Tag: Release_110_48
699    Description:
700    
701    New working version (110.48).  NEW BOOTFILES!
702    
703    ----------------------------------------------------------------------
704    Name: Allen Leung (leunga (at) reservoir (dot) com)
705    Date: 2004/08/09 12:21:00 EST
706    Tag: leunga-110_47-dijsktra
707    Description:
708    
709    Bug fix from Carl Hauser:
710    
711    single_source_shortest_paths in dijkstra.sml was observed to get wrong
712    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
713    
714    The problem is that following the expression A.update(dist,s,Num.zero)
715    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
716    
717    ----------------------------------------------------------------------
718    Name: Matthias Blume (blume (at) tti - c (dot) org)
719    Date: 2004/08/06 18:10:00 CDT
720    Tag: blume-20040806-cmdline
721    Description:
722    
723    Fiddled with handling of command-line options:
724    
725      * sml now quits after processing the command line
726        if -H, -S, -h<n>, or -s<n> appears as the last
727        command-line argument
728      * a new option -q terminates the session when encountered on
729        the command line; subsequent arguments will be ignored
730      * bug fixes: short (erroneous) arguments are no longer ignored
731        completely
732    
733    ----------------------------------------------------------------------
734    Name: Allen Leung (leunga (at) reservoir (dot) com)
735    Date: 2004/08/04 18:17:00 EST
736    Tag: leunga-110_47-ppc-ibm-asm
737    Description:
738    
739      - Added minimal IBM assembly syntax support for PowerPC.
740    
741      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
742        been altered in the runtime.  We need an automatic way to keep the file
743        in sync.
744    
745    ----------------------------------------------------------------------
746    Name: Matthias Blume (blume (at) tti - c (dot) org)
747    Date: 2004/08/04 14:00:00 CDT
748    Tag: Release_110_47
749    Description:
750    
751    New working version (110.47).  NEW BOOTFILES!
752    
753    ----------------------------------------------------------------------
754    Name: Matthias Blume (blume (at) tti - c (dot) org)
755    Date: 2004/08/03 14:25:00 CDT
756    Tag: blume-20040803-callingconv
757    Description:
758    
759    Added low-level support for choosing C calling conventions by
760    twiddling the type of rawccall.  (See
761    src/compiler/Semant/types/cproto.sml for details.)
762    
763    ----------------------------------------------------------------------
764    Name: Matthias Blume (blume (at) tti - c (dot) org)
765    Date: 2004/08/02 15:55:00 CDT
766    Tag: blume-20040802-backout
767    Description:
768    
769    Backed out of change to win32-filesys.c.  The earlier patch to
770    get_file_time caused CM to produce files with the wrong time stamp.
771    
772    ----------------------------------------------------------------------
773    Name: Matthias Blume (blume (at) tti - c (dot) org)
774    Date: 2004/08/02 14:45:00 CDT
775    Tag: blume-20040802-nlffi-win32
776    Description:
777    
778    Added NLFFI support for Win32, adapted from a patch provided by David
779    Hansel.  This is currently completely untested.  Also, the issue
780    concerning stdcall vs. ccall is still unresolved.
781    
782    ----------------------------------------------------------------------
783    Name: Matthias Blume (blume (at) tti - c (dot) org)
784    Date: 2004/07/30 17:55:00 CDT
785    Tag: blume-20040730-various
786    Description:
787    
788    Gearing up towards 110.47...
789    
790    - various minor bugfixes to ml-nlffigen
791    - a beginning of a manual for nlffi
792    
793    - eliminated 'export name=value' in config/install.sh as this does
794      not work with certain versions of /bin/sh
795      (Thanks to David King at Motorola for catching this.)
796    
797    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
798      - added a test for tm==NULL to gmtime.c and localtime.c
799      - applied patch for incorrect GetFileTime under win32
800      - toSeconds -> toMilliseconds in Win32/win32-process.sml
801    
802    ----------------------------------------------------------------------
803    Name: Matthias Blume (blume (at) tti - c (dot) org)
804    Date: 2004/07/21 18:20:00 CDT
805    Tag: blume-20040721-nlffigen
806    Description:
807    
808    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
809      for a typedef to an incomplete type, but leave out the "typ" member.
810      (This is just for consistency.)
811    - Started to produce what is supposed to become better (i.e., comprehensive)
812      documentation of what ml-nlffigen does and produces.
813    
814    ----------------------------------------------------------------------
815    Name: Matthias Blume (blume (at) tti - c (dot) org)
816    Date: 2004/07/14 16:25:00 CDT
817    Tag: blume-20040714-union
818    Description:
819    
820    Added C_UNION to c-calls/c-types.sml and updated the machinery
821    (ml-nlffigen, cproto.sml) that conveys C function interface
822    information to the code generator.
823    
824    However, the actual architecture-specific implementation of function
825    arguments and results that are C unions is still not implemented.
826    
827    ----------------------------------------------------------------------
828    Name: Allen Leung (leunga (at) reservoir (dot) com)
829    Date: 2004/07/14 14:38:00 EST
830    Tag: leunga-110_46_1-ppc-lwzu
831    Description:
832    
833        Added these instructions to the PowerPC architecture:
834        LBZU(X), LHZU(X), LWZU(X),
835        STWU(X), STFDU, STFSU
836    
837        etc...
838    
839        Note: I haven't added their instruction encoding into the description.
840    
841    ----------------------------------------------------------------------
842    Name: Allen Leung (leunga (at) reservoir (dot) com)
843    Date: 2004/07/13 15:04:00 EST
844    Tag: leunga-110_46_1-ppc-lwarx
845    Description:
846    
847       Added the two instructions LWARX and STWCX to the PowerPC
848    instruction set.
849    
850       A (untested) rewrite of loop-structure.sml.  The old version
851    is completely broken.
852    
853    ----------------------------------------------------------------------
854    Name: Matthias Blume (blume (at) tti - c (dot) org)
855    Date: 2004/07/13 13:50:00 CDT
856    Tag: blume-20040713-nlffi
857    Description:
858    
859    - use paramAlloc to report c-calls with too many arguments
860      (for PPC version where parameter area is pre-allocated)
861    - added ccall_maxargspace to machspec (to implement the above)
862    - made "make" commend in CM's "make" tool configurable
863    - added option (default: on) for passing the name of the SML/NJ's "bin"
864      directory to "make"; the call looks like this:
865    
866         make <options> SMLNJ_BINDIR=<dir> <target>
867    
868      This can be used by the Makefile to, e.g., pick the "right" version
869      of ml-nlffigen.
870    - minor code tweaks
871    
872    ----------------------------------------------------------------------
873    Name: Matthias Blume (blume (at) tti - c (dot) org)
874    Date: 2004/07/12 22:50:00 CDT
875    Tag: blume-110_46_1-macosx-nlffi
876    Description:
877    
878    NLFFI under Mac OS X now working (sort of).  This is largely untested,
879    though.
880    
881    Note:  1. You have to make a new, clean build of the runtime system.
882           2. There are new BOOTFILES, you have to use them!
883              (Doing the bootstrap process yourself would be *very* painful!
884               If you absolutely have to do it, build the system under
885               a different architecture and then cross-compile.)
886    
887    Version bumped to 110.46.1 to account for runtime data format changes.
888    
889    ----------------------------------------------------------------------
890    Name: Matthias Blume (blume (at) tti - c (dot) org)
891    Date: 2004/06/18 14:30:00 CDT
892    Tag: blume-20040618-unix
893    Description:
894    
895    Changed the implementation of structure Unix so that the same stream
896    is returned every time one of the {text,bin}{In,Out}streamOf functions
897    is invoked on the same proc.  This is not what the spec currently
898    says -- although IMO it arguably should.  (See discussion below.)
899    
900    ----------------------------------------------------------------------
901    Name: Matthias Blume (blume (at) tti - c (dot) org)
902    Date: 2004/06/17 18:15:00 CDT
903    Tag: Release_110_46
904    Description:
905    
906    New working version (110.46).  NEW BOOTFILES!
907    
908    ----------------------------------------------------------------------
909    Name: Matthias Blume (blume (at) tti - c (dot) org)
910    Date: 2004/06/17 17:20:00 CDT
911    Tag: blume-20040617-timer-unix
912    Description:
913    
914    Changed the interface of structures Timer and Unix to match the most
915    recent Basis spec.
916    
917    In the case of Unix there still seems to be an open/weird issue:
918    
919       The {text,bin}{In,Out}streamOf functions are supposed to create
920       fresh streams whenever they are called -- as opposed to have them
921       return the same stream every time.  This design is supposed to
922       prevent space leaks caused by proc values hanging on to streams.
923    
924       The reap function, on the other hand, is supposed to close the
925       streams.  This cannot be done without having a handle on the
926       stream in proc after all...
927    
928    I took the liberty to implement the following stopgap solution:
929    
930      The proc value hangs on to the most recently created stream(s).
931      Reap closes those.  If either or both of the two streams hadn't
932      been created at all yet, then reap will close the corresponding
933      file descriptors directly.
934    
935    PS: I don't understand the original space leak argument anymore.  If
936    a proc hangs on to the imperative stream, then I/O operations on those
937    will advance the state of the cached stream and avoid the space leak.
938    
939    ----------------------------------------------------------------------
940    Name: Matthias Blume (blume (at) tti - c (dot) org)
941    Date: 2004/05/28 16:45:00 CDT
942    Tag: blume-20040528-basis
943    Description:
944    
945    Added signature PACK_REAL and exported functor PrimIO.
946    
947    ----------------------------------------------------------------------
948    Name: Matthias Blume (blume (at) tti - c (dot) org)
949    Date: 2004/05/25 16:00:00 CDT
950    Tag: blume-20040525-group-owner
951    Description:
952    
953    CM now ignores (but still accepts) the "owner" information in group
954    descriptions.  The owner of a group is its next enclosing
955    library. Each group must have a unique owner.  (There is a virtual
956    "toplevel" library that owns groups which are not nested within a real
957    library.)  Previously, each group had to explicitly declare its owner,
958    and CM would check that such a declaration is correct.  The new scheme
959    is to have CM check that for each group there is precisely one owning
960    library.
961    
962    The advantage of the new scheme is that the programmer no longer needs
963    to maintain the somewhat annoying owner information.  The downside is
964    that CM cannot enforce the ownership rule across multiple runs of
965    CM.make.  Fortunately, enclosing the same group in two different
966    libraries A and B which are not part of the same program does not
967    cause real problems.
968    
969    ----------------------------------------------------------------------
970    Name: Matthias Blume (blume (at) tti - c (dot) org)
971    Date: 2004/05/20 16:00:00 CDT
972    Tag: blume-20040520-win32
973    Description:
974    
975    Made the win32 version work again.  (Strangely, a misplaced comma had
976    slipped into win32-process.c which prevented the runtime from being
977    compiled correctly.)
978    
979    Also, included a minor addition to ml-build.bat analogous to what was
980    done in blume-20040519-ml-build.
981    
982    ----------------------------------------------------------------------
983    Name: Matthias Blume (blume (at) tti - c (dot) org)
984    Date: 2004/05/19 22:10:00 CDT
985    Tag: blume-20040519-ml-build
986    Description:
987    
988    Arranged for ml-build to clean up after itself a little bit better.
989    The script generates a temporary SML source file and compiles it using
990    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
991    gets rid of those at the end, so they don't accumulate under .cm.
992    
993    This required a minor change to install.sh because the name of the
994    metadata directory (default: .cm) is actually configurable at
995    installation time.
996    
997    ----------------------------------------------------------------------
998    Name: Matthias Blume (blume (at) tti - c (dot) org)
999    Date: 2004/05/18 15:50:00 CDT
1000    Tag: blume-20040518-mkreader
1001    Description:
1002    
1003    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1004    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1005    
1006    ----------------------------------------------------------------------
1007    Name: Matthias Blume (blume (at) tti - c (dot) org)
1008    Date: 2004/05/11 14:35:00 CDT
1009    Tag: blume-20040511-win32sock
1010    Description:
1011    
1012    Added previously missing support for many socket-related functions
1013    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1014    for the voluminous patch!
1015    
1016    (I have not tested this patch under win32 yet.)
1017    
1018    Here is David's e-mail:
1019    
1020    Hi,
1021    
1022    Attached to this email you find a diff against sml/nj 110.45
1023    that will enable socket support under Windows.
1024    
1025    To apply the patch (using unix or cygwin)
1026    1) gunzip runtime.diff.gz
1027    2) "cd" into "src/runtime" in the source tree of a fresh
1028       110.45 installation.
1029    3) patch -p 1 < [your/path/to]runtime.diff
1030    
1031    The code compiles fine but has NOT yet been extensively tested.
1032    I only ran a few tests for basic socket client functionality
1033    (which worked fine).  Especially the functions that use ioctl
1034    are not tested at all and might not work (see below).
1035    
1036    I implemented this since we want to move to a newer version of sml/nj
1037    but need socket support in order to use it.  This is the first time I
1038    even had a look at the sml/nj source,  so please review my changes
1039    before making this part of the distribution!  Here are a few issues
1040    that I think might be better for someone to solve who is more
1041    familiar with the sml/nj source (and socket programming):
1042    
1043    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1044      exception since I could not figure out what the windows equivalent
1045      of these functions is
1046    
1047    - In sockets-osdep.h there are a some #include statements that are
1048      only used in a few files that include sockets-osdep.h
1049    
1050    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1051      does not process its return value since I don't know how to
1052      report an error upwards.
1053    
1054    - It would probably be good to have a call to WSACleanup() when
1055      the library is unloaded (if there is such a possibility).
1056      Otherwise I think Windows will take care of this automatically
1057      when the process finishes.
1058    
1059    - I used ioctlsocket() as a replacement for ioctl() but I have
1060      no idea if that is actually the proper replacement on Windows.
1061    
1062    - All these issues are marked in the code by "FIXME" comments.
1063    
1064    We use sml/nj extensively in our products and are quite happy
1065    with it.  I hope this contribution will help you.
1066    
1067    Keep up the good work!
1068    
1069    David
1070    
1071    ----------------------------------------------------------------------
1072    Name: Matthias Blume (blume (at) tti - c (dot) org)
1073    Date: 2004/05/11 14:20:00 CDT
1074    Tag: blume-20040511-installml
1075    Description:
1076    
1077    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1078    the patch.)
1079    
1080    ----------------------------------------------------------------------
1081    Name: Matthias Blume (blume (at) tti - c (dot) org)
1082    Date: 2004/05/11 14:05:00 CDT
1083    Tag: blume-20040511-nlffi-netbsd
1084    Description:
1085    
1086    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1087    the patch.)
1088    
1089    ----------------------------------------------------------------------
1090    Name: Matthias Blume (blume (at) tti - c (dot) org)
1091    Date: 2004/05/11 12:05:00 CDT
1092    Tag: blume-20040511-exports
1093    Description:
1094    
1095    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1096    export lists in compiler-related .cm-files.
1097    
1098    ----------------------------------------------------------------------
1099    Name: Matthias Blume (blume (at) tti - c (dot) org)
1100    Date: 2004/05/11 11:35:00 CDT
1101    Tag: blume-20040511-allsource
1102    Description:
1103    
1104    The installer now honors the "src-smlnj" target again, although its meaning
1105    has changed from "all sources required for the compiler" to "all sources
1106    the installer knows about".  In other words, if you enable "src-smlnj"
1107    in the "targets" file, then the installer will pull in sources for
1108    everything.  (Notice that this refers to source code only.  Compiled
1109    code is still only installed for modules that were requested explicitly
1110    or which are required for other modules that were requested explicitly.)
1111    
1112    ----------------------------------------------------------------------
1113    Name: Matthias Blume (blume (at) tti - c (dot) org)
1114    Date: 2004/04/23 17:40:00 CDT
1115    Tag: blume-20040423-ieee-scan
1116    Description:
1117    
1118    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1119    in the exponent calculation we get INF or ZERO (depending on the mantissa
1120    and the sign of the exponent).
1121    
1122    ----------------------------------------------------------------------
1123    Name: Matthias Blume (blume (at) tti - c (dot) org)
1124    Date: 2004/04/23 10:40:00 CDT
1125    Tag: blume-20040423-ml-build
1126    Description:
1127    
1128    The ml-build script now terminates with a non-0 status when something
1129    goes wrong.
1130    
1131    ----------------------------------------------------------------------
1132    Name: Matthias Blume (blume (at) tti - c (dot) org)
1133    Date: 2004/04/22 16:35:00 CDT
1134    Tag: blume-20040422-Option
1135    Description:
1136    
1137    Made exception Option to be the same as exception Option.Option
1138    (as it should be).
1139    
1140    ----------------------------------------------------------------------
1141    Name: Allen Leung (leunga (at) reservoir (dot) com)
1142    Date: 2004/03/19 14:40:00 EST
1143    Tag: leunga-20040319-cygwin-nlffi
1144    Description:
1145    
1146    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1147    of SML/NJ.  The problem is that
1148    
1149       lib = dlopen(NULL, ...)
1150       f   = dlsym(lib, "malloc");
1151    
1152    does not work on Windows unless we explicitly export symbols
1153    such as 'malloc' during linking.  We fixed this by explicitly
1154    exporting the required symbols with the magic gcc incantation:
1155    
1156        -Wl,--export-all cygwin.def
1157    
1158    where cygwin.def is a file containing all the symbols that
1159    we wish to export.
1160    
1161    I suspect this is a Windows problem and we'll have to
1162    do the same (somehow with windows compilers) when
1163    we build the native win32 version with the system
1164    calls LoadLibrary/GetProcAddress.
1165    
1166    ----------------------------------------------------------------------
1167    Name: Matthias Blume (blume (at) tti - c (dot) org)
1168    Date: 2004/03/04 16:35:00 CST
1169    Tag: blume-20040304-intinf-fmt
1170    Description:
1171    
1172    Fixed problem with IntInf.fmt (sign would show up on the right instead
1173    of on the left for BIN, OCT, and HEX).
1174    
1175    ----------------------------------------------------------------------
1176    Name: Matthias Blume (blume (at) tti - c (dot) org)
1177    Date: 2004/03/04 11:25:00 CST
1178    Tag: blume-20040304-symlinks
1179    Description:
1180    
1181    Fixed problem with installer script (unix only) where bin/ml-yacc and
1182    friends pointed (via symlinks) to absolute locations instead of just
1183    .run-sml.  This was reported by Vesa A Norrman.
1184    
1185    ----------------------------------------------------------------------
1186    Name: Matthias Blume (blume (at) tti - c (dot) org)
1187    Date: 2004/02/13 14:50:00 CST
1188    Tag: Release_110_45
1189    Description:
1190    
1191    New working version (110.45). New bootfiles.
1192    
1193    ----------------------------------------------------------------------
1194    Name: Matthias Blume (blume (at) tti - c (dot) org)
1195    Date: 2004/01/26 15:15:15 CST
1196    Tag: blume-20040126-toplevel
1197    Description:
1198    
1199    Improved handling of exceptions at the interactive toplevel.
1200    
1201    ----------------------------------------------------------------------
1202    Name: Matthias Blume (blume (at) tti - c (dot) org)
1203    Date: 2004/01/26 11:25:00
1204    Tag: blume-20040126-app
1205    Description:
1206    
1207    Type of top-level "app" corrected.
1208    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1209    
1210    ----------------------------------------------------------------------
1211    Name: Matthias Blume (blume (at) tti - c (dot) org)
1212    Date: 2003/11/18 17:10 CST
1213    Tag: blume-20031118-basis-fiddle
1214    Description:
1215    
1216    - changed Timer interface to what might become the spec
1217    - POSIX_FLAGS -> BIT_FLAGS according to spec
1218    - some other minor discrepancies wrt. spec eliminated
1219    
1220    ----------------------------------------------------------------------
1221    Name: Matthias Blume (blume (at) tti - c (dot) org)
1222    Date: 2003/11/06 12:00:00 CST
1223    Tag: Release_110_44
1224    Description:
1225    
1226    New working version (110.44). New bootfiles.
1227    
1228    ----------------------------------------------------------------------
1229    Name: Matthias Blume (blume (at) tti - c (dot) org)
1230    Date: 2003/11/04 11:50:00 CST
1231    Tag: blume-20031104-move-libraries
1232    Description:
1233    
1234    Eliminated the "dont_move_libraries" directive in config/targets.
1235    (The mechanism was broken and could not be fixed easily.  Moreover,
1236    there does not seem to be any reason not to move all libraries into
1237    lib during installation.  I originally implemented this directive as a
1238    backward-compatibility feature when I first introduced the new CM.
1239    Now that things have been stable for a long time and going back to the
1240    old CM is not an option, there is no reason to keep it around.)
1241    
1242    ----------------------------------------------------------------------
1243    Name: Matthias Blume (blume (at) tti - c (dot) org)
1244    Date: 2003/11/03 16:00:00 CST
1245    Tag: blume-20031103-installdir
1246    Description:
1247    
1248    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1249    Richards for pointing out the problem and providing the solution.)
1250    
1251    ----------------------------------------------------------------------
1252    Name: Matthias Blume (blume (at) tti - c (dot) org)
1253    Date: 2003/10/01 17:05:00 CDT
1254    Tag: blume-20031001-lal-mlrisc
1255    Description:
1256    
1257    MLRISC bug fix from Lal.
1258    
1259    ----------------------------------------------------------------------
1260    Name: Matthias Blume (blume (at) tti - c (dot) org)
1261    Date: 2003/09/30 16:10:00 CDT
1262    Tag: blume-20030930-primio-bat
1263    Description:
1264    
1265    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1266    2. Improved .bat files (for Win32 port) to make things work under Win95.
1267       (thanks to Aaron S. Hawley for this one)
1268    
1269    ----------------------------------------------------------------------
1270    Name: Matthias Blume (blume (at) tti - c (dot) org)
1271    Date: 2003/09/26 16:05:00 CDT
1272    Tag: blume-20030926-wrappriv
1273    Description:
1274    
1275    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1276    
1277    ----------------------------------------------------------------------
1278    Name: Matthias Blume (blume (at) tti - c (dot) org)
1279    Date: 2003/09/26 15:00:00 CDT
1280    Tag: blume-20030926-110_43_3
1281    Description:
1282    
1283    - additional cleanup
1284    - version number bump, NEW BOOTFILES
1285    
1286    ----------------------------------------------------------------------
1287    Name: Matthias Blume (blume (at) tti - c (dot) org)
1288    Date: 2003/09/26 12:00:00 CDT
1289    Tag: blume-20030926-ppautoload
1290    Description:
1291    
1292    I modified the read-eval-print loop so that the autoloader gets
1293    invoked whenever the prettyprinter tries to look up a symbol that
1294    is not currently defined in the toplevel environment but which
1295    appears in CM's autoload registry.  As a result, we see far fewer of
1296    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1297    
1298    In addition to this I tried to clean up some pieces of the Basis
1299    implementation (e.g., Socket, Word8Array) in order to prevent other
1300    instances of these ?.Foo.Bar.xxx names from being printed.
1301    
1302    The mechanism that picks names for types still needs some work, though.
1303    (Right now it seems that if there is a type A.t which is defined to
1304    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1305    "?.B.u" although the perhaps more sensible solution would be to use
1306    "A.t" in this case.  In other words, the prettyprinter should follow
1307    a chain of DEFtycs not farther than there are corresponding toplevel
1308    names in the current environment.)
1309    
1310    ----------------------------------------------------------------------
1311    Name: Matthias Blume (blume (at) tti - c (dot) org)
1312    Date: 2003/09/24 16:31:00 CDT
1313    Tag: blume-20030924-installer
1314    Description:
1315    
1316    Another installer tweak:  All the ML code for the installer is now
1317    compiled during CMB.make and put into a little library called
1318    $smlnj/installer.cm.  The installation then simply invokes
1319    
1320       sml -m $smlnj/installer.cm
1321    
1322    and everything happens automagically.
1323    
1324    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1325    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1326            and BINDIR.
1327    
1328    The new scheme guarantees that the ML code responsible for the installation
1329    is in sync with the APIs of the main system.  Also, the installer is
1330    somewhat faster because the installer script is precompiled.
1331    
1332    ----------------------------------------------------------------------
1333    Name: Matthias Blume (blume (at) tti - c (dot) org)
1334    Date: 2003/09/24 15:35:00 CDT
1335    Tag: blume-20030924-synsock
1336    Description:
1337    
1338    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1339    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1340    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1341    implementations of non-blocking operations from CML's Socket
1342    structure.
1343    
1344    ----------------------------------------------------------------------
1345    Name: Matthias Blume (blume (at) tti - c (dot) org)
1346    Date: 2003/09/24 15:10:05 CDT
1347    Tag: blume-20030924-sockets
1348    Description:
1349    
1350    1. Fixed SOCKET API and implementation to match Basis spec.
1351       This required changing the internal representation of sockets to one
1352       that remembers (for each socket file descriptor) whether it is currently
1353       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1354       call is made only if the state actually needs to change).
1355    
1356    2. OS-specific details of sockets were moved into separate files, thus
1357       making it possible to unify the bulk of the socket implementations
1358       between Unix and Win32.
1359    
1360    3. CML's socket API changed accordingly.
1361       (Note that we need to remove non-blocking functions from this API
1362       since they are redundant in the case of CML!)
1363    
1364    4. CML's socket implementation now makes use of non-blocking functions
1365       provided by Basis, thus removing all OS-dependent code from this part
1366       of CML.
1367    
1368    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1369    
1370    ----------------------------------------------------------------------
1371    Name: Matthias Blume (blume (at) tti - c (dot) org)
1372    Date: 2003/09/22 12:10:00 CDT
1373    Tag: blume-20030922-110_43_2
1374    Description:
1375    
1376    Made a new interim version and bootfiles for developer's bootstrapping
1377    convenience.
1378    
1379       110.43.2 -- NEW BOOTFILES
1380    
1381    ----------------------------------------------------------------------
1382    Name: Matthias Blume (blume (at) tti - c (dot) org)
1383    Date: 2003/09/19 15:55:00 CDT
1384    Tag: blume-20030919-cmdir
1385    Description:
1386    
1387    1. new-install.sh -> install.sh
1388    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1389    3. tweaked installer so that another name instead of .cm can be chosen
1390       at install time (by setting the CM_DIR_ARC environment variable
1391       during installation); once installation is complete, the name is
1392       fixed
1393    
1394    ----------------------------------------------------------------------
1395    Name: Matthias Blume (blume (at) tti - c (dot) org)
1396    Date: 2003/09/18 16:00:00 CDT
1397    Tag: blume-20030918-110_43_1
1398    Description:
1399    
1400    Made a new interim version and bootfiles for developer's bootstrapping
1401    convenience.
1402    
1403       110.43.1 -- NEW BOOTFILES
1404    
1405    ----------------------------------------------------------------------
1406    Name: Matthias Blume (blume (at) tti - c (dot) org)
1407    Date: 2003/09/18 15:20:00 CDT
1408    Tag: blume-20030918-misc
1409    Description:
1410    
1411    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1412       this seems to be controversial at the moment) and used those in
1413       Posix.ProcEnv.times.
1414    
1415    2. Added Time.{from,to}Nanoseconds to Time.
1416    
1417    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1418       For example, fromLargeInt never needs to look at more than 3 "big
1419       digits" to get its 53 bits of precision.
1420    
1421    ----------------------------------------------------------------------
1422    Name: Matthias Blume (blume (at) tti - c (dot) org)
1423    Date: 2003/09/17 16:30:00 CDT
1424    Tag: blume-20030917-real32-slices
1425    Description:
1426    
1427    Added an entry to the primitive environment
1428    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1429    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1430    
1431    Removed some of the "magic" constants in real64.sml and replaced them
1432    with code that generates these values from their corresponding
1433    integer counterparts.
1434    
1435    Made all(?) the slice-related changes to the Basis and made everything
1436    compile again...
1437    
1438    ----------------------------------------------------------------------
1439    Name: Matthias Blume (blume (at) tti - c (dot) org)
1440    Date: 2003/09/15 17:45:00 CDT
1441    Tag: blume-20030915-rbase
1442    Description:
1443    
1444    Fixed bug in Real.fromLargeInt.
1445    
1446    ----------------------------------------------------------------------
1447    Name: Matthias Blume (blume (at) tti - c (dot) org)
1448    Date: 2003/09/13 18:11:00 CDT
1449    Tag: blume-20030913-libinstall
1450    Description:
1451    
1452    Minor bugfix in config/libinstall (set anchor with path to
1453    standalone tool after installing it, otherwise libraries that
1454    need ml-lex or ml-yacc won't compile the first time the installer
1455    runs).
1456    
1457    ----------------------------------------------------------------------
1458    Name: Matthias Blume (blume (at) tti - c (dot) org)
1459    Date: 2003/09/12 11:45:00 CDT
1460    Tag: blume-20030912-various
1461    Description:
1462    
1463     - fixed bug in Real.toLargeInt
1464     - fixed bug in Posix.ProcEnv.times
1465     - changed inputLine functions to return an option
1466     - minor installer improvements / bugfixes
1467     - changed default @SMLalloc parameter for x86/celeron to 64k
1468    
1469    ----------------------------------------------------------------------
1470    Name: Matthias Blume (blume (at) tti - c (dot) org)
1471    Date: 2003/09/09 22:00:00 CDT
1472    Tag: Release_110_43
1473    Description:
1474    
1475        New working release 110.43.  New bootfiles.
1476    
1477    ----------------------------------------------------------------------
1478    Name: Matthias Blume (blume (at) tti - c (dot) org)
1479    Date: 2003/09/09 19:20:00 CDT
1480    Tag: blume-20030909-installer
1481    Description:
1482    
1483    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1484    Modified config/install.bat to take advantage of it.  Also modified
1485    config/install.sh (and called it config/new-install.sh) to take advantage
1486    of it on Unix systems.  (The SML code is (supposed to be) platform-
1487    independent.)
1488    
1489    The installer can now install everything under Win32
1490    as well as under *nix as long as it compiles.
1491    
1492    Other changes:
1493    
1494      - made CML compile again under Win32
1495      - made eXene compile under Win32 (by providing a fake structure UnixSock
1496        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1497      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1498        same as type int;  under Win32 it isn't
1499      - fixed some slice-related problems in the win32-specific parts of CML
1500      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1501        (under Win32, the volume name is case-insensitive, and the
1502        OS.Path code compares volume names for equality)
1503    
1504    ----------------------------------------------------------------------
1505    Name: Matthias Blume (blume (at) tti - c (dot) org)
1506    Date: 2003/09/08 11:55:00 CDT
1507    Tag: blume-20030908-fullpath
1508    Description:
1509    
1510    Made Win32 version of OS.FileSys.fullPath return current directory
1511    when given an empty string.  This is what the spec says, and incidentally,
1512    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1513    cases when presented with the name of a non-existing .cm file.)
1514    
1515    ----------------------------------------------------------------------
1516    Name: Matthias Blume (blume (at) tti - c (dot) org)
1517    Date: 2003/09/04 16:30:00 CDT
1518    Tag: blume-20030905-slices-etc
1519    Description:
1520    
1521      1. Changed interface to vectors and arrays in Basis to match
1522         (draft) Basis spec.
1523      2. Added signatures and implementations of slices according to
1524         Basis spec.
1525      3. Edited source code throughout the system to make it compile again
1526         under 1. and 2.  (In some cases code had to be added to have it
1527         match the new signatures.)
1528      4. MLRISC should be backward-compatible: the copies of the originals
1529         of files that needed to change under 3. were retained, the .cm files
1530         check the compiler version number and use old versions when
1531         appropriate.
1532      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1533         dirstream -> string option (in accordance with Basis spec).
1534      6. When generating code that counts lines, ml-lex used function
1535         CharVector.foldli, taking advantage of its old interface.
1536         This has been replaced with the corresponding code from
1537         CharVectorSlice.  (html-lex must be re-lexed!)
1538      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1539         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1540      8. Removed temporary additions (fromInternal, toInternal) from the
1541         (now obsolete) IntInf in smlnj-lib/Util.
1542      9. Cleaned up structure Byte.
1543     10. Added localOffset, scan, and fromString to Date (according to spec).
1544         Cleaned/corrected implementation of Date.
1545         (Still need to check for correctness; implement better canonicalizeDate.)
1546     11. Added "scan" to signature IEEE_REAL.
1547     12. Some improvements to IntInf [in particular: efficiency-hack for
1548         mod and rem when second operand is 2 (for parity checks).]
1549     13. Changed representation of type Time.time, using a single IntInf.int
1550         value counting microseconds.  This considerably simplified the
1551         implementation of structure Time.  We now support negative time
1552         values; scan and fromString handle signs.
1553     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1554         ArraySlice).
1555    
1556    ----------------------------------------------------------------------
1557    Name: Matthias Blume (blume (at) tti - c (dot) org)
1558    Date: 2003/08/28 17:00:00 CDT
1559    Tag: blume-20030828-intinf
1560    Description:
1561    
1562    This is a major update which comes with a version number bump
1563    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1564    and an implementation of IntInf in the Basis.
1565    
1566    There are a fairly large number of related changes and updates throughout
1567    the system:
1568    
1569    Basis:
1570            - Implemented IntInf.
1571            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1572            - Added some missing Real64 operations, most notably Real.toLargeInt.
1573            - Added FixedInt as a synonym for Int32.
1574    
1575    compiler:
1576            * Added support for a built-in intinf type.
1577              - literals
1578              - pattern matching
1579              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1580              - overloading on literals and operations
1581    
1582            This required adding a primitive type intinf, some additional
1583            primops, and implementations for several non-trivial intinf
1584            operations in Core.  (The intinf type is completely abstract
1585            to the compiler; all operations get delegated back to the Core.)
1586    
1587            * Intinf equality is handled by polyequal.  However, the compiler
1588            does not print its usual warning in this case (since polyequal
1589            is the right thing to do there).
1590    
1591            * Improved the organization of structure InlineT.
1592    
1593            * A word about conversion primops:
1594            If conversions involving intinf do not cancel out during
1595            CPS contract, then the compiler must insert calls to Core functions.
1596            Since all core access must be resolved already during the FLINT
1597            translate phase, it would be too late a the time of CPS contract
1598            to add new Core calls.  For this reason, conversion primops
1599            for intinf carry two arguments:  1. the numeric argument that
1600            they are supposed to convert, and 2. the Core function that
1601            can help with this conversion if necessary.  If CPS contract
1602            eliminates a primop, then the associated Core function becomes
1603            dead and goes away.  Intinf conversion primops that do not get
1604            eliminated by CPS contract get rewritten into calls of their
1605            core functions by a separate, new phase.
1606    
1607    interactive system:
1608            - Control.Print.intinfDepth controls max length of intinf constants
1609            being printed. (Analogous to Control.Print.stringDepth.)
1610            - Cleanup in printutil and pputil: got rid of unused stuff and
1611            duplicates; replaced some of the code with code that makes better
1612            use of library functionality.
1613    
1614    CM:
1615            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1616            are no longer silent.
1617    
1618    CKIT:
1619            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1620            in favor of LargeInt -- which is now the same as IntInf.
1621            CKIT-knowledgable people should check whether this is what's
1622            intended and otherwise change things back to using Int32 or
1623            FixedInt.
1624    
1625    Throughout the code:
1626            Started using IntInf.int literals and built-in operations
1627            (e.g., comparison with 0) where this seems appropriate.
1628    
1629    
1630    ----------------------------------------------------------------------
1631    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1632    Date: 2003/08/13 11:36:00 CDT
1633    Tag: dbm-20030813-mcz-merge1
1634    Description:
1635    
1636    Merging changes from the mcz-branch development branch into trunk.
1637    These changes involve replacement of the emulated old prettyprinter
1638    interface with direct use of the SML/NJ Lib PP library, and fixing
1639    of a couple of bugs (895, 1186) relating to error messages.  A new
1640    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1641    has been added.
1642    
1643    ----------------------------------------------------------------------
1644    Name: Matthias Blume (blume (at) tti - c (dot) org)
1645    Date: 2003/08/11 15:45:00 CDT
1646    Tag: blume-20030811-windows
1647    Description:
1648    
1649    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1650    
1651        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1652    
1653    This patch restores SML/NJ's ability to run under win32.  There are a
1654    number of changes, including fixes for several bugs that had gone
1655    unnoticed until now:
1656    
1657      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1658        trying to run the win32 version from within cygwin.)
1659    
1660      - There are a number of simple .bat scripts that substitute for their
1661        corresponding Unix shell-scripts.  (See below.)
1662    
1663      - The internals of ml-build have been modified slightly.  The main
1664        difference is that instead of calling ".link-sml" (or link-sml.bat)
1665        using OS.Process.system, the ML process delegates this task back
1666        to the script.  Otherwise problems arise in mixed environments such
1667        as Cygwin where scripts look and work like Unix scripts, but
1668        where OS.Process.system cannot run them.
1669    
1670      - In CM, the srcpath pickler used native pathname syntax -- which
1671        is incorrect in the case of cross-compilation.  The new pickle format
1672        is independent of platform-specific naming conventions.
1673    
1674      - Path configuration files (such as lib/pathconfig) can now choose
1675        between native and standard syntax.  Placing a line of the form
1676    
1677            standard!
1678    
1679        into the file causes all subsequent paths to be interpreted using
1680        CM standard pathname syntax (= Unix conventions); a line
1681    
1682            native!
1683    
1684        switches back to native style.  This was needed so that
1685        path config files can be written portably, see src/system/pathconfig.
1686    
1687      - Runtime system:
1688    
1689          - win32-filesys.c:  get_file_time and set_file_time now
1690            access modification time, not creation time.
1691    
1692          - I/O code made aware of new array representation.
1693    
1694          - Bug fixes in X86.prim.masm.
1695    
1696       - src/system/makeml made aware of win32. (For use under cygwin
1697            and other Unix-environments for windows.)
1698    
1699       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1700         which caused BinIO.inputAll to fail consistently.
1701    
1702    .bat scripts:
1703    
1704       Windows .bat scripts assume that SMLNJ_HOME is defined.
1705    
1706       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1707         applications (sml, ml-yacc, ml-lex).
1708       - ml-build.bat: analogous to ml-build.
1709       - config\install.bat:  Analogous to config/install.sh.  This requires
1710         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1711         (nmake etc. must be on the path, and vcvars32 must have been run.)
1712         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1713         and the bootfile hierarchy must have been unpacked under
1714         sml.boot.x86-win32.
1715         The script is very primitive and does a poor job at error checking.
1716         It only installs the base system, ml-lex, and ml-yacc.  No other
1717         libraries are being installed (i.e., you get only those that
1718         are part of the compiler.)
1719       - link-sml.bat:  analogous to .link-sml, but not currently used
1720    
1721    Unrelated bug fixes:
1722    
1723      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1724        types.
1725      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1726        otherwise.)
1727    
1728    ----------------------------------------------------------------------
1729    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1730    Date: 2003/06/17
1731    Tag: macqueen-20030617-bug895
1732    Description:
1733    
1734    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1735    Tag will be used for new development branch (mcz-branch) for
1736    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1737    
1738    ----------------------------------------------------------------------
1739    Name: Matthias Blume (blume (at) tti - c (dot) org)
1740    Date: 2003/05/27 16:55:00 CDT
1741    Tag: blume-20030527-polyeq
1742    Description:
1743    
1744    Tried to eliminated most cases of polymorphic equality.
1745    
1746    ----------------------------------------------------------------------
1747    Name: Matthias Blume (blume (at) tti - c (dot) org)
1748    Date: 2003/05/21 17:45:00 CDT
1749    Tag: blume-20030517-complete
1750    Description:
1751    
1752    Two changes:
1753    
1754      1. Added a flag for controlling whether non-exhaustive bindings will
1755         be treated as errors (default is false).
1756      2. Cleaned up the *entire* source tree so that CMB.make goes through
1757         without a single non-exhaustive match- or bind warning.
1758    
1759    ----------------------------------------------------------------------
1760    Name: Matthias Blume (blume (at) tti - c (dot) org)
1761    Date: 2003/05/17 10:20:00 CDT
1762    Tag: blume-20030517-absyn
1763    Description:
1764    
1765    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1766    
1767       This mainly affects the quality of error messages.  However, some
1768       of the code is now more straightforward than before. (Treatment of
1769       the above four constructs in translate.sml is much simpler than
1770       the "macro-expansion" that was going on before.  Plus, the mach-
1771       compiler no longer gets invoked just to be able to compile an
1772       if-expression.)
1773    
1774    2. The ErrorMsg.Error exception is now caught and absorbed by the
1775       interactive loop.
1776    
1777    ----------------------------------------------------------------------
1778    Name: Allen Leung
1779    Date: 2003/05/16 13:05:00 CDT
1780    Tag: leunga-20030516-cygwin-runtime
1781    Description:
1782    
1783        Ported the runtime system to cygwin, which uses the unix
1784    x86-unix bin files.  Missing/buggy features:
1785    
1786       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1787         the Cygwin library.
1788       o Ctrl-C handling may be flaky.
1789       o Windows system calls and Windows I/O are not supported.
1790    
1791       A new set of binfiles is located at:
1792    
1793         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1794    
1795       This is only needed for bootstrapping the cygwin version of smlnj.
1796       Other x86 versions can use the existing binfiles.
1797    
1798    ----------------------------------------------------------------------
1799    Name: Matthias Blume
1800    Date: 2003/04/08 15:42:00 CDT
1801    Tag: blume-20030408-listpair
1802    Description:
1803    
1804    1. Added a target 'mlrisc' to installer.
1805    
1806    2. Added missing elements to structure ListPair.
1807    
1808    ----------------------------------------------------------------------
1809    Name: Allen Leung
1810    Date: 2003/01/07 10:40:00 EST
1811    Tag: leunga-20030107-int-rem
1812    Description:
1813    
1814    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1815    The arguments to the SUBL instruction were swapped.
1816    
1817    ----------------------------------------------------------------------
1818    Name: Matthias Blume
1819    Date: 2002/12/12 16:25:00 EST
1820    Tag: blume-20021212-risc-ra
1821    Description:
1822    
1823    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1824    sent the RA into an infinite loop when floating point registers get
1825    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1826    110.37 and 110.38.)
1827    There was another set of potential problems related to the handling of
1828    MLRISC annotations (but those did not yet cause real problems, apparently).
1829    
1830    ----------------------------------------------------------------------
1831    Name: Matthias Blume
1832    Date: 2002/12/06 22:40:00 EST
1833    Tag: blume-20021206-cm-fileid
1834    Description:
1835    
1836    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1837    This fixes the problem of CM getting confused by files that suddenly
1838    change their identity (e.g., by getting unlinked and recreated by some
1839    text editor such as vi).  There might be a better/cheaper/cleaner way
1840    of doing this, but for now this will have to do.
1841    
1842    ----------------------------------------------------------------------
1843    Name: Matthias Blume
1844    Date: 2002/10/28 09:50:00 EST
1845    Tag: blume-20021028-typecheck
1846    Description:
1847    
1848    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1849    
1850    ----------------------------------------------------------------------
1851    Name: Matthias Blume
1852    Date: 2002/10/17 09:10:00 EDT
1853    Tag: Release_110_42
1854    Description:
1855    
1856    In good old tradition, there has been a slight hiccup so that we have
1857    to patch 110.42 after the fact.   The old release tag has been replaced
1858    (see below).
1859    
1860    The change solves a problem with two competing approaches the
1861    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1862    each other's way.
1863    
1864    This change only affects the runtime system code and the installer script.
1865    (No new bootfiles.)
1866    
1867    ----------------------------------------------------------------------
1868    Name: Matthias Blume
1869    Date: 2002/10/16 12:00:00 EDT
1870    Tag: Release_110_42_removed
1871    Description:
1872    
1873    New working release.  New bootfiles.
1874    
1875    ----------------------------------------------------------------------
1876    Name: Matthias Blume
1877    Date: 2002/10/10 13:10:00 EDT
1878    Tag: blume-20021010-ppc-divs
1879    Description:
1880    
1881    The mltree operator DIVS must be implemented with an overflow check on
1882    the PPC because the hardware indicates divide-by-zero using "overflow" as
1883    well.
1884    
1885    ----------------------------------------------------------------------
1886    Name: Matthias Blume
1887    Date: 2002/07/23 11:45:00
1888    Tag: blume-20020723-smlnj-home
1889    Description:
1890    
1891    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1892    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1893    CM_PATHCONFIG is also set), the path configuration file is assumed
1894    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1895    the entire tree to some other place and everything will "just work".
1896    
1897    (Companion commands such as ml-build and ml-makedepend also sense this
1898    variable.)
1899    
1900    ----------------------------------------------------------------------
1901    Name: Matthias Blume
1902    Date: 2002/07/12 21:19:00 EDT
1903    Tag: blume-20020712-liveness
1904    Description:
1905    
1906    Exported two useful "step" functions from liveness module (MLRISC).
1907    
1908    ----------------------------------------------------------------------
1909    Name: Matthias Blume
1910    Date: 2002/07/05 16:00 EDT
1911    Tag: Release_110_41
1912    Description:
1913    
1914    New working release.  New bootfiles.
1915    
1916    ----------------------------------------------------------------------
1917    Name: Matthias Blume
1918    Date: 2002/07/05 10:25:00 EDT
1919    Tag: blume-20020705-btimp
1920    Description:
1921    
1922    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1923    other clients can set up backtracing support.
1924    
1925    ----------------------------------------------------------------------
1926    Name: Matthias Blume
1927    Date: 2002/06/25 17:23:00 EDT
1928    Tag: blume-20020625-fpmax
1929    Description:
1930    
1931    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1932    case.  (The sense of the isNaN test was reversed -- which made min and
1933    max always return their first argument.)
1934    
1935    ----------------------------------------------------------------------
1936    Name: Matthias Blume
1937    Date: 2002/06/11
1938    Tag: blume-20020611-unixpath
1939    Description:
1940    
1941    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1942    
1943    ----------------------------------------------------------------------
1944    Name: Matthias Blume
1945    Date: 2002/06/10 16:35:00 EDT
1946    Tag: blume-20020610-ieeereal
1947    Description:
1948    
1949    I back-ported my implementation of IEEEReal.fromString from the
1950    idlbasis-devel branch so that we can test it.
1951    
1952    Another small change is that ppDec tries to give more information
1953    than just "<sig>" in the case of functors.  However, this code is
1954    broken in some mysterious way if the functor's body's signature
1955    has not been declared by ascription but gets inferred from the
1956    implementation.  This needs fixing...
1957    
1958    ----------------------------------------------------------------------
1959    Name: Matthias Blume
1960    Date: 2002/05/31
1961    Tag: blume-20020531-btrace-mode
1962    Description:
1963    
1964    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1965    the wayside when I switched over to using Controls everywhere.)
1966    
1967    ----------------------------------------------------------------------
1968    Name: Lal George
1969    Date: 2002/05/23 12:21:40 EDT
1970    Tag: george-20020523-visual-labels
1971    Description:
1972    
1973    Labels are now displayed in the graphical output to make
1974    the fall-through and target blocks obvious.
1975    
1976    ----------------------------------------------------------------------
1977    Name: Matthias Blume
1978    Date: 2002/05/22 11:03:00 EDT
1979    Tag: blume-20020522-shrink
1980    Description:
1981    
1982    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1983    situation that comes up when allocating huge arrays.
1984    
1985    ----------------------------------------------------------------------
1986    Name: Matthias Blume
1987    Date: 2002/05/21 16:00:00 EDT
1988    Tag: Release_110_40
1989    Description:
1990    
1991    New working release (110.40).  New bootfiles.
1992    
1993    [Also: John Reppy fixed GC bug 1131.]
1994    
1995    ----------------------------------------------------------------------
1996    Name: Matthias Blume
1997    Date: 2002/05/21 12:35:00 EDT
1998    Tag: blume-20020521-cmdoc
1999    Description:
2000    
2001    CM documentation update.
2002    
2003    ----------------------------------------------------------------------
2004    Name: Matthias Blume
2005    Date: 2002/05/21 10:55:00 EDT
2006    Tag: blume-20020521-misc
2007    Description:
2008    
2009    - John tweaked runtime to be silent on heap export (except when
2010      GC messages are on).
2011    - I added a few more things (cross-compiling versions of CMB) to
2012      config/preloads (as suggestions).
2013    
2014    ----------------------------------------------------------------------
2015    Name: Matthias Blume
2016    Date: 2002/05/20 22:25:00 EDT
2017    Tag: blume-20020520-controls
2018    Description:
2019    
2020    - Added ControlUtil structure to control-lib.cm.
2021    - Use it throughout.
2022    - Used Controls facility to define MLRISC controls (as opposed to
2023      registering MLRISC control ref cells with Controls after the fact)
2024    - Fixed messed-up controls priorities.
2025    
2026    * Removed again all the stuff from config/preloads that one wouldn't
2027      be able to preload at the time the initial heap image is built.
2028      (Many libraries, e.g., CML, do not exist yet at this time.  The
2029       only libraries that can be preloaded via config/preloads are those
2030       that come bundled with the bootfiles.)
2031    
2032    ----------------------------------------------------------------------
2033    Name: Matthias Blume
2034    Date: 2002/05/20 10:59:00 EDT
2035    Tag: blume-20020520-preloads
2036    Description:
2037    
2038    Added a lot of commented-out suggestions for things to be included
2039    in config/preloads.
2040    
2041    ----------------------------------------------------------------------
2042    Name: Allen Leung
2043    Date: 2002/05/18 14:20:00 EDT
2044    Tag: leunga-20020518-mdl
2045    Description:
2046    
2047    o Made the mdl tool stuff compile and run again.
2048    o I've disabled all the stuff that depends on RTL specifications; they
2049      are all badly broken anyway.
2050    
2051    ----------------------------------------------------------------------
2052    Name: Matthias Blume
2053    Date: 2002/05/17 16:49:00 EDT
2054    Tag: blume-20020517-controls
2055    Description:
2056    
2057    0. John Reppy made several modifications to the SML/NJ library.
2058       In particular, there is a shiny new controls-lib.cm.
2059    
2060    1. Pushed new controls interface through compiler so that everything
2061       compiles again.
2062    
2063    2. Added FormatComb and FORMAT_COMB to the CML version of the
2064       SML/NJ library (so that CML compiles again).
2065    
2066    3. Modified init scripts because XXX_DEFAULT environment variables
2067       are no longer with us.  (Boot-time initialization is now done
2068       using the same environment variables that are also used for
2069       startup-time initialization of controls.)
2070    
2071    ----------------------------------------------------------------------
2072    Name: Lal George
2073    Date: 2002/05/15 09:20:10 EDT
2074    Tag: george-20020515-pseudo-op-decls
2075    Description:
2076    
2077            All pseudo-ops emitted before the first segment declaration
2078    such as TEXT, DATA, and BSS directives are assumed to be global
2079    declarations and are emitted first in the assembly file. This is
2080    useful in a number of situations where one has pseudo-ops that are not
2081    specific to any segment, and also works around the constraint that one
2082    cannot have client pseudo-ops in the TEXT segment.
2083    
2084            Because no segment is associated with these declarations it is
2085    an error to allocate any space or objects before the first segment
2086    directive and an exception will be raised. However, we cannot make
2087    this check for client pseudo-ops.
2088    
2089            These top level declarations are a field in the CFG graph_info.
2090    In theory you can continue to add to this field after the CFG has been
2091    built -- provided you know what you are doing;-)
2092    
2093    ----------------------------------------------------------------------
2094    Name: Matthias Blume
2095    Date: 2002/05/13 16:40:00 EDT
2096    Tag: blume-20020513-pp-etc
2097    Description:
2098    
2099    A few minor bugfixes:
2100    
2101      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2102        (Bogus printouts for redefined bindings still occur.  Compiler
2103        bug should no longer occur now.  We need to redo the prettyprinter
2104        from scratch.)
2105    
2106      - CM pathname printer now also adds escape sequences for ( and )
2107    
2108      - commend and docu fixes for ml-nlffi
2109    
2110    ----------------------------------------------------------------------
2111    Name: Matthias Blume
2112    Date: 2002/05/10 16:40:00 EDT
2113    Tag: blume-20020510-erg-textio
2114    Description:
2115    
2116    Applied the following bugfix provided by Emden Gansner:
2117    
2118        Output is corrupted when outputSubstr is used rather than output.
2119    
2120        The problem occurs when a substring
2121    
2122            ss = (s, dataStart, dataLen)
2123    
2124        where dataStart > 0, fills a stream buffer with avail bytes left.
2125        avail bytes of s, starting at index dataStart, are copied into the
2126        buffer, the buffer is flushed, and then the remaining dataLen-avail
2127        bytes of ss are copied into the beginning of the buffer. Instead of
2128        starting this copy at index dataStart+avail in s, the current code
2129        starts the copy at index avail.
2130    
2131        Fix:
2132        In text-io-fn.sml, change line 695 from
2133             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2134        to
2135             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2136    
2137    ----------------------------------------------------------------------
2138    Name: Matthias Blume
2139    Date: 2002/04/12 13:55:00 EDT
2140    Tag: blume-20020412-assyntax
2141    Description:
2142    
2143    1. Grabbed newer assyntax.h from the XFree86 project.
2144    2. Fiddled with how to compile X86.prim.asm without warnings.
2145    3. (Very) Minor cleanup in CM.
2146    
2147    ----------------------------------------------------------------------
2148    Name: Matthias Blume
2149    Date: 2002/04/01 (no joke!) 17:07:00 EST
2150    Tag: blume-20020401-x86div
2151    Description:
2152    
2153    Added full support for div/mod/rem/quot on the x86, using the machine
2154    instruction's two results (without clumsily recomputing the remainder)
2155    directly where appropriate.
2156    
2157    Some more extensive power-of-two support was added to the x86 instruction
2158    selector (avoiding expensive divs, mods, and muls where they can be
2159    replaced with cheaper shifts and masks).  However, this sort of thing
2160    ought to be done earlier, e.g., within the CPS optimizer so that
2161    all architectures benefit from it.
2162    
2163    The compiler compiles to a fixed point, but changes might be somewhat
2164    fragile nevertheless.  Please, report any strange things that you might
2165    see wrt. div/mod/quot/rem...
2166    
2167    ----------------------------------------------------------------------
2168    Name: Matthias Blume
2169    Date: 2002/03/29 17:22:00
2170    Tag: blume-20020329-div
2171    Description:
2172    
2173    Fixed my broken div/mod logic.  Unfortunately, this means that the
2174    inline code for div/mod now has one more comparison than before.
2175    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2176    The problem was with quotient = 0, because that alone does not tell
2177    us which way the rounding went.  One then has to look at whether
2178    remainder and divisor have the same sign...  :(
2179    
2180    Anyway, I replaced the bootfiles with fresh ones...
2181    
2182    ----------------------------------------------------------------------
2183    Name: Matthias Blume
2184    Date: 2002/03/29 14:10:00 EST
2185    Tag: blume-20020329-inlprims
2186    Description:
2187    
2188    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2189    
2190    Primops have changed. This means that the bin/boot-file formats have
2191    changed as well.
2192    
2193    To make sure that there is no confusion, I made a new version.
2194    
2195    
2196    CHANGES:
2197    
2198    * removed REMT from mltree (remainder should never overflow).
2199    
2200    * added primops to deal with divisions of all flavors to the frontend
2201    
2202    * handled these primops all the way through so they map to their respective
2203      MLRISC support
2204    
2205    * used these primops in the implementation of Int, Int32, Word, Word32
2206    
2207    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2208    
2209    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2210    
2211    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2212      floating point
2213    
2214    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2215      and Real (but Real.abs maps to a separate floating-point-only primop)
2216    
2217    
2218    TODO items:
2219    
2220    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2221      instructions because the machine instruction encoder cannot handle
2222      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2223      REM in the first place, and REM is supported by the x86 machine coder.)
2224    
2225    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2226      selection phase.  (The two can be streamlined because the hardware
2227      delivers both quotient and remainder at the same time anyway.)
2228    
2229    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2230      (Currently the behavior is inconsistent both across architectures and
2231      wrt. the draft Basis spec.)
2232    
2233    * Word8 should eventually be handled natively, too.
2234    
2235    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2236      as if there currently is no execution path that could trigger it in
2237      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2238      hold for things like multiplication and division.)
2239    
2240    ----------------------------------------------------------------------
2241    Name: Matthias Blume
2242    Date: 2002/03/27 16:27:00 EST
2243    Tag: blume-20020327-mlrisc-divisions
2244    Description:
2245    
2246    Added support for all four division operations (ML's div, mod, quot,
2247    and rem) to MLRISC.  In the course of doing so, I also rationalized
2248    the naming (no more annoying switch-around of DIV and QUOT), by
2249    parameterizing the operation by div_rounding_mode (which can be either
2250    DIV_TO_ZERO or DIV_TO_NEGINF).
2251    
2252    The generic MLTreeGen functor takes care of compiling all four
2253    operations down to only round-to-zero div.
2254    
2255    Missing pieces:
2256    
2257      * Doing something smarter than relying on MLTreeGen on architectures
2258        like, e.g., the x86 where hardware division delivers both quotient and
2259        remainder at the same time.  With this, the implementation of the
2260        round-to-neginf operations could be further streamlined.
2261    
2262      * Remove inlining support for div/mod/rem from the frontend and replace it
2263        with primops that get carried through to the backend.  Do this for all
2264        int and word types.
2265    
2266    ----------------------------------------------------------------------
2267    Name: Matthias Blume
2268    Date: 2002/03/25 17:25:00 EST
2269    Tag: blume-20020325-divmod
2270    Description:
2271    
2272    I improved (hopefully without breaking them) the implementation of Int.div,
2273    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2274    advantage of the following observations:
2275    
2276      Let  q = x quot y      r = x rem y
2277           d = x div  y      m = x mod y
2278    
2279    where "quot" is the round-to-zero version of integer division that
2280    hardware usually provides.  Then we have:
2281    
2282         r = x - q * y        where neither the * nor the - will overflow
2283         d = if q >= 0 orelse x = q * y then q else q - 1
2284                              where neither the * nor the - will overflow
2285         m = if q >= 0 orelse r = 0 then r else r + y
2286                              where the + will not overflow
2287    
2288    This results in substantial simplification of the generated code.
2289    The following table shows the number of CFG nodes and edges generated
2290    for
2291            fun f (x, y) = x OPER y
2292            (* with OPER \in div, mod, quot, rem *)
2293    
2294    
2295        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2296        --------------------------------------------------------
2297         div |         24 |         39 |         12 |         16
2298         mod |         41 |         71 |         12 |         16
2299        quot |          8 |         10 |          8 |         10
2300         rem |         10 |         14 |          8 |         10
2301    
2302    
2303    ----------------------------------------------------------------------
2304    Name: Matthias Blume
2305    Date: 2002/03/25 22:06:00 EST
2306    Tag: blume-20020325-cprotobug
2307    Description:
2308    
2309    Fixed a bug in cproto (c prototype decoder).
2310    
2311    ----------------------------------------------------------------------
2312    Name: Matthias Blume
2313    Date: 2002/03/25 16:00:00 EST
2314    Tag: blume-20020325-raw-primops
2315    Description:
2316    
2317    I did some cleanup to Allen's new primop code and
2318    replaced yesterday's bootfiles with new ones.
2319    (But they are stored in the same place.)
2320    
2321    ----------------------------------------------------------------------
2322    Name: Matthias Blume
2323    Date: 2002/03/24 22:40:00 EST
2324    Tag: blume-20020324-bootfiles
2325    Description:
2326    
2327    Made the bootfiles that Allen asked for.
2328    
2329    ----------------------------------------------------------------------
2330    Name: Allen Leung
2331    Date: 2002/03/23 15:50:00 EST
2332    Tag: leunga-20020323-flint-cps-rcc-primops
2333    Description:
2334    
2335      1. Changes to FLINT primops:
2336    
2337        (* make a call to a C-function;
2338         * The primop carries C function prototype information and specifies
2339         * which of its (ML-) arguments are floating point. C prototype
2340         * information is for use by the backend, ML information is for
2341         * use by the CPS converter. *)
2342      | RAW_CCALL of { c_proto: CTypes.c_proto,
2343                       ml_args: ccall_type list,
2344                       ml_res_opt: ccall_type option,
2345                       reentrant : bool
2346                     } option
2347       (* Allocate uninitialized storage on the heap.
2348        * The record is meant to hold short-lived C objects, i.e., they
2349        * are not ML pointers.  With the tag, the representation is
2350        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2351        *)
2352      | RAW_RECORD of {tag:bool,sz:int}
2353      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2354    
2355      2.  These CPS primops are now overloaded:
2356    
2357           rawload of {kind:numkind}
2358           rawstore of {kind:numkind}
2359    
2360          The one argument form is:
2361    
2362             rawload {kind} address
2363    
2364          The two argument form is:
2365    
2366             rawload {kind} [ml object, byte-offset]
2367    
2368      3. RAW_CCALL/RCC now takes two extra arguments:
2369    
2370         a. The first is whether the C call is reentrant, i.e., whether
2371            ML state should be saved and restored.
2372         b. The second argument is a string argument specifying the name of
2373            library and the C function.
2374    
2375         These things are currently not handled in the code generator, yet.
2376    
2377      4. In CProto,
2378    
2379         An encoding type of "bool" means "ml object" and is mapped into
2380         C prototype of PTR.  Note that "bool" is different than "string",
2381         even though "string" is also mapped into PTR, because "bool"
2382         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2383    
2384      5. Pickler/unpicker
2385    
2386         Changed to handle RAW_RECORD and newest RAW_CCALL
2387    
2388      6. MLRiscGen,
2389    
2390         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2391         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2392            in the file CodeGen/cpscompile/cps-c-calls.sml
2393    
2394      7. Added the conditional move operator
2395    
2396             condmove of branch
2397    
2398         to cps.  Generation of this is still buggy so it is currently
2399         disabled.
2400    
2401    ----------------------------------------------------------------------
2402    Name: Lal George
2403    Date: 2002/03/22 14:18:25 EST
2404    Tag: george-20020322-cps-branch-prob
2405    Description:
2406    
2407    Implemented the Ball-Larus branch prediction-heuristics, and
2408    incorporated graphical viewers for control flow graphs.
2409    
2410    Ball-Larus Heuristics:
2411    ---------------------
2412    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2413    
2414    By design it uses the Dempster-Shafer theory for combining
2415    probabilities.  For example, in the function:
2416    
2417        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2418    
2419    the ball-larus heuristics predicts that the n=0 is unlikely
2420    (OH-heuristic), and the 'then' branch is unlikely because of the
2421    RH-heuristic -- giving the 'then' branch an even lower combined
2422    probability using the Dempster-Shafer theory.
2423    
2424    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2425    probability of the 'then' branch because of the loop in the else
2426    branch.
2427    
2428    
2429    Graphical Viewing:
2430    ------------------
2431    I merely plugged in Allen's graphical viewers into the compiler. The
2432    additional code is not much. At the top level, saying:
2433    
2434            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2435    
2436    will display the graphical view of the control flow graph just before
2437    back-patching.  daVinci must be in your path for this to work. If
2438    daVinci is not available, then the default viewer can be changed
2439    using:
2440    
2441            Control.MLRISC.getString "viewer"
2442    
2443    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2444    course, these viewers must be in your path.
2445    
2446    The above will display the compilation unit at the level of clusters,
2447    many of which are small, boring, and un-interesting. Also setting:
2448    
2449            Control.MLRISC.getInt "cfg-graphical-view_size"
2450    
2451    will display clusters that are larger than the value set by the above.
2452    
2453    
2454    ----------------------------------------------------------------------
2455    Name: Matthias Blume
2456    Date: 2002/03/21 22:20:00 EST
2457    Tag: blume-20020321-kmp-bugfix
2458    Description:
2459    
2460    Changed the interface to the KMP routine in PreString and fixed
2461    a minor bug in one place where it was used.
2462    
2463    ----------------------------------------------------------------------
2464    Name: Allen Leung
2465    Date: 2002/03/21 20:30:00 EST
2466    Tag: leunga-20020321-cfg
2467    Description:
2468    
2469      Fixed a potential problem in cfg edge splitting.
2470    
2471    ----------------------------------------------------------------------
2472    Name: Allen Leung
2473    Date: 2002/03/21 17:15:00 EST
2474    Tag: leunga-20020321-x86-fp-cfg
2475    Description:
2476    
2477      1. Recoded the buggy parts of x86-fp.
2478    
2479         a. All the block reordering code has been removed.
2480            We now depend on the block placement phases to do this work.
2481    
2482         b. Critical edge splitting code has been simplified and moved into the
2483            CFG modules, as where they belong.
2484    
2485         Both of these were quite buggy and complex.  The code is now much, much
2486         simpler.
2487    
2488      2. X86 backend.
2489    
2490         a. Added instructions for 64-bit support.  Instruction selection for
2491            64-bit has not been committed, however, since that
2492            requires changes to MLTREE which haven't been approved by
2493            Lal and John.
2494    
2495         b. Added support for FUCOMI and FUCOMIP when generating code for
2496            PentiumPro and above.  We only generate these instructions in
2497            the fast-fp mode.
2498    
2499         c. Added cases for JP and JNP in X86FreqProps.
2500    
2501      3. CFG
2502    
2503         CFG now has a bunch of methods for edge splitting and merging.
2504    
2505      4. Machine description.
2506    
2507         John's simplification of MLTREE_BASIS.fcond broke a few machine
2508         description things:
2509    
2510         rtl-build.{sig,sml} and hppa.mdl fixed.
2511    
2512         NOTE: the machine description stuff in the repository is still broken.
2513               Again, I can't put my fixes in because that involves
2514               changes to MLTREE.
2515    
2516    ----------------------------------------------------------------------
2517    Name: Matthias Blume
2518    Date: 2002/03/20 15:55:00 EST
2519    Tag: blume-20020320-kmp
2520    Description:
2521    
2522    Implemented Knuth-Morris-Pratt string matching in PreString and used
2523    it for String.isSubstring, Substring.isSubstring, and
2524    Substring.position.
2525    
2526    (Might need some stress-testing.  Simple examples worked fine.)
2527    
2528    ----------------------------------------------------------------------
2529    Name: Matthias Blume
2530    Date: 2002/03/19 16:37:00 EST
2531    Tag: blume-20020319-witnesses
2532    Description:
2533    
2534    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2535    
2536    This implements a generic mechanism for changing constness qualifiers
2537    anywhere within big C types without resorting to outright "casts".
2538    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2539    modify the constness at the outermost level.)
2540    The implementation of "convert" is based on the idea of "witness"
2541    values -- values that are not used by the operation but whose types
2542    "testify" to their applicability.  On the implementation side, "convert"
2543    is simply a projection (returning its second curried argument).  With
2544    cross-module inlining, it should not result in any machine code being
2545    generated.
2546    
2547    ----------------------------------------------------------------------
2548    Name: Matthias Blume
2549    Date: 2002/03/15 16:40:00 EST
2550    Tag: blume-20020315-basis
2551    Description:
2552    
2553    Provided (preliminary?) implementations for
2554    
2555      {String,Substring}.{concatWith,isSuffix,isSubstring}
2556    
2557    and
2558    
2559      Substring.full
2560    
2561    Those are in the Basis spec but they were missing in SML/NJ.
2562    
2563    ----------------------------------------------------------------------
2564    Name: Matthias Blume
2565    Date: 2002/03/14 21:30:00 EST
2566    Tag: blume-20020314-controls
2567    Description:
2568    
2569    Controls:
2570    ---------
2571    
2572    1. Factored out the recently-added Controls : CONTROLS stuff and put
2573       it into its own library $/controls-lib.cm.  The source tree for
2574       this is under src/smlnj-lib/Controls.
2575    
2576    2. Changed the names of types and functions in this interface, so they
2577       make a bit more "sense":
2578    
2579          module -> registry
2580          'a registry -> 'a group
2581    
2582    3. The interface now deals in ref cells only.  The getter/setter interface
2583       is (mostly) gone.
2584    
2585    4. Added a function that lets one register an already-existing ref cell.
2586    
2587    5. Made the corresponding modifications to the rest of the code so that
2588       everything compiles again.
2589    
2590    6. Changed the implementation of Controls.MLRISC back to something closer
2591       to the original.  In particular, this module (and therefore MLRISC)
2592       does not depend on Controls.  There now is some link-time code in
2593       int-sys.sml that registers the MLRISC controls with the Controls
2594       module.
2595    
2596    CM:
2597    ---
2598    
2599      * One can now specify the lambda-split aggressiveness in init.cmi.
2600    
2601    ----------------------------------------------------------------------
2602    Name: Allen Leung
2603    Date: 2002/03/13 17:30:00 EST
2604    Tag: leunga-20020313-x86-fp-unary
2605    Description:
2606    
2607    Bug fix for:
2608    
2609    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2610    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2611    > - fun f(x,(y,z)) = Real.~ y;
2612    > [autoloading]
2613    > [autoloading done]
2614    >       fchsl   (%eax), 184(%esp)
2615    > Error: MLRisc bug: X86MCEmitter.emitInstr
2616    >
2617    > uncaught exception Error
2618    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2619    
2620    The problem was that the code generator did not generate any fp registers
2621    in this case, and the ra didn't know that it needed to run the X86FP phase to
2622    translate the pseudo fp instruction.   This only happened with unary fp
2623    operators in certain situations.
2624    
2625    ----------------------------------------------------------------------
2626    Name: Matthias Blume
2627    Date: 2002/03/13 14:00:00 EST
2628    Tag: blume-20020313-overload-etc
2629    Description:
2630    
2631    1. Added _overload as a synonym for overload for backward compatibility.
2632       (Control.overloadKW must be true for either version to be accepted.)
2633    
2634    2. Fixed bug in install script that caused more things to be installed
2635       than what was requested in config/targets.
2636    
2637    3. Made CM aware of the (_)overload construct so that autoloading
2638       works.
2639    
2640    ----------------------------------------------------------------------
2641    Name: Matthias Blume
2642    Date: 2002/03/12 22:03:00 EST
2643    Tag: blume-20020312-url
2644    Description:
2645    
2646    Forgot to update BOOT and srcarchiveurl.
2647    
2648    ----------------------------------------------------------------------
2649    Name: Matthias Blume
2650    Date: 2002/03/12 17:30:00 EST
2651    Tag: blume-20020312-version110392
2652    Description:
2653    
2654    Yet another version number bump (because of small changes to the
2655    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2656    
2657    Changes:
2658    
2659      The new pid generation scheme described a few weeks ago was overly
2660      complicated.  I implemented a new mechanism that is simpler and
2661      provides a bit more "stability":  Once CM has seen a compilation
2662      unit, it keeps its identity constant (as long as you do not delete
2663      those crucial CM/GUID/* files).  This means that when you change
2664      an interface, compile, then go back to the old interface, and
2665      compile again, you arrive at the original pid.
2666    
2667      There now also is a mechanism that instructs CM to use the plain
2668      environment hash as a module's pid (effectively making its GUID
2669      the empty string).  For this, "noguid" must be specified as an
2670      option to the .sml file in question within its .cm file.
2671      This is most useful for code that is being generated by tools such
2672      as ml-nlffigen (because during development programmers tend to
2673      erase the tool's entire output directory tree including CM's cached
2674      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2675      revert to the old, broken behavior of SML/NJ, but in specific cases
2676      where there is no danger of interface confusion, its use is ok
2677      (I think).
2678    
2679      ml-nlffigen by default generates "noguid" annotations.  They can be
2680      turned off by specifying -guid in its command line.
2681    
2682    ----------------------------------------------------------------------
2683    Name: Lal George
2684    Date: 2002/03/12 12 14:42:36 EST
2685    Tag: george-20020312-frequency-computation
2686    Description:
2687    
2688    Integrated jump chaining and static block frequency into the
2689    compiler. More details and numbers later.
2690    
2691    ----------------------------------------------------------------------
2692    Name: Lal George
2693    Date: 2002/03/11 11 22:38:53 EST
2694    Tag: george-20020311-jump-chain-elim
2695    Description:
2696    
2697    Tested the jump chain elimination on all architectures (except the
2698    hppa).  This is on by default right now and is profitable for the
2699    alpha and x86, however, it may not be profitable for the sparc and ppc
2700    when compiling the compiler.
2701    
2702    The gc test will typically jump to a label at the end of the cluster,
2703    where there is another jump to an external cluster containing the actual
2704    code to invoke gc. This is to allow factoring of common gc invocation
2705    sequences. That is to say, we generate:
2706    
2707            f:
2708               testgc
2709               ja   L1      % jump if above to L1
2710    
2711            L1:
2712               jmp L2
2713    
2714    
2715    After jump chain elimination the 'ja L1' instructions is converted to
2716    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2717    up being implemented in their long form (if L2 is far away) using:
2718    
2719            jbe     L3      % jump if below or equal to L3
2720            jmp     L2
2721         L3:
2722            ...
2723    
2724    
2725    For large compilation units L2  may be far away.
2726    
2727    
2728    ----------------------------------------------------------------------
2729    Name: Matthias Blume
2730    Date: 2002/03/11 13:30:00 EST
2731    Tag: blume-20020311-mltreeeval
2732    Description:
2733    
2734    A functor parameter was missing.
2735    
2736    ----------------------------------------------------------------------
2737    Name: Allen Leung
2738    Date: 2002/03/11 10:30:00 EST
2739    Tag: leunga-20020311-runtime-string0
2740    Description:
2741    
2742       The representation of the empty string now points to a
2743    legal null terminated C string instead of unit.  It is now possible
2744    to convert an ML string into C string with InlineT.CharVector.getData.
2745    This compiles into one single machine instruction.
2746    
2747    ----------------------------------------------------------------------
2748    Name: Allen Leung
2749    Date: 2002/03/10 23:55:00 EST
2750    Tag: leunga-20020310-x86-call
2751    Description:
2752    
2753       Added machine generation for CALL instruction (relative displacement mode)
2754    
2755    ----------------------------------------------------------------------
2756    Name: Matthias Blume
2757    Date: 2002/03/08 16:05:00
2758    Tag: blume-20020308-entrypoints
2759    Description:
2760    
2761    Version number bumped to 110.39.1.  NEW BOOTFILES!
2762    
2763    Entrypoints: non-zero offset into a code object where execution should begin.
2764    
2765    - Added the notion of an entrypoint to CodeObj.
2766    - Added reading/writing of entrypoint info to Binfile.
2767    - Made runtime system bootloader aware of entrypoints.
2768    - Use the address of the label of the first function given to mlriscGen
2769      as the entrypoint.  This address is currently always 0, but it will
2770      not be 0 once we turn on block placement.
2771    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2772      with entry points) from mlriscGen.
2773    
2774    ----------------------------------------------------------------------
2775    Name: Allen Leung
2776    Date: 2002/03/07 20:45:00 EST
2777    Tag: leunga-20020307-x86-cmov
2778    Description:
2779    
2780       Bug fixes for CMOVcc on x86.
2781    
2782       1. Added machine code generation for CMOVcc
2783       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2784       3. CMOVcc cannot have an immediate operand as argument.
2785    
2786    ----------------------------------------------------------------------
2787    Name: Matthias Blume
2788    Date: 2002/03/07 16:15:00 EST
2789    Tag: blume-20020307-controls
2790    Description:
2791    
2792    This is a very large but mostly boring patch which makes (almost)
2793    every tuneable compiler knob (i.e., pretty much everything under
2794    Control.* plus a few other things) configurable via both the command
2795    line and environment variables in the style CM did its configuration
2796    until now.
2797    
2798    Try starting sml with '-h' (or, if you are brave, '-H')
2799    
2800    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2801    implements the underlying generic mechanism.
2802    
2803    The interface to some of the existing such facilities has changed somewhat.
2804    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2805    (The getFoo interface is still there for backward-compatibility, but its
2806    use is deprecated.)
2807    
2808    The ml-build script passes -Cxxx=yyy command-line arguments through so
2809    that one can now twiddle the compiler settings when using this "batch"
2810    compiler.
2811    
2812    TODO items:
2813    
2814    We should go through and throw out all controls that are no longer
2815    connected to anything.  Moreover, we should go through and provide
2816    meaningful (and correct!) documentation strings for those controls
2817    that still are connected.
2818    
2819    Currently, multiple calls to Controls.new are accepted (only the first
2820    has any effect).  Eventually we should make sure that every control
2821    is being made (via Controls.new) exactly once.  Future access can then
2822    be done using Controls.acc.
2823    
2824    Finally, it would probably be a good idea to use the getter-setter
2825    interface to controls rather than ref cells.  For the time being, both
2826    styles are provided by the Controls module, but getter-setter pairs are
2827    better if thread-safety is of any concern because they can be wrapped.
2828    
2829    *****************************************
2830    
2831    One bug fix: The function blockPlacement in three of the MLRISC
2832    backpatch files used to be hard-wired to one of two possibilities at
2833    link time (according to the value of the placementFlag).  But (I
2834    think) it should rather sense the flag every time.
2835    
2836    *****************************************
2837    
2838    Other assorted changes (by other people who did not supply a HISTORY entry):
2839    
2840    1. the cross-module inliner now works much better (Monnier)
2841    2. representation of weights, frequencies, and probabilities in MLRISC
2842       changed in preparation of using those for weighted block placement
2843       (Reppy, George)
2844    
2845    ----------------------------------------------------------------------
2846    Name: Lal George
2847    Date: 2002/03/07 14:44:24 EST 2002
2848    Tag: george-20020307-weighted-block-placement
2849    
2850    Tested the weighted block placement optimization on all architectures
2851    (except the hppa) using AMPL to generate the block and edge frequencies.
2852    Changes were required in the machine properties to correctly
2853    categorize trap instructions. There is an MLRISC flag
2854    "weighted-block-placement" that can be used to enable weighted block
2855    placement, but this will be ineffective without block/edge
2856    frequencies (coming soon).
2857    
2858    
2859    ----------------------------------------------------------------------
2860    Name: Lal George
2861    Date: 2002/03/05 17:24:48 EST
2862    Tag: george-20020305-linkage-cluster
2863    
2864    In order to support the block placement optimization, a new cluster
2865    is generated as the very first cluster (called the linkage cluster).
2866    It contains a single jump to the 'real' entry point for the compilation
2867    unit. Block placement has no effect on the linkage cluster itself, but
2868    all the other clusters  have full freedom in the manner in which they
2869    reorder blocks or functions.
2870    
2871    On the x86 the typical linkage code that is generated is:
2872       ----------------------
2873            .align 2
2874       L0:
2875            addl    $L1-L0, 72(%esp)
2876            jmp     L1
2877    
2878    
2879            .align  2
2880       L1:
2881       ----------------------
2882    
2883    72(%esp) is the memory location for the stdlink register. This
2884    must contain the address of the CPS function being called. In the
2885    above example, it contains the address of  L0; before
2886    calling L1 (the real entry point for the compilation unit), it
2887    must contain the address for L1, and hence
2888    
2889            addl $L1-L0, 72(%esp)
2890    
2891    I have tested this on all architectures except the hppa.The increase
2892    in code size is of course negligible
2893    
2894    ----------------------------------------------------------------------
2895    Name: Allen Leung
2896    Date: 2002/03/03 13:20:00 EST
2897    Tag: leunga-20020303-mlrisc-tools
2898    
2899      Added #[ ... ] expressions to mlrisc tools
2900    
2901    ----------------------------------------------------------------------
2902    Name: Matthias Blume
2903    Date: 2002/02/27 12:29:00 EST
2904    Tag: blume-20020227-cdebug
2905    Description:
2906    
2907    - made types in structure C and C_Debug to be equal
2908    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2909    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2910    
2911    ----------------------------------------------------------------------
2912    Name: Matthias Blume
2913    Date: 2002/02/26 12:00:00 EST
2914    Tag: blume-20020226-ffi
2915    Description:
2916    
2917    1. Fixed a minor bug in CM's "noweb" tool:
2918       If numbering is turned off, then truly don't number (i.e., do not
2919       supply the -L option to noweb).  The previous behavior was to supply
2920       -L'' -- which caused noweb to use the "default" line numbering scheme.
2921       Thanks to Chris Richards for pointing this out (and supplying the fix).
2922    
2923    2. Once again, I reworked some aspects of the FFI:
2924    
2925       A. The incomplete/complete type business:
2926    
2927       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2928         gone!
2929       - ML types representing an incomplete type are now *equal* to
2930         ML types representing their corresponding complete types (just like
2931         in C).  This is still safe because ml-nlffigen will not generate
2932         RTTI for incomplete types, nor will it generate functions that
2933         require access to such RTTI.   But when ML code generated from both
2934         incomplete and complete versions of the C type meet, the ML types
2935         are trivially interoperable.
2936    
2937         NOTE:  These changes restore the full generality of the translation
2938         (which was previously lost when I eliminated functorization)!
2939    
2940       B. Enum types:
2941    
2942       - Structure C now has a type constructor "enum" that is similar to
2943         how the "su" constructor works.  However, "enum" is not a phantom
2944         type because each "T enum" has values (and is isomorphic to
2945         MLRep.Signed.int).
2946       - There are generic access operations for enum objects (using
2947         MLRep.Signed.int).
2948       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2949         * The structure contains the definition of type "mlrep" (the ML-side
2950         representation type of the enum).  Normally, mlrep is the same
2951         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2952         then mlrep will be defined as a datatype -- thus facilitating
2953         pattern matching on mlrep values.
2954         ("-ec" will be suppressed if there are duplicate values in an
2955          enumeration.)
2956         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2957         will be generated for each C enum constant xxx.
2958         * Conversion functions m2i and i2m convert between mlrep and
2959         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2960         * Coversion functions c and ml convert between mlrep and "tag enum".
2961         * Access functions (get/set) fetch and store mlrep values.
2962       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2963         enumerations are merged into one single enumeration represented by
2964         structure E_'.
2965    
2966    ----------------------------------------------------------------------
2967    Name: Allen Leung
2968    Date: 2002/02/25 04:45:00 EST
2969    Tag: leunga-20020225-cps-spill
2970    
2971    This is a new implementation of the CPS spill phase.
2972    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2973    In case of problems, replace it with the old file spill.sml
2974    
2975    The current compiler runs into some serious performance problems when
2976    constructing a large record.  This can happen when we try to compile a
2977    structure with many items.  Even a very simple structure like the following
2978    makes the compiler slow down.
2979    
2980        structure Foo = struct
2981           val x_1 = 0w1 : Word32.int
2982           val x_2 = 0w2 : Word32.int
2983           val x_3 = 0w3 : Word32.int
2984           ...
2985           val x_N = 0wN : Word32.int
2986        end
2987    
2988    The following table shows the compile time, from N=1000 to N=4000,
2989    with the old compiler:
2990    
2991    N
2992    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2993           MLRISC ra                               0.06u  0.00s  0.05g
2994              (spills = 0 reloads = 0)
2995           TOTAL                                   0.63u  0.07s  0.21g
2996    
2997    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2998           MLRISC ra                               5.68u  0.59s  3.93g
2999              (spills = 0 reloads = 0)
3000           TOTAL                                   14.71u  0.99s  4.81g
3001    
3002    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3003           MLRISC ra                               5.54u  0.65s  3.91g
3004              (spills = 543 reloads = 1082)
3005           TOTAL                                   65.40u  3.13s  6.00g
3006    
3007    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3008           MLRISC ra                               0.80u  0.10s  0.55g
3009              (spills = 42 reloads = 84)
3010           TOTAL                                   129.42u  5.10s  4.13g
3011    
3012    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3013           MLRISC ra                               2.69u  0.27s  1.38g
3014              (spills = 62 reloads = 124)
3015           TOTAL                                   682.48u  19.61s  13.99g
3016    
3017    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3018           MLRISC ra                               4.96u  0.27s  2.72g
3019              (spills = 85 reloads = 170)
3020           TOTAL                                   2375.26u  57.21s  48.00g
3021    
3022    As you can see the old cps spill module suffers from some serious
3023    performance problem.  But since I cannot decipher the old code fully,
3024    instead of patching the problems up, I'm reimplementing it
3025    with a different algorithm.  The new code is more modular,
3026    smaller when compiled, and substantially faster
3027    (O(n log n) time and O(n) space).  Timing of the new spill module:
3028    
3029    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3030          MLRISC ra                               0.25u  0.02s  0.15g
3031             (spills=1 reloads=3)
3032          TOTAL                                   7.74u  0.34s  1.62g
3033    
3034    Implementation details:
3035    
3036    As far as I can tell, the purpose of the CPS spill module is to make sure the
3037    number of live variables at any program point (the bandwidth)
3038    does not exceed a certain limit, which is determined by the
3039    size of the spill area.
3040    
3041    When the bandwidth is too large, we decrease the register pressure by
3042    packing live variables into spill records.  How we achieve this is
3043    completely different than what we did in the old code.
3044    
3045    First, there is something about the MLRiscGen code generator
3046    that we should be aware of:
3047    
3048    o MLRiscGen performs code motion!
3049    
3050       In particular, it will move floating point computations and
3051       address computations involving only the heap pointer to
3052       their use sites (if there is only a single use).
3053       What this means is that if we have a CPS record construction
3054       statement
3055    
3056           RECORD(k,vl,w,e)
3057    
3058       we should never count the new record address w as live if w
3059       has only one use (which is often the case).
3060    
3061       We should do something similar to floating point, but the transformation
3062       there is much more complex, so I won't deal with that.
3063    
3064    Secondly, there are now two new cps primops at our disposal:
3065    
3066     1. rawrecord of record_kind option
3067        This pure operator allocates some uninitialized storage from the heap.
3068        There are two forms:
3069    
3070         rawrecord NONE [INT n]  allocates a tagless record of length n
3071         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3072                                     and initializes the tag.
3073    
3074     2. rawupdate of cty
3075          rawupdate cty (v,i,x)
3076          Assigns to x to the ith component of record v.
3077          The storelist is not updated.
3078    
3079    We use these new primops for both spilling and increment record construction.
3080    
3081     1. Spilling.
3082    
3083        This is implemented with a linear scan algorithm (but generalized
3084        to trees).  The algorithm will create a single spill record at the
3085        beginning of the cps function and use rawupdate to spill to it,
3086        and SELECT or SELp to reload from it.  So both spills and reloads
3087        are fine-grain operations.  In contrast, in the old algorithm
3088        "spills" have to be bundled together in records.
3089    
3090        Ideally, we should sink the spill record construction to where
3091        it is needed.  We can even split the spill record into multiple ones
3092        at the places where they are needed.  But CPS is not a good
3093        representation for global code motion, so I'll keep it simple and
3094        am not attempting this.
3095    
3096     2. Incremental record construction (aka record splitting).
3097    
3098        Long records with many component values which are simulatenously live
3099        (recall that single use record addresses are not considered to
3100         be live) are constructed with rawrecord and rawupdate.
3101        We allocate space on the heap with rawrecord first, then gradually
3102        fill it in with rawupdate.  This is the technique suggested to me
3103        by Matthias.
3104    
3105        Some restrictions on when this is applicable:
3106        1. It is not a VECTOR record.  The code generator currently does not handle
3107           this case. VECTOR record uses double indirection like arrays.
3108        2. All the record component values are defined in the same "basic block"
3109           as the record constructor.  This is to prevent speculative
3110           record construction.
3111    
3112    ----------------------------------------------------------------------
3113    Name: Allen Leung
3114    Date: 2002/02/22 01:02:00 EST
3115    Tag: leunga-20020222-mlrisc-tools
3116    
3117    Minor bug fixes in the parser and rewriter
3118    
3119    ----------------------------------------------------------------------
3120    Name: Allen Leung
3121    Date: 2002/02/21 20:20:00 EST
3122    Tag: leunga-20020221-peephole
3123    
3124    Regenerated the peephole files.  Some contained typos in the specification
3125    and some didn't compile because of pretty printing bugs in the old version
3126    of 'nowhere'.
3127    
3128    ----------------------------------------------------------------------
3129    Name: Allen Leung
3130    Date: 2002/02/19 20:20:00 EST
3131    Tag: leunga-20020219-mlrisc-tools
3132    Description:
3133    
3134       Minor bug fixes to the mlrisc-tools library:
3135    
3136       1.  Fixed up parsing colon suffixed keywords
3137       2.  Added the ability to shut the error messages up
3138       3.  Reimplemented the pretty printer and fixed up/improved
3139           the pretty printing of handle and -> types.
3140       4.  Fixed up generation of literal symbols in the nowhere tool.
3141       5.  Added some SML keywords to to sml.sty
3142    
3143    ----------------------------------------------------------------------
3144    Name: Matthias Blume
3145    Date: 2002/02/19 16:20:00 EST
3146    Tag: blume-20020219-cmffi
3147    Description:
3148    
3149    A wild mix of changes, some minor, some major:
3150    
3151    * All C FFI-related libraries are now anchored under $c:
3152        $/c.cm      --> $c/c.cm
3153        $/c-int.cm  --> $c/internals/c-int.cm
3154        $/memory.cm --> $c/memory/memory.cm
3155    
3156    * "make" tool (in CM) now treats its argument pathname slightly
3157      differently:
3158        1. If the native expansion is an absolute name, then before invoking
3159           the "make" command on it, CM will apply OS.Path.mkRelative
3160           (with relativeTo = OS.FileSys.getDir()) to it.
3161        2. The argument will be passed through to subsequent phases of CM
3162           processing without "going native".  In particular, if the argument
3163           was an anchored path, then "make" will not lose track of that anchor.
3164    
3165    * Compiler backends now "know" their respective C calling conventions
3166      instead of having to be told about it by ml-nlffigen.  This relieves
3167      ml-nlffigen from one of its burdens.
3168    
3169    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3170    
3171    * Export C_DEBUG and C_Debug from $c/c.cm.
3172    
3173    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3174      subtyping relationship between incomplete pointers and their complete
3175      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3176      with the convention of instantiating 'o with ('t, 'c) obj whenever
3177      the pointer target type is complete.  In the incomplete case, 'o
3178      will be instantiated with some "'c iobj" -- a type obtained by
3179      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3180    
3181      Operations that work on both incomplete and complete pointer types are
3182      typed as taking an 'o ptr while operations that require the target to
3183      be known are typed as taking some ('t, 'c) obj ptr.
3184    
3185      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3186      where void is an eqtype without any values.  This makes it possible
3187      to work on voidptr values using functions meant to operate on light
3188      incomplete pointers.
3189    
3190    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3191      been vastly simplified.
3192    
3193    ----------------------------------------------------------------------
3194    Name: Matthias Blume
3195    Date: 2002/02/19 10:48:00 EST
3196    Tag: blume-20020219-pqfix
3197    Description:
3198    
3199    Applied Chris Okasaki's bug fix for priority queues.
3200    
3201    ----------------------------------------------------------------------
3202    Name: Matthias Blume
3203    Date: 2002/02/15 17:05:00
3204    Tag: Release_110_39
3205    Description:
3206    
3207    Last-minute retagging is becoming a tradition... :-(
3208    
3209    This is the working release 110.39.
3210    
3211    ----------------------------------------------------------------------
3212    Name: Matthias Blume
3213    Date: 2002/02/15 16:00:00 EST
3214    Tag: Release_110_39-orig
3215    Description:
3216    
3217    Working release 110.39.  New bootfiles.
3218    
3219    (Update: There was a small bug in the installer so it wouldn't work
3220    with all shells.  So I retagged. -Matthias)
3221    
3222    ----------------------------------------------------------------------
3223    Name: Matthias Blume
3224    Date: 2002/02/15 14:17:00 EST
3225    Tag: blume-20020215-showbindings
3226    Description:
3227    
3228    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3229    the latter can be useful for exploring what bindings are available at
3230    the interactive prompt.  (The first function returns only the list
3231    of symbols that are really bound, the second prints those but also the
3232    ones that CM's autoloading mechanism knows about.)
3233    
3234    ----------------------------------------------------------------------
3235    Name: Matthias Blume
3236    Date: 2002/02/15 12:08:00 EST
3237    Tag: blume-20020215-iptrs
3238    Description:
3239    
3240    Two improvements to ml-nlffigen:
3241    
3242      1. Write files only if they do not exist or if their current contents
3243         do not coincide with what's being written.  (That is, avoid messing
3244         with the time stamps unless absolutely necessary.)
3245    
3246      2. Implement a "repository" mechanism for generated files related
3247         to "incomplete pointer types".   See the README file for details.
3248    
3249    ----------------------------------------------------------------------
3250    Name: Matthias Blume
3251    Date: 2002/02/14 11:50:00 EST
3252    Tag: blume-20020214-quote
3253    Description:
3254    
3255    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3256    because of the new and improved tag generation scheme.  (Thanks to Allen
3257    Leung for pointing it out.)
3258    
3259    ----------------------------------------------------------------------
3260    Name: Lal George
3261    Date: 2002/02/14 09:55:27 EST 2002
3262    Tag: george-20020214-isabelle-bug
3263    Description:
3264    
3265    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3266    of Isabelle on the x86.
3267    
3268    From Allen:
3269    -----------
3270     I've found the problem:
3271    
3272         in ra-core.sml, I use the counter "blocked" to keep track of the
3273         true number of elements in the freeze queue.  When the counter goes
3274         to zero, I skip examining the queue.  But I've messed up the
3275         bookkeeping in combine():
3276    
3277             else ();
3278             case !ucol of
3279               PSEUDO => (if !cntv > 0 then
3280                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3281                                        ^^^^^^^^^^^^^^^^^^^^^^^
3282                      moveu := mergeMoveList(!movev, !moveu)
3283                     )
3284                  else ();
3285    
3286         combine() is called to coalesce two nodes u and v.
3287         I think I was thinking that if the move counts of u and v are both
3288         greater than zero then after they are coalesced then one node is
3289         removed from the freeze queue.  Apparently I was thinking that
3290         both u and v are of low degree, but that's clearly not necessarily true.
3291    
3292    
3293    02/12/2002:
3294        Here's the patch.  HOL now compiles.
3295    
3296        I don't know how this impact on performance (compile
3297        time or runtime).  This bug caused the RA (especially on the x86)
3298        to go thru the potential spill phase when there are still nodes on the
3299        freeze queue.
3300    
3301    
3302    
3303    
3304    ----------------------------------------------------------------------
3305    Name: Matthias Blume
3306    Date: 2002/02/13 22:40:00 EST
3307    Tag: blume-20020213-fptr-rtti
3308    Description:
3309    
3310    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3311    updates.
3312    
3313    ----------------------------------------------------------------------
3314    Name: Matthias Blume
3315    Date: 2002/02/13 16:41:00 EST
3316    Tag: blume-20020213-cmlpq
3317    Description:
3318    
3319    Added new priority queue export symbols (which have just been added to
3320    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3321    not compile and the installer would choke.)
3322    
3323    ----------------------------------------------------------------------
3324    Name: Matthias Blume
3325    Date: 2002/02/13 16:15:00 EST
3326    Tag: blume-20020213-various
3327    Description:
3328    
3329    1. More tweaks to ml-nlffigen:
3330    
3331       - better internal datastructures (resulting in slight speedup)
3332       - "-match" option requires exact match
3333       - "localized" gensym counters (untagged structs/unions nested within
3334         other structs/unions or within typedefs get a fresh counter; their
3335         tag will be prefixed by a concatenation of their parents' tags)
3336       - bug fixes (related to calculation of transitive closure of types
3337         to be included in the output)
3338    
3339    2. Minor Basis updates:
3340    
3341       - added implementations for List.collate and Option.app
3342    
3343    ----------------------------------------------------------------------
3344    Name: Matthias Blume
3345    Date: 2002/02/11 15:55:00 EST
3346    Tag: blume-20020211-gensym
3347    Description:
3348    
3349    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3350    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3351    names (ML structure names that correspond to unnamed C structs, unions,
3352    and enums), so that separate runs of ml-nlffigen do not clash.
3353    
3354    ----------------------------------------------------------------------
3355    Name: Matthias Blume
3356    Date: 2002/02/11 12:05:00 EST
3357    Tag: blume-20020211-gensml
3358    Description:
3359    
3360    A quick fix for a problem with GenSML (in the pgraph-util library):
3361    Make generation of toplevel "local" optional.  (Strictly speaking,
3362    signature definitions within "local" are not legal SML.)
3363    
3364    Other than that: updates to INSTALL and cm/TODO.
3365    
3366    ----------------------------------------------------------------------
3367    Name: Matthias Blume
3368    Date: 2002/02/08 15:00:00 EST
3369    Tag: blume-20020208-uniquepid
3370    Description:
3371    
3372    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3373    
3374    1. The installer (config/install.sh) has gotten smarter:
3375    
3376         - Configuration options are a bit easier to specify now
3377           (in config/targets).
3378         - Bug in recognizing .tar.bz2 files fixed.
3379         - Installer automatically resolves dependencies between
3380           configuration options (e.g., if you ask for eXene, you will
3381           also get cml -- regardless whether you asked for it or not).
3382         - Installer can run in "quieter mode" by setting the environment
3383           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3384           "completely silent", though.
3385         - Build HashCons library as part of smlnj-lib.
3386    
3387    2. A new scheme for assigning persistent identifiers to compilation
3388       units (and, by extension, to types etc.) has been put into place.
3389       This fixes a long-standing bug where types and even dynamic values
3390       can get internally confused, thereby compromising type safety
3391       (abstraction) and dynamic correctness.  See
3392    
3393         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3394    
3395       for an example of how things could go wrong until now.
3396    
3397       The downside of the new scheme is that pids are not quite as
3398       persistent as they used to be: CM will generate a fresh pid
3399       for every compilation unit that it thinks it sees for the first
3400       time.  That means that if you compile starting from a clean, fresh
3401       source tree at two different times, you end up with different
3402       binaries.
3403    
3404       Cutoff recompilation, however, has not been compromised because
3405       CM keeps pid information in special caches between runs.
3406    
3407    ----------------------------------------------------------------------
3408    Name: Lal George
3409    Date: 2002/02/07 15:34:13 EST 2002
3410    Tag: <none>
3411    Description:
3412    
3413    Compilers that generate assembly code may produce  global labels
3414    whose value is resolved at link time. The various peephole optimization
3415    modules did not take this in account.
3416    
3417    TODO. The Labels.addrOf function should really return an option
3418    type so that clients are forced to deal with this issue, rather
3419    than an exception being raised.
3420    
3421    ----------------------------------------------------------------------
3422    Name: Lal George
3423    Date: 2002/02/06 13:55:02 EST
3424    Tag: george-20020206-ra-breakup
3425    Description:
3426    
3427    1. A bug fix from Allen.
3428    
3429        A typo causes extra fstp %st(0)'s to be generated at compensation
3430        edges, which might cause stack underflow traps at runtime.  This
3431        occurs in fft where there are extraneous fstps right before the 'into'
3432        trap instruction (in this case they are harmless since none of the
3433        integers overflow.)
3434    
3435    2. Pulled out various utility modules that were embedded in the modules
3436       of the register allocator. I need these modules for other purposes, but
3437       they are not complete enough to put into a library (just yet).
3438    ----------------------------------------------------------------------
3439    Name: Matthias Blume
3440    Date: 2002/01/31 16:05:00 EST
3441    Tag: blume-20020131-sparc-ccalls
3442    Description:
3443    
3444    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3445       of extra stack space by mistake.  Fixed.
3446    
3447    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3448    
3449    ----------------------------------------------------------------------
3450    Name: Allen Leung
3451    Date: 2002/01/30
3452    Tag: leunga-20020130-nowhere-bug-fix
3453    Description:
3454    
3455       MLRISC bug fixes:
3456       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3457       2. MachineInt.fromString was negating its value.
3458    
3459    ----------------------------------------------------------------------
3460    Name: Matthias Blume
3461    Date: 2002/01/29
3462    Tag: blume-20020129-INSTALL
3463    Description:
3464    
3465    - Added somewhat detailed installation instructions (file INSTALL).
3466    - Fixed curl-detection bug in config/install.sh.
3467    - It is now possible to select the URL getter using the URLGETTER
3468      environment variable:
3469    
3470          not set / "unknown"      --> automatic detection (script tries wget,
3471                                       curl, and lynx)
3472          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3473                                       how to properly invoke them)
3474          other                    --> use $URLGETTER directly, it must take
3475                                       precisely two command-line arguments
3476                                       (source URL and destination file name)
3477    
3478    ----------------------------------------------------------------------
3479    Name: Matthias Blume
3480    Date: 2002/01/28
3481    Tag: blume-20020128-sparc-ccalls
3482    Description:
3483    
3484    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3485    - Make use of the allocParam argument in sparc-c-calls.
3486    
3487    ----------------------------------------------------------------------
3488  Name: Matthias Blume  Name: Matthias Blume
3489  Date: 2002/01/28  Date: 2002/01/28
3490  Tag: blume-20020128-allocParam  Tag: blume-20020128-allocParam
# Line 584  Line 4056 
4056    
4057  ----------------------------------------------------------------------  ----------------------------------------------------------------------
4058  Name: Matthias Blume  Name: Matthias Blume
 >>>>>>> 1.169  
4059  Date: 2001/09/18 15:35:00 EDT  Date: 2001/09/18 15:35:00 EDT
4060  Tag: blume-20010918-readme11036  Tag: blume-20010918-readme11036
4061  Description:  Description:

Legend:
Removed from v.1044  
changed lines
  Added in v.1724

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