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 1733, Thu Dec 16 05:37:04 2004 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/12/15 23:40:00 CST
18    Tag: blume-20041215-hashSubstring
19    Description:
20    
21    - HashString.hashString' -> HashString.hashSubstring
22    - corresponding changes in atom.sml
23    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
24      added some comments
25    
26    ----------------------------------------------------------------------
27    Name: Matthias Blume (blume (at) tti - c (dot) org)
28    Date: 2004/12/15 15:30:00 CST
29    Tag: Release_110_52
30    Description:
31    
32    New working version (110.52).  NEW BOOTFILES!
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume (blume (at) tti - c (dot) org)
36    Date: 2004/12/15 12:45:00 CST
37    Tag: blume-20041215-spaces
38    Description:
39    
40    More on the space problem (this time for Win32).
41    
42    ----------------------------------------------------------------------
43    Name: Matthias Blume (blume (at) tti - c (dot) org)
44    Date: 2004/12/14 17:30:00 CST
45    Tag: blume-20041214-spaces
46    Description:
47    
48    Hacked some of the scripts (in particular: the installer) to cope with
49    spaces in filenames a bit better.  But beware: the current "solution"
50    is likely still full of bugs and inherently incomplete.  (We need to
51    do away with those shell scripts for a comprehensive solution.)
52    
53    ----------------------------------------------------------------------
54    Name: Matthias Blume (blume (at) tti - c (dot) org)
55    Date: 2004/12/13 14:45:00 CST
56    Tag: blume-20041213-ml-makedepend
57    Description:
58    
59    Fixed bug in code for ml-makedepend.
60    
61    ----------------------------------------------------------------------
62    Name: Matthias Blume (blume (at) tti - c (dot) org)
63    Date: 2004/12/09 16:30:00 CST
64    Tag: blume-20041209-statistics
65    Description:
66    
67    Added two simple but potentially useful statistics modules to SML/NJ Library.
68    (See CHANGES file there.)
69    
70    ----------------------------------------------------------------------
71    Name: Matthias Blume (blume (at) tti - c (dot) org)
72    Date: 2004/12/01 16:50:00 CST
73    Tag: blume-20041201-atom
74    Description:
75    
76    smlnj-lib:
77    
78      Added function HashString.hashString' for substrings.
79      Hand-inlined CharVector.foldl into HashString (for speed).
80      Modified implementation of structure Atom to avoid extracting
81      strings from substrings unless necessary.
82      (Also see CHANGES file for smlnj-lib.)
83    
84    ----------------------------------------------------------------------
85    Name: Matthias Blume (blume (at) tti - c (dot) org)
86    Date: 2004/11/24 22:15:00 CST
87    Tag: blume-20041124-cml
88    Description:
89    
90    Made sure CML compiles when Position = Int64.
91    
92    ----------------------------------------------------------------------
93    Name: Matthias Blume (blume (at) tti - c (dot) org)
94    Date: 2004/11/24 14:30:00 CST
95    Tag: blume-20041124-position
96    Description:
97    
98    The compiler can now be compiled in a mode that makes structure Position
99    equal to Int64.  The default, however, is unchanged (Position = Int31)
100    for the time being.
101    
102    To enable 64-bit positions, use the following procedure:
103    
104    1. Start sml
105    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
106    3. Type
107          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
108    4. Run CMB.make() as usual.
109    
110    This is barely tested.  The only test so far was a little SML program
111    counting the number of characters in an 8-gigabyte file by
112    reading it character-by-character.  That test was successful.
113    
114    In support of 64-bit positions, a number of new functions have been
115    added to the runtime system.
116    
117    ----------------------------------------------------------------------
118    Name: Matthias Blume (blume (at) tti - c (dot) org)
119    Date: 2004/11/23 14:45:00 CST
120    Tag: blume-20041123-useFile
121    Description:
122    
123    Fixed a problem with unhelpful error messages related to problems with
124    .cm- or .sml files that appear as part of the sml command line.
125    
126    ----------------------------------------------------------------------
127    Name: Matthias Blume (blume (at) tti - c (dot) org)
128    Date: 2004/11/18 15:40:00 CST
129    Tag: Release_110_51
130    Description:
131    
132    New working version (110.51).  NEW BOOTFILES!
133    
134    ----------------------------------------------------------------------
135    Name: Matthias Blume (blume (at) tti - c (dot) org)
136    Date: 2004/11/18 15:35:00 CST
137    Tag: <none>
138    Description:
139    
140    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
141    Joemann).
142    
143    ----------------------------------------------------------------------
144    Name: Allen Leung (leunga (at) reservoir (dot) com)
145    Date: 2004/11/17 16:05:21 EST 2004
146    Tag: leunga-20041117-mlrisc-live-kill
147    Description:
148    
149       Added support for MLTree constructs LIVE and KILL
150    to all the architectures.
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume (blume (at) tti - c (dot) org)
154    Date: 2004/11/13 00:20:00 CST
155    Tag: blume-20041113-versiontool
156    Description:
157    
158    - Stripped down the versiontool: It now only handles the version number.
159      The date string is generated at bootstrap time (during makeml).
160    
161    - In a previous commit, fixed a minor issue with how polyequal is being
162      translated.  In particular, the code now "looks through" abstractions.
163      This results in slightly fewer polyEqual warnings and hopefully slightly
164      more efficient code.  Important examples for where this matters are
165      the new int64 and word64 types.
166    
167    ----------------------------------------------------------------------
168    Name: Matthias Blume (blume (at) tti - c (dot) org)
169    Date: 2004/11/12 00:30:00 CST
170    Tag: blume-20041112-int64
171    Description:
172    
173    Structure Int64 fully hooked in.  (The implementation is not very
174    efficient, though.)
175    
176    ----------------------------------------------------------------------
177    Name: Matthias Blume (blume (at) tti - c (dot) org)
178    Date: 2004/11/11 17:30:00 CST
179    Tag: blume-20041111-more64
180    Description:
181    
182    All the pieces of Word64 are now there, with the exception of the
183    conversions from and to LargeWord. (Eventually these need to be identities,
184    but for the time being they don't even make sense because LargeWord is
185    32-bit wide.)
186    
187    Also started to add similar support for Int64, but major pieces of that
188    are still missing.
189    
190    ----------------------------------------------------------------------
191    Name: Matthias Blume (blume (at) tti - c (dot) org)
192    Date: 2004/11/11 00:15:00 CST
193    Tag: blume-20041111-word64
194    Description:
195    
196    Structure Word64 is now (almost) complete, word literals and patterns
197    seem to work.  There are a few odd pieces missing.  In particular,
198    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
199    at the moment.
200    
201    Making Word64 official would mean that LargeWord becomes Word64.  But
202    this requires extreme care because most word-word conversions have to
203    go through LargeWord, so making a mistake means loss of efficiency or
204    worse.  Eventually there will be a solution similar to (but actually
205    simpler than) what I did with IntInf.
206    
207    ----------------------------------------------------------------------
208    Name: Matthias Blume (blume (at) tti - c (dot) org)
209    Date: 2004/11/10 18:12:00 CST
210    Tag: blume-20041110-64bit
211    Description:
212    
213    More 64-bit hacking (but still not even half-way there yet).
214    Also, some assorted improvements to the handling of 8-bit words.
215    
216    ----------------------------------------------------------------------
217    Name: Matthias Blume (blume (at) tti - c (dot) org)
218    Date: 2004/11/09 17:50:00 CST
219    Tag: <none>
220    Description:
221    
222    Started adding some infrastructure for supporting 64-bit int- and
223    word-types.  (Still in its very early stages.)
224    
225    ----------------------------------------------------------------------
226    Name: Matthias Blume (blume (at) tti - c (dot) org)
227    Date: 2004/10/28 10:45:00 CDT
228    Tag: Release_110_50
229    Description:
230    
231    New working version (110.50).  NEW BOOTFILES!
232    
233                 =====================
234    
235    Also:
236    
237      - Changend config/srcarchiveurl from a file just
238        containing the URL string into a file containing
239        shell script code.  The code has access to the $VERSION variable.
240      - Made corresponding changes to config/install.sh and config/unpack.
241      - Default contents of config/srcarchiveurl uses $VERSION and
242        normally does not have to be edited to reflect a version change.
243    
244      (As a result, a version change can be done by just editing
245       config/version, the rest is now automatic.)
246    
247    ----------------------------------------------------------------------
248    Name: Matthias Blume (blume (at) tti - c (dot) org)
249    Date: 2004/10/27 17:50:00 CDT
250    Tag: blume-20041027-btrace-msg
251    Description:
252    
253    BackTrace.monitor now also reports the source of the exception that
254    triggered the trace.
255    
256    ----------------------------------------------------------------------
257    Name: Matthias Blume (blume (at) tti - c (dot) org)
258    Date: 2004/10/27 17:20:00 CDT
259    Tag: blume-20041027-x86-c-calls
260    Description:
261    
262    This is the HISTORY entry for two earlier commits, both concerning
263    the x86 c-calls code in MLRISC:
264    
265      - added a missing LOAD in the code that deals with struct arguments
266      - made sure the caller does not add the wrong number of bytes to the
267        stack pointer after a call of a function returning a struct
268        (the callee already pops the implicit argument which points to
269         the space reserved for the result)
270    
271    ----------------------------------------------------------------------
272    Name: Allen Leung (leunga (at) reservoir (dot) com)
273    Date: 2004/10/24 14:00:00 EST
274    Tag: leunga-20041024-x86-gas-fucomip
275    Description:
276    
277       John discovered a bug in the syntax of fucomip.
278       The opcodes FU?COMIP? have been changed to
279    
280           fu?comip? %st(i), %st
281    
282    ----------------------------------------------------------------------
283    Name: Matthias Blume (blume (at) tti - c (dot) org)
284    Date: 2004/10/20 15:06:00 CDT
285    Tag: blume-20041020-standalone-backtrace
286    Description:
287    
288    Added a mechanism for getting back-trace information from standalone
289    programs.  Here is how it works:
290    
291       1. The part of the program from which you want to get backtrace
292          information (usually the whole program) should be wrapped with
293          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
294          main program could be modified from something like
295    
296             fun main (pgm, args) = ...
297    
298         to
299    
300            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
301    
302       2. To be able to access BackTrace.monitor, you have to add
303          library $smlnj-tdp/plugins.cm to the .cm file that contains your
304          main function.
305    
306       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
307          CM might have created in the past for your project).
308    
309       4. Build the system using this command line:
310    
311             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
312                      myprog.cm MyProg.main myprog
313    
314          instead of the usual
315    
316             ml-build myprog.cm MyProg.main myprog
317    
318    I changed a library name:
319    
320       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
321    
322    New libraries:
323    
324       $smlnj-tdp/back-trace.cm
325          -- when loaded causes the back-trace plugin to be installed
326       $smlnj-tdp/coverage.cm
327          -- when loaded causes the coverage plugin to be installed
328    
329    ----------------------------------------------------------------------
330    Name: Matthias Blume (blume (at) tti - c (dot) org)
331    Date: 2004/10/18 16:45:00 CDT
332    Tag: blume-20041018-groupowner
333    Description:
334    
335    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
336    
337    Eliminated group owner specs from .cm files throughout the source tree.
338    
339    ----------------------------------------------------------------------
340    Name: Matthias Blume (blume (at) tti - c (dot) org)
341    Date: 2004/10/15 15:45:00 CDT
342    Tag: blume-20041015-coverage
343    Description:
344    
345    * Test coverage tool added!
346    * Further reorganization of tracing-, debugging-, and profiling support:
347    
348        - moved original BTImp -- now called BackTrace -- into a separate
349          library called $/trace-debug-profile.cm
350        - eliminated all mentions of BTrace from SMLofNJ.Internals
351        - only the instrumentation mechanism is now left in the compiler proper
352        - BackTrace module is a plugin which is NOT plugged in by default
353        - Coverage module is another such plugin
354    
355    To get the benefits of any of these plugin modules, the code in
356    question must be compiled with tdp instrumentation turned on.  This
357    can be done by setting SMLofNJ.Internals.TDP.mode to true.
358    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
359    
360    Plugins are selected at link time.  (Pre-compiled instrumented code
361    can be re-loaded with different plugins in effect.)  When an
362    instrumented module is linked, whatever plugins are at that time
363    enabled will come into effect for that module.
364    
365    To enable the back-trace plugin, load library $/trace-debug-profile.cm
366    and invoke BackTrace.install() (e.g., from the interactive prompt).
367    To enable the coverage plugin, load the same library and invoke
368    Coverage.install().
369    
370    Back-traces are generated automatically on uncaught exceptions and
371    when the code in question explicitly invokes BackTrace.trigger().
372    
373    Coverage (and execution frequency-) information must be queried
374    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
375    
376    ----------------------------------------------------------------------
377    Name: Matthias Blume (blume (at) tti - c (dot) org)
378    Date: 2004/10/14 17:40:00 CDT
379    Tag: blume-20041014-tdp-core
380    Description:
381    
382    Snapshot of a significant overhaul of how the trace/debug/profile support
383    is hooked into the system (specifically: Core and SMLofNJ.Internals).
384    
385    ----------------------------------------------------------------------
386    Name: Matthias Blume (blume (at) tti - c (dot) org)
387    Date: 2004/10/13 16:34:00 CDT
388    Tag: blume-20041013-tdp
389    Description:
390    
391    Some rationalization of names:
392    
393       structure BTrace -> structure TDPInstrument
394       etc.
395    
396    This is is preparation of using the original back-trace
397    instrumentation for other purposes.  "TDP" stands for
398    Trace/Debug/Profile.
399    
400    The control flag controlling whether instrumentation is on or off is now
401    registered under a different name, so instead of running sml as
402    
403     sml -Cinstrument.btrace-mode=true
404    
405    one has to say
406    
407     sml -Ctdp.instrument=true
408    
409    ----------------------------------------------------------------------
410    Name: Matthias Blume (blume (at) tti - c (dot) org)
411    Date: 2004/10/11 16:37:00 CDT
412    Tag: blume-20041011-regions
413    Description:
414    
415    Made some minor modifications to elabcore.sml to have source regions
416    be propagated more tightly -- resulting in better (i.e., smaller)
417    regions being reported in error- and debug messages.
418    
419    ----------------------------------------------------------------------
420    Name: Matthias Blume (blume (at) tti - c (dot) org)
421    Date: 2004/10/08 22:50:00 CDT
422    Tag: blume-20041008-cmkw
423    Description:
424    
425    Fixed handling of keywords in .cm files: After seeing "is" the lexer
426    treats subsequent occurrences of "group", "library", "source", "is",
427    "*", and "-" as ordinary identifiers rather than keywords.
428    
429    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
430    is supposed to accept a tool argument called "source", but this did
431    not work because of the clash with the keyword.
432    
433    ----------------------------------------------------------------------
434    Name: Matthias Blume (blume (at) tti - c (dot) org)
435    Date: 2004/10/07 16:00:00 CDT
436    Tag: blume-20041007-cleanup
437    Description:
438    
439    Assorted cleanup work:
440    
441      - got rid of intstrmap in favor of using the library's
442        hash table implementation
443      - threw out most of the pathnames stuff, as it was not used anyway
444      - simplified tokentable implementation
445      - fixed some minor spelling errors
446    
447    ----------------------------------------------------------------------
448    Name: Matthias Blume (blume (at) tti - c (dot) org)
449    Date: 2004/10/06 15:15:15 CDT
450    Tag: blume-20041006-handler
451    Description:
452    
453    Cleaned up the absyn to reflect the invariant that HANDLE always
454    carries a FNexp as part of the type definition.  This eliminates some
455    superfluous sanity checks at runtime down the road.
456    
457    Some minor cleanup of the btrace code.
458    
459    ----------------------------------------------------------------------
460    Name: Matthias Blume (blume (at) tti - c (dot) org)
461    Date: 2004/10/01 10:20:30 CDT
462    Tag: blume-20041001-slave
463    Description:
464    
465    Added hack to make slave mode work in the presence of the version
466    tool.  (Still, since the master does two passes over the code for
467    CMB.make, the release number gets bumped twice when slaves are
468    attached.  I don't know if this is worth fixing...)
469    
470    ----------------------------------------------------------------------
471    Name: Matthias Blume (blume (at) tti - c (dot) org)
472    Date: 2004/09/30 10:55:00 CDT
473    Tag: blume-20040930-version
474    Description:
475    
476    * Moved the "version" magic into its own little library under
477      src/system/smlnj/internal.  This avoids expensive reconstruction of
478      a stable src/compiler/core.cm.
479    
480    * At the same time, structure CompilerVersion is now known as
481      structure SMLNJVersion.
482    
483    * Arranged for the version tool to NOT kick in when rebuilding the system
484      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
485      Also, loading the versiontool does not work when rebuilding the system
486      because CM is not properly initialized at that time.
487    
488    ----------------------------------------------------------------------
489    Name: Matthias Blume (blume (at) tti - c (dot) org)
490    Date: 2004/09/29 14:00:00 CDT
491    Tag: blume-20040929-autoversion
492    Description:
493    
494    Implemented some CM magic to have
495        file src/compiler/TopLevel/main/version.sml
496    generated automagically.
497    The version is taken from two files: config/version and config/release.
498    The first is expected to contain a two-part version number such as 110.49.
499    The second should contain a single number, but it may be missing.
500    
501    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
502    time the version tool is loaded (which is the first time you say CMB.make),
503    then the tool will increment the value stored in config/release every
504    time CMB.make is invoked.
505    
506    The binfile format is now insensitive to anything beyond the first
507    two components of a version number, so bumping the release does not render
508    binfiles incompatible.  Auto-bumping can be used to keep track of versions
509    during development without invalidating existing binfiles.
510    
511    In any case, every CMB.make updates the date information in version.sml.
512    (This is the date that is printed in the banner.)
513    
514    ----------------------------------------------------------------------
515    Name: Matthias Blume (blume (at) tti - c (dot) org)
516    Date: 2004/09/28 10:53:00 CDT
517    Tag: blume-20040928-controls
518    Description:
519    
520    Some cleanup of the controls code.
521    
522    ----------------------------------------------------------------------
523    Name: Matthias Blume (blume (at) tti - c (dot) org)
524    Date: 2004/09/27 22:08:00 CDT
525    Tag: blume-20040927-controls
526    Description:
527    
528    Added two pieces of functionality to the Controls interface:
529    
530    1.   val save'restore: 'a control -> unit -> unit
531    
532       grabs the current value of the control in stage 1 and restores it
533       in stage 2.
534    
535    2.   val set' : 'a control * 'a -> unit -> unit
536    
537       stores the given value into the control in stage 2 (i.e., delayed)
538       but does all error checking in stage 1.
539       (This is for string controls that need to do parse their argument --
540        something that might fail.  In some cases, notably in CM, one
541        already knows the intended argument but wants to delay the actual
542        assignment until a time when error recovery would be more difficult.)
543    
544    Changed the handling of controls in tool arguments to classes "sml" and
545    "lazysml":
546       - use Controls.save'restore as a more robust way of restoring the
547         old value (in particular: without having to re-parse the string)
548       - use controls to handle the "overload" keyword in the init group
549         (I believe this change actually fixes a long-standing obscure bug.)
550    
551    ----------------------------------------------------------------------
552    Name: Matthias Blume (blume (at) tti - c (dot) org)
553    Date: 2004/09/27 17:00:00 CDT
554    Tag: blume-20040927-lazysml
555    Description:
556    
557    Added a new tool class called "lazysml" to CM's tool chest.  The only
558    difference to "sml" is that compilation is done with Control.lazysml
559    set to true.  A source of class "lazysml" is automatically recognized
560    by a file name suffix of ".lml".
561    
562    In addition to the above feature, the original class "sml" now also
563    supports a tool argument "lazy" which has the same effect.  As a
564    result, the following three lines are equivalent:
565    
566        foo.sml : lazysml
567        foo.sml : sml (lazy)
568        foo.sml (lazy)
569    
570    The setting goes into effect both during parsing and during
571    compilation.  The original setting is restored right after parsing and
572    after compilation, respectively.
573    
574    In addition to all the above, there is also a general mechanism to set
575    ANY of the "controls" that are available at the command line via
576    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
577    apply as well. (In fact, "lazy" is implemented as a special case of
578    the general mechanism.)
579    
580    The .cm file syntax uses a new keyword tool argument called "with".
581    There are several ways of indicating the desired settings:
582    
583        foo.sml (with:parser.quotations=true)
584        foo.sml (with:(name:parser.quotations value:true))
585        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
586        foo.sml (with:(name1=value1 name2=value2 ...))
587        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
588    
589        etc.
590    
591    Another possible abbreviation is to leave out the =v or value:v part
592    if the name refers to a boolean control (in which case the value is
593    taken to be true).  Thus, one could get lazy sml also by saying:
594    
595        foo.sml (with:parser.lazy-keyword=true)
596        foo.sml (with:parser.lazy-keyword)
597        foo.sml (with:(name:parser.lazy-keyword value:true))
598        foo.sml (with:(name:parser.lazy-keyword))
599    
600    ----------------------------------------------------------------------
601    Name: Matthias Blume (blume (at) tti - c (dot) org)
602    Date: 2004/09/24 16:22:00 CDT
603    Tag: blume-20040924-ppc-long-branch
604    Description:
605    
606    Turned message about "emiting long form of branch" off by default.
607    Added a control flag to turn it back on when desired.
608    
609    ----------------------------------------------------------------------
610    Name: Matthias Blume (blume (at) tti - c (dot) org)
611    Date: 2004/09/24 16:05:00 CDT
612    Tag: blume-20040924-rounding
613    Description:
614    
615    Applied patch for setting rounding modes under Mac OS X.  Thanks to
616    Melissa O'Neill for providing the code!
617    
618    ----------------------------------------------------------------------
619    Name: Matthias Blume (blume (at) tti - c (dot) org)
620    Date: 2004/09/23 17:30:00 CDT
621    Tag: blume-20040923-envvars
622    Description:
623    
624    1. Changed definition of type ControlRegistry.registry_tree to
625       include control_info (i.e., the name of the controlling environment
626       variable).
627    
628    2. Added command-line flags -e and -E to print the names of environment
629       variables that can be used to control internal settings.
630       (This uses the new API mentioned in 1.)
631    
632    ----------------------------------------------------------------------
633    Name: Matthias Blume (blume (at) tti - c (dot) org)
634    Date: 2004/09/13 16:50:00 CDT
635    Tag: Release_110_49
636    Description:
637    
638    New working version (110.49).  NEW BOOTFILES!
639    
640    ----------------------------------------------------------------------
641    Name: Matthias Blume
642    Date: 2004/09/13 16:20:00 CDT
643    Tag: blume-20040913-config-mlrisc
644    Description:
645    
646    Put target "mlrisc" back into the default list.
647    (There is no harm in having it, and some users have expressed their
648    wish to have "mlrisc" included by default.)
649    
650    ----------------------------------------------------------------------
651    Name: John Reppy
652    Date: 2004/09/13
653    Tag: jhr-20040913-signals
654    Description:
655    
656    Fixed the signal masking code to properly nest mask/unmask operations
657    on a per-signal basis.
658    
659    ----------------------------------------------------------------------
660    Name: Matthias Blume (blume (at) tti - c (dot) org)
661    Date: 2004/09/08 13:20:00 CDT
662    Tag: blume-20040908-heap-magic
663    Description:
664    
665    Bumped the heap macig number to 0x09082004 to account for the changed
666    layout of the ML frame under MacOS X.
667    
668    ----------------------------------------------------------------------
669    Name: Allen Leung (leunga (at) reservoir (dot) com)
670    Date: 2004/09/03 11:26:00 EST
671    Tag: leunga-20040903-cygwin-install
672    Description:
673    
674       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
675    The Cygwin runtime is turned on by setting the environment
676    variable SMLNJ_CYGWIN_RUNTIME to 1.
677    
678    ----------------------------------------------------------------------
679    Name: Matthias Blume (blume (at) tti - c (dot) org)
680    Date: 2004/08/31 17:15:00 CDT
681    Tag: blume-20040831-core
682    Description:
683    
684    Added some exports to src/compiler/core.cm upon request by J. Joemann.
685    
686    ----------------------------------------------------------------------
687    Name: Matthias Blume (blume (at) tti - c (dot) org)
688    Date: 2004/08/30 17:55:00 CDT
689    Tag: blume-20040830-installer
690    Description:
691    
692    Upon request by Johannes Joemann:
693    
694    - improved ML code of installer to fall back to coping when renaming
695      fails (i.e., when source and target are on different file systems);
696      the code compiles but has yet to be tested in anger
697    - removed mlrisc from list of default targets (config/targets)
698    
699    ----------------------------------------------------------------------
700    Name: Matthias Blume (blume (at) tti - c (dot) org)
701    Date: 2004/08/27 17:20:00 CDT
702    Tag: blume-20040827-ptreql
703    Description:
704    
705    Added ptreql primop to structure InlineT (upon request from Larry
706    Paulson).
707    
708    ----------------------------------------------------------------------
709    Name: Allen Leung (leunga (at) reservoir (dot) com)
710    Date: 2004/08/15 21:21:00 EST
711    Tag: leunga-110_48-udgraph
712    Description:
713    
714    Another bug fix from Carl Hauser:
715    
716    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
717    > 48c48
718    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
719    > e::L)
720    > ---
721    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
722    > Without this, any deletion of an edge in an undirected graph does severe
723    > violence to the graph.
724    
725    ----------------------------------------------------------------------
726    Name: Allen Leung (leunga (at) reservoir (dot) com)
727    Date: 2004/08/10 23:35:00 EST
728    Tag: leunga-110_48-ppc
729    Description:
730    
731       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
732    
733    ----------------------------------------------------------------------
734    Name: Matthias Blume (blume (at) tti - c (dot) org)
735    Date: 2004/08/10 12:00:00 CDT
736    Tag: Release_110_48
737    Description:
738    
739    New working version (110.48).  NEW BOOTFILES!
740    
741    ----------------------------------------------------------------------
742    Name: Allen Leung (leunga (at) reservoir (dot) com)
743    Date: 2004/08/09 12:21:00 EST
744    Tag: leunga-110_47-dijsktra
745    Description:
746    
747    Bug fix from Carl Hauser:
748    
749    single_source_shortest_paths in dijkstra.sml was observed to get wrong
750    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
751    
752    The problem is that following the expression A.update(dist,s,Num.zero)
753    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
754    
755    ----------------------------------------------------------------------
756    Name: Matthias Blume (blume (at) tti - c (dot) org)
757    Date: 2004/08/06 18:10:00 CDT
758    Tag: blume-20040806-cmdline
759    Description:
760    
761    Fiddled with handling of command-line options:
762    
763      * sml now quits after processing the command line
764        if -H, -S, -h<n>, or -s<n> appears as the last
765        command-line argument
766      * a new option -q terminates the session when encountered on
767        the command line; subsequent arguments will be ignored
768      * bug fixes: short (erroneous) arguments are no longer ignored
769        completely
770    
771    ----------------------------------------------------------------------
772    Name: Allen Leung (leunga (at) reservoir (dot) com)
773    Date: 2004/08/04 18:17:00 EST
774    Tag: leunga-110_47-ppc-ibm-asm
775    Description:
776    
777      - Added minimal IBM assembly syntax support for PowerPC.
778    
779      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
780        been altered in the runtime.  We need an automatic way to keep the file
781        in sync.
782    
783    ----------------------------------------------------------------------
784    Name: Matthias Blume (blume (at) tti - c (dot) org)
785    Date: 2004/08/04 14:00:00 CDT
786    Tag: Release_110_47
787    Description:
788    
789    New working version (110.47).  NEW BOOTFILES!
790    
791    ----------------------------------------------------------------------
792    Name: Matthias Blume (blume (at) tti - c (dot) org)
793    Date: 2004/08/03 14:25:00 CDT
794    Tag: blume-20040803-callingconv
795    Description:
796    
797    Added low-level support for choosing C calling conventions by
798    twiddling the type of rawccall.  (See
799    src/compiler/Semant/types/cproto.sml for details.)
800    
801    ----------------------------------------------------------------------
802    Name: Matthias Blume (blume (at) tti - c (dot) org)
803    Date: 2004/08/02 15:55:00 CDT
804    Tag: blume-20040802-backout
805    Description:
806    
807    Backed out of change to win32-filesys.c.  The earlier patch to
808    get_file_time caused CM to produce files with the wrong time stamp.
809    
810    ----------------------------------------------------------------------
811    Name: Matthias Blume (blume (at) tti - c (dot) org)
812    Date: 2004/08/02 14:45:00 CDT
813    Tag: blume-20040802-nlffi-win32
814    Description:
815    
816    Added NLFFI support for Win32, adapted from a patch provided by David
817    Hansel.  This is currently completely untested.  Also, the issue
818    concerning stdcall vs. ccall is still unresolved.
819    
820    ----------------------------------------------------------------------
821    Name: Matthias Blume (blume (at) tti - c (dot) org)
822    Date: 2004/07/30 17:55:00 CDT
823    Tag: blume-20040730-various
824    Description:
825    
826    Gearing up towards 110.47...
827    
828    - various minor bugfixes to ml-nlffigen
829    - a beginning of a manual for nlffi
830    
831    - eliminated 'export name=value' in config/install.sh as this does
832      not work with certain versions of /bin/sh
833      (Thanks to David King at Motorola for catching this.)
834    
835    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
836      - added a test for tm==NULL to gmtime.c and localtime.c
837      - applied patch for incorrect GetFileTime under win32
838      - toSeconds -> toMilliseconds in Win32/win32-process.sml
839    
840    ----------------------------------------------------------------------
841    Name: Matthias Blume (blume (at) tti - c (dot) org)
842    Date: 2004/07/21 18:20:00 CDT
843    Tag: blume-20040721-nlffigen
844    Description:
845    
846    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
847      for a typedef to an incomplete type, but leave out the "typ" member.
848      (This is just for consistency.)
849    - Started to produce what is supposed to become better (i.e., comprehensive)
850      documentation of what ml-nlffigen does and produces.
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume (blume (at) tti - c (dot) org)
854    Date: 2004/07/14 16:25:00 CDT
855    Tag: blume-20040714-union
856    Description:
857    
858    Added C_UNION to c-calls/c-types.sml and updated the machinery
859    (ml-nlffigen, cproto.sml) that conveys C function interface
860    information to the code generator.
861    
862    However, the actual architecture-specific implementation of function
863    arguments and results that are C unions is still not implemented.
864    
865    ----------------------------------------------------------------------
866    Name: Allen Leung (leunga (at) reservoir (dot) com)
867    Date: 2004/07/14 14:38:00 EST
868    Tag: leunga-110_46_1-ppc-lwzu
869    Description:
870    
871        Added these instructions to the PowerPC architecture:
872        LBZU(X), LHZU(X), LWZU(X),
873        STWU(X), STFDU, STFSU
874    
875        etc...
876    
877        Note: I haven't added their instruction encoding into the description.
878    
879    ----------------------------------------------------------------------
880    Name: Allen Leung (leunga (at) reservoir (dot) com)
881    Date: 2004/07/13 15:04:00 EST
882    Tag: leunga-110_46_1-ppc-lwarx
883    Description:
884    
885       Added the two instructions LWARX and STWCX to the PowerPC
886    instruction set.
887    
888       A (untested) rewrite of loop-structure.sml.  The old version
889    is completely broken.
890    
891    ----------------------------------------------------------------------
892    Name: Matthias Blume (blume (at) tti - c (dot) org)
893    Date: 2004/07/13 13:50:00 CDT
894    Tag: blume-20040713-nlffi
895    Description:
896    
897    - use paramAlloc to report c-calls with too many arguments
898      (for PPC version where parameter area is pre-allocated)
899    - added ccall_maxargspace to machspec (to implement the above)
900    - made "make" commend in CM's "make" tool configurable
901    - added option (default: on) for passing the name of the SML/NJ's "bin"
902      directory to "make"; the call looks like this:
903    
904         make <options> SMLNJ_BINDIR=<dir> <target>
905    
906      This can be used by the Makefile to, e.g., pick the "right" version
907      of ml-nlffigen.
908    - minor code tweaks
909    
910    ----------------------------------------------------------------------
911    Name: Matthias Blume (blume (at) tti - c (dot) org)
912    Date: 2004/07/12 22:50:00 CDT
913    Tag: blume-110_46_1-macosx-nlffi
914    Description:
915    
916    NLFFI under Mac OS X now working (sort of).  This is largely untested,
917    though.
918    
919    Note:  1. You have to make a new, clean build of the runtime system.
920           2. There are new BOOTFILES, you have to use them!
921              (Doing the bootstrap process yourself would be *very* painful!
922               If you absolutely have to do it, build the system under
923               a different architecture and then cross-compile.)
924    
925    Version bumped to 110.46.1 to account for runtime data format changes.
926    
927    ----------------------------------------------------------------------
928    Name: Matthias Blume (blume (at) tti - c (dot) org)
929    Date: 2004/06/18 14:30:00 CDT
930    Tag: blume-20040618-unix
931    Description:
932    
933    Changed the implementation of structure Unix so that the same stream
934    is returned every time one of the {text,bin}{In,Out}streamOf functions
935    is invoked on the same proc.  This is not what the spec currently
936    says -- although IMO it arguably should.  (See discussion below.)
937    
938    ----------------------------------------------------------------------
939    Name: Matthias Blume (blume (at) tti - c (dot) org)
940    Date: 2004/06/17 18:15:00 CDT
941    Tag: Release_110_46
942    Description:
943    
944    New working version (110.46).  NEW BOOTFILES!
945    
946    ----------------------------------------------------------------------
947    Name: Matthias Blume (blume (at) tti - c (dot) org)
948    Date: 2004/06/17 17:20:00 CDT
949    Tag: blume-20040617-timer-unix
950    Description:
951    
952    Changed the interface of structures Timer and Unix to match the most
953    recent Basis spec.
954    
955    In the case of Unix there still seems to be an open/weird issue:
956    
957       The {text,bin}{In,Out}streamOf functions are supposed to create
958       fresh streams whenever they are called -- as opposed to have them
959       return the same stream every time.  This design is supposed to
960       prevent space leaks caused by proc values hanging on to streams.
961    
962       The reap function, on the other hand, is supposed to close the
963       streams.  This cannot be done without having a handle on the
964       stream in proc after all...
965    
966    I took the liberty to implement the following stopgap solution:
967    
968      The proc value hangs on to the most recently created stream(s).
969      Reap closes those.  If either or both of the two streams hadn't
970      been created at all yet, then reap will close the corresponding
971      file descriptors directly.
972    
973    PS: I don't understand the original space leak argument anymore.  If
974    a proc hangs on to the imperative stream, then I/O operations on those
975    will advance the state of the cached stream and avoid the space leak.
976    
977    ----------------------------------------------------------------------
978    Name: Matthias Blume (blume (at) tti - c (dot) org)
979    Date: 2004/05/28 16:45:00 CDT
980    Tag: blume-20040528-basis
981    Description:
982    
983    Added signature PACK_REAL and exported functor PrimIO.
984    
985    ----------------------------------------------------------------------
986    Name: Matthias Blume (blume (at) tti - c (dot) org)
987    Date: 2004/05/25 16:00:00 CDT
988    Tag: blume-20040525-group-owner
989    Description:
990    
991    CM now ignores (but still accepts) the "owner" information in group
992    descriptions.  The owner of a group is its next enclosing
993    library. Each group must have a unique owner.  (There is a virtual
994    "toplevel" library that owns groups which are not nested within a real
995    library.)  Previously, each group had to explicitly declare its owner,
996    and CM would check that such a declaration is correct.  The new scheme
997    is to have CM check that for each group there is precisely one owning
998    library.
999    
1000    The advantage of the new scheme is that the programmer no longer needs
1001    to maintain the somewhat annoying owner information.  The downside is
1002    that CM cannot enforce the ownership rule across multiple runs of
1003    CM.make.  Fortunately, enclosing the same group in two different
1004    libraries A and B which are not part of the same program does not
1005    cause real problems.
1006    
1007    ----------------------------------------------------------------------
1008    Name: Matthias Blume (blume (at) tti - c (dot) org)
1009    Date: 2004/05/20 16:00:00 CDT
1010    Tag: blume-20040520-win32
1011    Description:
1012    
1013    Made the win32 version work again.  (Strangely, a misplaced comma had
1014    slipped into win32-process.c which prevented the runtime from being
1015    compiled correctly.)
1016    
1017    Also, included a minor addition to ml-build.bat analogous to what was
1018    done in blume-20040519-ml-build.
1019    
1020    ----------------------------------------------------------------------
1021    Name: Matthias Blume (blume (at) tti - c (dot) org)
1022    Date: 2004/05/19 22:10:00 CDT
1023    Tag: blume-20040519-ml-build
1024    Description:
1025    
1026    Arranged for ml-build to clean up after itself a little bit better.
1027    The script generates a temporary SML source file and compiles it using
1028    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1029    gets rid of those at the end, so they don't accumulate under .cm.
1030    
1031    This required a minor change to install.sh because the name of the
1032    metadata directory (default: .cm) is actually configurable at
1033    installation time.
1034    
1035    ----------------------------------------------------------------------
1036    Name: Matthias Blume (blume (at) tti - c (dot) org)
1037    Date: 2004/05/18 15:50:00 CDT
1038    Tag: blume-20040518-mkreader
1039    Description:
1040    
1041    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1042    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1043    
1044    ----------------------------------------------------------------------
1045    Name: Matthias Blume (blume (at) tti - c (dot) org)
1046    Date: 2004/05/11 14:35:00 CDT
1047    Tag: blume-20040511-win32sock
1048    Description:
1049    
1050    Added previously missing support for many socket-related functions
1051    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1052    for the voluminous patch!
1053    
1054    (I have not tested this patch under win32 yet.)
1055    
1056    Here is David's e-mail:
1057    
1058    Hi,
1059    
1060    Attached to this email you find a diff against sml/nj 110.45
1061    that will enable socket support under Windows.
1062    
1063    To apply the patch (using unix or cygwin)
1064    1) gunzip runtime.diff.gz
1065    2) "cd" into "src/runtime" in the source tree of a fresh
1066       110.45 installation.
1067    3) patch -p 1 < [your/path/to]runtime.diff
1068    
1069    The code compiles fine but has NOT yet been extensively tested.
1070    I only ran a few tests for basic socket client functionality
1071    (which worked fine).  Especially the functions that use ioctl
1072    are not tested at all and might not work (see below).
1073    
1074    I implemented this since we want to move to a newer version of sml/nj
1075    but need socket support in order to use it.  This is the first time I
1076    even had a look at the sml/nj source,  so please review my changes
1077    before making this part of the distribution!  Here are a few issues
1078    that I think might be better for someone to solve who is more
1079    familiar with the sml/nj source (and socket programming):
1080    
1081    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1082      exception since I could not figure out what the windows equivalent
1083      of these functions is
1084    
1085    - In sockets-osdep.h there are a some #include statements that are
1086      only used in a few files that include sockets-osdep.h
1087    
1088    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1089      does not process its return value since I don't know how to
1090      report an error upwards.
1091    
1092    - It would probably be good to have a call to WSACleanup() when
1093      the library is unloaded (if there is such a possibility).
1094      Otherwise I think Windows will take care of this automatically
1095      when the process finishes.
1096    
1097    - I used ioctlsocket() as a replacement for ioctl() but I have
1098      no idea if that is actually the proper replacement on Windows.
1099    
1100    - All these issues are marked in the code by "FIXME" comments.
1101    
1102    We use sml/nj extensively in our products and are quite happy
1103    with it.  I hope this contribution will help you.
1104    
1105    Keep up the good work!
1106    
1107    David
1108    
1109    ----------------------------------------------------------------------
1110    Name: Matthias Blume (blume (at) tti - c (dot) org)
1111    Date: 2004/05/11 14:20:00 CDT
1112    Tag: blume-20040511-installml
1113    Description:
1114    
1115    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1116    the patch.)
1117    
1118    ----------------------------------------------------------------------
1119    Name: Matthias Blume (blume (at) tti - c (dot) org)
1120    Date: 2004/05/11 14:05:00 CDT
1121    Tag: blume-20040511-nlffi-netbsd
1122    Description:
1123    
1124    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1125    the patch.)
1126    
1127    ----------------------------------------------------------------------
1128    Name: Matthias Blume (blume (at) tti - c (dot) org)
1129    Date: 2004/05/11 12:05:00 CDT
1130    Tag: blume-20040511-exports
1131    Description:
1132    
1133    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1134    export lists in compiler-related .cm-files.
1135    
1136    ----------------------------------------------------------------------
1137    Name: Matthias Blume (blume (at) tti - c (dot) org)
1138    Date: 2004/05/11 11:35:00 CDT
1139    Tag: blume-20040511-allsource
1140    Description:
1141    
1142    The installer now honors the "src-smlnj" target again, although its meaning
1143    has changed from "all sources required for the compiler" to "all sources
1144    the installer knows about".  In other words, if you enable "src-smlnj"
1145    in the "targets" file, then the installer will pull in sources for
1146    everything.  (Notice that this refers to source code only.  Compiled
1147    code is still only installed for modules that were requested explicitly
1148    or which are required for other modules that were requested explicitly.)
1149    
1150    ----------------------------------------------------------------------
1151    Name: Matthias Blume (blume (at) tti - c (dot) org)
1152    Date: 2004/04/23 17:40:00 CDT
1153    Tag: blume-20040423-ieee-scan
1154    Description:
1155    
1156    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1157    in the exponent calculation we get INF or ZERO (depending on the mantissa
1158    and the sign of the exponent).
1159    
1160    ----------------------------------------------------------------------
1161    Name: Matthias Blume (blume (at) tti - c (dot) org)
1162    Date: 2004/04/23 10:40:00 CDT
1163    Tag: blume-20040423-ml-build
1164    Description:
1165    
1166    The ml-build script now terminates with a non-0 status when something
1167    goes wrong.
1168    
1169    ----------------------------------------------------------------------
1170    Name: Matthias Blume (blume (at) tti - c (dot) org)
1171    Date: 2004/04/22 16:35:00 CDT
1172    Tag: blume-20040422-Option
1173    Description:
1174    
1175    Made exception Option to be the same as exception Option.Option
1176    (as it should be).
1177    
1178    ----------------------------------------------------------------------
1179    Name: Allen Leung (leunga (at) reservoir (dot) com)
1180    Date: 2004/03/19 14:40:00 EST
1181    Tag: leunga-20040319-cygwin-nlffi
1182    Description:
1183    
1184    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1185    of SML/NJ.  The problem is that
1186    
1187       lib = dlopen(NULL, ...)
1188       f   = dlsym(lib, "malloc");
1189    
1190    does not work on Windows unless we explicitly export symbols
1191    such as 'malloc' during linking.  We fixed this by explicitly
1192    exporting the required symbols with the magic gcc incantation:
1193    
1194        -Wl,--export-all cygwin.def
1195    
1196    where cygwin.def is a file containing all the symbols that
1197    we wish to export.
1198    
1199    I suspect this is a Windows problem and we'll have to
1200    do the same (somehow with windows compilers) when
1201    we build the native win32 version with the system
1202    calls LoadLibrary/GetProcAddress.
1203    
1204    ----------------------------------------------------------------------
1205    Name: Matthias Blume (blume (at) tti - c (dot) org)
1206    Date: 2004/03/04 16:35:00 CST
1207    Tag: blume-20040304-intinf-fmt
1208    Description:
1209    
1210    Fixed problem with IntInf.fmt (sign would show up on the right instead
1211    of on the left for BIN, OCT, and HEX).
1212    
1213    ----------------------------------------------------------------------
1214    Name: Matthias Blume (blume (at) tti - c (dot) org)
1215    Date: 2004/03/04 11:25:00 CST
1216    Tag: blume-20040304-symlinks
1217    Description:
1218    
1219    Fixed problem with installer script (unix only) where bin/ml-yacc and
1220    friends pointed (via symlinks) to absolute locations instead of just
1221    .run-sml.  This was reported by Vesa A Norrman.
1222    
1223    ----------------------------------------------------------------------
1224    Name: Matthias Blume (blume (at) tti - c (dot) org)
1225    Date: 2004/02/13 14:50:00 CST
1226    Tag: Release_110_45
1227    Description:
1228    
1229    New working version (110.45). New bootfiles.
1230    
1231    ----------------------------------------------------------------------
1232    Name: Matthias Blume (blume (at) tti - c (dot) org)
1233    Date: 2004/01/26 15:15:15 CST
1234    Tag: blume-20040126-toplevel
1235    Description:
1236    
1237    Improved handling of exceptions at the interactive toplevel.
1238    
1239    ----------------------------------------------------------------------
1240    Name: Matthias Blume (blume (at) tti - c (dot) org)
1241    Date: 2004/01/26 11:25:00
1242    Tag: blume-20040126-app
1243    Description:
1244    
1245    Type of top-level "app" corrected.
1246    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1247    
1248    ----------------------------------------------------------------------
1249    Name: Matthias Blume (blume (at) tti - c (dot) org)
1250    Date: 2003/11/18 17:10 CST
1251    Tag: blume-20031118-basis-fiddle
1252    Description:
1253    
1254    - changed Timer interface to what might become the spec
1255    - POSIX_FLAGS -> BIT_FLAGS according to spec
1256    - some other minor discrepancies wrt. spec eliminated
1257    
1258    ----------------------------------------------------------------------
1259    Name: Matthias Blume (blume (at) tti - c (dot) org)
1260    Date: 2003/11/06 12:00:00 CST
1261    Tag: Release_110_44
1262    Description:
1263    
1264    New working version (110.44). New bootfiles.
1265    
1266    ----------------------------------------------------------------------
1267    Name: Matthias Blume (blume (at) tti - c (dot) org)
1268    Date: 2003/11/04 11:50:00 CST
1269    Tag: blume-20031104-move-libraries
1270    Description:
1271    
1272    Eliminated the "dont_move_libraries" directive in config/targets.
1273    (The mechanism was broken and could not be fixed easily.  Moreover,
1274    there does not seem to be any reason not to move all libraries into
1275    lib during installation.  I originally implemented this directive as a
1276    backward-compatibility feature when I first introduced the new CM.
1277    Now that things have been stable for a long time and going back to the
1278    old CM is not an option, there is no reason to keep it around.)
1279    
1280    ----------------------------------------------------------------------
1281    Name: Matthias Blume (blume (at) tti - c (dot) org)
1282    Date: 2003/11/03 16:00:00 CST
1283    Tag: blume-20031103-installdir
1284    Description:
1285    
1286    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1287    Richards for pointing out the problem and providing the solution.)
1288    
1289    ----------------------------------------------------------------------
1290    Name: Matthias Blume (blume (at) tti - c (dot) org)
1291    Date: 2003/10/01 17:05:00 CDT
1292    Tag: blume-20031001-lal-mlrisc
1293    Description:
1294    
1295    MLRISC bug fix from Lal.
1296    
1297    ----------------------------------------------------------------------
1298    Name: Matthias Blume (blume (at) tti - c (dot) org)
1299    Date: 2003/09/30 16:10:00 CDT
1300    Tag: blume-20030930-primio-bat
1301    Description:
1302    
1303    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1304    2. Improved .bat files (for Win32 port) to make things work under Win95.
1305       (thanks to Aaron S. Hawley for this one)
1306    
1307    ----------------------------------------------------------------------
1308    Name: Matthias Blume (blume (at) tti - c (dot) org)
1309    Date: 2003/09/26 16:05:00 CDT
1310    Tag: blume-20030926-wrappriv
1311    Description:
1312    
1313    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1314    
1315    ----------------------------------------------------------------------
1316    Name: Matthias Blume (blume (at) tti - c (dot) org)
1317    Date: 2003/09/26 15:00:00 CDT
1318    Tag: blume-20030926-110_43_3
1319    Description:
1320    
1321    - additional cleanup
1322    - version number bump, NEW BOOTFILES
1323    
1324    ----------------------------------------------------------------------
1325    Name: Matthias Blume (blume (at) tti - c (dot) org)
1326    Date: 2003/09/26 12:00:00 CDT
1327    Tag: blume-20030926-ppautoload
1328    Description:
1329    
1330    I modified the read-eval-print loop so that the autoloader gets
1331    invoked whenever the prettyprinter tries to look up a symbol that
1332    is not currently defined in the toplevel environment but which
1333    appears in CM's autoload registry.  As a result, we see far fewer of
1334    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1335    
1336    In addition to this I tried to clean up some pieces of the Basis
1337    implementation (e.g., Socket, Word8Array) in order to prevent other
1338    instances of these ?.Foo.Bar.xxx names from being printed.
1339    
1340    The mechanism that picks names for types still needs some work, though.
1341    (Right now it seems that if there is a type A.t which is defined to
1342    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1343    "?.B.u" although the perhaps more sensible solution would be to use
1344    "A.t" in this case.  In other words, the prettyprinter should follow
1345    a chain of DEFtycs not farther than there are corresponding toplevel
1346    names in the current environment.)
1347    
1348    ----------------------------------------------------------------------
1349    Name: Matthias Blume (blume (at) tti - c (dot) org)
1350    Date: 2003/09/24 16:31:00 CDT
1351    Tag: blume-20030924-installer
1352    Description:
1353    
1354    Another installer tweak:  All the ML code for the installer is now
1355    compiled during CMB.make and put into a little library called
1356    $smlnj/installer.cm.  The installation then simply invokes
1357    
1358       sml -m $smlnj/installer.cm
1359    
1360    and everything happens automagically.
1361    
1362    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1363    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1364            and BINDIR.
1365    
1366    The new scheme guarantees that the ML code responsible for the installation
1367    is in sync with the APIs of the main system.  Also, the installer is
1368    somewhat faster because the installer script is precompiled.
1369    
1370    ----------------------------------------------------------------------
1371    Name: Matthias Blume (blume (at) tti - c (dot) org)
1372    Date: 2003/09/24 15:35:00 CDT
1373    Tag: blume-20030924-synsock
1374    Description:
1375    
1376    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1377    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1378    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1379    implementations of non-blocking operations from CML's Socket
1380    structure.
1381    
1382    ----------------------------------------------------------------------
1383    Name: Matthias Blume (blume (at) tti - c (dot) org)
1384    Date: 2003/09/24 15:10:05 CDT
1385    Tag: blume-20030924-sockets
1386    Description:
1387    
1388    1. Fixed SOCKET API and implementation to match Basis spec.
1389       This required changing the internal representation of sockets to one
1390       that remembers (for each socket file descriptor) whether it is currently
1391       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1392       call is made only if the state actually needs to change).
1393    
1394    2. OS-specific details of sockets were moved into separate files, thus
1395       making it possible to unify the bulk of the socket implementations
1396       between Unix and Win32.
1397    
1398    3. CML's socket API changed accordingly.
1399       (Note that we need to remove non-blocking functions from this API
1400       since they are redundant in the case of CML!)
1401    
1402    4. CML's socket implementation now makes use of non-blocking functions
1403       provided by Basis, thus removing all OS-dependent code from this part
1404       of CML.
1405    
1406    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1407    
1408    ----------------------------------------------------------------------
1409    Name: Matthias Blume (blume (at) tti - c (dot) org)
1410    Date: 2003/09/22 12:10:00 CDT
1411    Tag: blume-20030922-110_43_2
1412    Description:
1413    
1414    Made a new interim version and bootfiles for developer's bootstrapping
1415    convenience.
1416    
1417       110.43.2 -- NEW BOOTFILES
1418    
1419    ----------------------------------------------------------------------
1420    Name: Matthias Blume (blume (at) tti - c (dot) org)
1421    Date: 2003/09/19 15:55:00 CDT
1422    Tag: blume-20030919-cmdir
1423    Description:
1424    
1425    1. new-install.sh -> install.sh
1426    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1427    3. tweaked installer so that another name instead of .cm can be chosen
1428       at install time (by setting the CM_DIR_ARC environment variable
1429       during installation); once installation is complete, the name is
1430       fixed
1431    
1432    ----------------------------------------------------------------------
1433    Name: Matthias Blume (blume (at) tti - c (dot) org)
1434    Date: 2003/09/18 16:00:00 CDT
1435    Tag: blume-20030918-110_43_1
1436    Description:
1437    
1438    Made a new interim version and bootfiles for developer's bootstrapping
1439    convenience.
1440    
1441       110.43.1 -- NEW BOOTFILES
1442    
1443    ----------------------------------------------------------------------
1444    Name: Matthias Blume (blume (at) tti - c (dot) org)
1445    Date: 2003/09/18 15:20:00 CDT
1446    Tag: blume-20030918-misc
1447    Description:
1448    
1449    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1450       this seems to be controversial at the moment) and used those in
1451       Posix.ProcEnv.times.
1452    
1453    2. Added Time.{from,to}Nanoseconds to Time.
1454    
1455    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1456       For example, fromLargeInt never needs to look at more than 3 "big
1457       digits" to get its 53 bits of precision.
1458    
1459    ----------------------------------------------------------------------
1460    Name: Matthias Blume (blume (at) tti - c (dot) org)
1461    Date: 2003/09/17 16:30:00 CDT
1462    Tag: blume-20030917-real32-slices
1463    Description:
1464    
1465    Added an entry to the primitive environment
1466    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1467    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1468    
1469    Removed some of the "magic" constants in real64.sml and replaced them
1470    with code that generates these values from their corresponding
1471    integer counterparts.
1472    
1473    Made all(?) the slice-related changes to the Basis and made everything
1474    compile again...
1475    
1476    ----------------------------------------------------------------------
1477    Name: Matthias Blume (blume (at) tti - c (dot) org)
1478    Date: 2003/09/15 17:45:00 CDT
1479    Tag: blume-20030915-rbase
1480    Description:
1481    
1482    Fixed bug in Real.fromLargeInt.
1483    
1484    ----------------------------------------------------------------------
1485    Name: Matthias Blume (blume (at) tti - c (dot) org)
1486    Date: 2003/09/13 18:11:00 CDT
1487    Tag: blume-20030913-libinstall
1488    Description:
1489    
1490    Minor bugfix in config/libinstall (set anchor with path to
1491    standalone tool after installing it, otherwise libraries that
1492    need ml-lex or ml-yacc won't compile the first time the installer
1493    runs).
1494    
1495    ----------------------------------------------------------------------
1496    Name: Matthias Blume (blume (at) tti - c (dot) org)
1497    Date: 2003/09/12 11:45:00 CDT
1498    Tag: blume-20030912-various
1499    Description:
1500    
1501     - fixed bug in Real.toLargeInt
1502     - fixed bug in Posix.ProcEnv.times
1503     - changed inputLine functions to return an option
1504     - minor installer improvements / bugfixes
1505     - changed default @SMLalloc parameter for x86/celeron to 64k
1506    
1507    ----------------------------------------------------------------------
1508    Name: Matthias Blume (blume (at) tti - c (dot) org)
1509    Date: 2003/09/09 22:00:00 CDT
1510    Tag: Release_110_43
1511    Description:
1512    
1513        New working release 110.43.  New bootfiles.
1514    
1515    ----------------------------------------------------------------------
1516    Name: Matthias Blume (blume (at) tti - c (dot) org)
1517    Date: 2003/09/09 19:20:00 CDT
1518    Tag: blume-20030909-installer
1519    Description:
1520    
1521    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1522    Modified config/install.bat to take advantage of it.  Also modified
1523    config/install.sh (and called it config/new-install.sh) to take advantage
1524    of it on Unix systems.  (The SML code is (supposed to be) platform-
1525    independent.)
1526    
1527    The installer can now install everything under Win32
1528    as well as under *nix as long as it compiles.
1529    
1530    Other changes:
1531    
1532      - made CML compile again under Win32
1533      - made eXene compile under Win32 (by providing a fake structure UnixSock
1534        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1535      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1536        same as type int;  under Win32 it isn't
1537      - fixed some slice-related problems in the win32-specific parts of CML
1538      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1539        (under Win32, the volume name is case-insensitive, and the
1540        OS.Path code compares volume names for equality)
1541    
1542    ----------------------------------------------------------------------
1543    Name: Matthias Blume (blume (at) tti - c (dot) org)
1544    Date: 2003/09/08 11:55:00 CDT
1545    Tag: blume-20030908-fullpath
1546    Description:
1547    
1548    Made Win32 version of OS.FileSys.fullPath return current directory
1549    when given an empty string.  This is what the spec says, and incidentally,
1550    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1551    cases when presented with the name of a non-existing .cm file.)
1552    
1553    ----------------------------------------------------------------------
1554    Name: Matthias Blume (blume (at) tti - c (dot) org)
1555    Date: 2003/09/04 16:30:00 CDT
1556    Tag: blume-20030905-slices-etc
1557    Description:
1558    
1559      1. Changed interface to vectors and arrays in Basis to match
1560         (draft) Basis spec.
1561      2. Added signatures and implementations of slices according to
1562         Basis spec.
1563      3. Edited source code throughout the system to make it compile again
1564         under 1. and 2.  (In some cases code had to be added to have it
1565         match the new signatures.)
1566      4. MLRISC should be backward-compatible: the copies of the originals
1567         of files that needed to change under 3. were retained, the .cm files
1568         check the compiler version number and use old versions when
1569         appropriate.
1570      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1571         dirstream -> string option (in accordance with Basis spec).
1572      6. When generating code that counts lines, ml-lex used function
1573         CharVector.foldli, taking advantage of its old interface.
1574         This has been replaced with the corresponding code from
1575         CharVectorSlice.  (html-lex must be re-lexed!)
1576      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1577         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1578      8. Removed temporary additions (fromInternal, toInternal) from the
1579         (now obsolete) IntInf in smlnj-lib/Util.
1580      9. Cleaned up structure Byte.
1581     10. Added localOffset, scan, and fromString to Date (according to spec).
1582         Cleaned/corrected implementation of Date.
1583         (Still need to check for correctness; implement better canonicalizeDate.)
1584     11. Added "scan" to signature IEEE_REAL.
1585     12. Some improvements to IntInf [in particular: efficiency-hack for
1586         mod and rem when second operand is 2 (for parity checks).]
1587     13. Changed representation of type Time.time, using a single IntInf.int
1588         value counting microseconds.  This considerably simplified the
1589         implementation of structure Time.  We now support negative time
1590         values; scan and fromString handle signs.
1591     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1592         ArraySlice).
1593    
1594    ----------------------------------------------------------------------
1595    Name: Matthias Blume (blume (at) tti - c (dot) org)
1596    Date: 2003/08/28 17:00:00 CDT
1597    Tag: blume-20030828-intinf
1598    Description:
1599    
1600    This is a major update which comes with a version number bump
1601    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1602    and an implementation of IntInf in the Basis.
1603    
1604    There are a fairly large number of related changes and updates throughout
1605    the system:
1606    
1607    Basis:
1608            - Implemented IntInf.
1609            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1610            - Added some missing Real64 operations, most notably Real.toLargeInt.
1611            - Added FixedInt as a synonym for Int32.
1612    
1613    compiler:
1614            * Added support for a built-in intinf type.
1615              - literals
1616              - pattern matching
1617              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1618              - overloading on literals and operations
1619    
1620            This required adding a primitive type intinf, some additional
1621            primops, and implementations for several non-trivial intinf
1622            operations in Core.  (The intinf type is completely abstract
1623            to the compiler; all operations get delegated back to the Core.)
1624    
1625            * Intinf equality is handled by polyequal.  However, the compiler
1626            does not print its usual warning in this case (since polyequal
1627            is the right thing to do there).
1628    
1629            * Improved the organization of structure InlineT.
1630    
1631            * A word about conversion primops:
1632            If conversions involving intinf do not cancel out during
1633            CPS contract, then the compiler must insert calls to Core functions.
1634            Since all core access must be resolved already during the FLINT
1635            translate phase, it would be too late a the time of CPS contract
1636            to add new Core calls.  For this reason, conversion primops
1637            for intinf carry two arguments:  1. the numeric argument that
1638            they are supposed to convert, and 2. the Core function that
1639            can help with this conversion if necessary.  If CPS contract
1640            eliminates a primop, then the associated Core function becomes
1641            dead and goes away.  Intinf conversion primops that do not get
1642            eliminated by CPS contract get rewritten into calls of their
1643            core functions by a separate, new phase.
1644    
1645    interactive system:
1646            - Control.Print.intinfDepth controls max length of intinf constants
1647            being printed. (Analogous to Control.Print.stringDepth.)
1648            - Cleanup in printutil and pputil: got rid of unused stuff and
1649            duplicates; replaced some of the code with code that makes better
1650            use of library functionality.
1651    
1652    CM:
1653            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1654            are no longer silent.
1655    
1656    CKIT:
1657            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1658            in favor of LargeInt -- which is now the same as IntInf.
1659            CKIT-knowledgable people should check whether this is what's
1660            intended and otherwise change things back to using Int32 or
1661            FixedInt.
1662    
1663    Throughout the code:
1664            Started using IntInf.int literals and built-in operations
1665            (e.g., comparison with 0) where this seems appropriate.
1666    
1667    
1668    ----------------------------------------------------------------------
1669    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1670    Date: 2003/08/13 11:36:00 CDT
1671    Tag: dbm-20030813-mcz-merge1
1672    Description:
1673    
1674    Merging changes from the mcz-branch development branch into trunk.
1675    These changes involve replacement of the emulated old prettyprinter
1676    interface with direct use of the SML/NJ Lib PP library, and fixing
1677    of a couple of bugs (895, 1186) relating to error messages.  A new
1678    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1679    has been added.
1680    
1681    ----------------------------------------------------------------------
1682    Name: Matthias Blume (blume (at) tti - c (dot) org)
1683    Date: 2003/08/11 15:45:00 CDT
1684    Tag: blume-20030811-windows
1685    Description:
1686    
1687    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1688    
1689        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1690    
1691    This patch restores SML/NJ's ability to run under win32.  There are a
1692    number of changes, including fixes for several bugs that had gone
1693    unnoticed until now:
1694    
1695      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1696        trying to run the win32 version from within cygwin.)
1697    
1698      - There are a number of simple .bat scripts that substitute for their
1699        corresponding Unix shell-scripts.  (See below.)
1700    
1701      - The internals of ml-build have been modified slightly.  The main
1702        difference is that instead of calling ".link-sml" (or link-sml.bat)
1703        using OS.Process.system, the ML process delegates this task back
1704        to the script.  Otherwise problems arise in mixed environments such
1705        as Cygwin where scripts look and work like Unix scripts, but
1706        where OS.Process.system cannot run them.
1707    
1708      - In CM, the srcpath pickler used native pathname syntax -- which
1709        is incorrect in the case of cross-compilation.  The new pickle format
1710        is independent of platform-specific naming conventions.
1711    
1712      - Path configuration files (such as lib/pathconfig) can now choose
1713        between native and standard syntax.  Placing a line of the form
1714    
1715            standard!
1716    
1717        into the file causes all subsequent paths to be interpreted using
1718        CM standard pathname syntax (= Unix conventions); a line
1719    
1720            native!
1721    
1722        switches back to native style.  This was needed so that
1723        path config files can be written portably, see src/system/pathconfig.
1724    
1725      - Runtime system:
1726    
1727          - win32-filesys.c:  get_file_time and set_file_time now
1728            access modification time, not creation time.
1729    
1730          - I/O code made aware of new array representation.
1731    
1732          - Bug fixes in X86.prim.masm.
1733    
1734       - src/system/makeml made aware of win32. (For use under cygwin
1735            and other Unix-environments for windows.)
1736    
1737       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1738         which caused BinIO.inputAll to fail consistently.
1739    
1740    .bat scripts:
1741    
1742       Windows .bat scripts assume that SMLNJ_HOME is defined.
1743    
1744       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1745         applications (sml, ml-yacc, ml-lex).
1746       - ml-build.bat: analogous to ml-build.
1747       - config\install.bat:  Analogous to config/install.sh.  This requires
1748         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1749         (nmake etc. must be on the path, and vcvars32 must have been run.)
1750         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1751         and the bootfile hierarchy must have been unpacked under
1752         sml.boot.x86-win32.
1753         The script is very primitive and does a poor job at error checking.
1754         It only installs the base system, ml-lex, and ml-yacc.  No other
1755         libraries are being installed (i.e., you get only those that
1756         are part of the compiler.)
1757       - link-sml.bat:  analogous to .link-sml, but not currently used
1758    
1759    Unrelated bug fixes:
1760    
1761      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1762        types.
1763      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1764        otherwise.)
1765    
1766    ----------------------------------------------------------------------
1767    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1768    Date: 2003/06/17
1769    Tag: macqueen-20030617-bug895
1770    Description:
1771    
1772    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1773    Tag will be used for new development branch (mcz-branch) for
1774    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1775    
1776    ----------------------------------------------------------------------
1777    Name: Matthias Blume (blume (at) tti - c (dot) org)
1778    Date: 2003/05/27 16:55:00 CDT
1779    Tag: blume-20030527-polyeq
1780    Description:
1781    
1782    Tried to eliminated most cases of polymorphic equality.
1783    
1784    ----------------------------------------------------------------------
1785    Name: Matthias Blume (blume (at) tti - c (dot) org)
1786    Date: 2003/05/21 17:45:00 CDT
1787    Tag: blume-20030517-complete
1788    Description:
1789    
1790    Two changes:
1791    
1792      1. Added a flag for controlling whether non-exhaustive bindings will
1793         be treated as errors (default is false).
1794      2. Cleaned up the *entire* source tree so that CMB.make goes through
1795         without a single non-exhaustive match- or bind warning.
1796    
1797    ----------------------------------------------------------------------
1798    Name: Matthias Blume (blume (at) tti - c (dot) org)
1799    Date: 2003/05/17 10:20:00 CDT
1800    Tag: blume-20030517-absyn
1801    Description:
1802    
1803    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1804    
1805       This mainly affects the quality of error messages.  However, some
1806       of the code is now more straightforward than before. (Treatment of
1807       the above four constructs in translate.sml is much simpler than
1808       the "macro-expansion" that was going on before.  Plus, the mach-
1809       compiler no longer gets invoked just to be able to compile an
1810       if-expression.)
1811    
1812    2. The ErrorMsg.Error exception is now caught and absorbed by the
1813       interactive loop.
1814    
1815    ----------------------------------------------------------------------
1816    Name: Allen Leung
1817    Date: 2003/05/16 13:05:00 CDT
1818    Tag: leunga-20030516-cygwin-runtime
1819    Description:
1820    
1821        Ported the runtime system to cygwin, which uses the unix
1822    x86-unix bin files.  Missing/buggy features:
1823    
1824       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1825         the Cygwin library.
1826       o Ctrl-C handling may be flaky.
1827       o Windows system calls and Windows I/O are not supported.
1828    
1829       A new set of binfiles is located at:
1830    
1831         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1832    
1833       This is only needed for bootstrapping the cygwin version of smlnj.
1834       Other x86 versions can use the existing binfiles.
1835    
1836    ----------------------------------------------------------------------
1837    Name: Matthias Blume
1838    Date: 2003/04/08 15:42:00 CDT
1839    Tag: blume-20030408-listpair
1840    Description:
1841    
1842    1. Added a target 'mlrisc' to installer.
1843    
1844    2. Added missing elements to structure ListPair.
1845    
1846    ----------------------------------------------------------------------
1847    Name: Allen Leung
1848    Date: 2003/01/07 10:40:00 EST
1849    Tag: leunga-20030107-int-rem
1850    Description:
1851    
1852    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1853    The arguments to the SUBL instruction were swapped.
1854    
1855    ----------------------------------------------------------------------
1856    Name: Matthias Blume
1857    Date: 2002/12/12 16:25:00 EST
1858    Tag: blume-20021212-risc-ra
1859    Description:
1860    
1861    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1862    sent the RA into an infinite loop when floating point registers get
1863    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1864    110.37 and 110.38.)
1865    There was another set of potential problems related to the handling of
1866    MLRISC annotations (but those did not yet cause real problems, apparently).
1867    
1868    ----------------------------------------------------------------------
1869    Name: Matthias Blume
1870    Date: 2002/12/06 22:40:00 EST
1871    Tag: blume-20021206-cm-fileid
1872    Description:
1873    
1874    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1875    This fixes the problem of CM getting confused by files that suddenly
1876    change their identity (e.g., by getting unlinked and recreated by some
1877    text editor such as vi).  There might be a better/cheaper/cleaner way
1878    of doing this, but for now this will have to do.
1879    
1880    ----------------------------------------------------------------------
1881    Name: Matthias Blume
1882    Date: 2002/10/28 09:50:00 EST
1883    Tag: blume-20021028-typecheck
1884    Description:
1885    
1886    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1887    
1888    ----------------------------------------------------------------------
1889    Name: Matthias Blume
1890    Date: 2002/10/17 09:10:00 EDT
1891    Tag: Release_110_42
1892    Description:
1893    
1894    In good old tradition, there has been a slight hiccup so that we have
1895    to patch 110.42 after the fact.   The old release tag has been replaced
1896    (see below).
1897    
1898    The change solves a problem with two competing approaches the
1899    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1900    each other's way.
1901    
1902    This change only affects the runtime system code and the installer script.
1903    (No new bootfiles.)
1904    
1905    ----------------------------------------------------------------------
1906    Name: Matthias Blume
1907    Date: 2002/10/16 12:00:00 EDT
1908    Tag: Release_110_42_removed
1909    Description:
1910    
1911    New working release.  New bootfiles.
1912    
1913    ----------------------------------------------------------------------
1914    Name: Matthias Blume
1915    Date: 2002/10/10 13:10:00 EDT
1916    Tag: blume-20021010-ppc-divs
1917    Description:
1918    
1919    The mltree operator DIVS must be implemented with an overflow check on
1920    the PPC because the hardware indicates divide-by-zero using "overflow" as
1921    well.
1922    
1923    ----------------------------------------------------------------------
1924    Name: Matthias Blume
1925    Date: 2002/07/23 11:45:00
1926    Tag: blume-20020723-smlnj-home
1927    Description:
1928    
1929    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1930    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1931    CM_PATHCONFIG is also set), the path configuration file is assumed
1932    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1933    the entire tree to some other place and everything will "just work".
1934    
1935    (Companion commands such as ml-build and ml-makedepend also sense this
1936    variable.)
1937    
1938    ----------------------------------------------------------------------
1939    Name: Matthias Blume
1940    Date: 2002/07/12 21:19:00 EDT
1941    Tag: blume-20020712-liveness
1942    Description:
1943    
1944    Exported two useful "step" functions from liveness module (MLRISC).
1945    
1946    ----------------------------------------------------------------------
1947    Name: Matthias Blume
1948    Date: 2002/07/05 16:00 EDT
1949    Tag: Release_110_41
1950    Description:
1951    
1952    New working release.  New bootfiles.
1953    
1954    ----------------------------------------------------------------------
1955    Name: Matthias Blume
1956    Date: 2002/07/05 10:25:00 EDT
1957    Tag: blume-20020705-btimp
1958    Description:
1959    
1960    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1961    other clients can set up backtracing support.
1962    
1963    ----------------------------------------------------------------------
1964    Name: Matthias Blume
1965    Date: 2002/06/25 17:23:00 EDT
1966    Tag: blume-20020625-fpmax
1967    Description:
1968    
1969    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1970    case.  (The sense of the isNaN test was reversed -- which made min and
1971    max always return their first argument.)
1972    
1973    ----------------------------------------------------------------------
1974    Name: Matthias Blume
1975    Date: 2002/06/11
1976    Tag: blume-20020611-unixpath
1977    Description:
1978    
1979    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1980    
1981    ----------------------------------------------------------------------
1982    Name: Matthias Blume
1983    Date: 2002/06/10 16:35:00 EDT
1984    Tag: blume-20020610-ieeereal
1985    Description:
1986    
1987    I back-ported my implementation of IEEEReal.fromString from the
1988    idlbasis-devel branch so that we can test it.
1989    
1990    Another small change is that ppDec tries to give more information
1991    than just "<sig>" in the case of functors.  However, this code is
1992    broken in some mysterious way if the functor's body's signature
1993    has not been declared by ascription but gets inferred from the
1994    implementation.  This needs fixing...
1995    
1996    ----------------------------------------------------------------------
1997    Name: Matthias Blume
1998    Date: 2002/05/31
1999    Tag: blume-20020531-btrace-mode
2000    Description:
2001    
2002    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2003    the wayside when I switched over to using Controls everywhere.)
2004    
2005    ----------------------------------------------------------------------
2006    Name: Lal George
2007    Date: 2002/05/23 12:21:40 EDT
2008    Tag: george-20020523-visual-labels
2009    Description:
2010    
2011    Labels are now displayed in the graphical output to make
2012    the fall-through and target blocks obvious.
2013    
2014    ----------------------------------------------------------------------
2015    Name: Matthias Blume
2016    Date: 2002/05/22 11:03:00 EDT
2017    Tag: blume-20020522-shrink
2018    Description:
2019    
2020    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2021    situation that comes up when allocating huge arrays.
2022    
2023    ----------------------------------------------------------------------
2024    Name: Matthias Blume
2025    Date: 2002/05/21 16:00:00 EDT
2026    Tag: Release_110_40
2027    Description:
2028    
2029    New working release (110.40).  New bootfiles.
2030    
2031    [Also: John Reppy fixed GC bug 1131.]
2032    
2033    ----------------------------------------------------------------------
2034    Name: Matthias Blume
2035    Date: 2002/05/21 12:35:00 EDT
2036    Tag: blume-20020521-cmdoc
2037    Description:
2038    
2039    CM documentation update.
2040    
2041    ----------------------------------------------------------------------
2042    Name: Matthias Blume
2043    Date: 2002/05/21 10:55:00 EDT
2044    Tag: blume-20020521-misc
2045    Description:
2046    
2047    - John tweaked runtime to be silent on heap export (except when
2048      GC messages are on).
2049    - I added a few more things (cross-compiling versions of CMB) to
2050      config/preloads (as suggestions).
2051    
2052    ----------------------------------------------------------------------
2053    Name: Matthias Blume
2054    Date: 2002/05/20 22:25:00 EDT
2055    Tag: blume-20020520-controls
2056    Description:
2057    
2058    - Added ControlUtil structure to control-lib.cm.
2059    - Use it throughout.
2060    - Used Controls facility to define MLRISC controls (as opposed to
2061      registering MLRISC control ref cells with Controls after the fact)
2062    - Fixed messed-up controls priorities.
2063    
2064    * Removed again all the stuff from config/preloads that one wouldn't
2065      be able to preload at the time the initial heap image is built.
2066      (Many libraries, e.g., CML, do not exist yet at this time.  The
2067       only libraries that can be preloaded via config/preloads are those
2068       that come bundled with the bootfiles.)
2069    
2070    ----------------------------------------------------------------------
2071    Name: Matthias Blume
2072    Date: 2002/05/20 10:59:00 EDT
2073    Tag: blume-20020520-preloads
2074    Description:
2075    
2076    Added a lot of commented-out suggestions for things to be included
2077    in config/preloads.
2078    
2079    ----------------------------------------------------------------------
2080    Name: Allen Leung
2081    Date: 2002/05/18 14:20:00 EDT
2082    Tag: leunga-20020518-mdl
2083    Description:
2084    
2085    o Made the mdl tool stuff compile and run again.
2086    o I've disabled all the stuff that depends on RTL specifications; they
2087      are all badly broken anyway.
2088    
2089    ----------------------------------------------------------------------
2090    Name: Matthias Blume
2091    Date: 2002/05/17 16:49:00 EDT
2092    Tag: blume-20020517-controls
2093    Description:
2094    
2095    0. John Reppy made several modifications to the SML/NJ library.
2096       In particular, there is a shiny new controls-lib.cm.
2097    
2098    1. Pushed new controls interface through compiler so that everything
2099       compiles again.
2100    
2101    2. Added FormatComb and FORMAT_COMB to the CML version of the
2102       SML/NJ library (so that CML compiles again).
2103    
2104    3. Modified init scripts because XXX_DEFAULT environment variables
2105       are no longer with us.  (Boot-time initialization is now done
2106       using the same environment variables that are also used for
2107       startup-time initialization of controls.)
2108    
2109    ----------------------------------------------------------------------
2110    Name: Lal George
2111    Date: 2002/05/15 09:20:10 EDT
2112    Tag: george-20020515-pseudo-op-decls
2113    Description:
2114    
2115            All pseudo-ops emitted before the first segment declaration
2116    such as TEXT, DATA, and BSS directives are assumed to be global
2117    declarations and are emitted first in the assembly file. This is
2118    useful in a number of situations where one has pseudo-ops that are not
2119    specific to any segment, and also works around the constraint that one
2120    cannot have client pseudo-ops in the TEXT segment.
2121    
2122            Because no segment is associated with these declarations it is
2123    an error to allocate any space or objects before the first segment
2124    directive and an exception will be raised. However, we cannot make
2125    this check for client pseudo-ops.
2126    
2127            These top level declarations are a field in the CFG graph_info.
2128    In theory you can continue to add to this field after the CFG has been
2129    built -- provided you know what you are doing;-)
2130    
2131    ----------------------------------------------------------------------
2132    Name: Matthias Blume
2133    Date: 2002/05/13 16:40:00 EDT
2134    Tag: blume-20020513-pp-etc
2135    Description:
2136    
2137    A few minor bugfixes:
2138    
2139      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2140        (Bogus printouts for redefined bindings still occur.  Compiler
2141        bug should no longer occur now.  We need to redo the prettyprinter
2142        from scratch.)
2143    
2144      - CM pathname printer now also adds escape sequences for ( and )
2145    
2146      - commend and docu fixes for ml-nlffi
2147    
2148    ----------------------------------------------------------------------
2149    Name: Matthias Blume
2150    Date: 2002/05/10 16:40:00 EDT
2151    Tag: blume-20020510-erg-textio
2152    Description:
2153    
2154    Applied the following bugfix provided by Emden Gansner:
2155    
2156        Output is corrupted when outputSubstr is used rather than output.
2157    
2158        The problem occurs when a substring
2159    
2160            ss = (s, dataStart, dataLen)
2161    
2162        where dataStart > 0, fills a stream buffer with avail bytes left.
2163        avail bytes of s, starting at index dataStart, are copied into the
2164        buffer, the buffer is flushed, and then the remaining dataLen-avail
2165        bytes of ss are copied into the beginning of the buffer. Instead of
2166        starting this copy at index dataStart+avail in s, the current code
2167        starts the copy at index avail.
2168    
2169        Fix:
2170        In text-io-fn.sml, change line 695 from
2171             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2172        to
2173             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2174    
2175    ----------------------------------------------------------------------
2176    Name: Matthias Blume
2177    Date: 2002/04/12 13:55:00 EDT
2178    Tag: blume-20020412-assyntax
2179    Description:
2180    
2181    1. Grabbed newer assyntax.h from the XFree86 project.
2182    2. Fiddled with how to compile X86.prim.asm without warnings.
2183    3. (Very) Minor cleanup in CM.
2184    
2185    ----------------------------------------------------------------------
2186    Name: Matthias Blume
2187    Date: 2002/04/01 (no joke!) 17:07:00 EST
2188    Tag: blume-20020401-x86div
2189    Description:
2190    
2191    Added full support for div/mod/rem/quot on the x86, using the machine
2192    instruction's two results (without clumsily recomputing the remainder)
2193    directly where appropriate.
2194    
2195    Some more extensive power-of-two support was added to the x86 instruction
2196    selector (avoiding expensive divs, mods, and muls where they can be
2197    replaced with cheaper shifts and masks).  However, this sort of thing
2198    ought to be done earlier, e.g., within the CPS optimizer so that
2199    all architectures benefit from it.
2200    
2201    The compiler compiles to a fixed point, but changes might be somewhat
2202    fragile nevertheless.  Please, report any strange things that you might
2203    see wrt. div/mod/quot/rem...
2204    
2205    ----------------------------------------------------------------------
2206    Name: Matthias Blume
2207    Date: 2002/03/29 17:22:00
2208    Tag: blume-20020329-div
2209    Description:
2210    
2211    Fixed my broken div/mod logic.  Unfortunately, this means that the
2212    inline code for div/mod now has one more comparison than before.
2213    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2214    The problem was with quotient = 0, because that alone does not tell
2215    us which way the rounding went.  One then has to look at whether
2216    remainder and divisor have the same sign...  :(
2217    
2218    Anyway, I replaced the bootfiles with fresh ones...
2219    
2220    ----------------------------------------------------------------------
2221    Name: Matthias Blume
2222    Date: 2002/03/29 14:10:00 EST
2223    Tag: blume-20020329-inlprims
2224    Description:
2225    
2226    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2227    
2228    Primops have changed. This means that the bin/boot-file formats have
2229    changed as well.
2230    
2231    To make sure that there is no confusion, I made a new version.
2232    
2233    
2234    CHANGES:
2235    
2236    * removed REMT from mltree (remainder should never overflow).
2237    
2238    * added primops to deal with divisions of all flavors to the frontend
2239    
2240    * handled these primops all the way through so they map to their respective
2241      MLRISC support
2242    
2243    * used these primops in the implementation of Int, Int32, Word, Word32
2244    
2245    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2246    
2247    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2248    
2249    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2250      floating point
2251    
2252    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2253      and Real (but Real.abs maps to a separate floating-point-only primop)
2254    
2255    
2256    TODO items:
2257    
2258    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2259      instructions because the machine instruction encoder cannot handle
2260      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2261      REM in the first place, and REM is supported by the x86 machine coder.)
2262    
2263    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2264      selection phase.  (The two can be streamlined because the hardware
2265      delivers both quotient and remainder at the same time anyway.)
2266    
2267    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2268      (Currently the behavior is inconsistent both across architectures and
2269      wrt. the draft Basis spec.)
2270    
2271    * Word8 should eventually be handled natively, too.
2272    
2273    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2274      as if there currently is no execution path that could trigger it in
2275      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2276      hold for things like multiplication and division.)
2277    
2278    ----------------------------------------------------------------------
2279    Name: Matthias Blume
2280    Date: 2002/03/27 16:27:00 EST
2281    Tag: blume-20020327-mlrisc-divisions
2282    Description:
2283    
2284    Added support for all four division operations (ML's div, mod, quot,
2285    and rem) to MLRISC.  In the course of doing so, I also rationalized
2286    the naming (no more annoying switch-around of DIV and QUOT), by
2287    parameterizing the operation by div_rounding_mode (which can be either
2288    DIV_TO_ZERO or DIV_TO_NEGINF).
2289    
2290    The generic MLTreeGen functor takes care of compiling all four
2291    operations down to only round-to-zero div.
2292    
2293    Missing pieces:
2294    
2295      * Doing something smarter than relying on MLTreeGen on architectures
2296        like, e.g., the x86 where hardware division delivers both quotient and
2297        remainder at the same time.  With this, the implementation of the
2298        round-to-neginf operations could be further streamlined.
2299    
2300      * Remove inlining support for div/mod/rem from the frontend and replace it
2301        with primops that get carried through to the backend.  Do this for all
2302        int and word types.
2303    
2304    ----------------------------------------------------------------------
2305    Name: Matthias Blume
2306    Date: 2002/03/25 17:25:00 EST
2307    Tag: blume-20020325-divmod
2308    Description:
2309    
2310    I improved (hopefully without breaking them) the implementation of Int.div,
2311    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2312    advantage of the following observations:
2313    
2314      Let  q = x quot y      r = x rem y
2315           d = x div  y      m = x mod y
2316    
2317    where "quot" is the round-to-zero version of integer division that
2318    hardware usually provides.  Then we have:
2319    
2320         r = x - q * y        where neither the * nor the - will overflow
2321         d = if q >= 0 orelse x = q * y then q else q - 1
2322                              where neither the * nor the - will overflow
2323         m = if q >= 0 orelse r = 0 then r else r + y
2324                              where the + will not overflow
2325    
2326    This results in substantial simplification of the generated code.
2327    The following table shows the number of CFG nodes and edges generated
2328    for
2329            fun f (x, y) = x OPER y
2330            (* with OPER \in div, mod, quot, rem *)
2331    
2332    
2333        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2334        --------------------------------------------------------
2335         div |         24 |         39 |         12 |         16
2336         mod |         41 |         71 |         12 |         16
2337        quot |          8 |         10 |          8 |         10
2338         rem |         10 |         14 |          8 |         10
2339    
2340    
2341    ----------------------------------------------------------------------
2342    Name: Matthias Blume
2343    Date: 2002/03/25 22:06:00 EST
2344    Tag: blume-20020325-cprotobug
2345    Description:
2346    
2347    Fixed a bug in cproto (c prototype decoder).
2348    
2349    ----------------------------------------------------------------------
2350    Name: Matthias Blume
2351    Date: 2002/03/25 16:00:00 EST
2352    Tag: blume-20020325-raw-primops
2353    Description:
2354    
2355    I did some cleanup to Allen's new primop code and
2356    replaced yesterday's bootfiles with new ones.
2357    (But they are stored in the same place.)
2358    
2359    ----------------------------------------------------------------------
2360    Name: Matthias Blume
2361    Date: 2002/03/24 22:40:00 EST
2362    Tag: blume-20020324-bootfiles
2363    Description:
2364    
2365    Made the bootfiles that Allen asked for.
2366    
2367    ----------------------------------------------------------------------
2368    Name: Allen Leung
2369    Date: 2002/03/23 15:50:00 EST
2370    Tag: leunga-20020323-flint-cps-rcc-primops
2371    Description:
2372    
2373      1. Changes to FLINT primops:
2374    
2375        (* make a call to a C-function;
2376         * The primop carries C function prototype information and specifies
2377         * which of its (ML-) arguments are floating point. C prototype
2378         * information is for use by the backend, ML information is for
2379         * use by the CPS converter. *)
2380      | RAW_CCALL of { c_proto: CTypes.c_proto,
2381                       ml_args: ccall_type list,
2382                       ml_res_opt: ccall_type option,
2383                       reentrant : bool
2384                     } option
2385       (* Allocate uninitialized storage on the heap.
2386        * The record is meant to hold short-lived C objects, i.e., they
2387        * are not ML pointers.  With the tag, the representation is
2388        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2389        *)
2390      | RAW_RECORD of {tag:bool,sz:int}
2391      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2392    
2393      2.  These CPS primops are now overloaded:
2394    
2395           rawload of {kind:numkind}
2396           rawstore of {kind:numkind}
2397    
2398          The one argument form is:
2399    
2400             rawload {kind} address
2401    
2402          The two argument form is:
2403    
2404             rawload {kind} [ml object, byte-offset]
2405    
2406      3. RAW_CCALL/RCC now takes two extra arguments:
2407    
2408         a. The first is whether the C call is reentrant, i.e., whether
2409            ML state should be saved and restored.
2410         b. The second argument is a string argument specifying the name of
2411            library and the C function.
2412    
2413         These things are currently not handled in the code generator, yet.
2414    
2415      4. In CProto,
2416    
2417         An encoding type of "bool" means "ml object" and is mapped into
2418         C prototype of PTR.  Note that "bool" is different than "string",
2419         even though "string" is also mapped into PTR, because "bool"
2420         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2421    
2422      5. Pickler/unpicker
2423    
2424         Changed to handle RAW_RECORD and newest RAW_CCALL
2425    
2426      6. MLRiscGen,
2427    
2428         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2429         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2430            in the file CodeGen/cpscompile/cps-c-calls.sml
2431    
2432      7. Added the conditional move operator
2433    
2434             condmove of branch
2435    
2436         to cps.  Generation of this is still buggy so it is currently
2437         disabled.
2438    
2439    ----------------------------------------------------------------------
2440    Name: Lal George
2441    Date: 2002/03/22 14:18:25 EST
2442    Tag: george-20020322-cps-branch-prob
2443    Description:
2444    
2445    Implemented the Ball-Larus branch prediction-heuristics, and
2446    incorporated graphical viewers for control flow graphs.
2447    
2448    Ball-Larus Heuristics:
2449    ---------------------
2450    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2451    
2452    By design it uses the Dempster-Shafer theory for combining
2453    probabilities.  For example, in the function:
2454    
2455        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2456    
2457    the ball-larus heuristics predicts that the n=0 is unlikely
2458    (OH-heuristic), and the 'then' branch is unlikely because of the
2459    RH-heuristic -- giving the 'then' branch an even lower combined
2460    probability using the Dempster-Shafer theory.
2461    
2462    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2463    probability of the 'then' branch because of the loop in the else
2464    branch.
2465    
2466    
2467    Graphical Viewing:
2468    ------------------
2469    I merely plugged in Allen's graphical viewers into the compiler. The
2470    additional code is not much. At the top level, saying:
2471    
2472            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2473    
2474    will display the graphical view of the control flow graph just before
2475    back-patching.  daVinci must be in your path for this to work. If
2476    daVinci is not available, then the default viewer can be changed
2477    using:
2478    
2479            Control.MLRISC.getString "viewer"
2480    
2481    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2482    course, these viewers must be in your path.
2483    
2484    The above will display the compilation unit at the level of clusters,
2485    many of which are small, boring, and un-interesting. Also setting:
2486    
2487            Control.MLRISC.getInt "cfg-graphical-view_size"
2488    
2489    will display clusters that are larger than the value set by the above.
2490    
2491    
2492    ----------------------------------------------------------------------
2493    Name: Matthias Blume
2494    Date: 2002/03/21 22:20:00 EST
2495    Tag: blume-20020321-kmp-bugfix
2496    Description:
2497    
2498    Changed the interface to the KMP routine in PreString and fixed
2499    a minor bug in one place where it was used.
2500    
2501    ----------------------------------------------------------------------
2502    Name: Allen Leung
2503    Date: 2002/03/21 20:30:00 EST
2504    Tag: leunga-20020321-cfg
2505    Description:
2506    
2507      Fixed a potential problem in cfg edge splitting.
2508    
2509    ----------------------------------------------------------------------
2510    Name: Allen Leung
2511    Date: 2002/03/21 17:15:00 EST
2512    Tag: leunga-20020321-x86-fp-cfg
2513    Description:
2514    
2515      1. Recoded the buggy parts of x86-fp.
2516    
2517         a. All the block reordering code has been removed.
2518            We now depend on the block placement phases to do this work.
2519    
2520         b. Critical edge splitting code has been simplified and moved into the
2521            CFG modules, as where they belong.
2522    
2523         Both of these were quite buggy and complex.  The code is now much, much
2524         simpler.
2525    
2526      2. X86 backend.
2527    
2528         a. Added instructions for 64-bit support.  Instruction selection for
2529            64-bit has not been committed, however, since that
2530            requires changes to MLTREE which haven't been approved by
2531            Lal and John.
2532    
2533         b. Added support for FUCOMI and FUCOMIP when generating code for
2534            PentiumPro and above.  We only generate these instructions in
2535            the fast-fp mode.
2536    
2537         c. Added cases for JP and JNP in X86FreqProps.
2538    
2539      3. CFG
2540    
2541         CFG now has a bunch of methods for edge splitting and merging.
2542    
2543      4. Machine description.
2544    
2545         John's simplification of MLTREE_BASIS.fcond broke a few machine
2546         description things:
2547    
2548         rtl-build.{sig,sml} and hppa.mdl fixed.
2549    
2550         NOTE: the machine description stuff in the repository is still broken.
2551               Again, I can't put my fixes in because that involves
2552               changes to MLTREE.
2553    
2554    ----------------------------------------------------------------------
2555    Name: Matthias Blume
2556    Date: 2002/03/20 15:55:00 EST
2557    Tag: blume-20020320-kmp
2558    Description:
2559    
2560    Implemented Knuth-Morris-Pratt string matching in PreString and used
2561    it for String.isSubstring, Substring.isSubstring, and
2562    Substring.position.
2563    
2564    (Might need some stress-testing.  Simple examples worked fine.)
2565    
2566    ----------------------------------------------------------------------
2567    Name: Matthias Blume
2568    Date: 2002/03/19 16:37:00 EST
2569    Tag: blume-20020319-witnesses
2570    Description:
2571    
2572    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2573    
2574    This implements a generic mechanism for changing constness qualifiers
2575    anywhere within big C types without resorting to outright "casts".
2576    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2577    modify the constness at the outermost level.)
2578    The implementation of "convert" is based on the idea of "witness"
2579    values -- values that are not used by the operation but whose types
2580    "testify" to their applicability.  On the implementation side, "convert"
2581    is simply a projection (returning its second curried argument).  With
2582    cross-module inlining, it should not result in any machine code being
2583    generated.
2584    
2585    ----------------------------------------------------------------------
2586    Name: Matthias Blume
2587    Date: 2002/03/15 16:40:00 EST
2588    Tag: blume-20020315-basis
2589    Description:
2590    
2591    Provided (preliminary?) implementations for
2592    
2593      {String,Substring}.{concatWith,isSuffix,isSubstring}
2594    
2595    and
2596    
2597      Substring.full
2598    
2599    Those are in the Basis spec but they were missing in SML/NJ.
2600    
2601    ----------------------------------------------------------------------
2602    Name: Matthias Blume
2603    Date: 2002/03/14 21:30:00 EST
2604    Tag: blume-20020314-controls
2605    Description:
2606    
2607    Controls:
2608    ---------
2609    
2610    1. Factored out the recently-added Controls : CONTROLS stuff and put
2611       it into its own library $/controls-lib.cm.  The source tree for
2612       this is under src/smlnj-lib/Controls.
2613    
2614    2. Changed the names of types and functions in this interface, so they
2615       make a bit more "sense":
2616    
2617          module -> registry
2618          'a registry -> 'a group
2619    
2620    3. The interface now deals in ref cells only.  The getter/setter interface
2621       is (mostly) gone.
2622    
2623    4. Added a function that lets one register an already-existing ref cell.
2624    
2625    5. Made the corresponding modifications to the rest of the code so that
2626       everything compiles again.
2627    
2628    6. Changed the implementation of Controls.MLRISC back to something closer
2629       to the original.  In particular, this module (and therefore MLRISC)
2630       does not depend on Controls.  There now is some link-time code in
2631       int-sys.sml that registers the MLRISC controls with the Controls
2632       module.
2633    
2634    CM:
2635    ---
2636    
2637      * One can now specify the lambda-split aggressiveness in init.cmi.
2638    
2639    ----------------------------------------------------------------------
2640    Name: Allen Leung
2641    Date: 2002/03/13 17:30:00 EST
2642    Tag: leunga-20020313-x86-fp-unary
2643    Description:
2644    
2645    Bug fix for:
2646    
2647    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2648    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2649    > - fun f(x,(y,z)) = Real.~ y;
2650    > [autoloading]
2651    > [autoloading done]
2652    >       fchsl   (%eax), 184(%esp)
2653    > Error: MLRisc bug: X86MCEmitter.emitInstr
2654    >
2655    > uncaught exception Error
2656    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2657    
2658    The problem was that the code generator did not generate any fp registers
2659    in this case, and the ra didn't know that it needed to run the X86FP phase to
2660    translate the pseudo fp instruction.   This only happened with unary fp
2661    operators in certain situations.
2662    
2663    ----------------------------------------------------------------------
2664    Name: Matthias Blume
2665    Date: 2002/03/13 14:00:00 EST
2666    Tag: blume-20020313-overload-etc
2667    Description:
2668    
2669    1. Added _overload as a synonym for overload for backward compatibility.
2670       (Control.overloadKW must be true for either version to be accepted.)
2671    
2672    2. Fixed bug in install script that caused more things to be installed
2673       than what was requested in config/targets.
2674    
2675    3. Made CM aware of the (_)overload construct so that autoloading
2676       works.
2677    
2678    ----------------------------------------------------------------------
2679    Name: Matthias Blume
2680    Date: 2002/03/12 22:03:00 EST
2681    Tag: blume-20020312-url
2682    Description:
2683    
2684    Forgot to update BOOT and srcarchiveurl.
2685    
2686    ----------------------------------------------------------------------
2687    Name: Matthias Blume
2688    Date: 2002/03/12 17:30:00 EST
2689    Tag: blume-20020312-version110392
2690    Description:
2691    
2692    Yet another version number bump (because of small changes to the
2693    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2694    
2695    Changes:
2696    
2697      The new pid generation scheme described a few weeks ago was overly
2698      complicated.  I implemented a new mechanism that is simpler and
2699      provides a bit more "stability":  Once CM has seen a compilation
2700      unit, it keeps its identity constant (as long as you do not delete
2701      those crucial CM/GUID/* files).  This means that when you change
2702      an interface, compile, then go back to the old interface, and
2703      compile again, you arrive at the original pid.
2704    
2705      There now also is a mechanism that instructs CM to use the plain
2706      environment hash as a module's pid (effectively making its GUID
2707      the empty string).  For this, "noguid" must be specified as an
2708      option to the .sml file in question within its .cm file.
2709      This is most useful for code that is being generated by tools such
2710      as ml-nlffigen (because during development programmers tend to
2711      erase the tool's entire output directory tree including CM's cached
2712      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2713      revert to the old, broken behavior of SML/NJ, but in specific cases
2714      where there is no danger of interface confusion, its use is ok
2715      (I think).
2716    
2717      ml-nlffigen by default generates "noguid" annotations.  They can be
2718      turned off by specifying -guid in its command line.
2719    
2720    ----------------------------------------------------------------------
2721    Name: Lal George
2722    Date: 2002/03/12 12 14:42:36 EST
2723    Tag: george-20020312-frequency-computation
2724    Description:
2725    
2726    Integrated jump chaining and static block frequency into the
2727    compiler. More details and numbers later.
2728    
2729    ----------------------------------------------------------------------
2730    Name: Lal George
2731    Date: 2002/03/11 11 22:38:53 EST
2732    Tag: george-20020311-jump-chain-elim
2733    Description:
2734    
2735    Tested the jump chain elimination on all architectures (except the
2736    hppa).  This is on by default right now and is profitable for the
2737    alpha and x86, however, it may not be profitable for the sparc and ppc
2738    when compiling the compiler.
2739    
2740    The gc test will typically jump to a label at the end of the cluster,
2741    where there is another jump to an external cluster containing the actual
2742    code to invoke gc. This is to allow factoring of common gc invocation
2743    sequences. That is to say, we generate:
2744    
2745            f:
2746               testgc
2747               ja   L1      % jump if above to L1
2748    
2749            L1:
2750               jmp L2
2751    
2752    
2753    After jump chain elimination the 'ja L1' instructions is converted to
2754    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2755    up being implemented in their long form (if L2 is far away) using:
2756    
2757            jbe     L3      % jump if below or equal to L3
2758            jmp     L2
2759         L3:
2760            ...
2761    
2762    
2763    For large compilation units L2  may be far away.
2764    
2765    
2766    ----------------------------------------------------------------------
2767    Name: Matthias Blume
2768    Date: 2002/03/11 13:30:00 EST
2769    Tag: blume-20020311-mltreeeval
2770    Description:
2771    
2772    A functor parameter was missing.
2773    
2774    ----------------------------------------------------------------------
2775    Name: Allen Leung
2776    Date: 2002/03/11 10:30:00 EST
2777    Tag: leunga-20020311-runtime-string0
2778    Description:
2779    
2780       The representation of the empty string now points to a
2781    legal null terminated C string instead of unit.  It is now possible
2782    to convert an ML string into C string with InlineT.CharVector.getData.
2783    This compiles into one single machine instruction.
2784    
2785    ----------------------------------------------------------------------
2786    Name: Allen Leung
2787    Date: 2002/03/10 23:55:00 EST
2788    Tag: leunga-20020310-x86-call
2789    Description:
2790    
2791       Added machine generation for CALL instruction (relative displacement mode)
2792    
2793    ----------------------------------------------------------------------
2794    Name: Matthias Blume
2795    Date: 2002/03/08 16:05:00
2796    Tag: blume-20020308-entrypoints
2797    Description:
2798    
2799    Version number bumped to 110.39.1.  NEW BOOTFILES!
2800    
2801    Entrypoints: non-zero offset into a code object where execution should begin.
2802    
2803    - Added the notion of an entrypoint to CodeObj.
2804    - Added reading/writing of entrypoint info to Binfile.
2805    - Made runtime system bootloader aware of entrypoints.
2806    - Use the address of the label of the first function given to mlriscGen
2807      as the entrypoint.  This address is currently always 0, but it will
2808      not be 0 once we turn on block placement.
2809    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2810      with entry points) from mlriscGen.
2811    
2812    ----------------------------------------------------------------------
2813    Name: Allen Leung
2814    Date: 2002/03/07 20:45:00 EST
2815    Tag: leunga-20020307-x86-cmov
2816    Description:
2817    
2818       Bug fixes for CMOVcc on x86.
2819    
2820       1. Added machine code generation for CMOVcc
2821       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2822       3. CMOVcc cannot have an immediate operand as argument.
2823    
2824    ----------------------------------------------------------------------
2825    Name: Matthias Blume
2826    Date: 2002/03/07 16:15:00 EST
2827    Tag: blume-20020307-controls
2828    Description:
2829    
2830    This is a very large but mostly boring patch which makes (almost)
2831    every tuneable compiler knob (i.e., pretty much everything under
2832    Control.* plus a few other things) configurable via both the command
2833    line and environment variables in the style CM did its configuration
2834    until now.
2835    
2836    Try starting sml with '-h' (or, if you are brave, '-H')
2837    
2838    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2839    implements the underlying generic mechanism.
2840    
2841    The interface to some of the existing such facilities has changed somewhat.
2842    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2843    (The getFoo interface is still there for backward-compatibility, but its
2844    use is deprecated.)
2845    
2846    The ml-build script passes -Cxxx=yyy command-line arguments through so
2847    that one can now twiddle the compiler settings when using this "batch"
2848    compiler.
2849    
2850    TODO items:
2851    
2852    We should go through and throw out all controls that are no longer
2853    connected to anything.  Moreover, we should go through and provide
2854    meaningful (and correct!) documentation strings for those controls
2855    that still are connected.
2856    
2857    Currently, multiple calls to Controls.new are accepted (only the first
2858    has any effect).  Eventually we should make sure that every control
2859    is being made (via Controls.new) exactly once.  Future access can then
2860    be done using Controls.acc.
2861    
2862    Finally, it would probably be a good idea to use the getter-setter
2863    interface to controls rather than ref cells.  For the time being, both
2864    styles are provided by the Controls module, but getter-setter pairs are
2865    better if thread-safety is of any concern because they can be wrapped.
2866    
2867    *****************************************
2868    
2869    One bug fix: The function blockPlacement in three of the MLRISC
2870    backpatch files used to be hard-wired to one of two possibilities at
2871    link time (according to the value of the placementFlag).  But (I
2872    think) it should rather sense the flag every time.
2873    
2874    *****************************************
2875    
2876    Other assorted changes (by other people who did not supply a HISTORY entry):
2877    
2878    1. the cross-module inliner now works much better (Monnier)
2879    2. representation of weights, frequencies, and probabilities in MLRISC
2880       changed in preparation of using those for weighted block placement
2881       (Reppy, George)
2882    
2883    ----------------------------------------------------------------------
2884    Name: Lal George
2885    Date: 2002/03/07 14:44:24 EST 2002
2886    Tag: george-20020307-weighted-block-placement
2887    
2888    Tested the weighted block placement optimization on all architectures
2889    (except the hppa) using AMPL to generate the block and edge frequencies.
2890    Changes were required in the machine properties to correctly
2891    categorize trap instructions. There is an MLRISC flag
2892    "weighted-block-placement" that can be used to enable weighted block
2893    placement, but this will be ineffective without block/edge
2894    frequencies (coming soon).
2895    
2896    
2897    ----------------------------------------------------------------------
2898    Name: Lal George
2899    Date: 2002/03/05 17:24:48 EST
2900    Tag: george-20020305-linkage-cluster
2901    
2902    In order to support the block placement optimization, a new cluster
2903    is generated as the very first cluster (called the linkage cluster).
2904    It contains a single jump to the 'real' entry point for the compilation
2905    unit. Block placement has no effect on the linkage cluster itself, but
2906    all the other clusters  have full freedom in the manner in which they
2907    reorder blocks or functions.
2908    
2909    On the x86 the typical linkage code that is generated is:
2910       ----------------------
2911            .align 2
2912       L0:
2913            addl    $L1-L0, 72(%esp)
2914            jmp     L1
2915    
2916    
2917            .align  2
2918       L1:
2919       ----------------------
2920    
2921    72(%esp) is the memory location for the stdlink register. This
2922    must contain the address of the CPS function being called. In the
2923    above example, it contains the address of  L0; before
2924    calling L1 (the real entry point for the compilation unit), it
2925    must contain the address for L1, and hence
2926    
2927            addl $L1-L0, 72(%esp)
2928    
2929    I have tested this on all architectures except the hppa.The increase
2930    in code size is of course negligible
2931    
2932    ----------------------------------------------------------------------
2933    Name: Allen Leung
2934    Date: 2002/03/03 13:20:00 EST
2935    Tag: leunga-20020303-mlrisc-tools
2936    
2937      Added #[ ... ] expressions to mlrisc tools
2938    
2939    ----------------------------------------------------------------------
2940    Name: Matthias Blume
2941    Date: 2002/02/27 12:29:00 EST
2942    Tag: blume-20020227-cdebug
2943    Description:
2944    
2945    - made types in structure C and C_Debug to be equal
2946    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2947    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2948    
2949    ----------------------------------------------------------------------
2950    Name: Matthias Blume
2951    Date: 2002/02/26 12:00:00 EST
2952    Tag: blume-20020226-ffi
2953    Description:
2954    
2955    1. Fixed a minor bug in CM's "noweb" tool:
2956       If numbering is turned off, then truly don't number (i.e., do not
2957       supply the -L option to noweb).  The previous behavior was to supply
2958       -L'' -- which caused noweb to use the "default" line numbering scheme.
2959       Thanks to Chris Richards for pointing this out (and supplying the fix).
2960    
2961    2. Once again, I reworked some aspects of the FFI:
2962    
2963       A. The incomplete/complete type business:
2964    
2965       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2966         gone!
2967       - ML types representing an incomplete type are now *equal* to
2968         ML types representing their corresponding complete types (just like
2969         in C).  This is still safe because ml-nlffigen will not generate
2970         RTTI for incomplete types, nor will it generate functions that
2971         require access to such RTTI.   But when ML code generated from both
2972         incomplete and complete versions of the C type meet, the ML types
2973         are trivially interoperable.
2974    
2975         NOTE:  These changes restore the full generality of the translation
2976         (which was previously lost when I eliminated functorization)!
2977    
2978       B. Enum types:
2979    
2980       - Structure C now has a type constructor "enum" that is similar to
2981         how the "su" constructor works.  However, "enum" is not a phantom
2982         type because each "T enum" has values (and is isomorphic to
2983         MLRep.Signed.int).
2984       - There are generic access operations for enum objects (using
2985         MLRep.Signed.int).
2986       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2987         * The structure contains the definition of type "mlrep" (the ML-side
2988         representation type of the enum).  Normally, mlrep is the same
2989         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2990         then mlrep will be defined as a datatype -- thus facilitating
2991         pattern matching on mlrep values.
2992         ("-ec" will be suppressed if there are duplicate values in an
2993          enumeration.)
2994         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2995         will be generated for each C enum constant xxx.
2996         * Conversion functions m2i and i2m convert between mlrep and
2997         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2998         * Coversion functions c and ml convert between mlrep and "tag enum".
2999         * Access functions (get/set) fetch and store mlrep values.
3000       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3001         enumerations are merged into one single enumeration represented by
3002         structure E_'.
3003    
3004    ----------------------------------------------------------------------
3005    Name: Allen Leung
3006    Date: 2002/02/25 04:45:00 EST
3007    Tag: leunga-20020225-cps-spill
3008    
3009    This is a new implementation of the CPS spill phase.
3010    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3011    In case of problems, replace it with the old file spill.sml
3012    
3013    The current compiler runs into some serious performance problems when
3014    constructing a large record.  This can happen when we try to compile a
3015    structure with many items.  Even a very simple structure like the following
3016    makes the compiler slow down.
3017    
3018        structure Foo = struct
3019           val x_1 = 0w1 : Word32.int
3020           val x_2 = 0w2 : Word32.int
3021           val x_3 = 0w3 : Word32.int
3022           ...
3023           val x_N = 0wN : Word32.int
3024        end
3025    
3026    The following table shows the compile time, from N=1000 to N=4000,
3027    with the old compiler:
3028    
3029    N
3030    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3031           MLRISC ra                               0.06u  0.00s  0.05g
3032              (spills = 0 reloads = 0)
3033           TOTAL                                   0.63u  0.07s  0.21g
3034    
3035    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3036           MLRISC ra                               5.68u  0.59s  3.93g
3037              (spills = 0 reloads = 0)
3038           TOTAL                                   14.71u  0.99s  4.81g
3039    
3040    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3041           MLRISC ra                               5.54u  0.65s  3.91g
3042              (spills = 543 reloads = 1082)
3043           TOTAL                                   65.40u  3.13s  6.00g
3044    
3045    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3046           MLRISC ra                               0.80u  0.10s  0.55g
3047              (spills = 42 reloads = 84)
3048           TOTAL                                   129.42u  5.10s  4.13g
3049    
3050    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3051           MLRISC ra                               2.69u  0.27s  1.38g
3052              (spills = 62 reloads = 124)
3053           TOTAL                                   682.48u  19.61s  13.99g
3054    
3055    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3056           MLRISC ra                               4.96u  0.27s  2.72g
3057              (spills = 85 reloads = 170)
3058           TOTAL                                   2375.26u  57.21s  48.00g
3059    
3060    As you can see the old cps spill module suffers from some serious
3061    performance problem.  But since I cannot decipher the old code fully,
3062    instead of patching the problems up, I'm reimplementing it
3063    with a different algorithm.  The new code is more modular,
3064    smaller when compiled, and substantially faster
3065    (O(n log n) time and O(n) space).  Timing of the new spill module:
3066    
3067    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3068          MLRISC ra                               0.25u  0.02s  0.15g
3069             (spills=1 reloads=3)
3070          TOTAL                                   7.74u  0.34s  1.62g
3071    
3072    Implementation details:
3073    
3074    As far as I can tell, the purpose of the CPS spill module is to make sure the
3075    number of live variables at any program point (the bandwidth)
3076    does not exceed a certain limit, which is determined by the
3077    size of the spill area.
3078    
3079    When the bandwidth is too large, we decrease the register pressure by
3080    packing live variables into spill records.  How we achieve this is
3081    completely different than what we did in the old code.
3082    
3083    First, there is something about the MLRiscGen code generator
3084    that we should be aware of:
3085    
3086    o MLRiscGen performs code motion!
3087    
3088       In particular, it will move floating point computations and
3089       address computations involving only the heap pointer to
3090       their use sites (if there is only a single use).
3091       What this means is that if we have a CPS record construction
3092       statement
3093    
3094           RECORD(k,vl,w,e)
3095    
3096       we should never count the new record address w as live if w
3097       has only one use (which is often the case).
3098    
3099       We should do something similar to floating point, but the transformation
3100       there is much more complex, so I won't deal with that.
3101    
3102    Secondly, there are now two new cps primops at our disposal:
3103    
3104     1. rawrecord of record_kind option
3105        This pure operator allocates some uninitialized storage from the heap.
3106        There are two forms:
3107    
3108         rawrecord NONE [INT n]  allocates a tagless record of length n
3109         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3110                                     and initializes the tag.
3111    
3112     2. rawupdate of cty
3113          rawupdate cty (v,i,x)
3114          Assigns to x to the ith component of record v.
3115          The storelist is not updated.
3116    
3117    We use these new primops for both spilling and increment record construction.
3118    
3119     1. Spilling.
3120    
3121        This is implemented with a linear scan algorithm (but generalized
3122        to trees).  The algorithm will create a single spill record at the
3123        beginning of the cps function and use rawupdate to spill to it,
3124        and SELECT or SELp to reload from it.  So both spills and reloads
3125        are fine-grain operations.  In contrast, in the old algorithm
3126        "spills" have to be bundled together in records.
3127    
3128        Ideally, we should sink the spill record construction to where
3129        it is needed.  We can even split the spill record into multiple ones
3130        at the places where they are needed.  But CPS is not a good
3131        representation for global code motion, so I'll keep it simple and
3132        am not attempting this.
3133    
3134     2. Incremental record construction (aka record splitting).
3135    
3136        Long records with many component values which are simulatenously live
3137        (recall that single use record addresses are not considered to
3138         be live) are constructed with rawrecord and rawupdate.
3139        We allocate space on the heap with rawrecord first, then gradually
3140        fill it in with rawupdate.  This is the technique suggested to me
3141        by Matthias.
3142    
3143        Some restrictions on when this is applicable:
3144        1. It is not a VECTOR record.  The code generator currently does not handle
3145           this case. VECTOR record uses double indirection like arrays.
3146        2. All the record component values are defined in the same "basic block"
3147           as the record constructor.  This is to prevent speculative
3148           record construction.
3149    
3150    ----------------------------------------------------------------------
3151    Name: Allen Leung
3152    Date: 2002/02/22 01:02:00 EST
3153    Tag: leunga-20020222-mlrisc-tools
3154    
3155    Minor bug fixes in the parser and rewriter
3156    
3157    ----------------------------------------------------------------------
3158    Name: Allen Leung
3159    Date: 2002/02/21 20:20:00 EST
3160    Tag: leunga-20020221-peephole
3161    
3162    Regenerated the peephole files.  Some contained typos in the specification
3163    and some didn't compile because of pretty printing bugs in the old version
3164    of 'nowhere'.
3165    
3166    ----------------------------------------------------------------------
3167    Name: Allen Leung
3168    Date: 2002/02/19 20:20:00 EST
3169    Tag: leunga-20020219-mlrisc-tools
3170    Description:
3171    
3172       Minor bug fixes to the mlrisc-tools library:
3173    
3174       1.  Fixed up parsing colon suffixed keywords
3175       2.  Added the ability to shut the error messages up
3176       3.  Reimplemented the pretty printer and fixed up/improved
3177           the pretty printing of handle and -> types.
3178       4.  Fixed up generation of literal symbols in the nowhere tool.
3179       5.  Added some SML keywords to to sml.sty
3180    
3181    ----------------------------------------------------------------------
3182    Name: Matthias Blume
3183    Date: 2002/02/19 16:20:00 EST
3184    Tag: blume-20020219-cmffi
3185    Description:
3186    
3187    A wild mix of changes, some minor, some major:
3188    
3189    * All C FFI-related libraries are now anchored under $c:
3190        $/c.cm      --> $c/c.cm
3191        $/c-int.cm  --> $c/internals/c-int.cm
3192        $/memory.cm --> $c/memory/memory.cm
3193    
3194    * "make" tool (in CM) now treats its argument pathname slightly
3195      differently:
3196        1. If the native expansion is an absolute name, then before invoking
3197           the "make" command on it, CM will apply OS.Path.mkRelative
3198           (with relativeTo = OS.FileSys.getDir()) to it.
3199        2. The argument will be passed through to subsequent phases of CM
3200           processing without "going native".  In particular, if the argument
3201           was an anchored path, then "make" will not lose track of that anchor.
3202    
3203    * Compiler backends now "know" their respective C calling conventions
3204      instead of having to be told about it by ml-nlffigen.  This relieves
3205      ml-nlffigen from one of its burdens.
3206    
3207    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3208    
3209    * Export C_DEBUG and C_Debug from $c/c.cm.
3210    
3211    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3212      subtyping relationship between incomplete pointers and their complete
3213      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3214      with the convention of instantiating 'o with ('t, 'c) obj whenever
3215      the pointer target type is complete.  In the incomplete case, 'o
3216      will be instantiated with some "'c iobj" -- a type obtained by
3217      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3218    
3219      Operations that work on both incomplete and complete pointer types are
3220      typed as taking an 'o ptr while operations that require the target to
3221      be known are typed as taking some ('t, 'c) obj ptr.
3222    
3223      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3224      where void is an eqtype without any values.  This makes it possible
3225      to work on voidptr values using functions meant to operate on light
3226      incomplete pointers.
3227    
3228    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3229      been vastly simplified.
3230    
3231    ----------------------------------------------------------------------
3232    Name: Matthias Blume
3233    Date: 2002/02/19 10:48:00 EST
3234    Tag: blume-20020219-pqfix
3235    Description:
3236    
3237    Applied Chris Okasaki's bug fix for priority queues.
3238    
3239    ----------------------------------------------------------------------
3240    Name: Matthias Blume
3241    Date: 2002/02/15 17:05:00
3242    Tag: Release_110_39
3243    Description:
3244    
3245    Last-minute retagging is becoming a tradition... :-(
3246    
3247    This is the working release 110.39.
3248    
3249    ----------------------------------------------------------------------
3250    Name: Matthias Blume
3251    Date: 2002/02/15 16:00:00 EST
3252    Tag: Release_110_39-orig
3253    Description:
3254    
3255    Working release 110.39.  New bootfiles.
3256    
3257    (Update: There was a small bug in the installer so it wouldn't work
3258    with all shells.  So I retagged. -Matthias)
3259    
3260    ----------------------------------------------------------------------
3261    Name: Matthias Blume
3262    Date: 2002/02/15 14:17:00 EST
3263    Tag: blume-20020215-showbindings
3264    Description:
3265    
3266    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3267    the latter can be useful for exploring what bindings are available at
3268    the interactive prompt.  (The first function returns only the list
3269    of symbols that are really bound, the second prints those but also the
3270    ones that CM's autoloading mechanism knows about.)
3271    
3272    ----------------------------------------------------------------------
3273    Name: Matthias Blume
3274    Date: 2002/02/15 12:08:00 EST
3275    Tag: blume-20020215-iptrs
3276    Description:
3277    
3278    Two improvements to ml-nlffigen:
3279    
3280      1. Write files only if they do not exist or if their current contents
3281         do not coincide with what's being written.  (That is, avoid messing
3282         with the time stamps unless absolutely necessary.)
3283    
3284      2. Implement a "repository" mechanism for generated files related
3285         to "incomplete pointer types".   See the README file for details.
3286    
3287    ----------------------------------------------------------------------
3288    Name: Matthias Blume
3289    Date: 2002/02/14 11:50:00 EST
3290    Tag: blume-20020214-quote
3291    Description:
3292    
3293    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3294    because of the new and improved tag generation scheme.  (Thanks to Allen
3295    Leung for pointing it out.)
3296    
3297    ----------------------------------------------------------------------
3298    Name: Lal George
3299    Date: 2002/02/14 09:55:27 EST 2002
3300    Tag: george-20020214-isabelle-bug
3301    Description:
3302    
3303    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3304    of Isabelle on the x86.
3305    
3306    From Allen:
3307    -----------
3308     I've found the problem:
3309    
3310         in ra-core.sml, I use the counter "blocked" to keep track of the
3311         true number of elements in the freeze queue.  When the counter goes
3312         to zero, I skip examining the queue.  But I've messed up the
3313         bookkeeping in combine():
3314    
3315             else ();
3316             case !ucol of
3317               PSEUDO => (if !cntv > 0 then
3318                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3319                                        ^^^^^^^^^^^^^^^^^^^^^^^
3320                      moveu := mergeMoveList(!movev, !moveu)
3321                     )
3322                  else ();
3323    
3324         combine() is called to coalesce two nodes u and v.
3325         I think I was thinking that if the move counts of u and v are both
3326         greater than zero then after they are coalesced then one node is
3327         removed from the freeze queue.  Apparently I was thinking that
3328         both u and v are of low degree, but that's clearly not necessarily true.
3329    
3330    
3331    02/12/2002:
3332        Here's the patch.  HOL now compiles.
3333    
3334        I don't know how this impact on performance (compile
3335        time or runtime).  This bug caused the RA (especially on the x86)
3336        to go thru the potential spill phase when there are still nodes on the
3337        freeze queue.
3338    
3339    
3340    
3341    
3342    ----------------------------------------------------------------------
3343    Name: Matthias Blume
3344    Date: 2002/02/13 22:40:00 EST
3345    Tag: blume-20020213-fptr-rtti
3346    Description:
3347    
3348    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3349    updates.
3350    
3351    ----------------------------------------------------------------------
3352    Name: Matthias Blume
3353    Date: 2002/02/13 16:41:00 EST
3354    Tag: blume-20020213-cmlpq
3355    Description:
3356    
3357    Added new priority queue export symbols (which have just been added to
3358    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3359    not compile and the installer would choke.)
3360    
3361    ----------------------------------------------------------------------
3362    Name: Matthias Blume
3363    Date: 2002/02/13 16:15:00 EST
3364    Tag: blume-20020213-various
3365    Description:
3366    
3367    1. More tweaks to ml-nlffigen:
3368    
3369       - better internal datastructures (resulting in slight speedup)
3370       - "-match" option requires exact match
3371       - "localized" gensym counters (untagged structs/unions nested within
3372         other structs/unions or within typedefs get a fresh counter; their
3373         tag will be prefixed by a concatenation of their parents' tags)
3374       - bug fixes (related to calculation of transitive closure of types
3375         to be included in the output)
3376    
3377    2. Minor Basis updates:
3378    
3379       - added implementations for List.collate and Option.app
3380    
3381    ----------------------------------------------------------------------
3382    Name: Matthias Blume
3383    Date: 2002/02/11 15:55:00 EST
3384    Tag: blume-20020211-gensym
3385    Description:
3386    
3387    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3388    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3389    names (ML structure names that correspond to unnamed C structs, unions,
3390    and enums), so that separate runs of ml-nlffigen do not clash.
3391    
3392    ----------------------------------------------------------------------
3393    Name: Matthias Blume
3394    Date: 2002/02/11 12:05:00 EST
3395    Tag: blume-20020211-gensml
3396    Description:
3397    
3398    A quick fix for a problem with GenSML (in the pgraph-util library):
3399    Make generation of toplevel "local" optional.  (Strictly speaking,
3400    signature definitions within "local" are not legal SML.)
3401    
3402    Other than that: updates to INSTALL and cm/TODO.
3403    
3404    ----------------------------------------------------------------------
3405    Name: Matthias Blume
3406    Date: 2002/02/08 15:00:00 EST
3407    Tag: blume-20020208-uniquepid
3408    Description:
3409    
3410    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3411    
3412    1. The installer (config/install.sh) has gotten smarter:
3413    
3414         - Configuration options are a bit easier to specify now
3415           (in config/targets).
3416         - Bug in recognizing .tar.bz2 files fixed.
3417         - Installer automatically resolves dependencies between
3418           configuration options (e.g., if you ask for eXene, you will
3419           also get cml -- regardless whether you asked for it or not).
3420         - Installer can run in "quieter mode" by setting the environment
3421           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3422           "completely silent", though.
3423         - Build HashCons library as part of smlnj-lib.
3424    
3425    2. A new scheme for assigning persistent identifiers to compilation
3426       units (and, by extension, to types etc.) has been put into place.
3427       This fixes a long-standing bug where types and even dynamic values
3428       can get internally confused, thereby compromising type safety
3429       (abstraction) and dynamic correctness.  See
3430    
3431         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3432    
3433       for an example of how things could go wrong until now.
3434    
3435       The downside of the new scheme is that pids are not quite as
3436       persistent as they used to be: CM will generate a fresh pid
3437       for every compilation unit that it thinks it sees for the first
3438       time.  That means that if you compile starting from a clean, fresh
3439       source tree at two different times, you end up with different
3440       binaries.
3441    
3442       Cutoff recompilation, however, has not been compromised because
3443       CM keeps pid information in special caches between runs.
3444    
3445    ----------------------------------------------------------------------
3446    Name: Lal George
3447    Date: 2002/02/07 15:34:13 EST 2002
3448    Tag: <none>
3449    Description:
3450    
3451    Compilers that generate assembly code may produce  global labels
3452    whose value is resolved at link time. The various peephole optimization
3453    modules did not take this in account.
3454    
3455    TODO. The Labels.addrOf function should really return an option
3456    type so that clients are forced to deal with this issue, rather
3457    than an exception being raised.
3458    
3459    ----------------------------------------------------------------------
3460    Name: Lal George
3461    Date: 2002/02/06 13:55:02 EST
3462    Tag: george-20020206-ra-breakup
3463    Description:
3464    
3465    1. A bug fix from Allen.
3466    
3467        A typo causes extra fstp %st(0)'s to be generated at compensation
3468        edges, which might cause stack underflow traps at runtime.  This
3469        occurs in fft where there are extraneous fstps right before the 'into'
3470        trap instruction (in this case they are harmless since none of the
3471        integers overflow.)
3472    
3473    2. Pulled out various utility modules that were embedded in the modules
3474       of the register allocator. I need these modules for other purposes, but
3475       they are not complete enough to put into a library (just yet).
3476    ----------------------------------------------------------------------
3477    Name: Matthias Blume
3478    Date: 2002/01/31 16:05:00 EST
3479    Tag: blume-20020131-sparc-ccalls
3480    Description:
3481    
3482    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3483       of extra stack space by mistake.  Fixed.
3484    
3485    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3486    
3487    ----------------------------------------------------------------------
3488    Name: Allen Leung
3489    Date: 2002/01/30
3490    Tag: leunga-20020130-nowhere-bug-fix
3491    Description:
3492    
3493       MLRISC bug fixes:
3494       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3495       2. MachineInt.fromString was negating its value.
3496    
3497    ----------------------------------------------------------------------
3498    Name: Matthias Blume
3499    Date: 2002/01/29
3500    Tag: blume-20020129-INSTALL
3501    Description:
3502    
3503    - Added somewhat detailed installation instructions (file INSTALL).
3504    - Fixed curl-detection bug in config/install.sh.
3505    - It is now possible to select the URL getter using the URLGETTER
3506      environment variable:
3507    
3508          not set / "unknown"      --> automatic detection (script tries wget,
3509                                       curl, and lynx)
3510          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3511                                       how to properly invoke them)
3512          other                    --> use $URLGETTER directly, it must take
3513                                       precisely two command-line arguments
3514                                       (source URL and destination file name)
3515    
3516    ----------------------------------------------------------------------
3517    Name: Matthias Blume
3518    Date: 2002/01/28
3519    Tag: blume-20020128-sparc-ccalls
3520    Description:
3521    
3522    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3523    - Make use of the allocParam argument in sparc-c-calls.
3524    
3525    ----------------------------------------------------------------------
3526  Name: Matthias Blume  Name: Matthias Blume
3527  Date: 2002/01/28  Date: 2002/01/28
3528  Tag: blume-20020128-allocParam  Tag: blume-20020128-allocParam
# Line 584  Line 4094 
4094    
4095  ----------------------------------------------------------------------  ----------------------------------------------------------------------
4096  Name: Matthias Blume  Name: Matthias Blume
 >>>>>>> 1.169  
4097  Date: 2001/09/18 15:35:00 EDT  Date: 2001/09/18 15:35:00 EDT
4098  Tag: blume-20010918-readme11036  Tag: blume-20010918-readme11036
4099  Description:  Description:

Legend:
Removed from v.1044  
changed lines
  Added in v.1733

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