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 615, Wed Apr 12 04:52:09 2000 UTC revision 1718, Thu Dec 2 22:48:18 2004 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/12/01 16:50:00 CST
18    Tag: blume-20041201-atom
19    Description:
20    
21    smlnj-lib:
22    
23      Added function HashString.hashString' for substrings.
24      Hand-inlined CharVector.foldl into HashString (for speed).
25      Modified implementation of structure Atom to avoid extracting
26      strings from substrings unless necessary.
27      (Also see CHANGES file for smlnj-lib.)
28    
29    ----------------------------------------------------------------------
30    Name: Matthias Blume (blume (at) tti - c (dot) org)
31    Date: 2004/11/24 22:15:00 CST
32    Tag: blume-20041124-cml
33    Description:
34    
35    Made sure CML compiles when Position = Int64.
36    
37    ----------------------------------------------------------------------
38    Name: Matthias Blume (blume (at) tti - c (dot) org)
39    Date: 2004/11/24 14:30:00 CST
40    Tag: blume-20041124-position
41    Description:
42    
43    The compiler can now be compiled in a mode that makes structure Position
44    equal to Int64.  The default, however, is unchanged (Position = Int31)
45    for the time being.
46    
47    To enable 64-bit positions, use the following procedure:
48    
49    1. Start sml
50    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
51    3. Type
52          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
53    4. Run CMB.make() as usual.
54    
55    This is barely tested.  The only test so far was a little SML program
56    counting the number of characters in an 8-gigabyte file by
57    reading it character-by-character.  That test was successful.
58    
59    In support of 64-bit positions, a number of new functions have been
60    added to the runtime system.
61    
62    ----------------------------------------------------------------------
63    Name: Matthias Blume (blume (at) tti - c (dot) org)
64    Date: 2004/11/23 14:45:00 CST
65    Tag: blume-20041123-useFile
66    Description:
67    
68    Fixed a problem with unhelpful error messages related to problems with
69    .cm- or .sml files that appear as part of the sml command line.
70    
71    ----------------------------------------------------------------------
72    Name: Matthias Blume (blume (at) tti - c (dot) org)
73    Date: 2004/11/18 15:40:00 CST
74    Tag: Release_110_51
75    Description:
76    
77    New working version (110.51).  NEW BOOTFILES!
78    
79    ----------------------------------------------------------------------
80    Name: Matthias Blume (blume (at) tti - c (dot) org)
81    Date: 2004/11/18 15:35:00 CST
82    Tag: <none>
83    Description:
84    
85    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
86    Joemann).
87    
88    ----------------------------------------------------------------------
89    Name: Allen Leung (leunga (at) reservoir (dot) com)
90    Date: 2004/11/17 16:05:21 EST 2004
91    Tag: leunga-20041117-mlrisc-live-kill
92    Description:
93    
94       Added support for MLTree constructs LIVE and KILL
95    to all the architectures.
96    
97    ----------------------------------------------------------------------
98    Name: Matthias Blume (blume (at) tti - c (dot) org)
99    Date: 2004/11/13 00:20:00 CST
100    Tag: blume-20041113-versiontool
101    Description:
102    
103    - Stripped down the versiontool: It now only handles the version number.
104      The date string is generated at bootstrap time (during makeml).
105    
106    - In a previous commit, fixed a minor issue with how polyequal is being
107      translated.  In particular, the code now "looks through" abstractions.
108      This results in slightly fewer polyEqual warnings and hopefully slightly
109      more efficient code.  Important examples for where this matters are
110      the new int64 and word64 types.
111    
112    ----------------------------------------------------------------------
113    Name: Matthias Blume (blume (at) tti - c (dot) org)
114    Date: 2004/11/12 00:30:00 CST
115    Tag: blume-20041112-int64
116    Description:
117    
118    Structure Int64 fully hooked in.  (The implementation is not very
119    efficient, though.)
120    
121    ----------------------------------------------------------------------
122    Name: Matthias Blume (blume (at) tti - c (dot) org)
123    Date: 2004/11/11 17:30:00 CST
124    Tag: blume-20041111-more64
125    Description:
126    
127    All the pieces of Word64 are now there, with the exception of the
128    conversions from and to LargeWord. (Eventually these need to be identities,
129    but for the time being they don't even make sense because LargeWord is
130    32-bit wide.)
131    
132    Also started to add similar support for Int64, but major pieces of that
133    are still missing.
134    
135    ----------------------------------------------------------------------
136    Name: Matthias Blume (blume (at) tti - c (dot) org)
137    Date: 2004/11/11 00:15:00 CST
138    Tag: blume-20041111-word64
139    Description:
140    
141    Structure Word64 is now (almost) complete, word literals and patterns
142    seem to work.  There are a few odd pieces missing.  In particular,
143    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
144    at the moment.
145    
146    Making Word64 official would mean that LargeWord becomes Word64.  But
147    this requires extreme care because most word-word conversions have to
148    go through LargeWord, so making a mistake means loss of efficiency or
149    worse.  Eventually there will be a solution similar to (but actually
150    simpler than) what I did with IntInf.
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume (blume (at) tti - c (dot) org)
154    Date: 2004/11/10 18:12:00 CST
155    Tag: blume-20041110-64bit
156    Description:
157    
158    More 64-bit hacking (but still not even half-way there yet).
159    Also, some assorted improvements to the handling of 8-bit words.
160    
161    ----------------------------------------------------------------------
162    Name: Matthias Blume (blume (at) tti - c (dot) org)
163    Date: 2004/11/09 17:50:00 CST
164    Tag: <none>
165    Description:
166    
167    Started adding some infrastructure for supporting 64-bit int- and
168    word-types.  (Still in its very early stages.)
169    
170    ----------------------------------------------------------------------
171    Name: Matthias Blume (blume (at) tti - c (dot) org)
172    Date: 2004/10/28 10:45:00 CDT
173    Tag: Release_110_50
174    Description:
175    
176    New working version (110.50).  NEW BOOTFILES!
177    
178                 =====================
179    
180    Also:
181    
182      - Changend config/srcarchiveurl from a file just
183        containing the URL string into a file containing
184        shell script code.  The code has access to the $VERSION variable.
185      - Made corresponding changes to config/install.sh and config/unpack.
186      - Default contents of config/srcarchiveurl uses $VERSION and
187        normally does not have to be edited to reflect a version change.
188    
189      (As a result, a version change can be done by just editing
190       config/version, the rest is now automatic.)
191    
192    ----------------------------------------------------------------------
193    Name: Matthias Blume (blume (at) tti - c (dot) org)
194    Date: 2004/10/27 17:50:00 CDT
195    Tag: blume-20041027-btrace-msg
196    Description:
197    
198    BackTrace.monitor now also reports the source of the exception that
199    triggered the trace.
200    
201    ----------------------------------------------------------------------
202    Name: Matthias Blume (blume (at) tti - c (dot) org)
203    Date: 2004/10/27 17:20:00 CDT
204    Tag: blume-20041027-x86-c-calls
205    Description:
206    
207    This is the HISTORY entry for two earlier commits, both concerning
208    the x86 c-calls code in MLRISC:
209    
210      - added a missing LOAD in the code that deals with struct arguments
211      - made sure the caller does not add the wrong number of bytes to the
212        stack pointer after a call of a function returning a struct
213        (the callee already pops the implicit argument which points to
214         the space reserved for the result)
215    
216    ----------------------------------------------------------------------
217    Name: Allen Leung (leunga (at) reservoir (dot) com)
218    Date: 2004/10/24 14:00:00 EST
219    Tag: leunga-20041024-x86-gas-fucomip
220    Description:
221    
222       John discovered a bug in the syntax of fucomip.
223       The opcodes FU?COMIP? have been changed to
224    
225           fu?comip? %st(i), %st
226    
227    ----------------------------------------------------------------------
228    Name: Matthias Blume (blume (at) tti - c (dot) org)
229    Date: 2004/10/20 15:06:00 CDT
230    Tag: blume-20041020-standalone-backtrace
231    Description:
232    
233    Added a mechanism for getting back-trace information from standalone
234    programs.  Here is how it works:
235    
236       1. The part of the program from which you want to get backtrace
237          information (usually the whole program) should be wrapped with
238          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
239          main program could be modified from something like
240    
241             fun main (pgm, args) = ...
242    
243         to
244    
245            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
246    
247       2. To be able to access BackTrace.monitor, you have to add
248          library $smlnj-tdp/plugins.cm to the .cm file that contains your
249          main function.
250    
251       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
252          CM might have created in the past for your project).
253    
254       4. Build the system using this command line:
255    
256             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
257                      myprog.cm MyProg.main myprog
258    
259          instead of the usual
260    
261             ml-build myprog.cm MyProg.main myprog
262    
263    I changed a library name:
264    
265       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
266    
267    New libraries:
268    
269       $smlnj-tdp/back-trace.cm
270          -- when loaded causes the back-trace plugin to be installed
271       $smlnj-tdp/coverage.cm
272          -- when loaded causes the coverage plugin to be installed
273    
274    ----------------------------------------------------------------------
275    Name: Matthias Blume (blume (at) tti - c (dot) org)
276    Date: 2004/10/18 16:45:00 CDT
277    Tag: blume-20041018-groupowner
278    Description:
279    
280    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
281    
282    Eliminated group owner specs from .cm files throughout the source tree.
283    
284    ----------------------------------------------------------------------
285    Name: Matthias Blume (blume (at) tti - c (dot) org)
286    Date: 2004/10/15 15:45:00 CDT
287    Tag: blume-20041015-coverage
288    Description:
289    
290    * Test coverage tool added!
291    * Further reorganization of tracing-, debugging-, and profiling support:
292    
293        - moved original BTImp -- now called BackTrace -- into a separate
294          library called $/trace-debug-profile.cm
295        - eliminated all mentions of BTrace from SMLofNJ.Internals
296        - only the instrumentation mechanism is now left in the compiler proper
297        - BackTrace module is a plugin which is NOT plugged in by default
298        - Coverage module is another such plugin
299    
300    To get the benefits of any of these plugin modules, the code in
301    question must be compiled with tdp instrumentation turned on.  This
302    can be done by setting SMLofNJ.Internals.TDP.mode to true.
303    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
304    
305    Plugins are selected at link time.  (Pre-compiled instrumented code
306    can be re-loaded with different plugins in effect.)  When an
307    instrumented module is linked, whatever plugins are at that time
308    enabled will come into effect for that module.
309    
310    To enable the back-trace plugin, load library $/trace-debug-profile.cm
311    and invoke BackTrace.install() (e.g., from the interactive prompt).
312    To enable the coverage plugin, load the same library and invoke
313    Coverage.install().
314    
315    Back-traces are generated automatically on uncaught exceptions and
316    when the code in question explicitly invokes BackTrace.trigger().
317    
318    Coverage (and execution frequency-) information must be queried
319    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
320    
321    ----------------------------------------------------------------------
322    Name: Matthias Blume (blume (at) tti - c (dot) org)
323    Date: 2004/10/14 17:40:00 CDT
324    Tag: blume-20041014-tdp-core
325    Description:
326    
327    Snapshot of a significant overhaul of how the trace/debug/profile support
328    is hooked into the system (specifically: Core and SMLofNJ.Internals).
329    
330    ----------------------------------------------------------------------
331    Name: Matthias Blume (blume (at) tti - c (dot) org)
332    Date: 2004/10/13 16:34:00 CDT
333    Tag: blume-20041013-tdp
334    Description:
335    
336    Some rationalization of names:
337    
338       structure BTrace -> structure TDPInstrument
339       etc.
340    
341    This is is preparation of using the original back-trace
342    instrumentation for other purposes.  "TDP" stands for
343    Trace/Debug/Profile.
344    
345    The control flag controlling whether instrumentation is on or off is now
346    registered under a different name, so instead of running sml as
347    
348     sml -Cinstrument.btrace-mode=true
349    
350    one has to say
351    
352     sml -Ctdp.instrument=true
353    
354    ----------------------------------------------------------------------
355    Name: Matthias Blume (blume (at) tti - c (dot) org)
356    Date: 2004/10/11 16:37:00 CDT
357    Tag: blume-20041011-regions
358    Description:
359    
360    Made some minor modifications to elabcore.sml to have source regions
361    be propagated more tightly -- resulting in better (i.e., smaller)
362    regions being reported in error- and debug messages.
363    
364    ----------------------------------------------------------------------
365    Name: Matthias Blume (blume (at) tti - c (dot) org)
366    Date: 2004/10/08 22:50:00 CDT
367    Tag: blume-20041008-cmkw
368    Description:
369    
370    Fixed handling of keywords in .cm files: After seeing "is" the lexer
371    treats subsequent occurrences of "group", "library", "source", "is",
372    "*", and "-" as ordinary identifiers rather than keywords.
373    
374    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
375    is supposed to accept a tool argument called "source", but this did
376    not work because of the clash with the keyword.
377    
378    ----------------------------------------------------------------------
379    Name: Matthias Blume (blume (at) tti - c (dot) org)
380    Date: 2004/10/07 16:00:00 CDT
381    Tag: blume-20041007-cleanup
382    Description:
383    
384    Assorted cleanup work:
385    
386      - got rid of intstrmap in favor of using the library's
387        hash table implementation
388      - threw out most of the pathnames stuff, as it was not used anyway
389      - simplified tokentable implementation
390      - fixed some minor spelling errors
391    
392    ----------------------------------------------------------------------
393    Name: Matthias Blume (blume (at) tti - c (dot) org)
394    Date: 2004/10/06 15:15:15 CDT
395    Tag: blume-20041006-handler
396    Description:
397    
398    Cleaned up the absyn to reflect the invariant that HANDLE always
399    carries a FNexp as part of the type definition.  This eliminates some
400    superfluous sanity checks at runtime down the road.
401    
402    Some minor cleanup of the btrace code.
403    
404    ----------------------------------------------------------------------
405    Name: Matthias Blume (blume (at) tti - c (dot) org)
406    Date: 2004/10/01 10:20:30 CDT
407    Tag: blume-20041001-slave
408    Description:
409    
410    Added hack to make slave mode work in the presence of the version
411    tool.  (Still, since the master does two passes over the code for
412    CMB.make, the release number gets bumped twice when slaves are
413    attached.  I don't know if this is worth fixing...)
414    
415    ----------------------------------------------------------------------
416    Name: Matthias Blume (blume (at) tti - c (dot) org)
417    Date: 2004/09/30 10:55:00 CDT
418    Tag: blume-20040930-version
419    Description:
420    
421    * Moved the "version" magic into its own little library under
422      src/system/smlnj/internal.  This avoids expensive reconstruction of
423      a stable src/compiler/core.cm.
424    
425    * At the same time, structure CompilerVersion is now known as
426      structure SMLNJVersion.
427    
428    * Arranged for the version tool to NOT kick in when rebuilding the system
429      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
430      Also, loading the versiontool does not work when rebuilding the system
431      because CM is not properly initialized at that time.
432    
433    ----------------------------------------------------------------------
434    Name: Matthias Blume (blume (at) tti - c (dot) org)
435    Date: 2004/09/29 14:00:00 CDT
436    Tag: blume-20040929-autoversion
437    Description:
438    
439    Implemented some CM magic to have
440        file src/compiler/TopLevel/main/version.sml
441    generated automagically.
442    The version is taken from two files: config/version and config/release.
443    The first is expected to contain a two-part version number such as 110.49.
444    The second should contain a single number, but it may be missing.
445    
446    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
447    time the version tool is loaded (which is the first time you say CMB.make),
448    then the tool will increment the value stored in config/release every
449    time CMB.make is invoked.
450    
451    The binfile format is now insensitive to anything beyond the first
452    two components of a version number, so bumping the release does not render
453    binfiles incompatible.  Auto-bumping can be used to keep track of versions
454    during development without invalidating existing binfiles.
455    
456    In any case, every CMB.make updates the date information in version.sml.
457    (This is the date that is printed in the banner.)
458    
459    ----------------------------------------------------------------------
460    Name: Matthias Blume (blume (at) tti - c (dot) org)
461    Date: 2004/09/28 10:53:00 CDT
462    Tag: blume-20040928-controls
463    Description:
464    
465    Some cleanup of the controls code.
466    
467    ----------------------------------------------------------------------
468    Name: Matthias Blume (blume (at) tti - c (dot) org)
469    Date: 2004/09/27 22:08:00 CDT
470    Tag: blume-20040927-controls
471    Description:
472    
473    Added two pieces of functionality to the Controls interface:
474    
475    1.   val save'restore: 'a control -> unit -> unit
476    
477       grabs the current value of the control in stage 1 and restores it
478       in stage 2.
479    
480    2.   val set' : 'a control * 'a -> unit -> unit
481    
482       stores the given value into the control in stage 2 (i.e., delayed)
483       but does all error checking in stage 1.
484       (This is for string controls that need to do parse their argument --
485        something that might fail.  In some cases, notably in CM, one
486        already knows the intended argument but wants to delay the actual
487        assignment until a time when error recovery would be more difficult.)
488    
489    Changed the handling of controls in tool arguments to classes "sml" and
490    "lazysml":
491       - use Controls.save'restore as a more robust way of restoring the
492         old value (in particular: without having to re-parse the string)
493       - use controls to handle the "overload" keyword in the init group
494         (I believe this change actually fixes a long-standing obscure bug.)
495    
496    ----------------------------------------------------------------------
497    Name: Matthias Blume (blume (at) tti - c (dot) org)
498    Date: 2004/09/27 17:00:00 CDT
499    Tag: blume-20040927-lazysml
500    Description:
501    
502    Added a new tool class called "lazysml" to CM's tool chest.  The only
503    difference to "sml" is that compilation is done with Control.lazysml
504    set to true.  A source of class "lazysml" is automatically recognized
505    by a file name suffix of ".lml".
506    
507    In addition to the above feature, the original class "sml" now also
508    supports a tool argument "lazy" which has the same effect.  As a
509    result, the following three lines are equivalent:
510    
511        foo.sml : lazysml
512        foo.sml : sml (lazy)
513        foo.sml (lazy)
514    
515    The setting goes into effect both during parsing and during
516    compilation.  The original setting is restored right after parsing and
517    after compilation, respectively.
518    
519    In addition to all the above, there is also a general mechanism to set
520    ANY of the "controls" that are available at the command line via
521    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
522    apply as well. (In fact, "lazy" is implemented as a special case of
523    the general mechanism.)
524    
525    The .cm file syntax uses a new keyword tool argument called "with".
526    There are several ways of indicating the desired settings:
527    
528        foo.sml (with:parser.quotations=true)
529        foo.sml (with:(name:parser.quotations value:true))
530        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
531        foo.sml (with:(name1=value1 name2=value2 ...))
532        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
533    
534        etc.
535    
536    Another possible abbreviation is to leave out the =v or value:v part
537    if the name refers to a boolean control (in which case the value is
538    taken to be true).  Thus, one could get lazy sml also by saying:
539    
540        foo.sml (with:parser.lazy-keyword=true)
541        foo.sml (with:parser.lazy-keyword)
542        foo.sml (with:(name:parser.lazy-keyword value:true))
543        foo.sml (with:(name:parser.lazy-keyword))
544    
545    ----------------------------------------------------------------------
546    Name: Matthias Blume (blume (at) tti - c (dot) org)
547    Date: 2004/09/24 16:22:00 CDT
548    Tag: blume-20040924-ppc-long-branch
549    Description:
550    
551    Turned message about "emiting long form of branch" off by default.
552    Added a control flag to turn it back on when desired.
553    
554    ----------------------------------------------------------------------
555    Name: Matthias Blume (blume (at) tti - c (dot) org)
556    Date: 2004/09/24 16:05:00 CDT
557    Tag: blume-20040924-rounding
558    Description:
559    
560    Applied patch for setting rounding modes under Mac OS X.  Thanks to
561    Melissa O'Neill for providing the code!
562    
563    ----------------------------------------------------------------------
564    Name: Matthias Blume (blume (at) tti - c (dot) org)
565    Date: 2004/09/23 17:30:00 CDT
566    Tag: blume-20040923-envvars
567    Description:
568    
569    1. Changed definition of type ControlRegistry.registry_tree to
570       include control_info (i.e., the name of the controlling environment
571       variable).
572    
573    2. Added command-line flags -e and -E to print the names of environment
574       variables that can be used to control internal settings.
575       (This uses the new API mentioned in 1.)
576    
577    ----------------------------------------------------------------------
578    Name: Matthias Blume (blume (at) tti - c (dot) org)
579    Date: 2004/09/13 16:50:00 CDT
580    Tag: Release_110_49
581    Description:
582    
583    New working version (110.49).  NEW BOOTFILES!
584    
585    ----------------------------------------------------------------------
586    Name: Matthias Blume
587    Date: 2004/09/13 16:20:00 CDT
588    Tag: blume-20040913-config-mlrisc
589    Description:
590    
591    Put target "mlrisc" back into the default list.
592    (There is no harm in having it, and some users have expressed their
593    wish to have "mlrisc" included by default.)
594    
595    ----------------------------------------------------------------------
596    Name: John Reppy
597    Date: 2004/09/13
598    Tag: jhr-20040913-signals
599    Description:
600    
601    Fixed the signal masking code to properly nest mask/unmask operations
602    on a per-signal basis.
603    
604    ----------------------------------------------------------------------
605    Name: Matthias Blume (blume (at) tti - c (dot) org)
606    Date: 2004/09/08 13:20:00 CDT
607    Tag: blume-20040908-heap-magic
608    Description:
609    
610    Bumped the heap macig number to 0x09082004 to account for the changed
611    layout of the ML frame under MacOS X.
612    
613    ----------------------------------------------------------------------
614    Name: Allen Leung (leunga (at) reservoir (dot) com)
615    Date: 2004/09/03 11:26:00 EST
616    Tag: leunga-20040903-cygwin-install
617    Description:
618    
619       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
620    The Cygwin runtime is turned on by setting the environment
621    variable SMLNJ_CYGWIN_RUNTIME to 1.
622    
623    ----------------------------------------------------------------------
624    Name: Matthias Blume (blume (at) tti - c (dot) org)
625    Date: 2004/08/31 17:15:00 CDT
626    Tag: blume-20040831-core
627    Description:
628    
629    Added some exports to src/compiler/core.cm upon request by J. Joemann.
630    
631    ----------------------------------------------------------------------
632    Name: Matthias Blume (blume (at) tti - c (dot) org)
633    Date: 2004/08/30 17:55:00 CDT
634    Tag: blume-20040830-installer
635    Description:
636    
637    Upon request by Johannes Joemann:
638    
639    - improved ML code of installer to fall back to coping when renaming
640      fails (i.e., when source and target are on different file systems);
641      the code compiles but has yet to be tested in anger
642    - removed mlrisc from list of default targets (config/targets)
643    
644    ----------------------------------------------------------------------
645    Name: Matthias Blume (blume (at) tti - c (dot) org)
646    Date: 2004/08/27 17:20:00 CDT
647    Tag: blume-20040827-ptreql
648    Description:
649    
650    Added ptreql primop to structure InlineT (upon request from Larry
651    Paulson).
652    
653    ----------------------------------------------------------------------
654    Name: Allen Leung (leunga (at) reservoir (dot) com)
655    Date: 2004/08/15 21:21:00 EST
656    Tag: leunga-110_48-udgraph
657    Description:
658    
659    Another bug fix from Carl Hauser:
660    
661    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
662    > 48c48
663    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
664    > e::L)
665    > ---
666    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
667    > Without this, any deletion of an edge in an undirected graph does severe
668    > violence to the graph.
669    
670    ----------------------------------------------------------------------
671    Name: Allen Leung (leunga (at) reservoir (dot) com)
672    Date: 2004/08/10 23:35:00 EST
673    Tag: leunga-110_48-ppc
674    Description:
675    
676       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
677    
678    ----------------------------------------------------------------------
679    Name: Matthias Blume (blume (at) tti - c (dot) org)
680    Date: 2004/08/10 12:00:00 CDT
681    Tag: Release_110_48
682    Description:
683    
684    New working version (110.48).  NEW BOOTFILES!
685    
686    ----------------------------------------------------------------------
687    Name: Allen Leung (leunga (at) reservoir (dot) com)
688    Date: 2004/08/09 12:21:00 EST
689    Tag: leunga-110_47-dijsktra
690    Description:
691    
692    Bug fix from Carl Hauser:
693    
694    single_source_shortest_paths in dijkstra.sml was observed to get wrong
695    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
696    
697    The problem is that following the expression A.update(dist,s,Num.zero)
698    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
699    
700    ----------------------------------------------------------------------
701    Name: Matthias Blume (blume (at) tti - c (dot) org)
702    Date: 2004/08/06 18:10:00 CDT
703    Tag: blume-20040806-cmdline
704    Description:
705    
706    Fiddled with handling of command-line options:
707    
708      * sml now quits after processing the command line
709        if -H, -S, -h<n>, or -s<n> appears as the last
710        command-line argument
711      * a new option -q terminates the session when encountered on
712        the command line; subsequent arguments will be ignored
713      * bug fixes: short (erroneous) arguments are no longer ignored
714        completely
715    
716    ----------------------------------------------------------------------
717    Name: Allen Leung (leunga (at) reservoir (dot) com)
718    Date: 2004/08/04 18:17:00 EST
719    Tag: leunga-110_47-ppc-ibm-asm
720    Description:
721    
722      - Added minimal IBM assembly syntax support for PowerPC.
723    
724      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
725        been altered in the runtime.  We need an automatic way to keep the file
726        in sync.
727    
728    ----------------------------------------------------------------------
729    Name: Matthias Blume (blume (at) tti - c (dot) org)
730    Date: 2004/08/04 14:00:00 CDT
731    Tag: Release_110_47
732    Description:
733    
734    New working version (110.47).  NEW BOOTFILES!
735    
736    ----------------------------------------------------------------------
737    Name: Matthias Blume (blume (at) tti - c (dot) org)
738    Date: 2004/08/03 14:25:00 CDT
739    Tag: blume-20040803-callingconv
740    Description:
741    
742    Added low-level support for choosing C calling conventions by
743    twiddling the type of rawccall.  (See
744    src/compiler/Semant/types/cproto.sml for details.)
745    
746    ----------------------------------------------------------------------
747    Name: Matthias Blume (blume (at) tti - c (dot) org)
748    Date: 2004/08/02 15:55:00 CDT
749    Tag: blume-20040802-backout
750    Description:
751    
752    Backed out of change to win32-filesys.c.  The earlier patch to
753    get_file_time caused CM to produce files with the wrong time stamp.
754    
755    ----------------------------------------------------------------------
756    Name: Matthias Blume (blume (at) tti - c (dot) org)
757    Date: 2004/08/02 14:45:00 CDT
758    Tag: blume-20040802-nlffi-win32
759    Description:
760    
761    Added NLFFI support for Win32, adapted from a patch provided by David
762    Hansel.  This is currently completely untested.  Also, the issue
763    concerning stdcall vs. ccall is still unresolved.
764    
765    ----------------------------------------------------------------------
766    Name: Matthias Blume (blume (at) tti - c (dot) org)
767    Date: 2004/07/30 17:55:00 CDT
768    Tag: blume-20040730-various
769    Description:
770    
771    Gearing up towards 110.47...
772    
773    - various minor bugfixes to ml-nlffigen
774    - a beginning of a manual for nlffi
775    
776    - eliminated 'export name=value' in config/install.sh as this does
777      not work with certain versions of /bin/sh
778      (Thanks to David King at Motorola for catching this.)
779    
780    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
781      - added a test for tm==NULL to gmtime.c and localtime.c
782      - applied patch for incorrect GetFileTime under win32
783      - toSeconds -> toMilliseconds in Win32/win32-process.sml
784    
785    ----------------------------------------------------------------------
786    Name: Matthias Blume (blume (at) tti - c (dot) org)
787    Date: 2004/07/21 18:20:00 CDT
788    Tag: blume-20040721-nlffigen
789    Description:
790    
791    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
792      for a typedef to an incomplete type, but leave out the "typ" member.
793      (This is just for consistency.)
794    - Started to produce what is supposed to become better (i.e., comprehensive)
795      documentation of what ml-nlffigen does and produces.
796    
797    ----------------------------------------------------------------------
798    Name: Matthias Blume (blume (at) tti - c (dot) org)
799    Date: 2004/07/14 16:25:00 CDT
800    Tag: blume-20040714-union
801    Description:
802    
803    Added C_UNION to c-calls/c-types.sml and updated the machinery
804    (ml-nlffigen, cproto.sml) that conveys C function interface
805    information to the code generator.
806    
807    However, the actual architecture-specific implementation of function
808    arguments and results that are C unions is still not implemented.
809    
810    ----------------------------------------------------------------------
811    Name: Allen Leung (leunga (at) reservoir (dot) com)
812    Date: 2004/07/14 14:38:00 EST
813    Tag: leunga-110_46_1-ppc-lwzu
814    Description:
815    
816        Added these instructions to the PowerPC architecture:
817        LBZU(X), LHZU(X), LWZU(X),
818        STWU(X), STFDU, STFSU
819    
820        etc...
821    
822        Note: I haven't added their instruction encoding into the description.
823    
824    ----------------------------------------------------------------------
825    Name: Allen Leung (leunga (at) reservoir (dot) com)
826    Date: 2004/07/13 15:04:00 EST
827    Tag: leunga-110_46_1-ppc-lwarx
828    Description:
829    
830       Added the two instructions LWARX and STWCX to the PowerPC
831    instruction set.
832    
833       A (untested) rewrite of loop-structure.sml.  The old version
834    is completely broken.
835    
836    ----------------------------------------------------------------------
837    Name: Matthias Blume (blume (at) tti - c (dot) org)
838    Date: 2004/07/13 13:50:00 CDT
839    Tag: blume-20040713-nlffi
840    Description:
841    
842    - use paramAlloc to report c-calls with too many arguments
843      (for PPC version where parameter area is pre-allocated)
844    - added ccall_maxargspace to machspec (to implement the above)
845    - made "make" commend in CM's "make" tool configurable
846    - added option (default: on) for passing the name of the SML/NJ's "bin"
847      directory to "make"; the call looks like this:
848    
849         make <options> SMLNJ_BINDIR=<dir> <target>
850    
851      This can be used by the Makefile to, e.g., pick the "right" version
852      of ml-nlffigen.
853    - minor code tweaks
854    
855    ----------------------------------------------------------------------
856    Name: Matthias Blume (blume (at) tti - c (dot) org)
857    Date: 2004/07/12 22:50:00 CDT
858    Tag: blume-110_46_1-macosx-nlffi
859    Description:
860    
861    NLFFI under Mac OS X now working (sort of).  This is largely untested,
862    though.
863    
864    Note:  1. You have to make a new, clean build of the runtime system.
865           2. There are new BOOTFILES, you have to use them!
866              (Doing the bootstrap process yourself would be *very* painful!
867               If you absolutely have to do it, build the system under
868               a different architecture and then cross-compile.)
869    
870    Version bumped to 110.46.1 to account for runtime data format changes.
871    
872    ----------------------------------------------------------------------
873    Name: Matthias Blume (blume (at) tti - c (dot) org)
874    Date: 2004/06/18 14:30:00 CDT
875    Tag: blume-20040618-unix
876    Description:
877    
878    Changed the implementation of structure Unix so that the same stream
879    is returned every time one of the {text,bin}{In,Out}streamOf functions
880    is invoked on the same proc.  This is not what the spec currently
881    says -- although IMO it arguably should.  (See discussion below.)
882    
883    ----------------------------------------------------------------------
884    Name: Matthias Blume (blume (at) tti - c (dot) org)
885    Date: 2004/06/17 18:15:00 CDT
886    Tag: Release_110_46
887    Description:
888    
889    New working version (110.46).  NEW BOOTFILES!
890    
891    ----------------------------------------------------------------------
892    Name: Matthias Blume (blume (at) tti - c (dot) org)
893    Date: 2004/06/17 17:20:00 CDT
894    Tag: blume-20040617-timer-unix
895    Description:
896    
897    Changed the interface of structures Timer and Unix to match the most
898    recent Basis spec.
899    
900    In the case of Unix there still seems to be an open/weird issue:
901    
902       The {text,bin}{In,Out}streamOf functions are supposed to create
903       fresh streams whenever they are called -- as opposed to have them
904       return the same stream every time.  This design is supposed to
905       prevent space leaks caused by proc values hanging on to streams.
906    
907       The reap function, on the other hand, is supposed to close the
908       streams.  This cannot be done without having a handle on the
909       stream in proc after all...
910    
911    I took the liberty to implement the following stopgap solution:
912    
913      The proc value hangs on to the most recently created stream(s).
914      Reap closes those.  If either or both of the two streams hadn't
915      been created at all yet, then reap will close the corresponding
916      file descriptors directly.
917    
918    PS: I don't understand the original space leak argument anymore.  If
919    a proc hangs on to the imperative stream, then I/O operations on those
920    will advance the state of the cached stream and avoid the space leak.
921    
922    ----------------------------------------------------------------------
923    Name: Matthias Blume (blume (at) tti - c (dot) org)
924    Date: 2004/05/28 16:45:00 CDT
925    Tag: blume-20040528-basis
926    Description:
927    
928    Added signature PACK_REAL and exported functor PrimIO.
929    
930    ----------------------------------------------------------------------
931    Name: Matthias Blume (blume (at) tti - c (dot) org)
932    Date: 2004/05/25 16:00:00 CDT
933    Tag: blume-20040525-group-owner
934    Description:
935    
936    CM now ignores (but still accepts) the "owner" information in group
937    descriptions.  The owner of a group is its next enclosing
938    library. Each group must have a unique owner.  (There is a virtual
939    "toplevel" library that owns groups which are not nested within a real
940    library.)  Previously, each group had to explicitly declare its owner,
941    and CM would check that such a declaration is correct.  The new scheme
942    is to have CM check that for each group there is precisely one owning
943    library.
944    
945    The advantage of the new scheme is that the programmer no longer needs
946    to maintain the somewhat annoying owner information.  The downside is
947    that CM cannot enforce the ownership rule across multiple runs of
948    CM.make.  Fortunately, enclosing the same group in two different
949    libraries A and B which are not part of the same program does not
950    cause real problems.
951    
952    ----------------------------------------------------------------------
953    Name: Matthias Blume (blume (at) tti - c (dot) org)
954    Date: 2004/05/20 16:00:00 CDT
955    Tag: blume-20040520-win32
956    Description:
957    
958    Made the win32 version work again.  (Strangely, a misplaced comma had
959    slipped into win32-process.c which prevented the runtime from being
960    compiled correctly.)
961    
962    Also, included a minor addition to ml-build.bat analogous to what was
963    done in blume-20040519-ml-build.
964    
965    ----------------------------------------------------------------------
966    Name: Matthias Blume (blume (at) tti - c (dot) org)
967    Date: 2004/05/19 22:10:00 CDT
968    Tag: blume-20040519-ml-build
969    Description:
970    
971    Arranged for ml-build to clean up after itself a little bit better.
972    The script generates a temporary SML source file and compiles it using
973    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
974    gets rid of those at the end, so they don't accumulate under .cm.
975    
976    This required a minor change to install.sh because the name of the
977    metadata directory (default: .cm) is actually configurable at
978    installation time.
979    
980    ----------------------------------------------------------------------
981    Name: Matthias Blume (blume (at) tti - c (dot) org)
982    Date: 2004/05/18 15:50:00 CDT
983    Tag: blume-20040518-mkreader
984    Description:
985    
986    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
987    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
988    
989    ----------------------------------------------------------------------
990    Name: Matthias Blume (blume (at) tti - c (dot) org)
991    Date: 2004/05/11 14:35:00 CDT
992    Tag: blume-20040511-win32sock
993    Description:
994    
995    Added previously missing support for many socket-related functions
996    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
997    for the voluminous patch!
998    
999    (I have not tested this patch under win32 yet.)
1000    
1001    Here is David's e-mail:
1002    
1003    Hi,
1004    
1005    Attached to this email you find a diff against sml/nj 110.45
1006    that will enable socket support under Windows.
1007    
1008    To apply the patch (using unix or cygwin)
1009    1) gunzip runtime.diff.gz
1010    2) "cd" into "src/runtime" in the source tree of a fresh
1011       110.45 installation.
1012    3) patch -p 1 < [your/path/to]runtime.diff
1013    
1014    The code compiles fine but has NOT yet been extensively tested.
1015    I only ran a few tests for basic socket client functionality
1016    (which worked fine).  Especially the functions that use ioctl
1017    are not tested at all and might not work (see below).
1018    
1019    I implemented this since we want to move to a newer version of sml/nj
1020    but need socket support in order to use it.  This is the first time I
1021    even had a look at the sml/nj source,  so please review my changes
1022    before making this part of the distribution!  Here are a few issues
1023    that I think might be better for someone to solve who is more
1024    familiar with the sml/nj source (and socket programming):
1025    
1026    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1027      exception since I could not figure out what the windows equivalent
1028      of these functions is
1029    
1030    - In sockets-osdep.h there are a some #include statements that are
1031      only used in a few files that include sockets-osdep.h
1032    
1033    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1034      does not process its return value since I don't know how to
1035      report an error upwards.
1036    
1037    - It would probably be good to have a call to WSACleanup() when
1038      the library is unloaded (if there is such a possibility).
1039      Otherwise I think Windows will take care of this automatically
1040      when the process finishes.
1041    
1042    - I used ioctlsocket() as a replacement for ioctl() but I have
1043      no idea if that is actually the proper replacement on Windows.
1044    
1045    - All these issues are marked in the code by "FIXME" comments.
1046    
1047    We use sml/nj extensively in our products and are quite happy
1048    with it.  I hope this contribution will help you.
1049    
1050    Keep up the good work!
1051    
1052    David
1053    
1054    ----------------------------------------------------------------------
1055    Name: Matthias Blume (blume (at) tti - c (dot) org)
1056    Date: 2004/05/11 14:20:00 CDT
1057    Tag: blume-20040511-installml
1058    Description:
1059    
1060    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1061    the patch.)
1062    
1063    ----------------------------------------------------------------------
1064    Name: Matthias Blume (blume (at) tti - c (dot) org)
1065    Date: 2004/05/11 14:05:00 CDT
1066    Tag: blume-20040511-nlffi-netbsd
1067    Description:
1068    
1069    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1070    the patch.)
1071    
1072    ----------------------------------------------------------------------
1073    Name: Matthias Blume (blume (at) tti - c (dot) org)
1074    Date: 2004/05/11 12:05:00 CDT
1075    Tag: blume-20040511-exports
1076    Description:
1077    
1078    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1079    export lists in compiler-related .cm-files.
1080    
1081    ----------------------------------------------------------------------
1082    Name: Matthias Blume (blume (at) tti - c (dot) org)
1083    Date: 2004/05/11 11:35:00 CDT
1084    Tag: blume-20040511-allsource
1085    Description:
1086    
1087    The installer now honors the "src-smlnj" target again, although its meaning
1088    has changed from "all sources required for the compiler" to "all sources
1089    the installer knows about".  In other words, if you enable "src-smlnj"
1090    in the "targets" file, then the installer will pull in sources for
1091    everything.  (Notice that this refers to source code only.  Compiled
1092    code is still only installed for modules that were requested explicitly
1093    or which are required for other modules that were requested explicitly.)
1094    
1095    ----------------------------------------------------------------------
1096    Name: Matthias Blume (blume (at) tti - c (dot) org)
1097    Date: 2004/04/23 17:40:00 CDT
1098    Tag: blume-20040423-ieee-scan
1099    Description:
1100    
1101    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1102    in the exponent calculation we get INF or ZERO (depending on the mantissa
1103    and the sign of the exponent).
1104    
1105    ----------------------------------------------------------------------
1106    Name: Matthias Blume (blume (at) tti - c (dot) org)
1107    Date: 2004/04/23 10:40:00 CDT
1108    Tag: blume-20040423-ml-build
1109    Description:
1110    
1111    The ml-build script now terminates with a non-0 status when something
1112    goes wrong.
1113    
1114    ----------------------------------------------------------------------
1115    Name: Matthias Blume (blume (at) tti - c (dot) org)
1116    Date: 2004/04/22 16:35:00 CDT
1117    Tag: blume-20040422-Option
1118    Description:
1119    
1120    Made exception Option to be the same as exception Option.Option
1121    (as it should be).
1122    
1123    ----------------------------------------------------------------------
1124    Name: Allen Leung (leunga (at) reservoir (dot) com)
1125    Date: 2004/03/19 14:40:00 EST
1126    Tag: leunga-20040319-cygwin-nlffi
1127    Description:
1128    
1129    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1130    of SML/NJ.  The problem is that
1131    
1132       lib = dlopen(NULL, ...)
1133       f   = dlsym(lib, "malloc");
1134    
1135    does not work on Windows unless we explicitly export symbols
1136    such as 'malloc' during linking.  We fixed this by explicitly
1137    exporting the required symbols with the magic gcc incantation:
1138    
1139        -Wl,--export-all cygwin.def
1140    
1141    where cygwin.def is a file containing all the symbols that
1142    we wish to export.
1143    
1144    I suspect this is a Windows problem and we'll have to
1145    do the same (somehow with windows compilers) when
1146    we build the native win32 version with the system
1147    calls LoadLibrary/GetProcAddress.
1148    
1149    ----------------------------------------------------------------------
1150    Name: Matthias Blume (blume (at) tti - c (dot) org)
1151    Date: 2004/03/04 16:35:00 CST
1152    Tag: blume-20040304-intinf-fmt
1153    Description:
1154    
1155    Fixed problem with IntInf.fmt (sign would show up on the right instead
1156    of on the left for BIN, OCT, and HEX).
1157    
1158    ----------------------------------------------------------------------
1159    Name: Matthias Blume (blume (at) tti - c (dot) org)
1160    Date: 2004/03/04 11:25:00 CST
1161    Tag: blume-20040304-symlinks
1162    Description:
1163    
1164    Fixed problem with installer script (unix only) where bin/ml-yacc and
1165    friends pointed (via symlinks) to absolute locations instead of just
1166    .run-sml.  This was reported by Vesa A Norrman.
1167    
1168    ----------------------------------------------------------------------
1169    Name: Matthias Blume (blume (at) tti - c (dot) org)
1170    Date: 2004/02/13 14:50:00 CST
1171    Tag: Release_110_45
1172    Description:
1173    
1174    New working version (110.45). New bootfiles.
1175    
1176    ----------------------------------------------------------------------
1177    Name: Matthias Blume (blume (at) tti - c (dot) org)
1178    Date: 2004/01/26 15:15:15 CST
1179    Tag: blume-20040126-toplevel
1180    Description:
1181    
1182    Improved handling of exceptions at the interactive toplevel.
1183    
1184    ----------------------------------------------------------------------
1185    Name: Matthias Blume (blume (at) tti - c (dot) org)
1186    Date: 2004/01/26 11:25:00
1187    Tag: blume-20040126-app
1188    Description:
1189    
1190    Type of top-level "app" corrected.
1191    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1192    
1193    ----------------------------------------------------------------------
1194    Name: Matthias Blume (blume (at) tti - c (dot) org)
1195    Date: 2003/11/18 17:10 CST
1196    Tag: blume-20031118-basis-fiddle
1197    Description:
1198    
1199    - changed Timer interface to what might become the spec
1200    - POSIX_FLAGS -> BIT_FLAGS according to spec
1201    - some other minor discrepancies wrt. spec eliminated
1202    
1203    ----------------------------------------------------------------------
1204    Name: Matthias Blume (blume (at) tti - c (dot) org)
1205    Date: 2003/11/06 12:00:00 CST
1206    Tag: Release_110_44
1207    Description:
1208    
1209    New working version (110.44). New bootfiles.
1210    
1211    ----------------------------------------------------------------------
1212    Name: Matthias Blume (blume (at) tti - c (dot) org)
1213    Date: 2003/11/04 11:50:00 CST
1214    Tag: blume-20031104-move-libraries
1215    Description:
1216    
1217    Eliminated the "dont_move_libraries" directive in config/targets.
1218    (The mechanism was broken and could not be fixed easily.  Moreover,
1219    there does not seem to be any reason not to move all libraries into
1220    lib during installation.  I originally implemented this directive as a
1221    backward-compatibility feature when I first introduced the new CM.
1222    Now that things have been stable for a long time and going back to the
1223    old CM is not an option, there is no reason to keep it around.)
1224    
1225    ----------------------------------------------------------------------
1226    Name: Matthias Blume (blume (at) tti - c (dot) org)
1227    Date: 2003/11/03 16:00:00 CST
1228    Tag: blume-20031103-installdir
1229    Description:
1230    
1231    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1232    Richards for pointing out the problem and providing the solution.)
1233    
1234    ----------------------------------------------------------------------
1235    Name: Matthias Blume (blume (at) tti - c (dot) org)
1236    Date: 2003/10/01 17:05:00 CDT
1237    Tag: blume-20031001-lal-mlrisc
1238    Description:
1239    
1240    MLRISC bug fix from Lal.
1241    
1242    ----------------------------------------------------------------------
1243    Name: Matthias Blume (blume (at) tti - c (dot) org)
1244    Date: 2003/09/30 16:10:00 CDT
1245    Tag: blume-20030930-primio-bat
1246    Description:
1247    
1248    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1249    2. Improved .bat files (for Win32 port) to make things work under Win95.
1250       (thanks to Aaron S. Hawley for this one)
1251    
1252    ----------------------------------------------------------------------
1253    Name: Matthias Blume (blume (at) tti - c (dot) org)
1254    Date: 2003/09/26 16:05:00 CDT
1255    Tag: blume-20030926-wrappriv
1256    Description:
1257    
1258    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1259    
1260    ----------------------------------------------------------------------
1261    Name: Matthias Blume (blume (at) tti - c (dot) org)
1262    Date: 2003/09/26 15:00:00 CDT
1263    Tag: blume-20030926-110_43_3
1264    Description:
1265    
1266    - additional cleanup
1267    - version number bump, NEW BOOTFILES
1268    
1269    ----------------------------------------------------------------------
1270    Name: Matthias Blume (blume (at) tti - c (dot) org)
1271    Date: 2003/09/26 12:00:00 CDT
1272    Tag: blume-20030926-ppautoload
1273    Description:
1274    
1275    I modified the read-eval-print loop so that the autoloader gets
1276    invoked whenever the prettyprinter tries to look up a symbol that
1277    is not currently defined in the toplevel environment but which
1278    appears in CM's autoload registry.  As a result, we see far fewer of
1279    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1280    
1281    In addition to this I tried to clean up some pieces of the Basis
1282    implementation (e.g., Socket, Word8Array) in order to prevent other
1283    instances of these ?.Foo.Bar.xxx names from being printed.
1284    
1285    The mechanism that picks names for types still needs some work, though.
1286    (Right now it seems that if there is a type A.t which is defined to
1287    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1288    "?.B.u" although the perhaps more sensible solution would be to use
1289    "A.t" in this case.  In other words, the prettyprinter should follow
1290    a chain of DEFtycs not farther than there are corresponding toplevel
1291    names in the current environment.)
1292    
1293    ----------------------------------------------------------------------
1294    Name: Matthias Blume (blume (at) tti - c (dot) org)
1295    Date: 2003/09/24 16:31:00 CDT
1296    Tag: blume-20030924-installer
1297    Description:
1298    
1299    Another installer tweak:  All the ML code for the installer is now
1300    compiled during CMB.make and put into a little library called
1301    $smlnj/installer.cm.  The installation then simply invokes
1302    
1303       sml -m $smlnj/installer.cm
1304    
1305    and everything happens automagically.
1306    
1307    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1308    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1309            and BINDIR.
1310    
1311    The new scheme guarantees that the ML code responsible for the installation
1312    is in sync with the APIs of the main system.  Also, the installer is
1313    somewhat faster because the installer script is precompiled.
1314    
1315    ----------------------------------------------------------------------
1316    Name: Matthias Blume (blume (at) tti - c (dot) org)
1317    Date: 2003/09/24 15:35:00 CDT
1318    Tag: blume-20030924-synsock
1319    Description:
1320    
1321    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1322    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1323    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1324    implementations of non-blocking operations from CML's Socket
1325    structure.
1326    
1327    ----------------------------------------------------------------------
1328    Name: Matthias Blume (blume (at) tti - c (dot) org)
1329    Date: 2003/09/24 15:10:05 CDT
1330    Tag: blume-20030924-sockets
1331    Description:
1332    
1333    1. Fixed SOCKET API and implementation to match Basis spec.
1334       This required changing the internal representation of sockets to one
1335       that remembers (for each socket file descriptor) whether it is currently
1336       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1337       call is made only if the state actually needs to change).
1338    
1339    2. OS-specific details of sockets were moved into separate files, thus
1340       making it possible to unify the bulk of the socket implementations
1341       between Unix and Win32.
1342    
1343    3. CML's socket API changed accordingly.
1344       (Note that we need to remove non-blocking functions from this API
1345       since they are redundant in the case of CML!)
1346    
1347    4. CML's socket implementation now makes use of non-blocking functions
1348       provided by Basis, thus removing all OS-dependent code from this part
1349       of CML.
1350    
1351    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1352    
1353    ----------------------------------------------------------------------
1354    Name: Matthias Blume (blume (at) tti - c (dot) org)
1355    Date: 2003/09/22 12:10:00 CDT
1356    Tag: blume-20030922-110_43_2
1357    Description:
1358    
1359    Made a new interim version and bootfiles for developer's bootstrapping
1360    convenience.
1361    
1362       110.43.2 -- NEW BOOTFILES
1363    
1364    ----------------------------------------------------------------------
1365    Name: Matthias Blume (blume (at) tti - c (dot) org)
1366    Date: 2003/09/19 15:55:00 CDT
1367    Tag: blume-20030919-cmdir
1368    Description:
1369    
1370    1. new-install.sh -> install.sh
1371    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1372    3. tweaked installer so that another name instead of .cm can be chosen
1373       at install time (by setting the CM_DIR_ARC environment variable
1374       during installation); once installation is complete, the name is
1375       fixed
1376    
1377    ----------------------------------------------------------------------
1378    Name: Matthias Blume (blume (at) tti - c (dot) org)
1379    Date: 2003/09/18 16:00:00 CDT
1380    Tag: blume-20030918-110_43_1
1381    Description:
1382    
1383    Made a new interim version and bootfiles for developer's bootstrapping
1384    convenience.
1385    
1386       110.43.1 -- NEW BOOTFILES
1387    
1388    ----------------------------------------------------------------------
1389    Name: Matthias Blume (blume (at) tti - c (dot) org)
1390    Date: 2003/09/18 15:20:00 CDT
1391    Tag: blume-20030918-misc
1392    Description:
1393    
1394    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1395       this seems to be controversial at the moment) and used those in
1396       Posix.ProcEnv.times.
1397    
1398    2. Added Time.{from,to}Nanoseconds to Time.
1399    
1400    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1401       For example, fromLargeInt never needs to look at more than 3 "big
1402       digits" to get its 53 bits of precision.
1403    
1404    ----------------------------------------------------------------------
1405    Name: Matthias Blume (blume (at) tti - c (dot) org)
1406    Date: 2003/09/17 16:30:00 CDT
1407    Tag: blume-20030917-real32-slices
1408    Description:
1409    
1410    Added an entry to the primitive environment
1411    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1412    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1413    
1414    Removed some of the "magic" constants in real64.sml and replaced them
1415    with code that generates these values from their corresponding
1416    integer counterparts.
1417    
1418    Made all(?) the slice-related changes to the Basis and made everything
1419    compile again...
1420    
1421    ----------------------------------------------------------------------
1422    Name: Matthias Blume (blume (at) tti - c (dot) org)
1423    Date: 2003/09/15 17:45:00 CDT
1424    Tag: blume-20030915-rbase
1425    Description:
1426    
1427    Fixed bug in Real.fromLargeInt.
1428    
1429    ----------------------------------------------------------------------
1430    Name: Matthias Blume (blume (at) tti - c (dot) org)
1431    Date: 2003/09/13 18:11:00 CDT
1432    Tag: blume-20030913-libinstall
1433    Description:
1434    
1435    Minor bugfix in config/libinstall (set anchor with path to
1436    standalone tool after installing it, otherwise libraries that
1437    need ml-lex or ml-yacc won't compile the first time the installer
1438    runs).
1439    
1440    ----------------------------------------------------------------------
1441    Name: Matthias Blume (blume (at) tti - c (dot) org)
1442    Date: 2003/09/12 11:45:00 CDT
1443    Tag: blume-20030912-various
1444    Description:
1445    
1446     - fixed bug in Real.toLargeInt
1447     - fixed bug in Posix.ProcEnv.times
1448     - changed inputLine functions to return an option
1449     - minor installer improvements / bugfixes
1450     - changed default @SMLalloc parameter for x86/celeron to 64k
1451    
1452    ----------------------------------------------------------------------
1453    Name: Matthias Blume (blume (at) tti - c (dot) org)
1454    Date: 2003/09/09 22:00:00 CDT
1455    Tag: Release_110_43
1456    Description:
1457    
1458        New working release 110.43.  New bootfiles.
1459    
1460    ----------------------------------------------------------------------
1461    Name: Matthias Blume (blume (at) tti - c (dot) org)
1462    Date: 2003/09/09 19:20:00 CDT
1463    Tag: blume-20030909-installer
1464    Description:
1465    
1466    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1467    Modified config/install.bat to take advantage of it.  Also modified
1468    config/install.sh (and called it config/new-install.sh) to take advantage
1469    of it on Unix systems.  (The SML code is (supposed to be) platform-
1470    independent.)
1471    
1472    The installer can now install everything under Win32
1473    as well as under *nix as long as it compiles.
1474    
1475    Other changes:
1476    
1477      - made CML compile again under Win32
1478      - made eXene compile under Win32 (by providing a fake structure UnixSock
1479        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1480      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1481        same as type int;  under Win32 it isn't
1482      - fixed some slice-related problems in the win32-specific parts of CML
1483      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1484        (under Win32, the volume name is case-insensitive, and the
1485        OS.Path code compares volume names for equality)
1486    
1487    ----------------------------------------------------------------------
1488    Name: Matthias Blume (blume (at) tti - c (dot) org)
1489    Date: 2003/09/08 11:55:00 CDT
1490    Tag: blume-20030908-fullpath
1491    Description:
1492    
1493    Made Win32 version of OS.FileSys.fullPath return current directory
1494    when given an empty string.  This is what the spec says, and incidentally,
1495    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1496    cases when presented with the name of a non-existing .cm file.)
1497    
1498    ----------------------------------------------------------------------
1499    Name: Matthias Blume (blume (at) tti - c (dot) org)
1500    Date: 2003/09/04 16:30:00 CDT
1501    Tag: blume-20030905-slices-etc
1502    Description:
1503    
1504      1. Changed interface to vectors and arrays in Basis to match
1505         (draft) Basis spec.
1506      2. Added signatures and implementations of slices according to
1507         Basis spec.
1508      3. Edited source code throughout the system to make it compile again
1509         under 1. and 2.  (In some cases code had to be added to have it
1510         match the new signatures.)
1511      4. MLRISC should be backward-compatible: the copies of the originals
1512         of files that needed to change under 3. were retained, the .cm files
1513         check the compiler version number and use old versions when
1514         appropriate.
1515      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1516         dirstream -> string option (in accordance with Basis spec).
1517      6. When generating code that counts lines, ml-lex used function
1518         CharVector.foldli, taking advantage of its old interface.
1519         This has been replaced with the corresponding code from
1520         CharVectorSlice.  (html-lex must be re-lexed!)
1521      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1522         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1523      8. Removed temporary additions (fromInternal, toInternal) from the
1524         (now obsolete) IntInf in smlnj-lib/Util.
1525      9. Cleaned up structure Byte.
1526     10. Added localOffset, scan, and fromString to Date (according to spec).
1527         Cleaned/corrected implementation of Date.
1528         (Still need to check for correctness; implement better canonicalizeDate.)
1529     11. Added "scan" to signature IEEE_REAL.
1530     12. Some improvements to IntInf [in particular: efficiency-hack for
1531         mod and rem when second operand is 2 (for parity checks).]
1532     13. Changed representation of type Time.time, using a single IntInf.int
1533         value counting microseconds.  This considerably simplified the
1534         implementation of structure Time.  We now support negative time
1535         values; scan and fromString handle signs.
1536     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1537         ArraySlice).
1538    
1539    ----------------------------------------------------------------------
1540    Name: Matthias Blume (blume (at) tti - c (dot) org)
1541    Date: 2003/08/28 17:00:00 CDT
1542    Tag: blume-20030828-intinf
1543    Description:
1544    
1545    This is a major update which comes with a version number bump
1546    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1547    and an implementation of IntInf in the Basis.
1548    
1549    There are a fairly large number of related changes and updates throughout
1550    the system:
1551    
1552    Basis:
1553            - Implemented IntInf.
1554            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1555            - Added some missing Real64 operations, most notably Real.toLargeInt.
1556            - Added FixedInt as a synonym for Int32.
1557    
1558    compiler:
1559            * Added support for a built-in intinf type.
1560              - literals
1561              - pattern matching
1562              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1563              - overloading on literals and operations
1564    
1565            This required adding a primitive type intinf, some additional
1566            primops, and implementations for several non-trivial intinf
1567            operations in Core.  (The intinf type is completely abstract
1568            to the compiler; all operations get delegated back to the Core.)
1569    
1570            * Intinf equality is handled by polyequal.  However, the compiler
1571            does not print its usual warning in this case (since polyequal
1572            is the right thing to do there).
1573    
1574            * Improved the organization of structure InlineT.
1575    
1576            * A word about conversion primops:
1577            If conversions involving intinf do not cancel out during
1578            CPS contract, then the compiler must insert calls to Core functions.
1579            Since all core access must be resolved already during the FLINT
1580            translate phase, it would be too late a the time of CPS contract
1581            to add new Core calls.  For this reason, conversion primops
1582            for intinf carry two arguments:  1. the numeric argument that
1583            they are supposed to convert, and 2. the Core function that
1584            can help with this conversion if necessary.  If CPS contract
1585            eliminates a primop, then the associated Core function becomes
1586            dead and goes away.  Intinf conversion primops that do not get
1587            eliminated by CPS contract get rewritten into calls of their
1588            core functions by a separate, new phase.
1589    
1590    interactive system:
1591            - Control.Print.intinfDepth controls max length of intinf constants
1592            being printed. (Analogous to Control.Print.stringDepth.)
1593            - Cleanup in printutil and pputil: got rid of unused stuff and
1594            duplicates; replaced some of the code with code that makes better
1595            use of library functionality.
1596    
1597    CM:
1598            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1599            are no longer silent.
1600    
1601    CKIT:
1602            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1603            in favor of LargeInt -- which is now the same as IntInf.
1604            CKIT-knowledgable people should check whether this is what's
1605            intended and otherwise change things back to using Int32 or
1606            FixedInt.
1607    
1608    Throughout the code:
1609            Started using IntInf.int literals and built-in operations
1610            (e.g., comparison with 0) where this seems appropriate.
1611    
1612    
1613    ----------------------------------------------------------------------
1614    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1615    Date: 2003/08/13 11:36:00 CDT
1616    Tag: dbm-20030813-mcz-merge1
1617    Description:
1618    
1619    Merging changes from the mcz-branch development branch into trunk.
1620    These changes involve replacement of the emulated old prettyprinter
1621    interface with direct use of the SML/NJ Lib PP library, and fixing
1622    of a couple of bugs (895, 1186) relating to error messages.  A new
1623    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1624    has been added.
1625    
1626    ----------------------------------------------------------------------
1627    Name: Matthias Blume (blume (at) tti - c (dot) org)
1628    Date: 2003/08/11 15:45:00 CDT
1629    Tag: blume-20030811-windows
1630    Description:
1631    
1632    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1633    
1634        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1635    
1636    This patch restores SML/NJ's ability to run under win32.  There are a
1637    number of changes, including fixes for several bugs that had gone
1638    unnoticed until now:
1639    
1640      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1641        trying to run the win32 version from within cygwin.)
1642    
1643      - There are a number of simple .bat scripts that substitute for their
1644        corresponding Unix shell-scripts.  (See below.)
1645    
1646      - The internals of ml-build have been modified slightly.  The main
1647        difference is that instead of calling ".link-sml" (or link-sml.bat)
1648        using OS.Process.system, the ML process delegates this task back
1649        to the script.  Otherwise problems arise in mixed environments such
1650        as Cygwin where scripts look and work like Unix scripts, but
1651        where OS.Process.system cannot run them.
1652    
1653      - In CM, the srcpath pickler used native pathname syntax -- which
1654        is incorrect in the case of cross-compilation.  The new pickle format
1655        is independent of platform-specific naming conventions.
1656    
1657      - Path configuration files (such as lib/pathconfig) can now choose
1658        between native and standard syntax.  Placing a line of the form
1659    
1660            standard!
1661    
1662        into the file causes all subsequent paths to be interpreted using
1663        CM standard pathname syntax (= Unix conventions); a line
1664    
1665            native!
1666    
1667        switches back to native style.  This was needed so that
1668        path config files can be written portably, see src/system/pathconfig.
1669    
1670      - Runtime system:
1671    
1672          - win32-filesys.c:  get_file_time and set_file_time now
1673            access modification time, not creation time.
1674    
1675          - I/O code made aware of new array representation.
1676    
1677          - Bug fixes in X86.prim.masm.
1678    
1679       - src/system/makeml made aware of win32. (For use under cygwin
1680            and other Unix-environments for windows.)
1681    
1682       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1683         which caused BinIO.inputAll to fail consistently.
1684    
1685    .bat scripts:
1686    
1687       Windows .bat scripts assume that SMLNJ_HOME is defined.
1688    
1689       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1690         applications (sml, ml-yacc, ml-lex).
1691       - ml-build.bat: analogous to ml-build.
1692       - config\install.bat:  Analogous to config/install.sh.  This requires
1693         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1694         (nmake etc. must be on the path, and vcvars32 must have been run.)
1695         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1696         and the bootfile hierarchy must have been unpacked under
1697         sml.boot.x86-win32.
1698         The script is very primitive and does a poor job at error checking.
1699         It only installs the base system, ml-lex, and ml-yacc.  No other
1700         libraries are being installed (i.e., you get only those that
1701         are part of the compiler.)
1702       - link-sml.bat:  analogous to .link-sml, but not currently used
1703    
1704    Unrelated bug fixes:
1705    
1706      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1707        types.
1708      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1709        otherwise.)
1710    
1711    ----------------------------------------------------------------------
1712    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1713    Date: 2003/06/17
1714    Tag: macqueen-20030617-bug895
1715    Description:
1716    
1717    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1718    Tag will be used for new development branch (mcz-branch) for
1719    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1720    
1721    ----------------------------------------------------------------------
1722    Name: Matthias Blume (blume (at) tti - c (dot) org)
1723    Date: 2003/05/27 16:55:00 CDT
1724    Tag: blume-20030527-polyeq
1725    Description:
1726    
1727    Tried to eliminated most cases of polymorphic equality.
1728    
1729    ----------------------------------------------------------------------
1730    Name: Matthias Blume (blume (at) tti - c (dot) org)
1731    Date: 2003/05/21 17:45:00 CDT
1732    Tag: blume-20030517-complete
1733    Description:
1734    
1735    Two changes:
1736    
1737      1. Added a flag for controlling whether non-exhaustive bindings will
1738         be treated as errors (default is false).
1739      2. Cleaned up the *entire* source tree so that CMB.make goes through
1740         without a single non-exhaustive match- or bind warning.
1741    
1742    ----------------------------------------------------------------------
1743    Name: Matthias Blume (blume (at) tti - c (dot) org)
1744    Date: 2003/05/17 10:20:00 CDT
1745    Tag: blume-20030517-absyn
1746    Description:
1747    
1748    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1749    
1750       This mainly affects the quality of error messages.  However, some
1751       of the code is now more straightforward than before. (Treatment of
1752       the above four constructs in translate.sml is much simpler than
1753       the "macro-expansion" that was going on before.  Plus, the mach-
1754       compiler no longer gets invoked just to be able to compile an
1755       if-expression.)
1756    
1757    2. The ErrorMsg.Error exception is now caught and absorbed by the
1758       interactive loop.
1759    
1760    ----------------------------------------------------------------------
1761    Name: Allen Leung
1762    Date: 2003/05/16 13:05:00 CDT
1763    Tag: leunga-20030516-cygwin-runtime
1764    Description:
1765    
1766        Ported the runtime system to cygwin, which uses the unix
1767    x86-unix bin files.  Missing/buggy features:
1768    
1769       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1770         the Cygwin library.
1771       o Ctrl-C handling may be flaky.
1772       o Windows system calls and Windows I/O are not supported.
1773    
1774       A new set of binfiles is located at:
1775    
1776         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1777    
1778       This is only needed for bootstrapping the cygwin version of smlnj.
1779       Other x86 versions can use the existing binfiles.
1780    
1781    ----------------------------------------------------------------------
1782    Name: Matthias Blume
1783    Date: 2003/04/08 15:42:00 CDT
1784    Tag: blume-20030408-listpair
1785    Description:
1786    
1787    1. Added a target 'mlrisc' to installer.
1788    
1789    2. Added missing elements to structure ListPair.
1790    
1791    ----------------------------------------------------------------------
1792    Name: Allen Leung
1793    Date: 2003/01/07 10:40:00 EST
1794    Tag: leunga-20030107-int-rem
1795    Description:
1796    
1797    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1798    The arguments to the SUBL instruction were swapped.
1799    
1800    ----------------------------------------------------------------------
1801    Name: Matthias Blume
1802    Date: 2002/12/12 16:25:00 EST
1803    Tag: blume-20021212-risc-ra
1804    Description:
1805    
1806    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1807    sent the RA into an infinite loop when floating point registers get
1808    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1809    110.37 and 110.38.)
1810    There was another set of potential problems related to the handling of
1811    MLRISC annotations (but those did not yet cause real problems, apparently).
1812    
1813    ----------------------------------------------------------------------
1814    Name: Matthias Blume
1815    Date: 2002/12/06 22:40:00 EST
1816    Tag: blume-20021206-cm-fileid
1817    Description:
1818    
1819    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1820    This fixes the problem of CM getting confused by files that suddenly
1821    change their identity (e.g., by getting unlinked and recreated by some
1822    text editor such as vi).  There might be a better/cheaper/cleaner way
1823    of doing this, but for now this will have to do.
1824    
1825    ----------------------------------------------------------------------
1826    Name: Matthias Blume
1827    Date: 2002/10/28 09:50:00 EST
1828    Tag: blume-20021028-typecheck
1829    Description:
1830    
1831    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1832    
1833    ----------------------------------------------------------------------
1834    Name: Matthias Blume
1835    Date: 2002/10/17 09:10:00 EDT
1836    Tag: Release_110_42
1837    Description:
1838    
1839    In good old tradition, there has been a slight hiccup so that we have
1840    to patch 110.42 after the fact.   The old release tag has been replaced
1841    (see below).
1842    
1843    The change solves a problem with two competing approaches the
1844    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1845    each other's way.
1846    
1847    This change only affects the runtime system code and the installer script.
1848    (No new bootfiles.)
1849    
1850    ----------------------------------------------------------------------
1851    Name: Matthias Blume
1852    Date: 2002/10/16 12:00:00 EDT
1853    Tag: Release_110_42_removed
1854    Description:
1855    
1856    New working release.  New bootfiles.
1857    
1858    ----------------------------------------------------------------------
1859    Name: Matthias Blume
1860    Date: 2002/10/10 13:10:00 EDT
1861    Tag: blume-20021010-ppc-divs
1862    Description:
1863    
1864    The mltree operator DIVS must be implemented with an overflow check on
1865    the PPC because the hardware indicates divide-by-zero using "overflow" as
1866    well.
1867    
1868    ----------------------------------------------------------------------
1869    Name: Matthias Blume
1870    Date: 2002/07/23 11:45:00
1871    Tag: blume-20020723-smlnj-home
1872    Description:
1873    
1874    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1875    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1876    CM_PATHCONFIG is also set), the path configuration file is assumed
1877    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1878    the entire tree to some other place and everything will "just work".
1879    
1880    (Companion commands such as ml-build and ml-makedepend also sense this
1881    variable.)
1882    
1883    ----------------------------------------------------------------------
1884    Name: Matthias Blume
1885    Date: 2002/07/12 21:19:00 EDT
1886    Tag: blume-20020712-liveness
1887    Description:
1888    
1889    Exported two useful "step" functions from liveness module (MLRISC).
1890    
1891    ----------------------------------------------------------------------
1892    Name: Matthias Blume
1893    Date: 2002/07/05 16:00 EDT
1894    Tag: Release_110_41
1895    Description:
1896    
1897    New working release.  New bootfiles.
1898    
1899    ----------------------------------------------------------------------
1900    Name: Matthias Blume
1901    Date: 2002/07/05 10:25:00 EDT
1902    Tag: blume-20020705-btimp
1903    Description:
1904    
1905    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1906    other clients can set up backtracing support.
1907    
1908    ----------------------------------------------------------------------
1909    Name: Matthias Blume
1910    Date: 2002/06/25 17:23:00 EDT
1911    Tag: blume-20020625-fpmax
1912    Description:
1913    
1914    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1915    case.  (The sense of the isNaN test was reversed -- which made min and
1916    max always return their first argument.)
1917    
1918    ----------------------------------------------------------------------
1919    Name: Matthias Blume
1920    Date: 2002/06/11
1921    Tag: blume-20020611-unixpath
1922    Description:
1923    
1924    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1925    
1926    ----------------------------------------------------------------------
1927    Name: Matthias Blume
1928    Date: 2002/06/10 16:35:00 EDT
1929    Tag: blume-20020610-ieeereal
1930    Description:
1931    
1932    I back-ported my implementation of IEEEReal.fromString from the
1933    idlbasis-devel branch so that we can test it.
1934    
1935    Another small change is that ppDec tries to give more information
1936    than just "<sig>" in the case of functors.  However, this code is
1937    broken in some mysterious way if the functor's body's signature
1938    has not been declared by ascription but gets inferred from the
1939    implementation.  This needs fixing...
1940    
1941    ----------------------------------------------------------------------
1942    Name: Matthias Blume
1943    Date: 2002/05/31
1944    Tag: blume-20020531-btrace-mode
1945    Description:
1946    
1947    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1948    the wayside when I switched over to using Controls everywhere.)
1949    
1950    ----------------------------------------------------------------------
1951    Name: Lal George
1952    Date: 2002/05/23 12:21:40 EDT
1953    Tag: george-20020523-visual-labels
1954    Description:
1955    
1956    Labels are now displayed in the graphical output to make
1957    the fall-through and target blocks obvious.
1958    
1959    ----------------------------------------------------------------------
1960    Name: Matthias Blume
1961    Date: 2002/05/22 11:03:00 EDT
1962    Tag: blume-20020522-shrink
1963    Description:
1964    
1965    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1966    situation that comes up when allocating huge arrays.
1967    
1968    ----------------------------------------------------------------------
1969    Name: Matthias Blume
1970    Date: 2002/05/21 16:00:00 EDT
1971    Tag: Release_110_40
1972    Description:
1973    
1974    New working release (110.40).  New bootfiles.
1975    
1976    [Also: John Reppy fixed GC bug 1131.]
1977    
1978    ----------------------------------------------------------------------
1979    Name: Matthias Blume
1980    Date: 2002/05/21 12:35:00 EDT
1981    Tag: blume-20020521-cmdoc
1982    Description:
1983    
1984    CM documentation update.
1985    
1986    ----------------------------------------------------------------------
1987    Name: Matthias Blume
1988    Date: 2002/05/21 10:55:00 EDT
1989    Tag: blume-20020521-misc
1990    Description:
1991    
1992    - John tweaked runtime to be silent on heap export (except when
1993      GC messages are on).
1994    - I added a few more things (cross-compiling versions of CMB) to
1995      config/preloads (as suggestions).
1996    
1997    ----------------------------------------------------------------------
1998    Name: Matthias Blume
1999    Date: 2002/05/20 22:25:00 EDT
2000    Tag: blume-20020520-controls
2001    Description:
2002    
2003    - Added ControlUtil structure to control-lib.cm.
2004    - Use it throughout.
2005    - Used Controls facility to define MLRISC controls (as opposed to
2006      registering MLRISC control ref cells with Controls after the fact)
2007    - Fixed messed-up controls priorities.
2008    
2009    * Removed again all the stuff from config/preloads that one wouldn't
2010      be able to preload at the time the initial heap image is built.
2011      (Many libraries, e.g., CML, do not exist yet at this time.  The
2012       only libraries that can be preloaded via config/preloads are those
2013       that come bundled with the bootfiles.)
2014    
2015    ----------------------------------------------------------------------
2016    Name: Matthias Blume
2017    Date: 2002/05/20 10:59:00 EDT
2018    Tag: blume-20020520-preloads
2019    Description:
2020    
2021    Added a lot of commented-out suggestions for things to be included
2022    in config/preloads.
2023    
2024    ----------------------------------------------------------------------
2025    Name: Allen Leung
2026    Date: 2002/05/18 14:20:00 EDT
2027    Tag: leunga-20020518-mdl
2028    Description:
2029    
2030    o Made the mdl tool stuff compile and run again.
2031    o I've disabled all the stuff that depends on RTL specifications; they
2032      are all badly broken anyway.
2033    
2034    ----------------------------------------------------------------------
2035    Name: Matthias Blume
2036    Date: 2002/05/17 16:49:00 EDT
2037    Tag: blume-20020517-controls
2038    Description:
2039    
2040    0. John Reppy made several modifications to the SML/NJ library.
2041       In particular, there is a shiny new controls-lib.cm.
2042    
2043    1. Pushed new controls interface through compiler so that everything
2044       compiles again.
2045    
2046    2. Added FormatComb and FORMAT_COMB to the CML version of the
2047       SML/NJ library (so that CML compiles again).
2048    
2049    3. Modified init scripts because XXX_DEFAULT environment variables
2050       are no longer with us.  (Boot-time initialization is now done
2051       using the same environment variables that are also used for
2052       startup-time initialization of controls.)
2053    
2054    ----------------------------------------------------------------------
2055    Name: Lal George
2056    Date: 2002/05/15 09:20:10 EDT
2057    Tag: george-20020515-pseudo-op-decls
2058    Description:
2059    
2060            All pseudo-ops emitted before the first segment declaration
2061    such as TEXT, DATA, and BSS directives are assumed to be global
2062    declarations and are emitted first in the assembly file. This is
2063    useful in a number of situations where one has pseudo-ops that are not
2064    specific to any segment, and also works around the constraint that one
2065    cannot have client pseudo-ops in the TEXT segment.
2066    
2067            Because no segment is associated with these declarations it is
2068    an error to allocate any space or objects before the first segment
2069    directive and an exception will be raised. However, we cannot make
2070    this check for client pseudo-ops.
2071    
2072            These top level declarations are a field in the CFG graph_info.
2073    In theory you can continue to add to this field after the CFG has been
2074    built -- provided you know what you are doing;-)
2075    
2076    ----------------------------------------------------------------------
2077    Name: Matthias Blume
2078    Date: 2002/05/13 16:40:00 EDT
2079    Tag: blume-20020513-pp-etc
2080    Description:
2081    
2082    A few minor bugfixes:
2083    
2084      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2085        (Bogus printouts for redefined bindings still occur.  Compiler
2086        bug should no longer occur now.  We need to redo the prettyprinter
2087        from scratch.)
2088    
2089      - CM pathname printer now also adds escape sequences for ( and )
2090    
2091      - commend and docu fixes for ml-nlffi
2092    
2093    ----------------------------------------------------------------------
2094    Name: Matthias Blume
2095    Date: 2002/05/10 16:40:00 EDT
2096    Tag: blume-20020510-erg-textio
2097    Description:
2098    
2099    Applied the following bugfix provided by Emden Gansner:
2100    
2101        Output is corrupted when outputSubstr is used rather than output.
2102    
2103        The problem occurs when a substring
2104    
2105            ss = (s, dataStart, dataLen)
2106    
2107        where dataStart > 0, fills a stream buffer with avail bytes left.
2108        avail bytes of s, starting at index dataStart, are copied into the
2109        buffer, the buffer is flushed, and then the remaining dataLen-avail
2110        bytes of ss are copied into the beginning of the buffer. Instead of
2111        starting this copy at index dataStart+avail in s, the current code
2112        starts the copy at index avail.
2113    
2114        Fix:
2115        In text-io-fn.sml, change line 695 from
2116             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2117        to
2118             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2119    
2120    ----------------------------------------------------------------------
2121    Name: Matthias Blume
2122    Date: 2002/04/12 13:55:00 EDT
2123    Tag: blume-20020412-assyntax
2124    Description:
2125    
2126    1. Grabbed newer assyntax.h from the XFree86 project.
2127    2. Fiddled with how to compile X86.prim.asm without warnings.
2128    3. (Very) Minor cleanup in CM.
2129    
2130    ----------------------------------------------------------------------
2131    Name: Matthias Blume
2132    Date: 2002/04/01 (no joke!) 17:07:00 EST
2133    Tag: blume-20020401-x86div
2134    Description:
2135    
2136    Added full support for div/mod/rem/quot on the x86, using the machine
2137    instruction's two results (without clumsily recomputing the remainder)
2138    directly where appropriate.
2139    
2140    Some more extensive power-of-two support was added to the x86 instruction
2141    selector (avoiding expensive divs, mods, and muls where they can be
2142    replaced with cheaper shifts and masks).  However, this sort of thing
2143    ought to be done earlier, e.g., within the CPS optimizer so that
2144    all architectures benefit from it.
2145    
2146    The compiler compiles to a fixed point, but changes might be somewhat
2147    fragile nevertheless.  Please, report any strange things that you might
2148    see wrt. div/mod/quot/rem...
2149    
2150    ----------------------------------------------------------------------
2151    Name: Matthias Blume
2152    Date: 2002/03/29 17:22:00
2153    Tag: blume-20020329-div
2154    Description:
2155    
2156    Fixed my broken div/mod logic.  Unfortunately, this means that the
2157    inline code for div/mod now has one more comparison than before.
2158    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2159    The problem was with quotient = 0, because that alone does not tell
2160    us which way the rounding went.  One then has to look at whether
2161    remainder and divisor have the same sign...  :(
2162    
2163    Anyway, I replaced the bootfiles with fresh ones...
2164    
2165    ----------------------------------------------------------------------
2166    Name: Matthias Blume
2167    Date: 2002/03/29 14:10:00 EST
2168    Tag: blume-20020329-inlprims
2169    Description:
2170    
2171    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2172    
2173    Primops have changed. This means that the bin/boot-file formats have
2174    changed as well.
2175    
2176    To make sure that there is no confusion, I made a new version.
2177    
2178    
2179    CHANGES:
2180    
2181    * removed REMT from mltree (remainder should never overflow).
2182    
2183    * added primops to deal with divisions of all flavors to the frontend
2184    
2185    * handled these primops all the way through so they map to their respective
2186      MLRISC support
2187    
2188    * used these primops in the implementation of Int, Int32, Word, Word32
2189    
2190    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2191    
2192    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2193    
2194    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2195      floating point
2196    
2197    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2198      and Real (but Real.abs maps to a separate floating-point-only primop)
2199    
2200    
2201    TODO items:
2202    
2203    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2204      instructions because the machine instruction encoder cannot handle
2205      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2206      REM in the first place, and REM is supported by the x86 machine coder.)
2207    
2208    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2209      selection phase.  (The two can be streamlined because the hardware
2210      delivers both quotient and remainder at the same time anyway.)
2211    
2212    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2213      (Currently the behavior is inconsistent both across architectures and
2214      wrt. the draft Basis spec.)
2215    
2216    * Word8 should eventually be handled natively, too.
2217    
2218    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2219      as if there currently is no execution path that could trigger it in
2220      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2221      hold for things like multiplication and division.)
2222    
2223    ----------------------------------------------------------------------
2224    Name: Matthias Blume
2225    Date: 2002/03/27 16:27:00 EST
2226    Tag: blume-20020327-mlrisc-divisions
2227    Description:
2228    
2229    Added support for all four division operations (ML's div, mod, quot,
2230    and rem) to MLRISC.  In the course of doing so, I also rationalized
2231    the naming (no more annoying switch-around of DIV and QUOT), by
2232    parameterizing the operation by div_rounding_mode (which can be either
2233    DIV_TO_ZERO or DIV_TO_NEGINF).
2234    
2235    The generic MLTreeGen functor takes care of compiling all four
2236    operations down to only round-to-zero div.
2237    
2238    Missing pieces:
2239    
2240      * Doing something smarter than relying on MLTreeGen on architectures
2241        like, e.g., the x86 where hardware division delivers both quotient and
2242        remainder at the same time.  With this, the implementation of the
2243        round-to-neginf operations could be further streamlined.
2244    
2245      * Remove inlining support for div/mod/rem from the frontend and replace it
2246        with primops that get carried through to the backend.  Do this for all
2247        int and word types.
2248    
2249    ----------------------------------------------------------------------
2250    Name: Matthias Blume
2251    Date: 2002/03/25 17:25:00 EST
2252    Tag: blume-20020325-divmod
2253    Description:
2254    
2255    I improved (hopefully without breaking them) the implementation of Int.div,
2256    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2257    advantage of the following observations:
2258    
2259      Let  q = x quot y      r = x rem y
2260           d = x div  y      m = x mod y
2261    
2262    where "quot" is the round-to-zero version of integer division that
2263    hardware usually provides.  Then we have:
2264    
2265         r = x - q * y        where neither the * nor the - will overflow
2266         d = if q >= 0 orelse x = q * y then q else q - 1
2267                              where neither the * nor the - will overflow
2268         m = if q >= 0 orelse r = 0 then r else r + y
2269                              where the + will not overflow
2270    
2271    This results in substantial simplification of the generated code.
2272    The following table shows the number of CFG nodes and edges generated
2273    for
2274            fun f (x, y) = x OPER y
2275            (* with OPER \in div, mod, quot, rem *)
2276    
2277    
2278        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2279        --------------------------------------------------------
2280         div |         24 |         39 |         12 |         16
2281         mod |         41 |         71 |         12 |         16
2282        quot |          8 |         10 |          8 |         10
2283         rem |         10 |         14 |          8 |         10
2284    
2285    
2286    ----------------------------------------------------------------------
2287    Name: Matthias Blume
2288    Date: 2002/03/25 22:06:00 EST
2289    Tag: blume-20020325-cprotobug
2290    Description:
2291    
2292    Fixed a bug in cproto (c prototype decoder).
2293    
2294    ----------------------------------------------------------------------
2295    Name: Matthias Blume
2296    Date: 2002/03/25 16:00:00 EST
2297    Tag: blume-20020325-raw-primops
2298    Description:
2299    
2300    I did some cleanup to Allen's new primop code and
2301    replaced yesterday's bootfiles with new ones.
2302    (But they are stored in the same place.)
2303    
2304    ----------------------------------------------------------------------
2305    Name: Matthias Blume
2306    Date: 2002/03/24 22:40:00 EST
2307    Tag: blume-20020324-bootfiles
2308    Description:
2309    
2310    Made the bootfiles that Allen asked for.
2311    
2312    ----------------------------------------------------------------------
2313    Name: Allen Leung
2314    Date: 2002/03/23 15:50:00 EST
2315    Tag: leunga-20020323-flint-cps-rcc-primops
2316    Description:
2317    
2318      1. Changes to FLINT primops:
2319    
2320        (* make a call to a C-function;
2321         * The primop carries C function prototype information and specifies
2322         * which of its (ML-) arguments are floating point. C prototype
2323         * information is for use by the backend, ML information is for
2324         * use by the CPS converter. *)
2325      | RAW_CCALL of { c_proto: CTypes.c_proto,
2326                       ml_args: ccall_type list,
2327                       ml_res_opt: ccall_type option,
2328                       reentrant : bool
2329                     } option
2330       (* Allocate uninitialized storage on the heap.
2331        * The record is meant to hold short-lived C objects, i.e., they
2332        * are not ML pointers.  With the tag, the representation is
2333        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2334        *)
2335      | RAW_RECORD of {tag:bool,sz:int}
2336      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2337    
2338      2.  These CPS primops are now overloaded:
2339    
2340           rawload of {kind:numkind}
2341           rawstore of {kind:numkind}
2342    
2343          The one argument form is:
2344    
2345             rawload {kind} address
2346    
2347          The two argument form is:
2348    
2349             rawload {kind} [ml object, byte-offset]
2350    
2351      3. RAW_CCALL/RCC now takes two extra arguments:
2352    
2353         a. The first is whether the C call is reentrant, i.e., whether
2354            ML state should be saved and restored.
2355         b. The second argument is a string argument specifying the name of
2356            library and the C function.
2357    
2358         These things are currently not handled in the code generator, yet.
2359    
2360      4. In CProto,
2361    
2362         An encoding type of "bool" means "ml object" and is mapped into
2363         C prototype of PTR.  Note that "bool" is different than "string",
2364         even though "string" is also mapped into PTR, because "bool"
2365         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2366    
2367      5. Pickler/unpicker
2368    
2369         Changed to handle RAW_RECORD and newest RAW_CCALL
2370    
2371      6. MLRiscGen,
2372    
2373         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2374         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2375            in the file CodeGen/cpscompile/cps-c-calls.sml
2376    
2377      7. Added the conditional move operator
2378    
2379             condmove of branch
2380    
2381         to cps.  Generation of this is still buggy so it is currently
2382         disabled.
2383    
2384    ----------------------------------------------------------------------
2385    Name: Lal George
2386    Date: 2002/03/22 14:18:25 EST
2387    Tag: george-20020322-cps-branch-prob
2388    Description:
2389    
2390    Implemented the Ball-Larus branch prediction-heuristics, and
2391    incorporated graphical viewers for control flow graphs.
2392    
2393    Ball-Larus Heuristics:
2394    ---------------------
2395    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2396    
2397    By design it uses the Dempster-Shafer theory for combining
2398    probabilities.  For example, in the function:
2399    
2400        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2401    
2402    the ball-larus heuristics predicts that the n=0 is unlikely
2403    (OH-heuristic), and the 'then' branch is unlikely because of the
2404    RH-heuristic -- giving the 'then' branch an even lower combined
2405    probability using the Dempster-Shafer theory.
2406    
2407    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2408    probability of the 'then' branch because of the loop in the else
2409    branch.
2410    
2411    
2412    Graphical Viewing:
2413    ------------------
2414    I merely plugged in Allen's graphical viewers into the compiler. The
2415    additional code is not much. At the top level, saying:
2416    
2417            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2418    
2419    will display the graphical view of the control flow graph just before
2420    back-patching.  daVinci must be in your path for this to work. If
2421    daVinci is not available, then the default viewer can be changed
2422    using:
2423    
2424            Control.MLRISC.getString "viewer"
2425    
2426    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2427    course, these viewers must be in your path.
2428    
2429    The above will display the compilation unit at the level of clusters,
2430    many of which are small, boring, and un-interesting. Also setting:
2431    
2432            Control.MLRISC.getInt "cfg-graphical-view_size"
2433    
2434    will display clusters that are larger than the value set by the above.
2435    
2436    
2437    ----------------------------------------------------------------------
2438    Name: Matthias Blume
2439    Date: 2002/03/21 22:20:00 EST
2440    Tag: blume-20020321-kmp-bugfix
2441    Description:
2442    
2443    Changed the interface to the KMP routine in PreString and fixed
2444    a minor bug in one place where it was used.
2445    
2446    ----------------------------------------------------------------------
2447    Name: Allen Leung
2448    Date: 2002/03/21 20:30:00 EST
2449    Tag: leunga-20020321-cfg
2450    Description:
2451    
2452      Fixed a potential problem in cfg edge splitting.
2453    
2454    ----------------------------------------------------------------------
2455    Name: Allen Leung
2456    Date: 2002/03/21 17:15:00 EST
2457    Tag: leunga-20020321-x86-fp-cfg
2458    Description:
2459    
2460      1. Recoded the buggy parts of x86-fp.
2461    
2462         a. All the block reordering code has been removed.
2463            We now depend on the block placement phases to do this work.
2464    
2465         b. Critical edge splitting code has been simplified and moved into the
2466            CFG modules, as where they belong.
2467    
2468         Both of these were quite buggy and complex.  The code is now much, much
2469         simpler.
2470    
2471      2. X86 backend.
2472    
2473         a. Added instructions for 64-bit support.  Instruction selection for
2474            64-bit has not been committed, however, since that
2475            requires changes to MLTREE which haven't been approved by
2476            Lal and John.
2477    
2478         b. Added support for FUCOMI and FUCOMIP when generating code for
2479            PentiumPro and above.  We only generate these instructions in
2480            the fast-fp mode.
2481    
2482         c. Added cases for JP and JNP in X86FreqProps.
2483    
2484      3. CFG
2485    
2486         CFG now has a bunch of methods for edge splitting and merging.
2487    
2488      4. Machine description.
2489    
2490         John's simplification of MLTREE_BASIS.fcond broke a few machine
2491         description things:
2492    
2493         rtl-build.{sig,sml} and hppa.mdl fixed.
2494    
2495         NOTE: the machine description stuff in the repository is still broken.
2496               Again, I can't put my fixes in because that involves
2497               changes to MLTREE.
2498    
2499    ----------------------------------------------------------------------
2500    Name: Matthias Blume
2501    Date: 2002/03/20 15:55:00 EST
2502    Tag: blume-20020320-kmp
2503    Description:
2504    
2505    Implemented Knuth-Morris-Pratt string matching in PreString and used
2506    it for String.isSubstring, Substring.isSubstring, and
2507    Substring.position.
2508    
2509    (Might need some stress-testing.  Simple examples worked fine.)
2510    
2511    ----------------------------------------------------------------------
2512    Name: Matthias Blume
2513    Date: 2002/03/19 16:37:00 EST
2514    Tag: blume-20020319-witnesses
2515    Description:
2516    
2517    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2518    
2519    This implements a generic mechanism for changing constness qualifiers
2520    anywhere within big C types without resorting to outright "casts".
2521    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2522    modify the constness at the outermost level.)
2523    The implementation of "convert" is based on the idea of "witness"
2524    values -- values that are not used by the operation but whose types
2525    "testify" to their applicability.  On the implementation side, "convert"
2526    is simply a projection (returning its second curried argument).  With
2527    cross-module inlining, it should not result in any machine code being
2528    generated.
2529    
2530    ----------------------------------------------------------------------
2531    Name: Matthias Blume
2532    Date: 2002/03/15 16:40:00 EST
2533    Tag: blume-20020315-basis
2534    Description:
2535    
2536    Provided (preliminary?) implementations for
2537    
2538      {String,Substring}.{concatWith,isSuffix,isSubstring}
2539    
2540    and
2541    
2542      Substring.full
2543    
2544    Those are in the Basis spec but they were missing in SML/NJ.
2545    
2546    ----------------------------------------------------------------------
2547    Name: Matthias Blume
2548    Date: 2002/03/14 21:30:00 EST
2549    Tag: blume-20020314-controls
2550    Description:
2551    
2552    Controls:
2553    ---------
2554    
2555    1. Factored out the recently-added Controls : CONTROLS stuff and put
2556       it into its own library $/controls-lib.cm.  The source tree for
2557       this is under src/smlnj-lib/Controls.
2558    
2559    2. Changed the names of types and functions in this interface, so they
2560       make a bit more "sense":
2561    
2562          module -> registry
2563          'a registry -> 'a group
2564    
2565    3. The interface now deals in ref cells only.  The getter/setter interface
2566       is (mostly) gone.
2567    
2568    4. Added a function that lets one register an already-existing ref cell.
2569    
2570    5. Made the corresponding modifications to the rest of the code so that
2571       everything compiles again.
2572    
2573    6. Changed the implementation of Controls.MLRISC back to something closer
2574       to the original.  In particular, this module (and therefore MLRISC)
2575       does not depend on Controls.  There now is some link-time code in
2576       int-sys.sml that registers the MLRISC controls with the Controls
2577       module.
2578    
2579    CM:
2580    ---
2581    
2582      * One can now specify the lambda-split aggressiveness in init.cmi.
2583    
2584    ----------------------------------------------------------------------
2585    Name: Allen Leung
2586    Date: 2002/03/13 17:30:00 EST
2587    Tag: leunga-20020313-x86-fp-unary
2588    Description:
2589    
2590    Bug fix for:
2591    
2592    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2593    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2594    > - fun f(x,(y,z)) = Real.~ y;
2595    > [autoloading]
2596    > [autoloading done]
2597    >       fchsl   (%eax), 184(%esp)
2598    > Error: MLRisc bug: X86MCEmitter.emitInstr
2599    >
2600    > uncaught exception Error
2601    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2602    
2603    The problem was that the code generator did not generate any fp registers
2604    in this case, and the ra didn't know that it needed to run the X86FP phase to
2605    translate the pseudo fp instruction.   This only happened with unary fp
2606    operators in certain situations.
2607    
2608    ----------------------------------------------------------------------
2609    Name: Matthias Blume
2610    Date: 2002/03/13 14:00:00 EST
2611    Tag: blume-20020313-overload-etc
2612    Description:
2613    
2614    1. Added _overload as a synonym for overload for backward compatibility.
2615       (Control.overloadKW must be true for either version to be accepted.)
2616    
2617    2. Fixed bug in install script that caused more things to be installed
2618       than what was requested in config/targets.
2619    
2620    3. Made CM aware of the (_)overload construct so that autoloading
2621       works.
2622    
2623    ----------------------------------------------------------------------
2624    Name: Matthias Blume
2625    Date: 2002/03/12 22:03:00 EST
2626    Tag: blume-20020312-url
2627    Description:
2628    
2629    Forgot to update BOOT and srcarchiveurl.
2630    
2631    ----------------------------------------------------------------------
2632    Name: Matthias Blume
2633    Date: 2002/03/12 17:30:00 EST
2634    Tag: blume-20020312-version110392
2635    Description:
2636    
2637    Yet another version number bump (because of small changes to the
2638    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2639    
2640    Changes:
2641    
2642      The new pid generation scheme described a few weeks ago was overly
2643      complicated.  I implemented a new mechanism that is simpler and
2644      provides a bit more "stability":  Once CM has seen a compilation
2645      unit, it keeps its identity constant (as long as you do not delete
2646      those crucial CM/GUID/* files).  This means that when you change
2647      an interface, compile, then go back to the old interface, and
2648      compile again, you arrive at the original pid.
2649    
2650      There now also is a mechanism that instructs CM to use the plain
2651      environment hash as a module's pid (effectively making its GUID
2652      the empty string).  For this, "noguid" must be specified as an
2653      option to the .sml file in question within its .cm file.
2654      This is most useful for code that is being generated by tools such
2655      as ml-nlffigen (because during development programmers tend to
2656      erase the tool's entire output directory tree including CM's cached
2657      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2658      revert to the old, broken behavior of SML/NJ, but in specific cases
2659      where there is no danger of interface confusion, its use is ok
2660      (I think).
2661    
2662      ml-nlffigen by default generates "noguid" annotations.  They can be
2663      turned off by specifying -guid in its command line.
2664    
2665    ----------------------------------------------------------------------
2666    Name: Lal George
2667    Date: 2002/03/12 12 14:42:36 EST
2668    Tag: george-20020312-frequency-computation
2669    Description:
2670    
2671    Integrated jump chaining and static block frequency into the
2672    compiler. More details and numbers later.
2673    
2674    ----------------------------------------------------------------------
2675    Name: Lal George
2676    Date: 2002/03/11 11 22:38:53 EST
2677    Tag: george-20020311-jump-chain-elim
2678    Description:
2679    
2680    Tested the jump chain elimination on all architectures (except the
2681    hppa).  This is on by default right now and is profitable for the
2682    alpha and x86, however, it may not be profitable for the sparc and ppc
2683    when compiling the compiler.
2684    
2685    The gc test will typically jump to a label at the end of the cluster,
2686    where there is another jump to an external cluster containing the actual
2687    code to invoke gc. This is to allow factoring of common gc invocation
2688    sequences. That is to say, we generate:
2689    
2690            f:
2691               testgc
2692               ja   L1      % jump if above to L1
2693    
2694            L1:
2695               jmp L2
2696    
2697    
2698    After jump chain elimination the 'ja L1' instructions is converted to
2699    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2700    up being implemented in their long form (if L2 is far away) using:
2701    
2702            jbe     L3      % jump if below or equal to L3
2703            jmp     L2
2704         L3:
2705            ...
2706    
2707    
2708    For large compilation units L2  may be far away.
2709    
2710    
2711    ----------------------------------------------------------------------
2712    Name: Matthias Blume
2713    Date: 2002/03/11 13:30:00 EST
2714    Tag: blume-20020311-mltreeeval
2715    Description:
2716    
2717    A functor parameter was missing.
2718    
2719    ----------------------------------------------------------------------
2720    Name: Allen Leung
2721    Date: 2002/03/11 10:30:00 EST
2722    Tag: leunga-20020311-runtime-string0
2723    Description:
2724    
2725       The representation of the empty string now points to a
2726    legal null terminated C string instead of unit.  It is now possible
2727    to convert an ML string into C string with InlineT.CharVector.getData.
2728    This compiles into one single machine instruction.
2729    
2730    ----------------------------------------------------------------------
2731    Name: Allen Leung
2732    Date: 2002/03/10 23:55:00 EST
2733    Tag: leunga-20020310-x86-call
2734    Description:
2735    
2736       Added machine generation for CALL instruction (relative displacement mode)
2737    
2738    ----------------------------------------------------------------------
2739    Name: Matthias Blume
2740    Date: 2002/03/08 16:05:00
2741    Tag: blume-20020308-entrypoints
2742    Description:
2743    
2744    Version number bumped to 110.39.1.  NEW BOOTFILES!
2745    
2746    Entrypoints: non-zero offset into a code object where execution should begin.
2747    
2748    - Added the notion of an entrypoint to CodeObj.
2749    - Added reading/writing of entrypoint info to Binfile.
2750    - Made runtime system bootloader aware of entrypoints.
2751    - Use the address of the label of the first function given to mlriscGen
2752      as the entrypoint.  This address is currently always 0, but it will
2753      not be 0 once we turn on block placement.
2754    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2755      with entry points) from mlriscGen.
2756    
2757    ----------------------------------------------------------------------
2758    Name: Allen Leung
2759    Date: 2002/03/07 20:45:00 EST
2760    Tag: leunga-20020307-x86-cmov
2761    Description:
2762    
2763       Bug fixes for CMOVcc on x86.
2764    
2765       1. Added machine code generation for CMOVcc
2766       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2767       3. CMOVcc cannot have an immediate operand as argument.
2768    
2769    ----------------------------------------------------------------------
2770    Name: Matthias Blume
2771    Date: 2002/03/07 16:15:00 EST
2772    Tag: blume-20020307-controls
2773    Description:
2774    
2775    This is a very large but mostly boring patch which makes (almost)
2776    every tuneable compiler knob (i.e., pretty much everything under
2777    Control.* plus a few other things) configurable via both the command
2778    line and environment variables in the style CM did its configuration
2779    until now.
2780    
2781    Try starting sml with '-h' (or, if you are brave, '-H')
2782    
2783    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2784    implements the underlying generic mechanism.
2785    
2786    The interface to some of the existing such facilities has changed somewhat.
2787    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2788    (The getFoo interface is still there for backward-compatibility, but its
2789    use is deprecated.)
2790    
2791    The ml-build script passes -Cxxx=yyy command-line arguments through so
2792    that one can now twiddle the compiler settings when using this "batch"
2793    compiler.
2794    
2795    TODO items:
2796    
2797    We should go through and throw out all controls that are no longer
2798    connected to anything.  Moreover, we should go through and provide
2799    meaningful (and correct!) documentation strings for those controls
2800    that still are connected.
2801    
2802    Currently, multiple calls to Controls.new are accepted (only the first
2803    has any effect).  Eventually we should make sure that every control
2804    is being made (via Controls.new) exactly once.  Future access can then
2805    be done using Controls.acc.
2806    
2807    Finally, it would probably be a good idea to use the getter-setter
2808    interface to controls rather than ref cells.  For the time being, both
2809    styles are provided by the Controls module, but getter-setter pairs are
2810    better if thread-safety is of any concern because they can be wrapped.
2811    
2812    *****************************************
2813    
2814    One bug fix: The function blockPlacement in three of the MLRISC
2815    backpatch files used to be hard-wired to one of two possibilities at
2816    link time (according to the value of the placementFlag).  But (I
2817    think) it should rather sense the flag every time.
2818    
2819    *****************************************
2820    
2821    Other assorted changes (by other people who did not supply a HISTORY entry):
2822    
2823    1. the cross-module inliner now works much better (Monnier)
2824    2. representation of weights, frequencies, and probabilities in MLRISC
2825       changed in preparation of using those for weighted block placement
2826       (Reppy, George)
2827    
2828    ----------------------------------------------------------------------
2829    Name: Lal George
2830    Date: 2002/03/07 14:44:24 EST 2002
2831    Tag: george-20020307-weighted-block-placement
2832    
2833    Tested the weighted block placement optimization on all architectures
2834    (except the hppa) using AMPL to generate the block and edge frequencies.
2835    Changes were required in the machine properties to correctly
2836    categorize trap instructions. There is an MLRISC flag
2837    "weighted-block-placement" that can be used to enable weighted block
2838    placement, but this will be ineffective without block/edge
2839    frequencies (coming soon).
2840    
2841    
2842    ----------------------------------------------------------------------
2843    Name: Lal George
2844    Date: 2002/03/05 17:24:48 EST
2845    Tag: george-20020305-linkage-cluster
2846    
2847    In order to support the block placement optimization, a new cluster
2848    is generated as the very first cluster (called the linkage cluster).
2849    It contains a single jump to the 'real' entry point for the compilation
2850    unit. Block placement has no effect on the linkage cluster itself, but
2851    all the other clusters  have full freedom in the manner in which they
2852    reorder blocks or functions.
2853    
2854    On the x86 the typical linkage code that is generated is:
2855       ----------------------
2856            .align 2
2857       L0:
2858            addl    $L1-L0, 72(%esp)
2859            jmp     L1
2860    
2861    
2862            .align  2
2863       L1:
2864       ----------------------
2865    
2866    72(%esp) is the memory location for the stdlink register. This
2867    must contain the address of the CPS function being called. In the
2868    above example, it contains the address of  L0; before
2869    calling L1 (the real entry point for the compilation unit), it
2870    must contain the address for L1, and hence
2871    
2872            addl $L1-L0, 72(%esp)
2873    
2874    I have tested this on all architectures except the hppa.The increase
2875    in code size is of course negligible
2876    
2877    ----------------------------------------------------------------------
2878    Name: Allen Leung
2879    Date: 2002/03/03 13:20:00 EST
2880    Tag: leunga-20020303-mlrisc-tools
2881    
2882      Added #[ ... ] expressions to mlrisc tools
2883    
2884    ----------------------------------------------------------------------
2885    Name: Matthias Blume
2886    Date: 2002/02/27 12:29:00 EST
2887    Tag: blume-20020227-cdebug
2888    Description:
2889    
2890    - made types in structure C and C_Debug to be equal
2891    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2892    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2893    
2894    ----------------------------------------------------------------------
2895    Name: Matthias Blume
2896    Date: 2002/02/26 12:00:00 EST
2897    Tag: blume-20020226-ffi
2898    Description:
2899    
2900    1. Fixed a minor bug in CM's "noweb" tool:
2901       If numbering is turned off, then truly don't number (i.e., do not
2902       supply the -L option to noweb).  The previous behavior was to supply
2903       -L'' -- which caused noweb to use the "default" line numbering scheme.
2904       Thanks to Chris Richards for pointing this out (and supplying the fix).
2905    
2906    2. Once again, I reworked some aspects of the FFI:
2907    
2908       A. The incomplete/complete type business:
2909    
2910       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2911         gone!
2912       - ML types representing an incomplete type are now *equal* to
2913         ML types representing their corresponding complete types (just like
2914         in C).  This is still safe because ml-nlffigen will not generate
2915         RTTI for incomplete types, nor will it generate functions that
2916         require access to such RTTI.   But when ML code generated from both
2917         incomplete and complete versions of the C type meet, the ML types
2918         are trivially interoperable.
2919    
2920         NOTE:  These changes restore the full generality of the translation
2921         (which was previously lost when I eliminated functorization)!
2922    
2923       B. Enum types:
2924    
2925       - Structure C now has a type constructor "enum" that is similar to
2926         how the "su" constructor works.  However, "enum" is not a phantom
2927         type because each "T enum" has values (and is isomorphic to
2928         MLRep.Signed.int).
2929       - There are generic access operations for enum objects (using
2930         MLRep.Signed.int).
2931       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2932         * The structure contains the definition of type "mlrep" (the ML-side
2933         representation type of the enum).  Normally, mlrep is the same
2934         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2935         then mlrep will be defined as a datatype -- thus facilitating
2936         pattern matching on mlrep values.
2937         ("-ec" will be suppressed if there are duplicate values in an
2938          enumeration.)
2939         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2940         will be generated for each C enum constant xxx.
2941         * Conversion functions m2i and i2m convert between mlrep and
2942         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2943         * Coversion functions c and ml convert between mlrep and "tag enum".
2944         * Access functions (get/set) fetch and store mlrep values.
2945       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2946         enumerations are merged into one single enumeration represented by
2947         structure E_'.
2948    
2949    ----------------------------------------------------------------------
2950    Name: Allen Leung
2951    Date: 2002/02/25 04:45:00 EST
2952    Tag: leunga-20020225-cps-spill
2953    
2954    This is a new implementation of the CPS spill phase.
2955    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2956    In case of problems, replace it with the old file spill.sml
2957    
2958    The current compiler runs into some serious performance problems when
2959    constructing a large record.  This can happen when we try to compile a
2960    structure with many items.  Even a very simple structure like the following
2961    makes the compiler slow down.
2962    
2963        structure Foo = struct
2964           val x_1 = 0w1 : Word32.int
2965           val x_2 = 0w2 : Word32.int
2966           val x_3 = 0w3 : Word32.int
2967           ...
2968           val x_N = 0wN : Word32.int
2969        end
2970    
2971    The following table shows the compile time, from N=1000 to N=4000,
2972    with the old compiler:
2973    
2974    N
2975    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2976           MLRISC ra                               0.06u  0.00s  0.05g
2977              (spills = 0 reloads = 0)
2978           TOTAL                                   0.63u  0.07s  0.21g
2979    
2980    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2981           MLRISC ra                               5.68u  0.59s  3.93g
2982              (spills = 0 reloads = 0)
2983           TOTAL                                   14.71u  0.99s  4.81g
2984    
2985    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2986           MLRISC ra                               5.54u  0.65s  3.91g
2987              (spills = 543 reloads = 1082)
2988           TOTAL                                   65.40u  3.13s  6.00g
2989    
2990    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2991           MLRISC ra                               0.80u  0.10s  0.55g
2992              (spills = 42 reloads = 84)
2993           TOTAL                                   129.42u  5.10s  4.13g
2994    
2995    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2996           MLRISC ra                               2.69u  0.27s  1.38g
2997              (spills = 62 reloads = 124)
2998           TOTAL                                   682.48u  19.61s  13.99g
2999    
3000    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3001           MLRISC ra                               4.96u  0.27s  2.72g
3002              (spills = 85 reloads = 170)
3003           TOTAL                                   2375.26u  57.21s  48.00g
3004    
3005    As you can see the old cps spill module suffers from some serious
3006    performance problem.  But since I cannot decipher the old code fully,
3007    instead of patching the problems up, I'm reimplementing it
3008    with a different algorithm.  The new code is more modular,
3009    smaller when compiled, and substantially faster
3010    (O(n log n) time and O(n) space).  Timing of the new spill module:
3011    
3012    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3013          MLRISC ra                               0.25u  0.02s  0.15g
3014             (spills=1 reloads=3)
3015          TOTAL                                   7.74u  0.34s  1.62g
3016    
3017    Implementation details:
3018    
3019    As far as I can tell, the purpose of the CPS spill module is to make sure the
3020    number of live variables at any program point (the bandwidth)
3021    does not exceed a certain limit, which is determined by the
3022    size of the spill area.
3023    
3024    When the bandwidth is too large, we decrease the register pressure by
3025    packing live variables into spill records.  How we achieve this is
3026    completely different than what we did in the old code.
3027    
3028    First, there is something about the MLRiscGen code generator
3029    that we should be aware of:
3030    
3031    o MLRiscGen performs code motion!
3032    
3033       In particular, it will move floating point computations and
3034       address computations involving only the heap pointer to
3035       their use sites (if there is only a single use).
3036       What this means is that if we have a CPS record construction
3037       statement
3038    
3039           RECORD(k,vl,w,e)
3040    
3041       we should never count the new record address w as live if w
3042       has only one use (which is often the case).
3043    
3044       We should do something similar to floating point, but the transformation
3045       there is much more complex, so I won't deal with that.
3046    
3047    Secondly, there are now two new cps primops at our disposal:
3048    
3049     1. rawrecord of record_kind option
3050        This pure operator allocates some uninitialized storage from the heap.
3051        There are two forms:
3052    
3053         rawrecord NONE [INT n]  allocates a tagless record of length n
3054         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3055                                     and initializes the tag.
3056    
3057     2. rawupdate of cty
3058          rawupdate cty (v,i,x)
3059          Assigns to x to the ith component of record v.
3060          The storelist is not updated.
3061    
3062    We use these new primops for both spilling and increment record construction.
3063    
3064     1. Spilling.
3065    
3066        This is implemented with a linear scan algorithm (but generalized
3067        to trees).  The algorithm will create a single spill record at the
3068        beginning of the cps function and use rawupdate to spill to it,
3069        and SELECT or SELp to reload from it.  So both spills and reloads
3070        are fine-grain operations.  In contrast, in the old algorithm
3071        "spills" have to be bundled together in records.
3072    
3073        Ideally, we should sink the spill record construction to where
3074        it is needed.  We can even split the spill record into multiple ones
3075        at the places where they are needed.  But CPS is not a good
3076        representation for global code motion, so I'll keep it simple and
3077        am not attempting this.
3078    
3079     2. Incremental record construction (aka record splitting).
3080    
3081        Long records with many component values which are simulatenously live
3082        (recall that single use record addresses are not considered to
3083         be live) are constructed with rawrecord and rawupdate.
3084        We allocate space on the heap with rawrecord first, then gradually
3085        fill it in with rawupdate.  This is the technique suggested to me
3086        by Matthias.
3087    
3088        Some restrictions on when this is applicable:
3089        1. It is not a VECTOR record.  The code generator currently does not handle
3090           this case. VECTOR record uses double indirection like arrays.
3091        2. All the record component values are defined in the same "basic block"
3092           as the record constructor.  This is to prevent speculative
3093           record construction.
3094    
3095    ----------------------------------------------------------------------
3096    Name: Allen Leung
3097    Date: 2002/02/22 01:02:00 EST
3098    Tag: leunga-20020222-mlrisc-tools
3099    
3100    Minor bug fixes in the parser and rewriter
3101    
3102    ----------------------------------------------------------------------
3103    Name: Allen Leung
3104    Date: 2002/02/21 20:20:00 EST
3105    Tag: leunga-20020221-peephole
3106    
3107    Regenerated the peephole files.  Some contained typos in the specification
3108    and some didn't compile because of pretty printing bugs in the old version
3109    of 'nowhere'.
3110    
3111    ----------------------------------------------------------------------
3112    Name: Allen Leung
3113    Date: 2002/02/19 20:20:00 EST
3114    Tag: leunga-20020219-mlrisc-tools
3115    Description:
3116    
3117       Minor bug fixes to the mlrisc-tools library:
3118    
3119       1.  Fixed up parsing colon suffixed keywords
3120       2.  Added the ability to shut the error messages up
3121       3.  Reimplemented the pretty printer and fixed up/improved
3122           the pretty printing of handle and -> types.
3123       4.  Fixed up generation of literal symbols in the nowhere tool.
3124       5.  Added some SML keywords to to sml.sty
3125    
3126    ----------------------------------------------------------------------
3127    Name: Matthias Blume
3128    Date: 2002/02/19 16:20:00 EST
3129    Tag: blume-20020219-cmffi
3130    Description:
3131    
3132    A wild mix of changes, some minor, some major:
3133    
3134    * All C FFI-related libraries are now anchored under $c:
3135        $/c.cm      --> $c/c.cm
3136        $/c-int.cm  --> $c/internals/c-int.cm
3137        $/memory.cm --> $c/memory/memory.cm
3138    
3139    * "make" tool (in CM) now treats its argument pathname slightly
3140      differently:
3141        1. If the native expansion is an absolute name, then before invoking
3142           the "make" command on it, CM will apply OS.Path.mkRelative
3143           (with relativeTo = OS.FileSys.getDir()) to it.
3144        2. The argument will be passed through to subsequent phases of CM
3145           processing without "going native".  In particular, if the argument
3146           was an anchored path, then "make" will not lose track of that anchor.
3147    
3148    * Compiler backends now "know" their respective C calling conventions
3149      instead of having to be told about it by ml-nlffigen.  This relieves
3150      ml-nlffigen from one of its burdens.
3151    
3152    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3153    
3154    * Export C_DEBUG and C_Debug from $c/c.cm.
3155    
3156    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3157      subtyping relationship between incomplete pointers and their complete
3158      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3159      with the convention of instantiating 'o with ('t, 'c) obj whenever
3160      the pointer target type is complete.  In the incomplete case, 'o
3161      will be instantiated with some "'c iobj" -- a type obtained by
3162      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3163    
3164      Operations that work on both incomplete and complete pointer types are
3165      typed as taking an 'o ptr while operations that require the target to
3166      be known are typed as taking some ('t, 'c) obj ptr.
3167    
3168      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3169      where void is an eqtype without any values.  This makes it possible
3170      to work on voidptr values using functions meant to operate on light
3171      incomplete pointers.
3172    
3173    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3174      been vastly simplified.
3175    
3176    ----------------------------------------------------------------------
3177    Name: Matthias Blume
3178    Date: 2002/02/19 10:48:00 EST
3179    Tag: blume-20020219-pqfix
3180    Description:
3181    
3182    Applied Chris Okasaki's bug fix for priority queues.
3183    
3184    ----------------------------------------------------------------------
3185    Name: Matthias Blume
3186    Date: 2002/02/15 17:05:00
3187    Tag: Release_110_39
3188    Description:
3189    
3190    Last-minute retagging is becoming a tradition... :-(
3191    
3192    This is the working release 110.39.
3193    
3194    ----------------------------------------------------------------------
3195    Name: Matthias Blume
3196    Date: 2002/02/15 16:00:00 EST
3197    Tag: Release_110_39-orig
3198    Description:
3199    
3200    Working release 110.39.  New bootfiles.
3201    
3202    (Update: There was a small bug in the installer so it wouldn't work
3203    with all shells.  So I retagged. -Matthias)
3204    
3205    ----------------------------------------------------------------------
3206    Name: Matthias Blume
3207    Date: 2002/02/15 14:17:00 EST
3208    Tag: blume-20020215-showbindings
3209    Description:
3210    
3211    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3212    the latter can be useful for exploring what bindings are available at
3213    the interactive prompt.  (The first function returns only the list
3214    of symbols that are really bound, the second prints those but also the
3215    ones that CM's autoloading mechanism knows about.)
3216    
3217    ----------------------------------------------------------------------
3218    Name: Matthias Blume
3219    Date: 2002/02/15 12:08:00 EST
3220    Tag: blume-20020215-iptrs
3221    Description:
3222    
3223    Two improvements to ml-nlffigen:
3224    
3225      1. Write files only if they do not exist or if their current contents
3226         do not coincide with what's being written.  (That is, avoid messing
3227         with the time stamps unless absolutely necessary.)
3228    
3229      2. Implement a "repository" mechanism for generated files related
3230         to "incomplete pointer types".   See the README file for details.
3231    
3232    ----------------------------------------------------------------------
3233    Name: Matthias Blume
3234    Date: 2002/02/14 11:50:00 EST
3235    Tag: blume-20020214-quote
3236    Description:
3237    
3238    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3239    because of the new and improved tag generation scheme.  (Thanks to Allen
3240    Leung for pointing it out.)
3241    
3242    ----------------------------------------------------------------------
3243    Name: Lal George
3244    Date: 2002/02/14 09:55:27 EST 2002
3245    Tag: george-20020214-isabelle-bug
3246    Description:
3247    
3248    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3249    of Isabelle on the x86.
3250    
3251    From Allen:
3252    -----------
3253     I've found the problem:
3254    
3255         in ra-core.sml, I use the counter "blocked" to keep track of the
3256         true number of elements in the freeze queue.  When the counter goes
3257         to zero, I skip examining the queue.  But I've messed up the
3258         bookkeeping in combine():
3259    
3260             else ();
3261             case !ucol of
3262               PSEUDO => (if !cntv > 0 then
3263                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3264                                        ^^^^^^^^^^^^^^^^^^^^^^^
3265                      moveu := mergeMoveList(!movev, !moveu)
3266                     )
3267                  else ();
3268    
3269         combine() is called to coalesce two nodes u and v.
3270         I think I was thinking that if the move counts of u and v are both
3271         greater than zero then after they are coalesced then one node is
3272         removed from the freeze queue.  Apparently I was thinking that
3273         both u and v are of low degree, but that's clearly not necessarily true.
3274    
3275    
3276    02/12/2002:
3277        Here's the patch.  HOL now compiles.
3278    
3279        I don't know how this impact on performance (compile
3280        time or runtime).  This bug caused the RA (especially on the x86)
3281        to go thru the potential spill phase when there are still nodes on the
3282        freeze queue.
3283    
3284    
3285    
3286    
3287    ----------------------------------------------------------------------
3288    Name: Matthias Blume
3289    Date: 2002/02/13 22:40:00 EST
3290    Tag: blume-20020213-fptr-rtti
3291    Description:
3292    
3293    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3294    updates.
3295    
3296    ----------------------------------------------------------------------
3297    Name: Matthias Blume
3298    Date: 2002/02/13 16:41:00 EST
3299    Tag: blume-20020213-cmlpq
3300    Description:
3301    
3302    Added new priority queue export symbols (which have just been added to
3303    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3304    not compile and the installer would choke.)
3305    
3306    ----------------------------------------------------------------------
3307    Name: Matthias Blume
3308    Date: 2002/02/13 16:15:00 EST
3309    Tag: blume-20020213-various
3310    Description:
3311    
3312    1. More tweaks to ml-nlffigen:
3313    
3314       - better internal datastructures (resulting in slight speedup)
3315       - "-match" option requires exact match
3316       - "localized" gensym counters (untagged structs/unions nested within
3317         other structs/unions or within typedefs get a fresh counter; their
3318         tag will be prefixed by a concatenation of their parents' tags)
3319       - bug fixes (related to calculation of transitive closure of types
3320         to be included in the output)
3321    
3322    2. Minor Basis updates:
3323    
3324       - added implementations for List.collate and Option.app
3325    
3326    ----------------------------------------------------------------------
3327    Name: Matthias Blume
3328    Date: 2002/02/11 15:55:00 EST
3329    Tag: blume-20020211-gensym
3330    Description:
3331    
3332    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3333    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3334    names (ML structure names that correspond to unnamed C structs, unions,
3335    and enums), so that separate runs of ml-nlffigen do not clash.
3336    
3337    ----------------------------------------------------------------------
3338    Name: Matthias Blume
3339    Date: 2002/02/11 12:05:00 EST
3340    Tag: blume-20020211-gensml
3341    Description:
3342    
3343    A quick fix for a problem with GenSML (in the pgraph-util library):
3344    Make generation of toplevel "local" optional.  (Strictly speaking,
3345    signature definitions within "local" are not legal SML.)
3346    
3347    Other than that: updates to INSTALL and cm/TODO.
3348    
3349    ----------------------------------------------------------------------
3350    Name: Matthias Blume
3351    Date: 2002/02/08 15:00:00 EST
3352    Tag: blume-20020208-uniquepid
3353    Description:
3354    
3355    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3356    
3357    1. The installer (config/install.sh) has gotten smarter:
3358    
3359         - Configuration options are a bit easier to specify now
3360           (in config/targets).
3361         - Bug in recognizing .tar.bz2 files fixed.
3362         - Installer automatically resolves dependencies between
3363           configuration options (e.g., if you ask for eXene, you will
3364           also get cml -- regardless whether you asked for it or not).
3365         - Installer can run in "quieter mode" by setting the environment
3366           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3367           "completely silent", though.
3368         - Build HashCons library as part of smlnj-lib.
3369    
3370    2. A new scheme for assigning persistent identifiers to compilation
3371       units (and, by extension, to types etc.) has been put into place.
3372       This fixes a long-standing bug where types and even dynamic values
3373       can get internally confused, thereby compromising type safety
3374       (abstraction) and dynamic correctness.  See
3375    
3376         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3377    
3378       for an example of how things could go wrong until now.
3379    
3380       The downside of the new scheme is that pids are not quite as
3381       persistent as they used to be: CM will generate a fresh pid
3382       for every compilation unit that it thinks it sees for the first
3383       time.  That means that if you compile starting from a clean, fresh
3384       source tree at two different times, you end up with different
3385       binaries.
3386    
3387       Cutoff recompilation, however, has not been compromised because
3388       CM keeps pid information in special caches between runs.
3389    
3390    ----------------------------------------------------------------------
3391    Name: Lal George
3392    Date: 2002/02/07 15:34:13 EST 2002
3393    Tag: <none>
3394    Description:
3395    
3396    Compilers that generate assembly code may produce  global labels
3397    whose value is resolved at link time. The various peephole optimization
3398    modules did not take this in account.
3399    
3400    TODO. The Labels.addrOf function should really return an option
3401    type so that clients are forced to deal with this issue, rather
3402    than an exception being raised.
3403    
3404    ----------------------------------------------------------------------
3405    Name: Lal George
3406    Date: 2002/02/06 13:55:02 EST
3407    Tag: george-20020206-ra-breakup
3408    Description:
3409    
3410    1. A bug fix from Allen.
3411    
3412        A typo causes extra fstp %st(0)'s to be generated at compensation
3413        edges, which might cause stack underflow traps at runtime.  This
3414        occurs in fft where there are extraneous fstps right before the 'into'
3415        trap instruction (in this case they are harmless since none of the
3416        integers overflow.)
3417    
3418    2. Pulled out various utility modules that were embedded in the modules
3419       of the register allocator. I need these modules for other purposes, but
3420       they are not complete enough to put into a library (just yet).
3421    ----------------------------------------------------------------------
3422    Name: Matthias Blume
3423    Date: 2002/01/31 16:05:00 EST
3424    Tag: blume-20020131-sparc-ccalls
3425    Description:
3426    
3427    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3428       of extra stack space by mistake.  Fixed.
3429    
3430    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3431    
3432    ----------------------------------------------------------------------
3433    Name: Allen Leung
3434    Date: 2002/01/30
3435    Tag: leunga-20020130-nowhere-bug-fix
3436    Description:
3437    
3438       MLRISC bug fixes:
3439       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3440       2. MachineInt.fromString was negating its value.
3441    
3442    ----------------------------------------------------------------------
3443    Name: Matthias Blume
3444    Date: 2002/01/29
3445    Tag: blume-20020129-INSTALL
3446    Description:
3447    
3448    - Added somewhat detailed installation instructions (file INSTALL).
3449    - Fixed curl-detection bug in config/install.sh.
3450    - It is now possible to select the URL getter using the URLGETTER
3451      environment variable:
3452    
3453          not set / "unknown"      --> automatic detection (script tries wget,
3454                                       curl, and lynx)
3455          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3456                                       how to properly invoke them)
3457          other                    --> use $URLGETTER directly, it must take
3458                                       precisely two command-line arguments
3459                                       (source URL and destination file name)
3460    
3461    ----------------------------------------------------------------------
3462    Name: Matthias Blume
3463    Date: 2002/01/28
3464    Tag: blume-20020128-sparc-ccalls
3465    Description:
3466    
3467    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3468    - Make use of the allocParam argument in sparc-c-calls.
3469    
3470    ----------------------------------------------------------------------
3471    Name: Matthias Blume
3472    Date: 2002/01/28
3473    Tag: blume-20020128-allocParam
3474    Description:
3475    
3476    John Reppy:  Changes c-calls API to accept client-callback for
3477    allocating extra stack space.
3478    me: Corresponding changes to mlriscGen (using a dummy argument that
3479        does not change the current behavior).
3480    
3481    ----------------------------------------------------------------------
3482    Name: Matthias Blume
3483    Date: 2002/01/28 12:00:00
3484    Tag: Release_110_38
3485    Description:
3486    
3487    This time for real!!!
3488    
3489    ----------------------------------------------------------------------
3490    Name: Matthias Blume
3491    Date: 2002/01/28 10:56:00 EST
3492    Tag: blume-20020128-retraction
3493    Description:
3494    
3495    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3496       with blume-Release_110_38-retracted.)
3497    
3498    1. Fixed a problem with incorrect rounding modes in real64.sml.
3499       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3500    
3501    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3502       and enums fixed.  The naming of corresponding ML identifiers should
3503       now be consistent again.
3504    
3505    ----------------------------------------------------------------------
3506    Name: Allen Leung
3507    Date: 2002/01/27
3508    Tag: leunga-20020127-nowhere
3509    Description:
3510    
3511       Added a target called nowhere in the configuration scripts.
3512       Enabling this will build the MLRISC 'nowhere' tool (for translating
3513       programs with where-clauses into legal SML code) during installation.
3514    
3515    ----------------------------------------------------------------------
3516    Name: Matthias Blume
3517    Date: 2002/01/25 21:27:00 EST
3518    Tag: blume-Release_110_38-retracted
3519    Description:
3520    
3521    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3522    
3523    README will be added later.
3524    
3525    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3526    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3527    
3528    ----------------------------------------------------------------------
3529    Name: Matthias Blume
3530    Date: 2002/01/25
3531    Tag: blume-20020125-ffi
3532    Description:
3533    
3534    A large number of tweaks and improvements to ml-nlffi-lib and
3535    ml-nlffigen:
3536    
3537       - ML represenation types have been streamlined
3538       - getter and setter functions work with concrete values, not abstract
3539         ones where possible
3540       - ml-nlffigen command line more flexible (see README file there)
3541       - some bugs have been fixed (hopefully)
3542    
3543    ----------------------------------------------------------------------
3544    Name: Lal George
3545    Date: 2002/01/24
3546    Tag: george-20020124-risc-ra-interface
3547    Description:
3548    
3549       There is a dramatic simplification in the interface to the
3550       register allocator for RISC architectures as a result of making
3551       parallel copy instructions explicit.
3552    
3553    ----------------------------------------------------------------------
3554    Name: Matthias Blume
3555    Date: 2002/01/22
3556    Tag: blume-20020122-x86-ccalls
3557    Description:
3558    
3559    Bug fix for c-calls on x86 (having to do with how char- and
3560    short-arguments are being handled).
3561    
3562    ----------------------------------------------------------------------
3563    Name: Matthias Blume
3564    Date: 2002/01/21
3565    Tag: blume-20020121-ff
3566    Description:
3567    
3568    Another day of fiddling with the FFI...
3569    
3570    1. Bug fix/workaround:  CKIT does not complain about negative array
3571       dimensions, so ml-nlffigen has to guard itself against this possibility.
3572       (Otherwise a negative dimension would send it into an infinite loop.)
3573    
3574    2. Some of the abstract types (light objects, light pointers, most "base"
3575       types) in structure C are now eqtypes.
3576    
3577    3. Added constructors and test functions for NULL function pointers.
3578    
3579    ----------------------------------------------------------------------
3580    Name: Matthias Blume
3581    Date: 2002/01/18
3582    Tag: blume-20020118-ready-for-new-release
3583    Description:
3584    
3585    Made config/srcarchiveurl point to a new place.  (Will provide boot
3586    files shortly.)
3587    
3588    Maybe we christen this to be 110.38?
3589    
3590    ----------------------------------------------------------------------
3591    Name: Matthias Blume
3592    Date: 2002/01/18
3593    Tag: blume-20020118-more-ffifiddle
3594    Description:
3595    
3596    Today's FFI fiddling:
3597    
3598      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3599        C.Set.* that use concrete (MLRep.*) arguments and results instead
3600        of abstract ones.
3601    
3602      - Provided word-style bit operations etc. for "int" representation
3603        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3604        and Long.
3605    
3606    ----------------------------------------------------------------------
3607    Name: Matthias Blume
3608    Date: 2002/01/18
3609    Tag: blume-20020118-use-x86-fp
3610    Description:
3611    
3612    Now that x86-fast-fp seems to be working, I turned it back on again
3613    by default.  (Seems to work fine now, even with the FFI.)
3614    
3615    Other than that, I added some documentation about the FFI to
3616    src/ml-nlffigen/README and updated the FFI test examples in
3617    src/ml-nlffi-lib/Tests/*.
3618    
3619    ----------------------------------------------------------------------
3620    Name: Allen Leung
3621    Date: 2002/01/17
3622    Tag: leunga-20020117-x86-fast-fp-call
3623    Description:
3624    
3625       1. Fixed a problem with handling return fp values when x86's fast fp
3626          mode is turned on.
3627    
3628       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3629          of %f32.
3630    
3631       3. Added a constructor INT32lit to the ast of MLRISC tools.
3632    
3633    ----------------------------------------------------------------------
3634    Name: Matthias Blume
3635    Date: 2002/01/16
3636    Tag: blume-20020116-ffifiddle
3637    Description:
3638    
3639    More fiddling with the FFI interface:
3640    
3641     - Make constness 'c instead of rw wherever possible.  This eliminates
3642       the need for certain explicit coercions.  (However, due to ML's
3643       value polymorphism, there will still be many cases where explicit
3644       coercions are necessary.  Phantom types are not the whole answer
3645       to modeling a subtyping relationship in ML.)
3646    
3647     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3648    
3649     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3650       clumsy workaround for varargs problem.)
3651    
3652    ----------------------------------------------------------------------
3653    Name: Lal George
3654    Date: 2002/01/15
3655    Tag: <none>
3656    Description:
3657    
3658    1. Since COPY instructions are no longer native to the architecture,
3659       a generic functor can be used to implement the expandCopies function.
3660    
3661    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3662       TEXT segment.
3663    
3664    ----------------------------------------------------------------------
3665    Name: Matthias Blume
3666    Date: 2002/01/15
3667    Tag: blume-20020115-ffiupdates
3668    Description:
3669    
3670    1. Fix for bug resulting in single-precision float values being returned
3671       incorrectly from FFI calls.
3672    
3673    2. Small modifications to C FFI API:
3674    
3675        - memory-allocation routines return straight objects (no options)
3676          and raise an exception in out-of-memory situations
3677        - unsafe extensions to cast between function pointers and pointers
3678          from/to ints
3679        - added structure C_Debug as an alternative to structure C where
3680          pointer-dereferencing (|*| and |*!) always check for null-pointers
3681        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3682          but also takes a (possibly empty) list of existing library handles
3683          that the current library depends on
3684    
3685    ----------------------------------------------------------------------
3686    Name: Matthias Blume
3687    Date: 2002/01/10
3688    Tag: blume-20020110-newffigen
3689    Description:
3690    
3691    1. Updates to portable graph code.
3692    
3693    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3694       more scalable now so that even huge interfaces such as the one
3695       for GTK compile in finite time and space. :-)
3696       See src/ml-nlffigen/README for details on what's new.
3697    
3698    ----------------------------------------------------------------------
3699    Name: Lal George
3700    Date: 2001/01/09 14:31:35 EST 2002
3701    Tag: george-20011206-rm-native-copy
3702    Description:
3703    
3704            Removed the native COPY and FCOPY instructions
3705            from all the architectures and replaced it with the
3706            explicit COPY instruction from the previous commit.
3707    
3708            It is now possible to simplify many of the optimizations
3709            modules that manipulate copies. This has not been
3710            done in this change.
3711    
3712    ----------------------------------------------------------------------
3713    Name: Lal George
3714    Date: 2001/12/06 16:50:13 EST 2001
3715    Tag: george-20011206-mlrisc-instruction
3716    Description:
3717    
3718    Changed the representation of instructions from being fully abstract
3719    to being partially concrete. That is to say:
3720    
3721      from
3722            type instruction
3723    
3724      to
3725            type instr                              (* machine instruction *)
3726    
3727            datatype instruction =
3728                LIVE of {regs: C.cellset, spilled: C.cellset}
3729              | KILL of {regs: C.cellset, spilled: C.cellset}
3730              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3731              | ANNOTATION of {i: instruction, a: Annotations.annotation}
3732              | INSTR of instr
3733    
3734    This makes the handling of certain special instructions that appear on
3735    all architectures easier and uniform.
3736    
3737    LIVE and KILL say that a list of registers are live or killed at the
3738    program point where they appear. No spill code is generated when an
3739    element of the 'regs' field is spilled, but the register is moved to
3740    the 'spilled' (which is present, more for debugging than anything else).
3741    
3742    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3743    We used to generate:
3744    
3745            DEFFREG f1
3746            f1 := f2 + f3
3747            trapb
3748    
3749    but now generate:
3750    
3751            f1 := f2 + f3
3752            trapb
3753            LIVE {regs=[f1,f2,f3], spilled=[]}
3754    
3755    Furthermore, the DEFFREG (hack) required that all floating point instruction
3756    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3757    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3758    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3759    this architecture).
3760    
3761    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3762    all the architectures. This will result in further simplification of the
3763    register allocator that must be aware of them for coalescing purposes, and
3764    will also simplify certain aspects of the machine description that provides
3765    callbacks related to parallel copies.
3766    
3767    ANNOTATION should be obvious, and now INSTR represents the honest to God
3768    machine instruction set!
3769    
3770    The <arch>/instructions/<arch>Instr.sml files define certain utility
3771    functions for making porting easier -- essentially converting upper case
3772    to lower case. All machine instructions (of type instr) are in upper case,
3773    and the lower case form generates an MLRISC instruction. For example on
3774    the alpha we have:
3775    
3776      datatype instr =
3777         LDA of {r:cell, b:cell, d:operand}
3778       | ...
3779    
3780      val lda : {r:cell, b:cell, d:operand} -> instruction
3781        ...
3782    
3783    where lda is just (INSTR o LDA), etc.
3784    
3785    ----------------------------------------------------------------------
3786    Name: Matthias Blume
3787    Date: 2001/11/22 21:40:00 EST
3788    Tag: Release_110_37
3789    Description:
3790    
3791    Release 110.37.  This time for real.
3792    
3793    ----------------------------------------------------------------------
3794    Name: Matthias Blume
3795    Date: 2001/11/21 16:35:00 EST
3796    Tag: blume-20011121-foot-in-mouth
3797    Description:
3798    
3799    Removed the "Release_110_37" tag because of a serious bug.
3800    This will be re-tagged once the bug is fixed.
3801    
3802    ----------------------------------------------------------------------
3803    Name: Matthias Blume
3804    Date: 2001/11/21 16:14:00 EST
3805    Tag: blume-20011121-forgottenfile
3806    Description:
3807    
3808    Forgot to add a file.  (Just a .tex-file -- part of
3809    the CM manual source.)
3810    
3811    ----------------------------------------------------------------------
3812    Name: Matthias Blume
3813    Date: 2001/11/21 16:10:00 EST
3814    Tag: blume-20011121-invalid_110_37
3815    Description:
3816    
3817    Note: I removed the original tag "Release_110_37" from this commit
3818          because we found a serious bug in all non-x86 backends.
3819          - Matthias
3820    
3821    1. Modifications to the SML/NJ code generator and to the runtime system
3822       so that code object name strings are directly inserted into code
3823       objects at code generation time.  The only business the runtime system
3824       has with this is now to read the name strings on occasions.
3825       (The encoding of the name string has also changed somewhat.)
3826    
3827    2. CM now implements a simple "set calculus" for specifying export lists.
3828       In particular, it is now possible to refer to the export lists of
3829       other libraries/groups/sources and form unions as well as differences.
3830       See the latest CM manual for details.
3831    
3832    3. An separate notion of "proxy" libraries has again be eliminated from
3833       CM's model.  (Proxy libraries are now simply a special case of using
3834       the export list calculus.)
3835    
3836    4. Some of the existing libraries now take advantage of the new set
3837       calculus.
3838       (Notice that not all libraries have been converted because some
3839       of the existing .cm-files are supposed to be backward compatible
3840       with 110.0.x.)
3841    
3842    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3843       "General.exnMessage"!  The former relies on a certain hook to be
3844       initialized, and that often does not happen in the stand-alone case.)
3845    
3846    ----------------------------------------------------------------------
3847    Name: Lal George
3848    Date: 2001/11/21  13:56:18 EST
3849    Tag: george-2001121-pseudo-ops
3850    Description: