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 589, Fri Mar 31 16:14:36 2000 UTC revision 1747, Fri Jan 14 23:53:21 2005 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: 2005/01/14 18:00:00 CST
18    Tag: blume-20050114-heap2exec
19    Description:
20    
21    Implemented new (but still experimental) heap2exec facility.  This is
22    tested under Mac OS X and should work under Linux (will test shortly).
23    It will probably also work on the Sparc (will test some time later).
24    
25      - removed old "HACKED_STANDALONE" hack from runtime
26    
27    To be able to test this, uncomment the request for "heap2asm" in
28    config/targets prior to installation.  (Notice that this is different
29    from "heap2exec" mentioned below.  Not a typo.)
30    
31    To perform an actual test, run the command
32    
33       $ bin/heap2exec heapfile execfile
34    
35    (You can put heap2exec on your shell's path.)
36    
37    For example, run
38    
39      $ bin/heap2exec bin/.heap/ml-yacc.ppc-darwin mly
40    
41    This will create a standalone executable called "mly" which you can
42    then invoke directly as a command.
43    
44    ----------------------------------------------------------------------
45    Name: Matthias Blume (blume (at) tti - c (dot) org)
46    Date: 2005/01/07 11:44:00 CST
47    Tag: blume-20050107-mlstring
48    Description:
49    
50    fixed off-by-one error in ML_STRING macro (globals.c)
51    
52    ----------------------------------------------------------------------
53    Name: Matthias Blume (blume (at) tti - c (dot) org)
54    Date: 2004/12/23 18:00:00 CST
55    Tag: blume-20041223-santa
56    Description:
57    
58    Made ml-build script "smarter" (but only very little).
59    
60    ----------------------------------------------------------------------
61    Name: Matthias Blume (blume (at) tti - c (dot) org)
62    Date: 2004/12/21 15:05:00 CST
63    Tag: blume-20041221-longlong
64    Description:
65    
66    * Implemented access to signed and unsigned long long data in NLFFI.
67    (The parameter-passing part of the picture has not complete. But
68    data structure access seems to work.)
69    
70    * Fixed CM's incorrect assumption that the PPC is little-endian.
71      (On the Mac, it is big-endian.  And that's currently our only
72      PPC platform.)
73    
74    ----------------------------------------------------------------------
75    Name: Matthias Blume (blume (at) tti - c (dot) org)
76    Date: 2004/12/21 12:50:00 CST
77    Tag: blume-20041221-memory
78    Description:
79    
80    Some cleanup in the $c/memory.cm library: separated some concerns by
81    moving allocation code and memory access code each into their own
82    files.
83    
84    ----------------------------------------------------------------------
85    Name: Allen Leung (leunga (at) reservoir (dot) com)
86    Date: 2004/12/17 16:12:00 EST
87    Tag: leunga-20041217-cygwin-smlnj-home
88    Description:
89    
90        The Unix I/O library of SML/NJ on cygwin does not understand
91    Windows style pathname, so problems arise when SMLNJ_HOME is set
92    to a Windows style pathname.  _run-sml now converts SMLNJ_HOME
93    to a POSIX pathname on cygwin.
94    
95    ----------------------------------------------------------------------
96    Name: Matthias Blume (blume (at) tti - c (dot) org)
97    Date: 2004/12/16 13:35:00 CST
98    Tag: Release_110_52
99    Description:
100    
101    Last-minute changes incorporated into 110.52.  Release tag moved.
102    
103    The changes:
104       - HashString.hashString' -> HashString.hashSubstring
105       - bug fix in UnivariateStats
106    
107    ----------------------------------------------------------------------
108    Name: Matthias Blume (blume (at) tti - c (dot) org)
109    Date: 2004/12/15 23:40:00 CST
110    Tag: blume-20041215-hashSubstring
111    Description:
112    
113    - HashString.hashString' -> HashString.hashSubstring
114    - corresponding changes in atom.sml
115    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
116      added some comments
117    
118    ----------------------------------------------------------------------
119    Name: Matthias Blume (blume (at) tti - c (dot) org)
120    Date: 2004/12/15 15:30:00 CST
121    Tag: (Release_110_52)
122    Description:
123    
124    New working version (110.52).  NEW BOOTFILES!
125    
126    ----------------------------------------------------------------------
127    Name: Matthias Blume (blume (at) tti - c (dot) org)
128    Date: 2004/12/15 12:45:00 CST
129    Tag: blume-20041215-spaces
130    Description:
131    
132    More on the space problem (this time for Win32).
133    
134    ----------------------------------------------------------------------
135    Name: Matthias Blume (blume (at) tti - c (dot) org)
136    Date: 2004/12/14 17:30:00 CST
137    Tag: blume-20041214-spaces
138    Description:
139    
140    Hacked some of the scripts (in particular: the installer) to cope with
141    spaces in filenames a bit better.  But beware: the current "solution"
142    is likely still full of bugs and inherently incomplete.  (We need to
143    do away with those shell scripts for a comprehensive solution.)
144    
145    ----------------------------------------------------------------------
146    Name: Matthias Blume (blume (at) tti - c (dot) org)
147    Date: 2004/12/13 14:45:00 CST
148    Tag: blume-20041213-ml-makedepend
149    Description:
150    
151    Fixed bug in code for ml-makedepend.
152    
153    ----------------------------------------------------------------------
154    Name: Matthias Blume (blume (at) tti - c (dot) org)
155    Date: 2004/12/09 16:30:00 CST
156    Tag: blume-20041209-statistics
157    Description:
158    
159    Added two simple but potentially useful statistics modules to SML/NJ Library.
160    (See CHANGES file there.)
161    
162    ----------------------------------------------------------------------
163    Name: Matthias Blume (blume (at) tti - c (dot) org)
164    Date: 2004/12/01 16:50:00 CST
165    Tag: blume-20041201-atom
166    Description:
167    
168    smlnj-lib:
169    
170      Added function HashString.hashString' for substrings.
171      Hand-inlined CharVector.foldl into HashString (for speed).
172      Modified implementation of structure Atom to avoid extracting
173      strings from substrings unless necessary.
174      (Also see CHANGES file for smlnj-lib.)
175    
176    ----------------------------------------------------------------------
177    Name: Matthias Blume (blume (at) tti - c (dot) org)
178    Date: 2004/11/24 22:15:00 CST
179    Tag: blume-20041124-cml
180    Description:
181    
182    Made sure CML compiles when Position = Int64.
183    
184    ----------------------------------------------------------------------
185    Name: Matthias Blume (blume (at) tti - c (dot) org)
186    Date: 2004/11/24 14:30:00 CST
187    Tag: blume-20041124-position
188    Description:
189    
190    The compiler can now be compiled in a mode that makes structure Position
191    equal to Int64.  The default, however, is unchanged (Position = Int31)
192    for the time being.
193    
194    To enable 64-bit positions, use the following procedure:
195    
196    1. Start sml
197    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
198    3. Type
199          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
200    4. Run CMB.make() as usual.
201    
202    This is barely tested.  The only test so far was a little SML program
203    counting the number of characters in an 8-gigabyte file by
204    reading it character-by-character.  That test was successful.
205    
206    In support of 64-bit positions, a number of new functions have been
207    added to the runtime system.
208    
209    ----------------------------------------------------------------------
210    Name: Matthias Blume (blume (at) tti - c (dot) org)
211    Date: 2004/11/23 14:45:00 CST
212    Tag: blume-20041123-useFile
213    Description:
214    
215    Fixed a problem with unhelpful error messages related to problems with
216    .cm- or .sml files that appear as part of the sml command line.
217    
218    ----------------------------------------------------------------------
219    Name: Matthias Blume (blume (at) tti - c (dot) org)
220    Date: 2004/11/18 15:40:00 CST
221    Tag: Release_110_51
222    Description:
223    
224    New working version (110.51).  NEW BOOTFILES!
225    
226    ----------------------------------------------------------------------
227    Name: Matthias Blume (blume (at) tti - c (dot) org)
228    Date: 2004/11/18 15:35:00 CST
229    Tag: <none>
230    Description:
231    
232    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
233    Joemann).
234    
235    ----------------------------------------------------------------------
236    Name: Allen Leung (leunga (at) reservoir (dot) com)
237    Date: 2004/11/17 16:05:21 EST 2004
238    Tag: leunga-20041117-mlrisc-live-kill
239    Description:
240    
241       Added support for MLTree constructs LIVE and KILL
242    to all the architectures.
243    
244    ----------------------------------------------------------------------
245    Name: Matthias Blume (blume (at) tti - c (dot) org)
246    Date: 2004/11/13 00:20:00 CST
247    Tag: blume-20041113-versiontool
248    Description:
249    
250    - Stripped down the versiontool: It now only handles the version number.
251      The date string is generated at bootstrap time (during makeml).
252    
253    - In a previous commit, fixed a minor issue with how polyequal is being
254      translated.  In particular, the code now "looks through" abstractions.
255      This results in slightly fewer polyEqual warnings and hopefully slightly
256      more efficient code.  Important examples for where this matters are
257      the new int64 and word64 types.
258    
259    ----------------------------------------------------------------------
260    Name: Matthias Blume (blume (at) tti - c (dot) org)
261    Date: 2004/11/12 00:30:00 CST
262    Tag: blume-20041112-int64
263    Description:
264    
265    Structure Int64 fully hooked in.  (The implementation is not very
266    efficient, though.)
267    
268    ----------------------------------------------------------------------
269    Name: Matthias Blume (blume (at) tti - c (dot) org)
270    Date: 2004/11/11 17:30:00 CST
271    Tag: blume-20041111-more64
272    Description:
273    
274    All the pieces of Word64 are now there, with the exception of the
275    conversions from and to LargeWord. (Eventually these need to be identities,
276    but for the time being they don't even make sense because LargeWord is
277    32-bit wide.)
278    
279    Also started to add similar support for Int64, but major pieces of that
280    are still missing.
281    
282    ----------------------------------------------------------------------
283    Name: Matthias Blume (blume (at) tti - c (dot) org)
284    Date: 2004/11/11 00:15:00 CST
285    Tag: blume-20041111-word64
286    Description:
287    
288    Structure Word64 is now (almost) complete, word literals and patterns
289    seem to work.  There are a few odd pieces missing.  In particular,
290    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
291    at the moment.
292    
293    Making Word64 official would mean that LargeWord becomes Word64.  But
294    this requires extreme care because most word-word conversions have to
295    go through LargeWord, so making a mistake means loss of efficiency or
296    worse.  Eventually there will be a solution similar to (but actually
297    simpler than) what I did with IntInf.
298    
299    ----------------------------------------------------------------------
300    Name: Matthias Blume (blume (at) tti - c (dot) org)
301    Date: 2004/11/10 18:12:00 CST
302    Tag: blume-20041110-64bit
303    Description:
304    
305    More 64-bit hacking (but still not even half-way there yet).
306    Also, some assorted improvements to the handling of 8-bit words.
307    
308    ----------------------------------------------------------------------
309    Name: Matthias Blume (blume (at) tti - c (dot) org)
310    Date: 2004/11/09 17:50:00 CST
311    Tag: <none>
312    Description:
313    
314    Started adding some infrastructure for supporting 64-bit int- and
315    word-types.  (Still in its very early stages.)
316    
317    ----------------------------------------------------------------------
318    Name: Matthias Blume (blume (at) tti - c (dot) org)
319    Date: 2004/10/28 10:45:00 CDT
320    Tag: Release_110_50
321    Description:
322    
323    New working version (110.50).  NEW BOOTFILES!
324    
325                 =====================
326    
327    Also:
328    
329      - Changend config/srcarchiveurl from a file just
330        containing the URL string into a file containing
331        shell script code.  The code has access to the $VERSION variable.
332      - Made corresponding changes to config/install.sh and config/unpack.
333      - Default contents of config/srcarchiveurl uses $VERSION and
334        normally does not have to be edited to reflect a version change.
335    
336      (As a result, a version change can be done by just editing
337       config/version, the rest is now automatic.)
338    
339    ----------------------------------------------------------------------
340    Name: Matthias Blume (blume (at) tti - c (dot) org)
341    Date: 2004/10/27 17:50:00 CDT
342    Tag: blume-20041027-btrace-msg
343    Description:
344    
345    BackTrace.monitor now also reports the source of the exception that
346    triggered the trace.
347    
348    ----------------------------------------------------------------------
349    Name: Matthias Blume (blume (at) tti - c (dot) org)
350    Date: 2004/10/27 17:20:00 CDT
351    Tag: blume-20041027-x86-c-calls
352    Description:
353    
354    This is the HISTORY entry for two earlier commits, both concerning
355    the x86 c-calls code in MLRISC:
356    
357      - added a missing LOAD in the code that deals with struct arguments
358      - made sure the caller does not add the wrong number of bytes to the
359        stack pointer after a call of a function returning a struct
360        (the callee already pops the implicit argument which points to
361         the space reserved for the result)
362    
363    ----------------------------------------------------------------------
364    Name: Allen Leung (leunga (at) reservoir (dot) com)
365    Date: 2004/10/24 14:00:00 EST
366    Tag: leunga-20041024-x86-gas-fucomip
367    Description:
368    
369       John discovered a bug in the syntax of fucomip.
370       The opcodes FU?COMIP? have been changed to
371    
372           fu?comip? %st(i), %st
373    
374    ----------------------------------------------------------------------
375    Name: Matthias Blume (blume (at) tti - c (dot) org)
376    Date: 2004/10/20 15:06:00 CDT
377    Tag: blume-20041020-standalone-backtrace
378    Description:
379    
380    Added a mechanism for getting back-trace information from standalone
381    programs.  Here is how it works:
382    
383       1. The part of the program from which you want to get backtrace
384          information (usually the whole program) should be wrapped with
385          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
386          main program could be modified from something like
387    
388             fun main (pgm, args) = ...
389    
390         to
391    
392            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
393    
394       2. To be able to access BackTrace.monitor, you have to add
395          library $smlnj-tdp/plugins.cm to the .cm file that contains your
396          main function.
397    
398       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
399          CM might have created in the past for your project).
400    
401       4. Build the system using this command line:
402    
403             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
404                      myprog.cm MyProg.main myprog
405    
406          instead of the usual
407    
408             ml-build myprog.cm MyProg.main myprog
409    
410    I changed a library name:
411    
412       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
413    
414    New libraries:
415    
416       $smlnj-tdp/back-trace.cm
417          -- when loaded causes the back-trace plugin to be installed
418       $smlnj-tdp/coverage.cm
419          -- when loaded causes the coverage plugin to be installed
420    
421    ----------------------------------------------------------------------
422    Name: Matthias Blume (blume (at) tti - c (dot) org)
423    Date: 2004/10/18 16:45:00 CDT
424    Tag: blume-20041018-groupowner
425    Description:
426    
427    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
428    
429    Eliminated group owner specs from .cm files throughout the source tree.
430    
431    ----------------------------------------------------------------------
432    Name: Matthias Blume (blume (at) tti - c (dot) org)
433    Date: 2004/10/15 15:45:00 CDT
434    Tag: blume-20041015-coverage
435    Description:
436    
437    * Test coverage tool added!
438    * Further reorganization of tracing-, debugging-, and profiling support:
439    
440        - moved original BTImp -- now called BackTrace -- into a separate
441          library called $/trace-debug-profile.cm
442        - eliminated all mentions of BTrace from SMLofNJ.Internals
443        - only the instrumentation mechanism is now left in the compiler proper
444        - BackTrace module is a plugin which is NOT plugged in by default
445        - Coverage module is another such plugin
446    
447    To get the benefits of any of these plugin modules, the code in
448    question must be compiled with tdp instrumentation turned on.  This
449    can be done by setting SMLofNJ.Internals.TDP.mode to true.
450    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
451    
452    Plugins are selected at link time.  (Pre-compiled instrumented code
453    can be re-loaded with different plugins in effect.)  When an
454    instrumented module is linked, whatever plugins are at that time
455    enabled will come into effect for that module.
456    
457    To enable the back-trace plugin, load library $/trace-debug-profile.cm
458    and invoke BackTrace.install() (e.g., from the interactive prompt).
459    To enable the coverage plugin, load the same library and invoke
460    Coverage.install().
461    
462    Back-traces are generated automatically on uncaught exceptions and
463    when the code in question explicitly invokes BackTrace.trigger().
464    
465    Coverage (and execution frequency-) information must be queried
466    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
467    
468    ----------------------------------------------------------------------
469    Name: Matthias Blume (blume (at) tti - c (dot) org)
470    Date: 2004/10/14 17:40:00 CDT
471    Tag: blume-20041014-tdp-core
472    Description:
473    
474    Snapshot of a significant overhaul of how the trace/debug/profile support
475    is hooked into the system (specifically: Core and SMLofNJ.Internals).
476    
477    ----------------------------------------------------------------------
478    Name: Matthias Blume (blume (at) tti - c (dot) org)
479    Date: 2004/10/13 16:34:00 CDT
480    Tag: blume-20041013-tdp
481    Description:
482    
483    Some rationalization of names:
484    
485       structure BTrace -> structure TDPInstrument
486       etc.
487    
488    This is is preparation of using the original back-trace
489    instrumentation for other purposes.  "TDP" stands for
490    Trace/Debug/Profile.
491    
492    The control flag controlling whether instrumentation is on or off is now
493    registered under a different name, so instead of running sml as
494    
495     sml -Cinstrument.btrace-mode=true
496    
497    one has to say
498    
499     sml -Ctdp.instrument=true
500    
501    ----------------------------------------------------------------------
502    Name: Matthias Blume (blume (at) tti - c (dot) org)
503    Date: 2004/10/11 16:37:00 CDT
504    Tag: blume-20041011-regions
505    Description:
506    
507    Made some minor modifications to elabcore.sml to have source regions
508    be propagated more tightly -- resulting in better (i.e., smaller)
509    regions being reported in error- and debug messages.
510    
511    ----------------------------------------------------------------------
512    Name: Matthias Blume (blume (at) tti - c (dot) org)
513    Date: 2004/10/08 22:50:00 CDT
514    Tag: blume-20041008-cmkw
515    Description:
516    
517    Fixed handling of keywords in .cm files: After seeing "is" the lexer
518    treats subsequent occurrences of "group", "library", "source", "is",
519    "*", and "-" as ordinary identifiers rather than keywords.
520    
521    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
522    is supposed to accept a tool argument called "source", but this did
523    not work because of the clash with the keyword.
524    
525    ----------------------------------------------------------------------
526    Name: Matthias Blume (blume (at) tti - c (dot) org)
527    Date: 2004/10/07 16:00:00 CDT
528    Tag: blume-20041007-cleanup
529    Description:
530    
531    Assorted cleanup work:
532    
533      - got rid of intstrmap in favor of using the library's
534        hash table implementation
535      - threw out most of the pathnames stuff, as it was not used anyway
536      - simplified tokentable implementation
537      - fixed some minor spelling errors
538    
539    ----------------------------------------------------------------------
540    Name: Matthias Blume (blume (at) tti - c (dot) org)
541    Date: 2004/10/06 15:15:15 CDT
542    Tag: blume-20041006-handler
543    Description:
544    
545    Cleaned up the absyn to reflect the invariant that HANDLE always
546    carries a FNexp as part of the type definition.  This eliminates some
547    superfluous sanity checks at runtime down the road.
548    
549    Some minor cleanup of the btrace code.
550    
551    ----------------------------------------------------------------------
552    Name: Matthias Blume (blume (at) tti - c (dot) org)
553    Date: 2004/10/01 10:20:30 CDT
554    Tag: blume-20041001-slave
555    Description:
556    
557    Added hack to make slave mode work in the presence of the version
558    tool.  (Still, since the master does two passes over the code for
559    CMB.make, the release number gets bumped twice when slaves are
560    attached.  I don't know if this is worth fixing...)
561    
562    ----------------------------------------------------------------------
563    Name: Matthias Blume (blume (at) tti - c (dot) org)
564    Date: 2004/09/30 10:55:00 CDT
565    Tag: blume-20040930-version
566    Description:
567    
568    * Moved the "version" magic into its own little library under
569      src/system/smlnj/internal.  This avoids expensive reconstruction of
570      a stable src/compiler/core.cm.
571    
572    * At the same time, structure CompilerVersion is now known as
573      structure SMLNJVersion.
574    
575    * Arranged for the version tool to NOT kick in when rebuilding the system
576      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
577      Also, loading the versiontool does not work when rebuilding the system
578      because CM is not properly initialized at that time.
579    
580    ----------------------------------------------------------------------
581    Name: Matthias Blume (blume (at) tti - c (dot) org)
582    Date: 2004/09/29 14:00:00 CDT
583    Tag: blume-20040929-autoversion
584    Description:
585    
586    Implemented some CM magic to have
587        file src/compiler/TopLevel/main/version.sml
588    generated automagically.
589    The version is taken from two files: config/version and config/release.
590    The first is expected to contain a two-part version number such as 110.49.
591    The second should contain a single number, but it may be missing.
592    
593    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
594    time the version tool is loaded (which is the first time you say CMB.make),
595    then the tool will increment the value stored in config/release every
596    time CMB.make is invoked.
597    
598    The binfile format is now insensitive to anything beyond the first
599    two components of a version number, so bumping the release does not render
600    binfiles incompatible.  Auto-bumping can be used to keep track of versions
601    during development without invalidating existing binfiles.
602    
603    In any case, every CMB.make updates the date information in version.sml.
604    (This is the date that is printed in the banner.)
605    
606    ----------------------------------------------------------------------
607    Name: Matthias Blume (blume (at) tti - c (dot) org)
608    Date: 2004/09/28 10:53:00 CDT
609    Tag: blume-20040928-controls
610    Description:
611    
612    Some cleanup of the controls code.
613    
614    ----------------------------------------------------------------------
615    Name: Matthias Blume (blume (at) tti - c (dot) org)
616    Date: 2004/09/27 22:08:00 CDT
617    Tag: blume-20040927-controls
618    Description:
619    
620    Added two pieces of functionality to the Controls interface:
621    
622    1.   val save'restore: 'a control -> unit -> unit
623    
624       grabs the current value of the control in stage 1 and restores it
625       in stage 2.
626    
627    2.   val set' : 'a control * 'a -> unit -> unit
628    
629       stores the given value into the control in stage 2 (i.e., delayed)
630       but does all error checking in stage 1.
631       (This is for string controls that need to do parse their argument --
632        something that might fail.  In some cases, notably in CM, one
633        already knows the intended argument but wants to delay the actual
634        assignment until a time when error recovery would be more difficult.)
635    
636    Changed the handling of controls in tool arguments to classes "sml" and
637    "lazysml":
638       - use Controls.save'restore as a more robust way of restoring the
639         old value (in particular: without having to re-parse the string)
640       - use controls to handle the "overload" keyword in the init group
641         (I believe this change actually fixes a long-standing obscure bug.)
642    
643    ----------------------------------------------------------------------
644    Name: Matthias Blume (blume (at) tti - c (dot) org)
645    Date: 2004/09/27 17:00:00 CDT
646    Tag: blume-20040927-lazysml
647    Description:
648    
649    Added a new tool class called "lazysml" to CM's tool chest.  The only
650    difference to "sml" is that compilation is done with Control.lazysml
651    set to true.  A source of class "lazysml" is automatically recognized
652    by a file name suffix of ".lml".
653    
654    In addition to the above feature, the original class "sml" now also
655    supports a tool argument "lazy" which has the same effect.  As a
656    result, the following three lines are equivalent:
657    
658        foo.sml : lazysml
659        foo.sml : sml (lazy)
660        foo.sml (lazy)
661    
662    The setting goes into effect both during parsing and during
663    compilation.  The original setting is restored right after parsing and
664    after compilation, respectively.
665    
666    In addition to all the above, there is also a general mechanism to set
667    ANY of the "controls" that are available at the command line via
668    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
669    apply as well. (In fact, "lazy" is implemented as a special case of
670    the general mechanism.)
671    
672    The .cm file syntax uses a new keyword tool argument called "with".
673    There are several ways of indicating the desired settings:
674    
675        foo.sml (with:parser.quotations=true)
676        foo.sml (with:(name:parser.quotations value:true))
677        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
678        foo.sml (with:(name1=value1 name2=value2 ...))
679        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
680    
681        etc.
682    
683    Another possible abbreviation is to leave out the =v or value:v part
684    if the name refers to a boolean control (in which case the value is
685    taken to be true).  Thus, one could get lazy sml also by saying:
686    
687        foo.sml (with:parser.lazy-keyword=true)
688        foo.sml (with:parser.lazy-keyword)
689        foo.sml (with:(name:parser.lazy-keyword value:true))
690        foo.sml (with:(name:parser.lazy-keyword))
691    
692    ----------------------------------------------------------------------
693    Name: Matthias Blume (blume (at) tti - c (dot) org)
694    Date: 2004/09/24 16:22:00 CDT
695    Tag: blume-20040924-ppc-long-branch
696    Description:
697    
698    Turned message about "emiting long form of branch" off by default.
699    Added a control flag to turn it back on when desired.
700    
701    ----------------------------------------------------------------------
702    Name: Matthias Blume (blume (at) tti - c (dot) org)
703    Date: 2004/09/24 16:05:00 CDT
704    Tag: blume-20040924-rounding
705    Description:
706    
707    Applied patch for setting rounding modes under Mac OS X.  Thanks to
708    Melissa O'Neill for providing the code!
709    
710    ----------------------------------------------------------------------
711    Name: Matthias Blume (blume (at) tti - c (dot) org)
712    Date: 2004/09/23 17:30:00 CDT
713    Tag: blume-20040923-envvars
714    Description:
715    
716    1. Changed definition of type ControlRegistry.registry_tree to
717       include control_info (i.e., the name of the controlling environment
718       variable).
719    
720    2. Added command-line flags -e and -E to print the names of environment
721       variables that can be used to control internal settings.
722       (This uses the new API mentioned in 1.)
723    
724    ----------------------------------------------------------------------
725    Name: Matthias Blume (blume (at) tti - c (dot) org)
726    Date: 2004/09/13 16:50:00 CDT
727    Tag: Release_110_49
728    Description:
729    
730    New working version (110.49).  NEW BOOTFILES!
731    
732    ----------------------------------------------------------------------
733    Name: Matthias Blume
734    Date: 2004/09/13 16:20:00 CDT
735    Tag: blume-20040913-config-mlrisc
736    Description:
737    
738    Put target "mlrisc" back into the default list.
739    (There is no harm in having it, and some users have expressed their
740    wish to have "mlrisc" included by default.)
741    
742    ----------------------------------------------------------------------
743    Name: John Reppy
744    Date: 2004/09/13
745    Tag: jhr-20040913-signals
746    Description:
747    
748    Fixed the signal masking code to properly nest mask/unmask operations
749    on a per-signal basis.
750    
751    ----------------------------------------------------------------------
752    Name: Matthias Blume (blume (at) tti - c (dot) org)
753    Date: 2004/09/08 13:20:00 CDT
754    Tag: blume-20040908-heap-magic
755    Description:
756    
757    Bumped the heap macig number to 0x09082004 to account for the changed
758    layout of the ML frame under MacOS X.
759    
760    ----------------------------------------------------------------------
761    Name: Allen Leung (leunga (at) reservoir (dot) com)
762    Date: 2004/09/03 11:26:00 EST
763    Tag: leunga-20040903-cygwin-install
764    Description:
765    
766       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
767    The Cygwin runtime is turned on by setting the environment
768    variable SMLNJ_CYGWIN_RUNTIME to 1.
769    
770    ----------------------------------------------------------------------
771    Name: Matthias Blume (blume (at) tti - c (dot) org)
772    Date: 2004/08/31 17:15:00 CDT
773    Tag: blume-20040831-core
774    Description:
775    
776    Added some exports to src/compiler/core.cm upon request by J. Joemann.
777    
778    ----------------------------------------------------------------------
779    Name: Matthias Blume (blume (at) tti - c (dot) org)
780    Date: 2004/08/30 17:55:00 CDT
781    Tag: blume-20040830-installer
782    Description:
783    
784    Upon request by Johannes Joemann:
785    
786    - improved ML code of installer to fall back to coping when renaming
787      fails (i.e., when source and target are on different file systems);
788      the code compiles but has yet to be tested in anger
789    - removed mlrisc from list of default targets (config/targets)
790    
791    ----------------------------------------------------------------------
792    Name: Matthias Blume (blume (at) tti - c (dot) org)
793    Date: 2004/08/27 17:20:00 CDT
794    Tag: blume-20040827-ptreql
795    Description:
796    
797    Added ptreql primop to structure InlineT (upon request from Larry
798    Paulson).
799    
800    ----------------------------------------------------------------------
801    Name: Allen Leung (leunga (at) reservoir (dot) com)
802    Date: 2004/08/15 21:21:00 EST
803    Tag: leunga-110_48-udgraph
804    Description:
805    
806    Another bug fix from Carl Hauser:
807    
808    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
809    > 48c48
810    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
811    > e::L)
812    > ---
813    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
814    > Without this, any deletion of an edge in an undirected graph does severe
815    > violence to the graph.
816    
817    ----------------------------------------------------------------------
818    Name: Allen Leung (leunga (at) reservoir (dot) com)
819    Date: 2004/08/10 23:35:00 EST
820    Tag: leunga-110_48-ppc
821    Description:
822    
823       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
824    
825    ----------------------------------------------------------------------
826    Name: Matthias Blume (blume (at) tti - c (dot) org)
827    Date: 2004/08/10 12:00:00 CDT
828    Tag: Release_110_48
829    Description:
830    
831    New working version (110.48).  NEW BOOTFILES!
832    
833    ----------------------------------------------------------------------
834    Name: Allen Leung (leunga (at) reservoir (dot) com)
835    Date: 2004/08/09 12:21:00 EST
836    Tag: leunga-110_47-dijsktra
837    Description:
838    
839    Bug fix from Carl Hauser:
840    
841    single_source_shortest_paths in dijkstra.sml was observed to get wrong
842    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
843    
844    The problem is that following the expression A.update(dist,s,Num.zero)
845    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
846    
847    ----------------------------------------------------------------------
848    Name: Matthias Blume (blume (at) tti - c (dot) org)
849    Date: 2004/08/06 18:10:00 CDT
850    Tag: blume-20040806-cmdline
851    Description:
852    
853    Fiddled with handling of command-line options:
854    
855      * sml now quits after processing the command line
856        if -H, -S, -h<n>, or -s<n> appears as the last
857        command-line argument
858      * a new option -q terminates the session when encountered on
859        the command line; subsequent arguments will be ignored
860      * bug fixes: short (erroneous) arguments are no longer ignored
861        completely
862    
863    ----------------------------------------------------------------------
864    Name: Allen Leung (leunga (at) reservoir (dot) com)
865    Date: 2004/08/04 18:17:00 EST
866    Tag: leunga-110_47-ppc-ibm-asm
867    Description:
868    
869      - Added minimal IBM assembly syntax support for PowerPC.
870    
871      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
872        been altered in the runtime.  We need an automatic way to keep the file
873        in sync.
874    
875    ----------------------------------------------------------------------
876    Name: Matthias Blume (blume (at) tti - c (dot) org)
877    Date: 2004/08/04 14:00:00 CDT
878    Tag: Release_110_47
879    Description:
880    
881    New working version (110.47).  NEW BOOTFILES!
882    
883    ----------------------------------------------------------------------
884    Name: Matthias Blume (blume (at) tti - c (dot) org)
885    Date: 2004/08/03 14:25:00 CDT
886    Tag: blume-20040803-callingconv
887    Description:
888    
889    Added low-level support for choosing C calling conventions by
890    twiddling the type of rawccall.  (See
891    src/compiler/Semant/types/cproto.sml for details.)
892    
893    ----------------------------------------------------------------------
894    Name: Matthias Blume (blume (at) tti - c (dot) org)
895    Date: 2004/08/02 15:55:00 CDT
896    Tag: blume-20040802-backout
897    Description:
898    
899    Backed out of change to win32-filesys.c.  The earlier patch to
900    get_file_time caused CM to produce files with the wrong time stamp.
901    
902    ----------------------------------------------------------------------
903    Name: Matthias Blume (blume (at) tti - c (dot) org)
904    Date: 2004/08/02 14:45:00 CDT
905    Tag: blume-20040802-nlffi-win32
906    Description:
907    
908    Added NLFFI support for Win32, adapted from a patch provided by David
909    Hansel.  This is currently completely untested.  Also, the issue
910    concerning stdcall vs. ccall is still unresolved.
911    
912    ----------------------------------------------------------------------
913    Name: Matthias Blume (blume (at) tti - c (dot) org)
914    Date: 2004/07/30 17:55:00 CDT
915    Tag: blume-20040730-various
916    Description:
917    
918    Gearing up towards 110.47...
919    
920    - various minor bugfixes to ml-nlffigen
921    - a beginning of a manual for nlffi
922    
923    - eliminated 'export name=value' in config/install.sh as this does
924      not work with certain versions of /bin/sh
925      (Thanks to David King at Motorola for catching this.)
926    
927    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
928      - added a test for tm==NULL to gmtime.c and localtime.c
929      - applied patch for incorrect GetFileTime under win32
930      - toSeconds -> toMilliseconds in Win32/win32-process.sml
931    
932    ----------------------------------------------------------------------
933    Name: Matthias Blume (blume (at) tti - c (dot) org)
934    Date: 2004/07/21 18:20:00 CDT
935    Tag: blume-20040721-nlffigen
936    Description:
937    
938    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
939      for a typedef to an incomplete type, but leave out the "typ" member.
940      (This is just for consistency.)
941    - Started to produce what is supposed to become better (i.e., comprehensive)
942      documentation of what ml-nlffigen does and produces.
943    
944    ----------------------------------------------------------------------
945    Name: Matthias Blume (blume (at) tti - c (dot) org)
946    Date: 2004/07/14 16:25:00 CDT
947    Tag: blume-20040714-union
948    Description:
949    
950    Added C_UNION to c-calls/c-types.sml and updated the machinery
951    (ml-nlffigen, cproto.sml) that conveys C function interface
952    information to the code generator.
953    
954    However, the actual architecture-specific implementation of function
955    arguments and results that are C unions is still not implemented.
956    
957    ----------------------------------------------------------------------
958    Name: Allen Leung (leunga (at) reservoir (dot) com)
959    Date: 2004/07/14 14:38:00 EST
960    Tag: leunga-110_46_1-ppc-lwzu
961    Description:
962    
963        Added these instructions to the PowerPC architecture:
964        LBZU(X), LHZU(X), LWZU(X),
965        STWU(X), STFDU, STFSU
966    
967        etc...
968    
969        Note: I haven't added their instruction encoding into the description.
970    
971    ----------------------------------------------------------------------
972    Name: Allen Leung (leunga (at) reservoir (dot) com)
973    Date: 2004/07/13 15:04:00 EST
974    Tag: leunga-110_46_1-ppc-lwarx
975    Description:
976    
977       Added the two instructions LWARX and STWCX to the PowerPC
978    instruction set.
979    
980       A (untested) rewrite of loop-structure.sml.  The old version
981    is completely broken.
982    
983    ----------------------------------------------------------------------
984    Name: Matthias Blume (blume (at) tti - c (dot) org)
985    Date: 2004/07/13 13:50:00 CDT
986    Tag: blume-20040713-nlffi
987    Description:
988    
989    - use paramAlloc to report c-calls with too many arguments
990      (for PPC version where parameter area is pre-allocated)
991    - added ccall_maxargspace to machspec (to implement the above)
992    - made "make" commend in CM's "make" tool configurable
993    - added option (default: on) for passing the name of the SML/NJ's "bin"
994      directory to "make"; the call looks like this:
995    
996         make <options> SMLNJ_BINDIR=<dir> <target>
997    
998      This can be used by the Makefile to, e.g., pick the "right" version
999      of ml-nlffigen.
1000    - minor code tweaks
1001    
1002    ----------------------------------------------------------------------
1003    Name: Matthias Blume (blume (at) tti - c (dot) org)
1004    Date: 2004/07/12 22:50:00 CDT
1005    Tag: blume-110_46_1-macosx-nlffi
1006    Description:
1007    
1008    NLFFI under Mac OS X now working (sort of).  This is largely untested,
1009    though.
1010    
1011    Note:  1. You have to make a new, clean build of the runtime system.
1012           2. There are new BOOTFILES, you have to use them!
1013              (Doing the bootstrap process yourself would be *very* painful!
1014               If you absolutely have to do it, build the system under
1015               a different architecture and then cross-compile.)
1016    
1017    Version bumped to 110.46.1 to account for runtime data format changes.
1018    
1019    ----------------------------------------------------------------------
1020    Name: Matthias Blume (blume (at) tti - c (dot) org)
1021    Date: 2004/06/18 14:30:00 CDT
1022    Tag: blume-20040618-unix
1023    Description:
1024    
1025    Changed the implementation of structure Unix so that the same stream
1026    is returned every time one of the {text,bin}{In,Out}streamOf functions
1027    is invoked on the same proc.  This is not what the spec currently
1028    says -- although IMO it arguably should.  (See discussion below.)
1029    
1030    ----------------------------------------------------------------------
1031    Name: Matthias Blume (blume (at) tti - c (dot) org)
1032    Date: 2004/06/17 18:15:00 CDT
1033    Tag: Release_110_46
1034    Description:
1035    
1036    New working version (110.46).  NEW BOOTFILES!
1037    
1038    ----------------------------------------------------------------------
1039    Name: Matthias Blume (blume (at) tti - c (dot) org)
1040    Date: 2004/06/17 17:20:00 CDT
1041    Tag: blume-20040617-timer-unix
1042    Description:
1043    
1044    Changed the interface of structures Timer and Unix to match the most
1045    recent Basis spec.
1046    
1047    In the case of Unix there still seems to be an open/weird issue:
1048    
1049       The {text,bin}{In,Out}streamOf functions are supposed to create
1050       fresh streams whenever they are called -- as opposed to have them
1051       return the same stream every time.  This design is supposed to
1052       prevent space leaks caused by proc values hanging on to streams.
1053    
1054       The reap function, on the other hand, is supposed to close the
1055       streams.  This cannot be done without having a handle on the
1056       stream in proc after all...
1057    
1058    I took the liberty to implement the following stopgap solution:
1059    
1060      The proc value hangs on to the most recently created stream(s).
1061      Reap closes those.  If either or both of the two streams hadn't
1062      been created at all yet, then reap will close the corresponding
1063      file descriptors directly.
1064    
1065    PS: I don't understand the original space leak argument anymore.  If
1066    a proc hangs on to the imperative stream, then I/O operations on those
1067    will advance the state of the cached stream and avoid the space leak.
1068    
1069    ----------------------------------------------------------------------
1070    Name: Matthias Blume (blume (at) tti - c (dot) org)
1071    Date: 2004/05/28 16:45:00 CDT
1072    Tag: blume-20040528-basis
1073    Description:
1074    
1075    Added signature PACK_REAL and exported functor PrimIO.
1076    
1077    ----------------------------------------------------------------------
1078    Name: Matthias Blume (blume (at) tti - c (dot) org)
1079    Date: 2004/05/25 16:00:00 CDT
1080    Tag: blume-20040525-group-owner
1081    Description:
1082    
1083    CM now ignores (but still accepts) the "owner" information in group
1084    descriptions.  The owner of a group is its next enclosing
1085    library. Each group must have a unique owner.  (There is a virtual
1086    "toplevel" library that owns groups which are not nested within a real
1087    library.)  Previously, each group had to explicitly declare its owner,
1088    and CM would check that such a declaration is correct.  The new scheme
1089    is to have CM check that for each group there is precisely one owning
1090    library.
1091    
1092    The advantage of the new scheme is that the programmer no longer needs
1093    to maintain the somewhat annoying owner information.  The downside is
1094    that CM cannot enforce the ownership rule across multiple runs of
1095    CM.make.  Fortunately, enclosing the same group in two different
1096    libraries A and B which are not part of the same program does not
1097    cause real problems.
1098    
1099    ----------------------------------------------------------------------
1100    Name: Matthias Blume (blume (at) tti - c (dot) org)
1101    Date: 2004/05/20 16:00:00 CDT
1102    Tag: blume-20040520-win32
1103    Description:
1104    
1105    Made the win32 version work again.  (Strangely, a misplaced comma had
1106    slipped into win32-process.c which prevented the runtime from being
1107    compiled correctly.)
1108    
1109    Also, included a minor addition to ml-build.bat analogous to what was
1110    done in blume-20040519-ml-build.
1111    
1112    ----------------------------------------------------------------------
1113    Name: Matthias Blume (blume (at) tti - c (dot) org)
1114    Date: 2004/05/19 22:10:00 CDT
1115    Tag: blume-20040519-ml-build
1116    Description:
1117    
1118    Arranged for ml-build to clean up after itself a little bit better.
1119    The script generates a temporary SML source file and compiles it using
1120    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1121    gets rid of those at the end, so they don't accumulate under .cm.
1122    
1123    This required a minor change to install.sh because the name of the
1124    metadata directory (default: .cm) is actually configurable at
1125    installation time.
1126    
1127    ----------------------------------------------------------------------
1128    Name: Matthias Blume (blume (at) tti - c (dot) org)
1129    Date: 2004/05/18 15:50:00 CDT
1130    Tag: blume-20040518-mkreader
1131    Description:
1132    
1133    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1134    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1135    
1136    ----------------------------------------------------------------------
1137    Name: Matthias Blume (blume (at) tti - c (dot) org)
1138    Date: 2004/05/11 14:35:00 CDT
1139    Tag: blume-20040511-win32sock
1140    Description:
1141    
1142    Added previously missing support for many socket-related functions
1143    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1144    for the voluminous patch!
1145    
1146    (I have not tested this patch under win32 yet.)
1147    
1148    Here is David's e-mail:
1149    
1150    Hi,
1151    
1152    Attached to this email you find a diff against sml/nj 110.45
1153    that will enable socket support under Windows.
1154    
1155    To apply the patch (using unix or cygwin)
1156    1) gunzip runtime.diff.gz
1157    2) "cd" into "src/runtime" in the source tree of a fresh
1158       110.45 installation.
1159    3) patch -p 1 < [your/path/to]runtime.diff
1160    
1161    The code compiles fine but has NOT yet been extensively tested.
1162    I only ran a few tests for basic socket client functionality
1163    (which worked fine).  Especially the functions that use ioctl
1164    are not tested at all and might not work (see below).
1165    
1166    I implemented this since we want to move to a newer version of sml/nj
1167    but need socket support in order to use it.  This is the first time I
1168    even had a look at the sml/nj source,  so please review my changes
1169    before making this part of the distribution!  Here are a few issues
1170    that I think might be better for someone to solve who is more
1171    familiar with the sml/nj source (and socket programming):
1172    
1173    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1174      exception since I could not figure out what the windows equivalent
1175      of these functions is
1176    
1177    - In sockets-osdep.h there are a some #include statements that are
1178      only used in a few files that include sockets-osdep.h
1179    
1180    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1181      does not process its return value since I don't know how to
1182      report an error upwards.
1183    
1184    - It would probably be good to have a call to WSACleanup() when
1185      the library is unloaded (if there is such a possibility).
1186      Otherwise I think Windows will take care of this automatically
1187      when the process finishes.
1188    
1189    - I used ioctlsocket() as a replacement for ioctl() but I have
1190      no idea if that is actually the proper replacement on Windows.
1191    
1192    - All these issues are marked in the code by "FIXME" comments.
1193    
1194    We use sml/nj extensively in our products and are quite happy
1195    with it.  I hope this contribution will help you.
1196    
1197    Keep up the good work!
1198    
1199    David
1200    
1201    ----------------------------------------------------------------------
1202    Name: Matthias Blume (blume (at) tti - c (dot) org)
1203    Date: 2004/05/11 14:20:00 CDT
1204    Tag: blume-20040511-installml
1205    Description:
1206    
1207    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1208    the patch.)
1209    
1210    ----------------------------------------------------------------------
1211    Name: Matthias Blume (blume (at) tti - c (dot) org)
1212    Date: 2004/05/11 14:05:00 CDT
1213    Tag: blume-20040511-nlffi-netbsd
1214    Description:
1215    
1216    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1217    the patch.)
1218    
1219    ----------------------------------------------------------------------
1220    Name: Matthias Blume (blume (at) tti - c (dot) org)
1221    Date: 2004/05/11 12:05:00 CDT
1222    Tag: blume-20040511-exports
1223    Description:
1224    
1225    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1226    export lists in compiler-related .cm-files.
1227    
1228    ----------------------------------------------------------------------
1229    Name: Matthias Blume (blume (at) tti - c (dot) org)
1230    Date: 2004/05/11 11:35:00 CDT
1231    Tag: blume-20040511-allsource
1232    Description:
1233    
1234    The installer now honors the "src-smlnj" target again, although its meaning
1235    has changed from "all sources required for the compiler" to "all sources
1236    the installer knows about".  In other words, if you enable "src-smlnj"
1237    in the "targets" file, then the installer will pull in sources for
1238    everything.  (Notice that this refers to source code only.  Compiled
1239    code is still only installed for modules that were requested explicitly
1240    or which are required for other modules that were requested explicitly.)
1241    
1242    ----------------------------------------------------------------------
1243    Name: Matthias Blume (blume (at) tti - c (dot) org)
1244    Date: 2004/04/23 17:40:00 CDT
1245    Tag: blume-20040423-ieee-scan
1246    Description:
1247    
1248    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1249    in the exponent calculation we get INF or ZERO (depending on the mantissa
1250    and the sign of the exponent).
1251    
1252    ----------------------------------------------------------------------
1253    Name: Matthias Blume (blume (at) tti - c (dot) org)
1254    Date: 2004/04/23 10:40:00 CDT
1255    Tag: blume-20040423-ml-build
1256    Description:
1257    
1258    The ml-build script now terminates with a non-0 status when something
1259    goes wrong.
1260    
1261    ----------------------------------------------------------------------
1262    Name: Matthias Blume (blume (at) tti - c (dot) org)
1263    Date: 2004/04/22 16:35:00 CDT
1264    Tag: blume-20040422-Option
1265    Description:
1266    
1267    Made exception Option to be the same as exception Option.Option
1268    (as it should be).
1269    
1270    ----------------------------------------------------------------------
1271    Name: Allen Leung (leunga (at) reservoir (dot) com)
1272    Date: 2004/03/19 14:40:00 EST
1273    Tag: leunga-20040319-cygwin-nlffi
1274    Description:
1275    
1276    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1277    of SML/NJ.  The problem is that
1278    
1279       lib = dlopen(NULL, ...)
1280       f   = dlsym(lib, "malloc");
1281    
1282    does not work on Windows unless we explicitly export symbols
1283    such as 'malloc' during linking.  We fixed this by explicitly
1284    exporting the required symbols with the magic gcc incantation:
1285    
1286        -Wl,--export-all cygwin.def
1287    
1288    where cygwin.def is a file containing all the symbols that
1289    we wish to export.
1290    
1291    I suspect this is a Windows problem and we'll have to
1292    do the same (somehow with windows compilers) when
1293    we build the native win32 version with the system
1294    calls LoadLibrary/GetProcAddress.
1295    
1296    ----------------------------------------------------------------------
1297    Name: Matthias Blume (blume (at) tti - c (dot) org)
1298    Date: 2004/03/04 16:35:00 CST
1299    Tag: blume-20040304-intinf-fmt
1300    Description:
1301    
1302    Fixed problem with IntInf.fmt (sign would show up on the right instead
1303    of on the left for BIN, OCT, and HEX).
1304    
1305    ----------------------------------------------------------------------
1306    Name: Matthias Blume (blume (at) tti - c (dot) org)
1307    Date: 2004/03/04 11:25:00 CST
1308    Tag: blume-20040304-symlinks
1309    Description:
1310    
1311    Fixed problem with installer script (unix only) where bin/ml-yacc and
1312    friends pointed (via symlinks) to absolute locations instead of just
1313    .run-sml.  This was reported by Vesa A Norrman.
1314    
1315    ----------------------------------------------------------------------
1316    Name: Matthias Blume (blume (at) tti - c (dot) org)
1317    Date: 2004/02/13 14:50:00 CST
1318    Tag: Release_110_45
1319    Description:
1320    
1321    New working version (110.45). New bootfiles.
1322    
1323    ----------------------------------------------------------------------
1324    Name: Matthias Blume (blume (at) tti - c (dot) org)
1325    Date: 2004/01/26 15:15:15 CST
1326    Tag: blume-20040126-toplevel
1327    Description:
1328    
1329    Improved handling of exceptions at the interactive toplevel.
1330    
1331    ----------------------------------------------------------------------
1332    Name: Matthias Blume (blume (at) tti - c (dot) org)
1333    Date: 2004/01/26 11:25:00
1334    Tag: blume-20040126-app
1335    Description:
1336    
1337    Type of top-level "app" corrected.
1338    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1339    
1340    ----------------------------------------------------------------------
1341    Name: Matthias Blume (blume (at) tti - c (dot) org)
1342    Date: 2003/11/18 17:10 CST
1343    Tag: blume-20031118-basis-fiddle
1344    Description:
1345    
1346    - changed Timer interface to what might become the spec
1347    - POSIX_FLAGS -> BIT_FLAGS according to spec
1348    - some other minor discrepancies wrt. spec eliminated
1349    
1350    ----------------------------------------------------------------------
1351    Name: Matthias Blume (blume (at) tti - c (dot) org)
1352    Date: 2003/11/06 12:00:00 CST
1353    Tag: Release_110_44
1354    Description:
1355    
1356    New working version (110.44). New bootfiles.
1357    
1358    ----------------------------------------------------------------------
1359    Name: Matthias Blume (blume (at) tti - c (dot) org)
1360    Date: 2003/11/04 11:50:00 CST
1361    Tag: blume-20031104-move-libraries
1362    Description:
1363    
1364    Eliminated the "dont_move_libraries" directive in config/targets.
1365    (The mechanism was broken and could not be fixed easily.  Moreover,
1366    there does not seem to be any reason not to move all libraries into
1367    lib during installation.  I originally implemented this directive as a
1368    backward-compatibility feature when I first introduced the new CM.
1369    Now that things have been stable for a long time and going back to the
1370    old CM is not an option, there is no reason to keep it around.)
1371    
1372    ----------------------------------------------------------------------
1373    Name: Matthias Blume (blume (at) tti - c (dot) org)
1374    Date: 2003/11/03 16:00:00 CST
1375    Tag: blume-20031103-installdir
1376    Description:
1377    
1378    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1379    Richards for pointing out the problem and providing the solution.)
1380    
1381    ----------------------------------------------------------------------
1382    Name: Matthias Blume (blume (at) tti - c (dot) org)
1383    Date: 2003/10/01 17:05:00 CDT
1384    Tag: blume-20031001-lal-mlrisc
1385    Description:
1386    
1387    MLRISC bug fix from Lal.
1388    
1389    ----------------------------------------------------------------------
1390    Name: Matthias Blume (blume (at) tti - c (dot) org)
1391    Date: 2003/09/30 16:10:00 CDT
1392    Tag: blume-20030930-primio-bat
1393    Description:
1394    
1395    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1396    2. Improved .bat files (for Win32 port) to make things work under Win95.
1397       (thanks to Aaron S. Hawley for this one)
1398    
1399    ----------------------------------------------------------------------
1400    Name: Matthias Blume (blume (at) tti - c (dot) org)
1401    Date: 2003/09/26 16:05:00 CDT
1402    Tag: blume-20030926-wrappriv
1403    Description:
1404    
1405    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1406    
1407    ----------------------------------------------------------------------
1408    Name: Matthias Blume (blume (at) tti - c (dot) org)
1409    Date: 2003/09/26 15:00:00 CDT
1410    Tag: blume-20030926-110_43_3
1411    Description:
1412    
1413    - additional cleanup
1414    - version number bump, NEW BOOTFILES
1415    
1416    ----------------------------------------------------------------------
1417    Name: Matthias Blume (blume (at) tti - c (dot) org)
1418    Date: 2003/09/26 12:00:00 CDT
1419    Tag: blume-20030926-ppautoload
1420    Description:
1421    
1422    I modified the read-eval-print loop so that the autoloader gets
1423    invoked whenever the prettyprinter tries to look up a symbol that
1424    is not currently defined in the toplevel environment but which
1425    appears in CM's autoload registry.  As a result, we see far fewer of
1426    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1427    
1428    In addition to this I tried to clean up some pieces of the Basis
1429    implementation (e.g., Socket, Word8Array) in order to prevent other
1430    instances of these ?.Foo.Bar.xxx names from being printed.
1431    
1432    The mechanism that picks names for types still needs some work, though.
1433    (Right now it seems that if there is a type A.t which is defined to
1434    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1435    "?.B.u" although the perhaps more sensible solution would be to use
1436    "A.t" in this case.  In other words, the prettyprinter should follow
1437    a chain of DEFtycs not farther than there are corresponding toplevel
1438    names in the current environment.)
1439    
1440    ----------------------------------------------------------------------
1441    Name: Matthias Blume (blume (at) tti - c (dot) org)
1442    Date: 2003/09/24 16:31:00 CDT
1443    Tag: blume-20030924-installer
1444    Description:
1445    
1446    Another installer tweak:  All the ML code for the installer is now
1447    compiled during CMB.make and put into a little library called
1448    $smlnj/installer.cm.  The installation then simply invokes
1449    
1450       sml -m $smlnj/installer.cm
1451    
1452    and everything happens automagically.
1453    
1454    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1455    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1456            and BINDIR.
1457    
1458    The new scheme guarantees that the ML code responsible for the installation
1459    is in sync with the APIs of the main system.  Also, the installer is
1460    somewhat faster because the installer script is precompiled.
1461    
1462    ----------------------------------------------------------------------
1463    Name: Matthias Blume (blume (at) tti - c (dot) org)
1464    Date: 2003/09/24 15:35:00 CDT
1465    Tag: blume-20030924-synsock
1466    Description:
1467    
1468    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1469    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1470    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1471    implementations of non-blocking operations from CML's Socket
1472    structure.
1473    
1474    ----------------------------------------------------------------------
1475    Name: Matthias Blume (blume (at) tti - c (dot) org)
1476    Date: 2003/09/24 15:10:05 CDT
1477    Tag: blume-20030924-sockets
1478    Description:
1479    
1480    1. Fixed SOCKET API and implementation to match Basis spec.
1481       This required changing the internal representation of sockets to one
1482       that remembers (for each socket file descriptor) whether it is currently
1483       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1484       call is made only if the state actually needs to change).
1485    
1486    2. OS-specific details of sockets were moved into separate files, thus
1487       making it possible to unify the bulk of the socket implementations
1488       between Unix and Win32.
1489    
1490    3. CML's socket API changed accordingly.
1491       (Note that we need to remove non-blocking functions from this API
1492       since they are redundant in the case of CML!)
1493    
1494    4. CML's socket implementation now makes use of non-blocking functions
1495       provided by Basis, thus removing all OS-dependent code from this part
1496       of CML.
1497    
1498    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1499    
1500    ----------------------------------------------------------------------
1501    Name: Matthias Blume (blume (at) tti - c (dot) org)
1502    Date: 2003/09/22 12:10:00 CDT
1503    Tag: blume-20030922-110_43_2
1504    Description:
1505    
1506    Made a new interim version and bootfiles for developer's bootstrapping
1507    convenience.
1508    
1509       110.43.2 -- NEW BOOTFILES
1510    
1511    ----------------------------------------------------------------------
1512    Name: Matthias Blume (blume (at) tti - c (dot) org)
1513    Date: 2003/09/19 15:55:00 CDT
1514    Tag: blume-20030919-cmdir
1515    Description:
1516    
1517    1. new-install.sh -> install.sh
1518    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1519    3. tweaked installer so that another name instead of .cm can be chosen
1520       at install time (by setting the CM_DIR_ARC environment variable
1521       during installation); once installation is complete, the name is
1522       fixed
1523    
1524    ----------------------------------------------------------------------
1525    Name: Matthias Blume (blume (at) tti - c (dot) org)
1526    Date: 2003/09/18 16:00:00 CDT
1527    Tag: blume-20030918-110_43_1
1528    Description:
1529    
1530    Made a new interim version and bootfiles for developer's bootstrapping
1531    convenience.
1532    
1533       110.43.1 -- NEW BOOTFILES
1534    
1535    ----------------------------------------------------------------------
1536    Name: Matthias Blume (blume (at) tti - c (dot) org)
1537    Date: 2003/09/18 15:20:00 CDT
1538    Tag: blume-20030918-misc
1539    Description:
1540    
1541    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1542       this seems to be controversial at the moment) and used those in
1543       Posix.ProcEnv.times.
1544    
1545    2. Added Time.{from,to}Nanoseconds to Time.
1546    
1547    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1548       For example, fromLargeInt never needs to look at more than 3 "big
1549       digits" to get its 53 bits of precision.
1550    
1551    ----------------------------------------------------------------------
1552    Name: Matthias Blume (blume (at) tti - c (dot) org)
1553    Date: 2003/09/17 16:30:00 CDT
1554    Tag: blume-20030917-real32-slices
1555    Description:
1556    
1557    Added an entry to the primitive environment
1558    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1559    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1560    
1561    Removed some of the "magic" constants in real64.sml and replaced them
1562    with code that generates these values from their corresponding
1563    integer counterparts.
1564    
1565    Made all(?) the slice-related changes to the Basis and made everything
1566    compile again...
1567    
1568    ----------------------------------------------------------------------
1569    Name: Matthias Blume (blume (at) tti - c (dot) org)
1570    Date: 2003/09/15 17:45:00 CDT
1571    Tag: blume-20030915-rbase
1572    Description:
1573    
1574    Fixed bug in Real.fromLargeInt.
1575    
1576    ----------------------------------------------------------------------
1577    Name: Matthias Blume (blume (at) tti - c (dot) org)
1578    Date: 2003/09/13 18:11:00 CDT
1579    Tag: blume-20030913-libinstall
1580    Description:
1581    
1582    Minor bugfix in config/libinstall (set anchor with path to
1583    standalone tool after installing it, otherwise libraries that
1584    need ml-lex or ml-yacc won't compile the first time the installer
1585    runs).
1586    
1587    ----------------------------------------------------------------------
1588    Name: Matthias Blume (blume (at) tti - c (dot) org)
1589    Date: 2003/09/12 11:45:00 CDT
1590    Tag: blume-20030912-various
1591    Description:
1592    
1593     - fixed bug in Real.toLargeInt
1594     - fixed bug in Posix.ProcEnv.times
1595     - changed inputLine functions to return an option
1596     - minor installer improvements / bugfixes
1597     - changed default @SMLalloc parameter for x86/celeron to 64k
1598    
1599    ----------------------------------------------------------------------
1600    Name: Matthias Blume (blume (at) tti - c (dot) org)
1601    Date: 2003/09/09 22:00:00 CDT
1602    Tag: Release_110_43
1603    Description:
1604    
1605        New working release 110.43.  New bootfiles.
1606    
1607    ----------------------------------------------------------------------
1608    Name: Matthias Blume (blume (at) tti - c (dot) org)
1609    Date: 2003/09/09 19:20:00 CDT
1610    Tag: blume-20030909-installer
1611    Description:
1612    
1613    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1614    Modified config/install.bat to take advantage of it.  Also modified
1615    config/install.sh (and called it config/new-install.sh) to take advantage
1616    of it on Unix systems.  (The SML code is (supposed to be) platform-
1617    independent.)
1618    
1619    The installer can now install everything under Win32
1620    as well as under *nix as long as it compiles.
1621    
1622    Other changes:
1623    
1624      - made CML compile again under Win32
1625      - made eXene compile under Win32 (by providing a fake structure UnixSock
1626        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1627      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1628        same as type int;  under Win32 it isn't
1629      - fixed some slice-related problems in the win32-specific parts of CML
1630      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1631        (under Win32, the volume name is case-insensitive, and the
1632        OS.Path code compares volume names for equality)
1633    
1634    ----------------------------------------------------------------------
1635    Name: Matthias Blume (blume (at) tti - c (dot) org)
1636    Date: 2003/09/08 11:55:00 CDT
1637    Tag: blume-20030908-fullpath
1638    Description:
1639    
1640    Made Win32 version of OS.FileSys.fullPath return current directory
1641    when given an empty string.  This is what the spec says, and incidentally,
1642    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1643    cases when presented with the name of a non-existing .cm file.)
1644    
1645    ----------------------------------------------------------------------
1646    Name: Matthias Blume (blume (at) tti - c (dot) org)
1647    Date: 2003/09/04 16:30:00 CDT
1648    Tag: blume-20030905-slices-etc
1649    Description:
1650    
1651      1. Changed interface to vectors and arrays in Basis to match
1652         (draft) Basis spec.
1653      2. Added signatures and implementations of slices according to
1654         Basis spec.
1655      3. Edited source code throughout the system to make it compile again
1656         under 1. and 2.  (In some cases code had to be added to have it
1657         match the new signatures.)
1658      4. MLRISC should be backward-compatible: the copies of the originals
1659         of files that needed to change under 3. were retained, the .cm files
1660         check the compiler version number and use old versions when
1661         appropriate.
1662      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1663         dirstream -> string option (in accordance with Basis spec).
1664      6. When generating code that counts lines, ml-lex used function
1665         CharVector.foldli, taking advantage of its old interface.
1666         This has been replaced with the corresponding code from
1667         CharVectorSlice.  (html-lex must be re-lexed!)
1668      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1669         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1670      8. Removed temporary additions (fromInternal, toInternal) from the
1671         (now obsolete) IntInf in smlnj-lib/Util.
1672      9. Cleaned up structure Byte.
1673     10. Added localOffset, scan, and fromString to Date (according to spec).
1674         Cleaned/corrected implementation of Date.
1675         (Still need to check for correctness; implement better canonicalizeDate.)
1676     11. Added "scan" to signature IEEE_REAL.
1677     12. Some improvements to IntInf [in particular: efficiency-hack for
1678         mod and rem when second operand is 2 (for parity checks).]
1679     13. Changed representation of type Time.time, using a single IntInf.int
1680         value counting microseconds.  This considerably simplified the
1681         implementation of structure Time.  We now support negative time
1682         values; scan and fromString handle signs.
1683     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1684         ArraySlice).
1685    
1686    ----------------------------------------------------------------------
1687    Name: Matthias Blume (blume (at) tti - c (dot) org)
1688    Date: 2003/08/28 17:00:00 CDT
1689    Tag: blume-20030828-intinf
1690    Description:
1691    
1692    This is a major update which comes with a version number bump
1693    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1694    and an implementation of IntInf in the Basis.
1695    
1696    There are a fairly large number of related changes and updates throughout
1697    the system:
1698    
1699    Basis:
1700            - Implemented IntInf.
1701            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1702            - Added some missing Real64 operations, most notably Real.toLargeInt.
1703            - Added FixedInt as a synonym for Int32.
1704    
1705    compiler:
1706            * Added support for a built-in intinf type.
1707              - literals
1708              - pattern matching
1709              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1710              - overloading on literals and operations
1711    
1712            This required adding a primitive type intinf, some additional
1713            primops, and implementations for several non-trivial intinf
1714            operations in Core.  (The intinf type is completely abstract
1715            to the compiler; all operations get delegated back to the Core.)
1716    
1717            * Intinf equality is handled by polyequal.  However, the compiler
1718            does not print its usual warning in this case (since polyequal
1719            is the right thing to do there).
1720    
1721            * Improved the organization of structure InlineT.
1722    
1723            * A word about conversion primops:
1724            If conversions involving intinf do not cancel out during
1725            CPS contract, then the compiler must insert calls to Core functions.
1726            Since all core access must be resolved already during the FLINT
1727            translate phase, it would be too late a the time of CPS contract
1728            to add new Core calls.  For this reason, conversion primops
1729            for intinf carry two arguments:  1. the numeric argument that
1730            they are supposed to convert, and 2. the Core function that
1731            can help with this conversion if necessary.  If CPS contract
1732            eliminates a primop, then the associated Core function becomes
1733            dead and goes away.  Intinf conversion primops that do not get
1734            eliminated by CPS contract get rewritten into calls of their
1735            core functions by a separate, new phase.
1736    
1737    interactive system:
1738            - Control.Print.intinfDepth controls max length of intinf constants
1739            being printed. (Analogous to Control.Print.stringDepth.)
1740            - Cleanup in printutil and pputil: got rid of unused stuff and
1741            duplicates; replaced some of the code with code that makes better
1742            use of library functionality.
1743    
1744    CM:
1745            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1746            are no longer silent.
1747    
1748    CKIT:
1749            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1750            in favor of LargeInt -- which is now the same as IntInf.
1751            CKIT-knowledgable people should check whether this is what's
1752            intended and otherwise change things back to using Int32 or
1753            FixedInt.
1754    
1755    Throughout the code:
1756            Started using IntInf.int literals and built-in operations
1757            (e.g., comparison with 0) where this seems appropriate.
1758    
1759    
1760    ----------------------------------------------------------------------
1761    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1762    Date: 2003/08/13 11:36:00 CDT
1763    Tag: dbm-20030813-mcz-merge1
1764    Description:
1765    
1766    Merging changes from the mcz-branch development branch into trunk.
1767    These changes involve replacement of the emulated old prettyprinter
1768    interface with direct use of the SML/NJ Lib PP library, and fixing
1769    of a couple of bugs (895, 1186) relating to error messages.  A new
1770    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1771    has been added.
1772    
1773    ----------------------------------------------------------------------
1774    Name: Matthias Blume (blume (at) tti - c (dot) org)
1775    Date: 2003/08/11 15:45:00 CDT
1776    Tag: blume-20030811-windows
1777    Description:
1778    
1779    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1780    
1781        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1782    
1783    This patch restores SML/NJ's ability to run under win32.  There are a
1784    number of changes, including fixes for several bugs that had gone
1785    unnoticed until now:
1786    
1787      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1788        trying to run the win32 version from within cygwin.)
1789    
1790      - There are a number of simple .bat scripts that substitute for their
1791        corresponding Unix shell-scripts.  (See below.)
1792    
1793      - The internals of ml-build have been modified slightly.  The main
1794        difference is that instead of calling ".link-sml" (or link-sml.bat)
1795        using OS.Process.system, the ML process delegates this task back
1796        to the script.  Otherwise problems arise in mixed environments such
1797        as Cygwin where scripts look and work like Unix scripts, but
1798        where OS.Process.system cannot run them.
1799    
1800      - In CM, the srcpath pickler used native pathname syntax -- which
1801        is incorrect in the case of cross-compilation.  The new pickle format
1802        is independent of platform-specific naming conventions.
1803    
1804      - Path configuration files (such as lib/pathconfig) can now choose
1805        between native and standard syntax.  Placing a line of the form
1806    
1807            standard!
1808    
1809        into the file causes all subsequent paths to be interpreted using
1810        CM standard pathname syntax (= Unix conventions); a line
1811    
1812            native!
1813    
1814        switches back to native style.  This was needed so that
1815        path config files can be written portably, see src/system/pathconfig.
1816    
1817      - Runtime system:
1818    
1819          - win32-filesys.c:  get_file_time and set_file_time now
1820            access modification time, not creation time.
1821    
1822          - I/O code made aware of new array representation.
1823    
1824          - Bug fixes in X86.prim.masm.
1825    
1826       - src/system/makeml made aware of win32. (For use under cygwin
1827            and other Unix-environments for windows.)
1828    
1829       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1830         which caused BinIO.inputAll to fail consistently.
1831    
1832    .bat scripts:
1833    
1834       Windows .bat scripts assume that SMLNJ_HOME is defined.
1835    
1836       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1837         applications (sml, ml-yacc, ml-lex).
1838       - ml-build.bat: analogous to ml-build.
1839       - config\install.bat:  Analogous to config/install.sh.  This requires
1840         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1841         (nmake etc. must be on the path, and vcvars32 must have been run.)
1842         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1843         and the bootfile hierarchy must have been unpacked under
1844         sml.boot.x86-win32.
1845         The script is very primitive and does a poor job at error checking.
1846         It only installs the base system, ml-lex, and ml-yacc.  No other
1847         libraries are being installed (i.e., you get only those that
1848         are part of the compiler.)
1849       - link-sml.bat:  analogous to .link-sml, but not currently used
1850    
1851    Unrelated bug fixes:
1852    
1853      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1854        types.
1855      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1856        otherwise.)
1857    
1858    ----------------------------------------------------------------------
1859    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1860    Date: 2003/06/17
1861    Tag: macqueen-20030617-bug895
1862    Description:
1863    
1864    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1865    Tag will be used for new development branch (mcz-branch) for
1866    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1867    
1868    ----------------------------------------------------------------------
1869    Name: Matthias Blume (blume (at) tti - c (dot) org)
1870    Date: 2003/05/27 16:55:00 CDT
1871    Tag: blume-20030527-polyeq
1872    Description:
1873    
1874    Tried to eliminated most cases of polymorphic equality.
1875    
1876    ----------------------------------------------------------------------
1877    Name: Matthias Blume (blume (at) tti - c (dot) org)
1878    Date: 2003/05/21 17:45:00 CDT
1879    Tag: blume-20030517-complete
1880    Description:
1881    
1882    Two changes:
1883    
1884      1. Added a flag for controlling whether non-exhaustive bindings will
1885         be treated as errors (default is false).
1886      2. Cleaned up the *entire* source tree so that CMB.make goes through
1887         without a single non-exhaustive match- or bind warning.
1888    
1889    ----------------------------------------------------------------------
1890    Name: Matthias Blume (blume (at) tti - c (dot) org)
1891    Date: 2003/05/17 10:20:00 CDT
1892    Tag: blume-20030517-absyn
1893    Description:
1894    
1895    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1896    
1897       This mainly affects the quality of error messages.  However, some
1898       of the code is now more straightforward than before. (Treatment of
1899       the above four constructs in translate.sml is much simpler than
1900       the "macro-expansion" that was going on before.  Plus, the mach-
1901       compiler no longer gets invoked just to be able to compile an
1902       if-expression.)
1903    
1904    2. The ErrorMsg.Error exception is now caught and absorbed by the
1905       interactive loop.
1906    
1907    ----------------------------------------------------------------------
1908    Name: Allen Leung
1909    Date: 2003/05/16 13:05:00 CDT
1910    Tag: leunga-20030516-cygwin-runtime
1911    Description:
1912    
1913        Ported the runtime system to cygwin, which uses the unix
1914    x86-unix bin files.  Missing/buggy features:
1915    
1916       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1917         the Cygwin library.
1918       o Ctrl-C handling may be flaky.
1919       o Windows system calls and Windows I/O are not supported.
1920    
1921       A new set of binfiles is located at:
1922    
1923         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1924    
1925       This is only needed for bootstrapping the cygwin version of smlnj.
1926       Other x86 versions can use the existing binfiles.
1927    
1928    ----------------------------------------------------------------------
1929    Name: Matthias Blume
1930    Date: 2003/04/08 15:42:00 CDT
1931    Tag: blume-20030408-listpair
1932    Description:
1933    
1934    1. Added a target 'mlrisc' to installer.
1935    
1936    2. Added missing elements to structure ListPair.
1937    
1938    ----------------------------------------------------------------------
1939    Name: Allen Leung
1940    Date: 2003/01/07 10:40:00 EST
1941    Tag: leunga-20030107-int-rem
1942    Description:
1943    
1944    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1945    The arguments to the SUBL instruction were swapped.
1946    
1947    ----------------------------------------------------------------------
1948    Name: Matthias Blume
1949    Date: 2002/12/12 16:25:00 EST
1950    Tag: blume-20021212-risc-ra
1951    Description:
1952    
1953    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1954    sent the RA into an infinite loop when floating point registers get
1955    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1956    110.37 and 110.38.)
1957    There was another set of potential problems related to the handling of
1958    MLRISC annotations (but those did not yet cause real problems, apparently).
1959    
1960    ----------------------------------------------------------------------
1961    Name: Matthias Blume
1962    Date: 2002/12/06 22:40:00 EST
1963    Tag: blume-20021206-cm-fileid
1964    Description:
1965    
1966    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1967    This fixes the problem of CM getting confused by files that suddenly
1968    change their identity (e.g., by getting unlinked and recreated by some
1969    text editor such as vi).  There might be a better/cheaper/cleaner way
1970    of doing this, but for now this will have to do.
1971    
1972    ----------------------------------------------------------------------
1973    Name: Matthias Blume
1974    Date: 2002/10/28 09:50:00 EST
1975    Tag: blume-20021028-typecheck
1976    Description:
1977    
1978    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1979    
1980    ----------------------------------------------------------------------
1981    Name: Matthias Blume
1982    Date: 2002/10/17 09:10:00 EDT
1983    Tag: Release_110_42
1984    Description:
1985    
1986    In good old tradition, there has been a slight hiccup so that we have
1987    to patch 110.42 after the fact.   The old release tag has been replaced
1988    (see below).
1989    
1990    The change solves a problem with two competing approaches the
1991    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1992    each other's way.
1993    
1994    This change only affects the runtime system code and the installer script.
1995    (No new bootfiles.)
1996    
1997    ----------------------------------------------------------------------
1998    Name: Matthias Blume
1999    Date: 2002/10/16 12:00:00 EDT
2000    Tag: Release_110_42_removed
2001    Description:
2002    
2003    New working release.  New bootfiles.
2004    
2005    ----------------------------------------------------------------------
2006    Name: Matthias Blume
2007    Date: 2002/10/10 13:10:00 EDT
2008    Tag: blume-20021010-ppc-divs
2009    Description:
2010    
2011    The mltree operator DIVS must be implemented with an overflow check on
2012    the PPC because the hardware indicates divide-by-zero using "overflow" as
2013    well.
2014    
2015    ----------------------------------------------------------------------
2016    Name: Matthias Blume
2017    Date: 2002/07/23 11:45:00
2018    Tag: blume-20020723-smlnj-home
2019    Description:
2020    
2021    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
2022    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
2023    CM_PATHCONFIG is also set), the path configuration file is assumed
2024    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
2025    the entire tree to some other place and everything will "just work".
2026    
2027    (Companion commands such as ml-build and ml-makedepend also sense this
2028    variable.)
2029    
2030    ----------------------------------------------------------------------
2031    Name: Matthias Blume
2032    Date: 2002/07/12 21:19:00 EDT
2033    Tag: blume-20020712-liveness
2034    Description:
2035    
2036    Exported two useful "step" functions from liveness module (MLRISC).
2037    
2038    ----------------------------------------------------------------------
2039    Name: Matthias Blume
2040    Date: 2002/07/05 16:00 EDT
2041    Tag: Release_110_41
2042    Description:
2043    
2044    New working release.  New bootfiles.
2045    
2046    ----------------------------------------------------------------------
2047    Name: Matthias Blume
2048    Date: 2002/07/05 10:25:00 EDT
2049    Tag: blume-20020705-btimp
2050    Description:
2051    
2052    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
2053    other clients can set up backtracing support.
2054    
2055    ----------------------------------------------------------------------
2056    Name: Matthias Blume
2057    Date: 2002/06/25 17:23:00 EDT
2058    Tag: blume-20020625-fpmax
2059    Description:
2060    
2061    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
2062    case.  (The sense of the isNaN test was reversed -- which made min and
2063    max always return their first argument.)
2064    
2065    ----------------------------------------------------------------------
2066    Name: Matthias Blume
2067    Date: 2002/06/11
2068    Tag: blume-20020611-unixpath
2069    Description:
2070    
2071    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
2072    
2073    ----------------------------------------------------------------------
2074    Name: Matthias Blume
2075    Date: 2002/06/10 16:35:00 EDT
2076    Tag: blume-20020610-ieeereal
2077    Description:
2078    
2079    I back-ported my implementation of IEEEReal.fromString from the
2080    idlbasis-devel branch so that we can test it.
2081    
2082    Another small change is that ppDec tries to give more information
2083    than just "<sig>" in the case of functors.  However, this code is
2084    broken in some mysterious way if the functor's body's signature
2085    has not been declared by ascription but gets inferred from the
2086    implementation.  This needs fixing...
2087    
2088    ----------------------------------------------------------------------
2089    Name: Matthias Blume
2090    Date: 2002/05/31
2091    Tag: blume-20020531-btrace-mode
2092    Description:
2093    
2094    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2095    the wayside when I switched over to using Controls everywhere.)
2096    
2097    ----------------------------------------------------------------------
2098    Name: Lal George
2099    Date: 2002/05/23 12:21:40 EDT
2100    Tag: george-20020523-visual-labels
2101    Description:
2102    
2103    Labels are now displayed in the graphical output to make
2104    the fall-through and target blocks obvious.
2105    
2106    ----------------------------------------------------------------------
2107    Name: Matthias Blume
2108    Date: 2002/05/22 11:03:00 EDT
2109    Tag: blume-20020522-shrink
2110    Description:
2111    
2112    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2113    situation that comes up when allocating huge arrays.
2114    
2115    ----------------------------------------------------------------------
2116    Name: Matthias Blume
2117    Date: 2002/05/21 16:00:00 EDT
2118    Tag: Release_110_40
2119    Description:
2120    
2121    New working release (110.40).  New bootfiles.
2122    
2123    [Also: John Reppy fixed GC bug 1131.]
2124    
2125    ----------------------------------------------------------------------
2126    Name: Matthias Blume
2127    Date: 2002/05/21 12:35:00 EDT
2128    Tag: blume-20020521-cmdoc
2129    Description:
2130    
2131    CM documentation update.
2132    
2133    ----------------------------------------------------------------------
2134    Name: Matthias Blume
2135    Date: 2002/05/21 10:55:00 EDT
2136    Tag: blume-20020521-misc
2137    Description:
2138    
2139    - John tweaked runtime to be silent on heap export (except when
2140      GC messages are on).
2141    - I added a few more things (cross-compiling versions of CMB) to
2142      config/preloads (as suggestions).
2143    
2144    ----------------------------------------------------------------------
2145    Name: Matthias Blume
2146    Date: 2002/05/20 22:25:00 EDT
2147    Tag: blume-20020520-controls
2148    Description:
2149    
2150    - Added ControlUtil structure to control-lib.cm.
2151    - Use it throughout.
2152    - Used Controls facility to define MLRISC controls (as opposed to
2153      registering MLRISC control ref cells with Controls after the fact)
2154    - Fixed messed-up controls priorities.
2155    
2156    * Removed again all the stuff from config/preloads that one wouldn't
2157      be able to preload at the time the initial heap image is built.
2158      (Many libraries, e.g., CML, do not exist yet at this time.  The
2159       only libraries that can be preloaded via config/preloads are those
2160       that come bundled with the bootfiles.)
2161    
2162    ----------------------------------------------------------------------
2163    Name: Matthias Blume
2164    Date: 2002/05/20 10:59:00 EDT
2165    Tag: blume-20020520-preloads
2166    Description:
2167    
2168    Added a lot of commented-out suggestions for things to be included
2169    in config/preloads.
2170    
2171    ----------------------------------------------------------------------
2172    Name: Allen Leung
2173    Date: 2002/05/18 14:20:00 EDT
2174    Tag: leunga-20020518-mdl
2175    Description:
2176    
2177    o Made the mdl tool stuff compile and run again.
2178    o I've disabled all the stuff that depends on RTL specifications; they
2179      are all badly broken anyway.
2180    
2181    ----------------------------------------------------------------------
2182    Name: Matthias Blume
2183    Date: 2002/05/17 16:49:00 EDT
2184    Tag: blume-20020517-controls
2185    Description:
2186    
2187    0. John Reppy made several modifications to the SML/NJ library.
2188       In particular, there is a shiny new controls-lib.cm.
2189    
2190    1. Pushed new controls interface through compiler so that everything
2191       compiles again.
2192    
2193    2. Added FormatComb and FORMAT_COMB to the CML version of the
2194       SML/NJ library (so that CML compiles again).
2195    
2196    3. Modified init scripts because XXX_DEFAULT environment variables
2197       are no longer with us.  (Boot-time initialization is now done
2198       using the same environment variables that are also used for
2199       startup-time initialization of controls.)
2200    
2201    ----------------------------------------------------------------------
2202    Name: Lal George
2203    Date: 2002/05/15 09:20:10 EDT
2204    Tag: george-20020515-pseudo-op-decls
2205    Description:
2206    
2207            All pseudo-ops emitted before the first segment declaration
2208    such as TEXT, DATA, and BSS directives are assumed to be global
2209    declarations and are emitted first in the assembly file. This is
2210    useful in a number of situations where one has pseudo-ops that are not
2211    specific to any segment, and also works around the constraint that one
2212    cannot have client pseudo-ops in the TEXT segment.
2213    
2214            Because no segment is associated with these declarations it is
2215    an error to allocate any space or objects before the first segment
2216    directive and an exception will be raised. However, we cannot make
2217    this check for client pseudo-ops.
2218    
2219            These top level declarations are a field in the CFG graph_info.
2220    In theory you can continue to add to this field after the CFG has been
2221    built -- provided you know what you are doing;-)
2222    
2223    ----------------------------------------------------------------------
2224    Name: Matthias Blume
2225    Date: 2002/05/13 16:40:00 EDT
2226    Tag: blume-20020513-pp-etc
2227    Description:
2228    
2229    A few minor bugfixes:
2230    
2231      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2232        (Bogus printouts for redefined bindings still occur.  Compiler
2233        bug should no longer occur now.  We need to redo the prettyprinter
2234        from scratch.)
2235    
2236      - CM pathname printer now also adds escape sequences for ( and )
2237    
2238      - commend and docu fixes for ml-nlffi
2239    
2240    ----------------------------------------------------------------------
2241    Name: Matthias Blume
2242    Date: 2002/05/10 16:40:00 EDT
2243    Tag: blume-20020510-erg-textio
2244    Description:
2245    
2246    Applied the following bugfix provided by Emden Gansner:
2247    
2248        Output is corrupted when outputSubstr is used rather than output.
2249    
2250        The problem occurs when a substring
2251    
2252            ss = (s, dataStart, dataLen)
2253    
2254        where dataStart > 0, fills a stream buffer with avail bytes left.
2255        avail bytes of s, starting at index dataStart, are copied into the
2256        buffer, the buffer is flushed, and then the remaining dataLen-avail
2257        bytes of ss are copied into the beginning of the buffer. Instead of
2258        starting this copy at index dataStart+avail in s, the current code
2259        starts the copy at index avail.
2260    
2261        Fix:
2262        In text-io-fn.sml, change line 695 from
2263             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2264        to
2265             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2266    
2267    ----------------------------------------------------------------------
2268    Name: Matthias Blume
2269    Date: 2002/04/12 13:55:00 EDT
2270    Tag: blume-20020412-assyntax
2271    Description:
2272    
2273    1. Grabbed newer assyntax.h from the XFree86 project.
2274    2. Fiddled with how to compile X86.prim.asm without warnings.
2275    3. (Very) Minor cleanup in CM.
2276    
2277    ----------------------------------------------------------------------
2278    Name: Matthias Blume
2279    Date: 2002/04/01 (no joke!) 17:07:00 EST
2280    Tag: blume-20020401-x86div
2281    Description:
2282    
2283    Added full support for div/mod/rem/quot on the x86, using the machine
2284    instruction's two results (without clumsily recomputing the remainder)
2285    directly where appropriate.
2286    
2287    Some more extensive power-of-two support was added to the x86 instruction
2288    selector (avoiding expensive divs, mods, and muls where they can be
2289    replaced with cheaper shifts and masks).  However, this sort of thing
2290    ought to be done earlier, e.g., within the CPS optimizer so that
2291    all architectures benefit from it.
2292    
2293    The compiler compiles to a fixed point, but changes might be somewhat
2294    fragile nevertheless.  Please, report any strange things that you might
2295    see wrt. div/mod/quot/rem...
2296    
2297    ----------------------------------------------------------------------
2298    Name: Matthias Blume
2299    Date: 2002/03/29 17:22:00
2300    Tag: blume-20020329-div
2301    Description:
2302    
2303    Fixed my broken div/mod logic.  Unfortunately, this means that the
2304    inline code for div/mod now has one more comparison than before.
2305    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2306    The problem was with quotient = 0, because that alone does not tell
2307    us which way the rounding went.  One then has to look at whether
2308    remainder and divisor have the same sign...  :(
2309    
2310    Anyway, I replaced the bootfiles with fresh ones...
2311    
2312    ----------------------------------------------------------------------
2313    Name: Matthias Blume
2314    Date: 2002/03/29 14:10:00 EST
2315    Tag: blume-20020329-inlprims
2316    Description:
2317    
2318    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2319    
2320    Primops have changed. This means that the bin/boot-file formats have
2321    changed as well.
2322    
2323    To make sure that there is no confusion, I made a new version.
2324    
2325    
2326    CHANGES:
2327    
2328    * removed REMT from mltree (remainder should never overflow).
2329    
2330    * added primops to deal with divisions of all flavors to the frontend
2331    
2332    * handled these primops all the way through so they map to their respective
2333      MLRISC support
2334    
2335    * used these primops in the implementation of Int, Int32, Word, Word32
2336    
2337    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2338    
2339    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2340    
2341    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2342      floating point
2343    
2344    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2345      and Real (but Real.abs maps to a separate floating-point-only primop)
2346    
2347    
2348    TODO items:
2349    
2350    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2351      instructions because the machine instruction encoder cannot handle
2352      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2353      REM in the first place, and REM is supported by the x86 machine coder.)
2354    
2355    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2356      selection phase.  (The two can be streamlined because the hardware
2357      delivers both quotient and remainder at the same time anyway.)
2358    
2359    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2360      (Currently the behavior is inconsistent both across architectures and
2361      wrt. the draft Basis spec.)
2362    
2363    * Word8 should eventually be handled natively, too.
2364    
2365    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2366      as if there currently is no execution path that could trigger it in
2367      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2368      hold for things like multiplication and division.)
2369    
2370    ----------------------------------------------------------------------
2371    Name: Matthias Blume
2372    Date: 2002/03/27 16:27:00 EST
2373    Tag: blume-20020327-mlrisc-divisions
2374    Description:
2375    
2376    Added support for all four division operations (ML's div, mod, quot,
2377    and rem) to MLRISC.  In the course of doing so, I also rationalized
2378    the naming (no more annoying switch-around of DIV and QUOT), by
2379    parameterizing the operation by div_rounding_mode (which can be either
2380    DIV_TO_ZERO or DIV_TO_NEGINF).
2381    
2382    The generic MLTreeGen functor takes care of compiling all four
2383    operations down to only round-to-zero div.
2384    
2385    Missing pieces:
2386    
2387      * Doing something smarter than relying on MLTreeGen on architectures
2388        like, e.g., the x86 where hardware division delivers both quotient and
2389        remainder at the same time.  With this, the implementation of the
2390        round-to-neginf operations could be further streamlined.
2391    
2392      * Remove inlining support for div/mod/rem from the frontend and replace it
2393        with primops that get carried through to the backend.  Do this for all
2394        int and word types.
2395    
2396    ----------------------------------------------------------------------
2397    Name: Matthias Blume
2398    Date: 2002/03/25 17:25:00 EST
2399    Tag: blume-20020325-divmod
2400    Description:
2401    
2402    I improved (hopefully without breaking them) the implementation of Int.div,
2403    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2404    advantage of the following observations:
2405    
2406      Let  q = x quot y      r = x rem y
2407           d = x div  y      m = x mod y
2408    
2409    where "quot" is the round-to-zero version of integer division that
2410    hardware usually provides.  Then we have:
2411    
2412         r = x - q * y        where neither the * nor the - will overflow
2413         d = if q >= 0 orelse x = q * y then q else q - 1
2414                              where neither the * nor the - will overflow
2415         m = if q >= 0 orelse r = 0 then r else r + y
2416                              where the + will not overflow
2417    
2418    This results in substantial simplification of the generated code.
2419    The following table shows the number of CFG nodes and edges generated
2420    for
2421            fun f (x, y) = x OPER y
2422            (* with OPER \in div, mod, quot, rem *)
2423    
2424    
2425        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2426        --------------------------------------------------------
2427         div |         24 |         39 |         12 |         16
2428         mod |         41 |         71 |         12 |         16
2429        quot |          8 |         10 |          8 |         10
2430         rem |         10 |         14 |          8 |         10
2431    
2432    
2433    ----------------------------------------------------------------------
2434    Name: Matthias Blume
2435    Date: 2002/03/25 22:06:00 EST
2436    Tag: blume-20020325-cprotobug
2437    Description:
2438    
2439    Fixed a bug in cproto (c prototype decoder).
2440    
2441    ----------------------------------------------------------------------
2442    Name: Matthias Blume
2443    Date: 2002/03/25 16:00:00 EST
2444    Tag: blume-20020325-raw-primops
2445    Description:
2446    
2447    I did some cleanup to Allen's new primop code and
2448    replaced yesterday's bootfiles with new ones.
2449    (But they are stored in the same place.)
2450    
2451    ----------------------------------------------------------------------
2452    Name: Matthias Blume
2453    Date: 2002/03/24 22:40:00 EST
2454    Tag: blume-20020324-bootfiles
2455    Description:
2456    
2457    Made the bootfiles that Allen asked for.
2458    
2459    ----------------------------------------------------------------------
2460    Name: Allen Leung
2461    Date: 2002/03/23 15:50:00 EST
2462    Tag: leunga-20020323-flint-cps-rcc-primops
2463    Description:
2464    
2465      1. Changes to FLINT primops:
2466    
2467        (* make a call to a C-function;
2468         * The primop carries C function prototype information and specifies
2469         * which of its (ML-) arguments are floating point. C prototype
2470         * information is for use by the backend, ML information is for
2471         * use by the CPS converter. *)
2472      | RAW_CCALL of { c_proto: CTypes.c_proto,
2473                       ml_args: ccall_type list,
2474                       ml_res_opt: ccall_type option,
2475                       reentrant : bool
2476                     } option
2477       (* Allocate uninitialized storage on the heap.
2478        * The record is meant to hold short-lived C objects, i.e., they
2479        * are not ML pointers.  With the tag, the representation is
2480        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2481        *)
2482      | RAW_RECORD of {tag:bool,sz:int}
2483      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2484    
2485      2.  These CPS primops are now overloaded:
2486    
2487           rawload of {kind:numkind}
2488           rawstore of {kind:numkind}
2489    
2490          The one argument form is:
2491    
2492             rawload {kind} address
2493    
2494          The two argument form is:
2495    
2496             rawload {kind} [ml object, byte-offset]
2497    
2498      3. RAW_CCALL/RCC now takes two extra arguments:
2499    
2500         a. The first is whether the C call is reentrant, i.e., whether
2501            ML state should be saved and restored.
2502         b. The second argument is a string argument specifying the name of
2503            library and the C function.
2504    
2505         These things are currently not handled in the code generator, yet.
2506    
2507      4. In CProto,
2508    
2509         An encoding type of "bool" means "ml object" and is mapped into
2510         C prototype of PTR.  Note that "bool" is different than "string",
2511         even though "string" is also mapped into PTR, because "bool"
2512         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2513    
2514      5. Pickler/unpicker
2515    
2516         Changed to handle RAW_RECORD and newest RAW_CCALL
2517    
2518      6. MLRiscGen,
2519    
2520         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2521         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2522            in the file CodeGen/cpscompile/cps-c-calls.sml
2523    
2524      7. Added the conditional move operator
2525    
2526             condmove of branch
2527    
2528         to cps.  Generation of this is still buggy so it is currently
2529         disabled.
2530    
2531    ----------------------------------------------------------------------
2532    Name: Lal George
2533    Date: 2002/03/22 14:18:25 EST
2534    Tag: george-20020322-cps-branch-prob
2535    Description:
2536    
2537    Implemented the Ball-Larus branch prediction-heuristics, and
2538    incorporated graphical viewers for control flow graphs.
2539    
2540    Ball-Larus Heuristics:
2541    ---------------------
2542    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2543    
2544    By design it uses the Dempster-Shafer theory for combining
2545    probabilities.  For example, in the function:
2546    
2547        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2548    
2549    the ball-larus heuristics predicts that the n=0 is unlikely
2550    (OH-heuristic), and the 'then' branch is unlikely because of the
2551    RH-heuristic -- giving the 'then' branch an even lower combined
2552    probability using the Dempster-Shafer theory.
2553    
2554    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2555    probability of the 'then' branch because of the loop in the else
2556    branch.
2557    
2558    
2559    Graphical Viewing:
2560    ------------------
2561    I merely plugged in Allen's graphical viewers into the compiler. The
2562    additional code is not much. At the top level, saying:
2563    
2564            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2565    
2566    will display the graphical view of the control flow graph just before
2567    back-patching.  daVinci must be in your path for this to work. If
2568    daVinci is not available, then the default viewer can be changed
2569    using:
2570    
2571            Control.MLRISC.getString "viewer"
2572    
2573    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2574    course, these viewers must be in your path.
2575    
2576    The above will display the compilation unit at the level of clusters,
2577    many of which are small, boring, and un-interesting. Also setting:
2578    
2579            Control.MLRISC.getInt "cfg-graphical-view_size"
2580    
2581    will display clusters that are larger than the value set by the above.
2582    
2583    
2584    ----------------------------------------------------------------------
2585    Name: Matthias Blume
2586    Date: 2002/03/21 22:20:00 EST
2587    Tag: blume-20020321-kmp-bugfix
2588    Description:
2589    
2590    Changed the interface to the KMP routine in PreString and fixed
2591    a minor bug in one place where it was used.
2592    
2593    ----------------------------------------------------------------------
2594    Name: Allen Leung
2595    Date: 2002/03/21 20:30:00 EST
2596    Tag: leunga-20020321-cfg
2597    Description:
2598    
2599      Fixed a potential problem in cfg edge splitting.
2600    
2601    ----------------------------------------------------------------------
2602    Name: Allen Leung
2603    Date: 2002/03/21 17:15:00 EST
2604    Tag: leunga-20020321-x86-fp-cfg
2605    Description:
2606    
2607      1. Recoded the buggy parts of x86-fp.
2608    
2609         a. All the block reordering code has been removed.
2610            We now depend on the block placement phases to do this work.
2611    
2612         b. Critical edge splitting code has been simplified and moved into the
2613            CFG modules, as where they belong.
2614    
2615         Both of these were quite buggy and complex.  The code is now much, much
2616         simpler.
2617    
2618      2. X86 backend.
2619    
2620         a. Added instructions for 64-bit support.  Instruction selection for
2621            64-bit has not been committed, however, since that
2622            requires changes to MLTREE which haven't been approved by
2623            Lal and John.
2624    
2625         b. Added support for FUCOMI and FUCOMIP when generating code for
2626            PentiumPro and above.  We only generate these instructions in
2627            the fast-fp mode.
2628    
2629         c. Added cases for JP and JNP in X86FreqProps.
2630    
2631      3. CFG
2632    
2633         CFG now has a bunch of methods for edge splitting and merging.
2634    
2635      4. Machine description.
2636    
2637         John's simplification of MLTREE_BASIS.fcond broke a few machine
2638         description things:
2639    
2640         rtl-build.{sig,sml} and hppa.mdl fixed.
2641    
2642         NOTE: the machine description stuff in the repository is still broken.
2643               Again, I can't put my fixes in because that involves
2644               changes to MLTREE.
2645    
2646    ----------------------------------------------------------------------
2647    Name: Matthias Blume
2648    Date: 2002/03/20 15:55:00 EST
2649    Tag: blume-20020320-kmp
2650    Description:
2651    
2652    Implemented Knuth-Morris-Pratt string matching in PreString and used
2653    it for String.isSubstring, Substring.isSubstring, and
2654    Substring.position.
2655    
2656    (Might need some stress-testing.  Simple examples worked fine.)
2657    
2658    ----------------------------------------------------------------------
2659    Name: Matthias Blume
2660    Date: 2002/03/19 16:37:00 EST
2661    Tag: blume-20020319-witnesses
2662    Description:
2663    
2664    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2665    
2666    This implements a generic mechanism for changing constness qualifiers
2667    anywhere within big C types without resorting to outright "casts".
2668    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2669    modify the constness at the outermost level.)
2670    The implementation of "convert" is based on the idea of "witness"
2671    values -- values that are not used by the operation but whose types
2672    "testify" to their applicability.  On the implementation side, "convert"
2673    is simply a projection (returning its second curried argument).  With
2674    cross-module inlining, it should not result in any machine code being
2675    generated.
2676    
2677    ----------------------------------------------------------------------
2678    Name: Matthias Blume
2679    Date: 2002/03/15 16:40:00 EST
2680    Tag: blume-20020315-basis
2681    Description:
2682    
2683    Provided (preliminary?) implementations for
2684    
2685      {String,Substring}.{concatWith,isSuffix,isSubstring}
2686    
2687    and
2688    
2689      Substring.full
2690    
2691    Those are in the Basis spec but they were missing in SML/NJ.
2692    
2693    ----------------------------------------------------------------------
2694    Name: Matthias Blume
2695    Date: 2002/03/14 21:30:00 EST
2696    Tag: blume-20020314-controls
2697    Description:
2698    
2699    Controls:
2700    ---------
2701    
2702    1. Factored out the recently-added Controls : CONTROLS stuff and put
2703       it into its own library $/controls-lib.cm.  The source tree for
2704       this is under src/smlnj-lib/Controls.
2705    
2706    2. Changed the names of types and functions in this interface, so they
2707       make a bit more "sense":
2708    
2709          module -> registry
2710          'a registry -> 'a group
2711    
2712    3. The interface now deals in ref cells only.  The getter/setter interface
2713       is (mostly) gone.
2714    
2715    4. Added a function that lets one register an already-existing ref cell.
2716    
2717    5. Made the corresponding modifications to the rest of the code so that
2718       everything compiles again.
2719    
2720    6. Changed the implementation of Controls.MLRISC back to something closer
2721       to the original.  In particular, this module (and therefore MLRISC)
2722       does not depend on Controls.  There now is some link-time code in
2723       int-sys.sml that registers the MLRISC controls with the Controls
2724       module.
2725    
2726    CM:
2727    ---
2728    
2729      * One can now specify the lambda-split aggressiveness in init.cmi.
2730    
2731    ----------------------------------------------------------------------
2732    Name: Allen Leung
2733    Date: 2002/03/13 17:30:00 EST
2734    Tag: leunga-20020313-x86-fp-unary
2735    Description:
2736    
2737    Bug fix for:
2738    
2739    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2740    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2741    > - fun f(x,(y,z)) = Real.~ y;
2742    > [autoloading]
2743    > [autoloading done]
2744    >       fchsl   (%eax), 184(%esp)
2745    > Error: MLRisc bug: X86MCEmitter.emitInstr
2746    >
2747    > uncaught exception Error
2748    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2749    
2750    The problem was that the code generator did not generate any fp registers
2751    in this case, and the ra didn't know that it needed to run the X86FP phase to
2752    translate the pseudo fp instruction.   This only happened with unary fp
2753    operators in certain situations.
2754    
2755    ----------------------------------------------------------------------
2756    Name: Matthias Blume
2757    Date: 2002/03/13 14:00:00 EST
2758    Tag: blume-20020313-overload-etc
2759    Description:
2760    
2761    1. Added _overload as a synonym for overload for backward compatibility.
2762       (Control.overloadKW must be true for either version to be accepted.)
2763    
2764    2. Fixed bug in install script that caused more things to be installed
2765       than what was requested in config/targets.
2766    
2767    3. Made CM aware of the (_)overload construct so that autoloading
2768       works.
2769    
2770    ----------------------------------------------------------------------
2771    Name: Matthias Blume
2772    Date: 2002/03/12 22:03:00 EST
2773    Tag: blume-20020312-url
2774    Description:
2775    
2776    Forgot to update BOOT and srcarchiveurl.
2777    
2778    ----------------------------------------------------------------------
2779    Name: Matthias Blume
2780    Date: 2002/03/12 17:30:00 EST
2781    Tag: blume-20020312-version110392
2782    Description:
2783    
2784    Yet another version number bump (because of small changes to the
2785    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2786    
2787    Changes:
2788    
2789      The new pid generation scheme described a few weeks ago was overly
2790      complicated.  I implemented a new mechanism that is simpler and
2791      provides a bit more "stability":  Once CM has seen a compilation
2792      unit, it keeps its identity constant (as long as you do not delete
2793      those crucial CM/GUID/* files).  This means that when you change
2794      an interface, compile, then go back to the old interface, and
2795      compile again, you arrive at the original pid.
2796    
2797      There now also is a mechanism that instructs CM to use the plain
2798      environment hash as a module's pid (effectively making its GUID
2799      the empty string).  For this, "noguid" must be specified as an
2800      option to the .sml file in question within its .cm file.
2801      This is most useful for code that is being generated by tools such
2802      as ml-nlffigen (because during development programmers tend to
2803      erase the tool's entire output directory tree including CM's cached
2804      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2805      revert to the old, broken behavior of SML/NJ, but in specific cases
2806      where there is no danger of interface confusion, its use is ok
2807      (I think).
2808    
2809      ml-nlffigen by default generates "noguid" annotations.  They can be
2810      turned off by specifying -guid in its command line.
2811    
2812    ----------------------------------------------------------------------
2813    Name: Lal George
2814    Date: 2002/03/12 12 14:42:36 EST
2815    Tag: george-20020312-frequency-computation
2816    Description:
2817    
2818    Integrated jump chaining and static block frequency into the
2819    compiler. More details and numbers later.
2820    
2821    ----------------------------------------------------------------------
2822    Name: Lal George
2823    Date: 2002/03/11 11 22:38:53 EST
2824    Tag: george-20020311-jump-chain-elim
2825    Description:
2826    
2827    Tested the jump chain elimination on all architectures (except the
2828    hppa).  This is on by default right now and is profitable for the
2829    alpha and x86, however, it may not be profitable for the sparc and ppc
2830    when compiling the compiler.
2831    
2832    The gc test will typically jump to a label at the end of the cluster,
2833    where there is another jump to an external cluster containing the actual
2834    code to invoke gc. This is to allow factoring of common gc invocation
2835    sequences. That is to say, we generate:
2836    
2837            f:
2838               testgc
2839               ja   L1      % jump if above to L1
2840    
2841            L1:
2842               jmp L2
2843    
2844    
2845    After jump chain elimination the 'ja L1' instructions is converted to
2846    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2847    up being implemented in their long form (if L2 is far away) using:
2848    
2849            jbe     L3      % jump if below or equal to L3
2850            jmp     L2
2851         L3:
2852            ...
2853    
2854    
2855    For large compilation units L2  may be far away.
2856    
2857    
2858    ----------------------------------------------------------------------
2859    Name: Matthias Blume
2860    Date: 2002/03/11 13:30:00 EST
2861    Tag: blume-20020311-mltreeeval
2862    Description:
2863    
2864    A functor parameter was missing.
2865    
2866    ----------------------------------------------------------------------
2867    Name: Allen Leung
2868    Date: 2002/03/11 10:30:00 EST
2869    Tag: leunga-20020311-runtime-string0
2870    Description:
2871    
2872       The representation of the empty string now points to a
2873    legal null terminated C string instead of unit.  It is now possible
2874    to convert an ML string into C string with InlineT.CharVector.getData.
2875    This compiles into one single machine instruction.
2876    
2877    ----------------------------------------------------------------------
2878    Name: Allen Leung
2879    Date: 2002/03/10 23:55:00 EST
2880    Tag: leunga-20020310-x86-call
2881    Description:
2882    
2883       Added machine generation for CALL instruction (relative displacement mode)
2884    
2885    ----------------------------------------------------------------------
2886    Name: Matthias Blume
2887    Date: 2002/03/08 16:05:00
2888    Tag: blume-20020308-entrypoints
2889    Description:
2890    
2891    Version number bumped to 110.39.1.  NEW BOOTFILES!
2892    
2893    Entrypoints: non-zero offset into a code object where execution should begin.
2894    
2895    - Added the notion of an entrypoint to CodeObj.
2896    - Added reading/writing of entrypoint info to Binfile.
2897    - Made runtime system bootloader aware of entrypoints.
2898    - Use the address of the label of the first function given to mlriscGen
2899      as the entrypoint.  This address is currently always 0, but it will
2900      not be 0 once we turn on block placement.
2901    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2902      with entry points) from mlriscGen.
2903    
2904    ----------------------------------------------------------------------
2905    Name: Allen Leung
2906    Date: 2002/03/07 20:45:00 EST
2907    Tag: leunga-20020307-x86-cmov
2908    Description:
2909    
2910       Bug fixes for CMOVcc on x86.
2911    
2912       1. Added machine code generation for CMOVcc
2913       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2914       3. CMOVcc cannot have an immediate operand as argument.
2915    
2916    ----------------------------------------------------------------------
2917    Name: Matthias Blume
2918    Date: 2002/03/07 16:15:00 EST
2919    Tag: blume-20020307-controls
2920    Description:
2921    
2922    This is a very large but mostly boring patch which makes (almost)
2923    every tuneable compiler knob (i.e., pretty much everything under
2924    Control.* plus a few other things) configurable via both the command
2925    line and environment variables in the style CM did its configuration
2926    until now.
2927    
2928    Try starting sml with '-h' (or, if you are brave, '-H')
2929    
2930    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2931    implements the underlying generic mechanism.
2932    
2933    The interface to some of the existing such facilities has changed somewhat.
2934    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2935    (The getFoo interface is still there for backward-compatibility, but its
2936    use is deprecated.)
2937    
2938    The ml-build script passes -Cxxx=yyy command-line arguments through so
2939    that one can now twiddle the compiler settings when using this "batch"
2940    compiler.
2941    
2942    TODO items:
2943    
2944    We should go through and throw out all controls that are no longer
2945    connected to anything.  Moreover, we should go through and provide
2946    meaningful (and correct!) documentation strings for those controls
2947    that still are connected.
2948    
2949    Currently, multiple calls to Controls.new are accepted (only the first
2950    has any effect).  Eventually we should make sure that every control
2951    is being made (via Controls.new) exactly once.  Future access can then
2952    be done using Controls.acc.
2953    
2954    Finally, it would probably be a good idea to use the getter-setter
2955    interface to controls rather than ref cells.  For the time being, both
2956    styles are provided by the Controls module, but getter-setter pairs are
2957    better if thread-safety is of any concern because they can be wrapped.
2958    
2959    *****************************************
2960    
2961    One bug fix: The function blockPlacement in three of the MLRISC
2962    backpatch files used to be hard-wired to one of two possibilities at
2963    link time (according to the value of the placementFlag).  But (I
2964    think) it should rather sense the flag every time.
2965    
2966    *****************************************
2967    
2968    Other assorted changes (by other people who did not supply a HISTORY entry):
2969    
2970    1. the cross-module inliner now works much better (Monnier)
2971    2. representation of weights, frequencies, and probabilities in MLRISC
2972       changed in preparation of using those for weighted block placement
2973       (Reppy, George)
2974    
2975    ----------------------------------------------------------------------
2976    Name: Lal George
2977    Date: 2002/03/07 14:44:24 EST 2002
2978    Tag: george-20020307-weighted-block-placement
2979    
2980    Tested the weighted block placement optimization on all architectures
2981    (except the hppa) using AMPL to generate the block and edge frequencies.
2982    Changes were required in the machine properties to correctly
2983    categorize trap instructions. There is an MLRISC flag
2984    "weighted-block-placement" that can be used to enable weighted block
2985    placement, but this will be ineffective without block/edge
2986    frequencies (coming soon).
2987    
2988    
2989    ----------------------------------------------------------------------
2990    Name: Lal George
2991    Date: 2002/03/05 17:24:48 EST
2992    Tag: george-20020305-linkage-cluster
2993    
2994    In order to support the block placement optimization, a new cluster
2995    is generated as the very first cluster (called the linkage cluster).
2996    It contains a single jump to the 'real' entry point for the compilation
2997    unit. Block placement has no effect on the linkage cluster itself, but
2998    all the other clusters  have full freedom in the manner in which they
2999    reorder blocks or functions.
3000    
3001    On the x86 the typical linkage code that is generated is:
3002       ----------------------
3003            .align 2
3004       L0:
3005            addl    $L1-L0, 72(%esp)
3006            jmp     L1
3007    
3008    
3009            .align  2
3010       L1:
3011       ----------------------
3012    
3013    72(%esp) is the memory location for the stdlink register. This
3014    must contain the address of the CPS function being called. In the
3015    above example, it contains the address of  L0; before
3016    calling L1 (the real entry point for the compilation unit), it
3017    must contain the address for L1, and hence
3018    
3019            addl $L1-L0, 72(%esp)
3020    
3021    I have tested this on all architectures except the hppa.The increase
3022    in code size is of course negligible
3023    
3024    ----------------------------------------------------------------------
3025    Name: Allen Leung
3026    Date: 2002/03/03 13:20:00 EST
3027    Tag: leunga-20020303-mlrisc-tools
3028    
3029      Added #[ ... ] expressions to mlrisc tools
3030    
3031    ----------------------------------------------------------------------
3032    Name: Matthias Blume
3033    Date: 2002/02/27 12:29:00 EST
3034    Tag: blume-20020227-cdebug
3035    Description:
3036    
3037    - made types in structure C and C_Debug to be equal
3038    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
3039    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
3040    
3041    ----------------------------------------------------------------------
3042    Name: Matthias Blume
3043    Date: 2002/02/26 12:00:00 EST
3044    Tag: blume-20020226-ffi
3045    Description:
3046    
3047    1. Fixed a minor bug in CM's "noweb" tool:
3048       If numbering is turned off, then truly don't number (i.e., do not
3049       supply the -L option to noweb).  The previous behavior was to supply
3050       -L'' -- which caused noweb to use the "default" line numbering scheme.
3051       Thanks to Chris Richards for pointing this out (and supplying the fix).
3052    
3053    2. Once again, I reworked some aspects of the FFI:
3054    
3055       A. The incomplete/complete type business:
3056    
3057       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
3058         gone!
3059       - ML types representing an incomplete type are now *equal* to
3060         ML types representing their corresponding complete types (just like
3061         in C).  This is still safe because ml-nlffigen will not generate
3062         RTTI for incomplete types, nor will it generate functions that
3063         require access to such RTTI.   But when ML code generated from both
3064         incomplete and complete versions of the C type meet, the ML types
3065         are trivially interoperable.
3066    
3067         NOTE:  These changes restore the full generality of the translation
3068         (which was previously lost when I eliminated functorization)!
3069    
3070       B. Enum types:
3071    
3072       - Structure C now has a type constructor "enum" that is similar to
3073         how the "su" constructor works.  However, "enum" is not a phantom
3074         type because each "T enum" has values (and is isomorphic to
3075         MLRep.Signed.int).
3076       - There are generic access operations for enum objects (using
3077         MLRep.Signed.int).
3078       - ml-nlffigen will generate a structure E_foo for each "enum foo".
3079         * The structure contains the definition of type "mlrep" (the ML-side
3080         representation type of the enum).  Normally, mlrep is the same
3081         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3082         then mlrep will be defined as a datatype -- thus facilitating
3083         pattern matching on mlrep values.
3084         ("-ec" will be suppressed if there are duplicate values in an
3085          enumeration.)
3086         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3087         will be generated for each C enum constant xxx.
3088         * Conversion functions m2i and i2m convert between mlrep and
3089         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
3090         * Coversion functions c and ml convert between mlrep and "tag enum".
3091         * Access functions (get/set) fetch and store mlrep values.
3092       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3093         enumerations are merged into one single enumeration represented by
3094         structure E_'.
3095    
3096    ----------------------------------------------------------------------
3097    Name: Allen Leung
3098    Date: 2002/02/25 04:45:00 EST
3099    Tag: leunga-20020225-cps-spill
3100    
3101    This is a new implementation of the CPS spill phase.
3102    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3103    In case of problems, replace it with the old file spill.sml
3104    
3105    The current compiler runs into some serious performance problems when
3106    constructing a large record.  This can happen when we try to compile a
3107    structure with many items.  Even a very simple structure like the following
3108    makes the compiler slow down.
3109    
3110        structure Foo = struct
3111           val x_1 = 0w1 : Word32.int
3112           val x_2 = 0w2 : Word32.int
3113           val x_3 = 0w3 : Word32.int
3114           ...
3115           val x_N = 0wN : Word32.int
3116        end
3117    
3118    The following table shows the compile time, from N=1000 to N=4000,
3119    with the old compiler:
3120    
3121    N
3122    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3123           MLRISC ra                               0.06u  0.00s  0.05g
3124              (spills = 0 reloads = 0)
3125           TOTAL                                   0.63u  0.07s  0.21g
3126    
3127    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3128           MLRISC ra                               5.68u  0.59s  3.93g
3129              (spills = 0 reloads = 0)
3130           TOTAL                                   14.71u  0.99s  4.81g
3131    
3132    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3133           MLRISC ra                               5.54u  0.65s  3.91g
3134              (spills = 543 reloads = 1082)
3135           TOTAL                                   65.40u  3.13s  6.00g
3136    
3137    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3138           MLRISC ra                               0.80u  0.10s  0.55g
3139              (spills = 42 reloads = 84)
3140           TOTAL                                   129.42u  5.10s  4.13g
3141    
3142    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3143           MLRISC ra                               2.69u  0.27s  1.38g
3144              (spills = 62 reloads = 124)
3145           TOTAL                                   682.48u  19.61s  13.99g
3146    
3147    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3148           MLRISC ra                               4.96u  0.27s  2.72g
3149              (spills = 85 reloads = 170)
3150           TOTAL                                   2375.26u  57.21s  48.00g
3151    
3152    As you can see the old cps spill module suffers from some serious
3153    performance problem.  But since I cannot decipher the old code fully,
3154    instead of patching the problems up, I'm reimplementing it
3155    with a different algorithm.  The new code is more modular,
3156    smaller when compiled, and substantially faster
3157    (O(n log n) time and O(n) space).  Timing of the new spill module:
3158    
3159    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3160          MLRISC ra                               0.25u  0.02s  0.15g
3161             (spills=1 reloads=3)
3162          TOTAL                                   7.74u  0.34s  1.62g
3163    
3164    Implementation details:
3165    
3166    As far as I can tell, the purpose of the CPS spill module is to make sure the
3167    number of live variables at any program point (the bandwidth)
3168    does not exceed a certain limit, which is determined by the
3169    size of the spill area.
3170    
3171    When the bandwidth is too large, we decrease the register pressure by
3172    packing live variables into spill records.  How we achieve this is
3173    completely different than what we did in the old code.
3174    
3175    First, there is something about the MLRiscGen code generator
3176    that we should be aware of:
3177    
3178    o MLRiscGen performs code motion!
3179    
3180       In particular, it will move floating point computations and
3181       address computations involving only the heap pointer to
3182       their use sites (if there is only a single use).
3183       What this means is that if we have a CPS record construction
3184       statement
3185    
3186           RECORD(k,vl,w,e)
3187    
3188       we should never count the new record address w as live if w
3189       has only one use (which is often the case).
3190    
3191       We should do something similar to floating point, but the transformation
3192       there is much more complex, so I won't deal with that.
3193    
3194    Secondly, there are now two new cps primops at our disposal:
3195    
3196     1. rawrecord of record_kind option
3197        This pure operator allocates some uninitialized storage from the heap.
3198        There are two forms:
3199    
3200         rawrecord NONE [INT n]  allocates a tagless record of length n
3201         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3202                                     and initializes the tag.
3203    
3204     2. rawupdate of cty
3205          rawupdate cty (v,i,x)
3206          Assigns to x to the ith component of record v.
3207          The storelist is not updated.
3208    
3209    We use these new primops for both spilling and increment record construction.
3210    
3211     1. Spilling.
3212    
3213        This is implemented with a linear scan algorithm (but generalized
3214        to trees).  The algorithm will create a single spill record at the
3215        beginning of the cps function and use rawupdate to spill to it,
3216        and SELECT or SELp to reload from it.  So both spills and reloads
3217        are fine-grain operations.  In contrast, in the old algorithm
3218        "spills" have to be bundled together in records.
3219    
3220        Ideally, we should sink the spill record construction to where
3221        it is needed.  We can even split the spill record into multiple ones
3222        at the places where they are needed.  But CPS is not a good
3223        representation for global code motion, so I'll keep it simple and
3224        am not attempting this.
3225    
3226     2. Incremental record construction (aka record splitting).
3227    
3228        Long records with many component values which are simulatenously live
3229        (recall that single use record addresses are not considered to
3230         be live) are constructed with rawrecord and rawupdate.
3231        We allocate space on the heap with rawrecord first, then gradually
3232        fill it in with rawupdate.  This is the technique suggested to me
3233        by Matthias.
3234    
3235        Some restrictions on when this is applicable:
3236        1. It is not a VECTOR record.  The code generator currently does not handle
3237           this case. VECTOR record uses double indirection like arrays.
3238        2. All the record component values are defined in the same "basic block"
3239           as the record constructor.  This is to prevent speculative
3240           record construction.
3241    
3242    ----------------------------------------------------------------------
3243    Name: Allen Leung
3244    Date: 2002/02/22 01:02:00 EST
3245    Tag: leunga-20020222-mlrisc-tools
3246    
3247    Minor bug fixes in the parser and rewriter
3248    
3249    ----------------------------------------------------------------------
3250    Name: Allen Leung
3251    Date: 2002/02/21 20:20:00 EST
3252    Tag: leunga-20020221-peephole
3253    
3254    Regenerated the peephole files.  Some contained typos in the specification
3255    and some didn't compile because of pretty printing bugs in the old version
3256    of 'nowhere'.
3257    
3258    ----------------------------------------------------------------------
3259    Name: Allen Leung
3260    Date: 2002/02/19 20:20:00 EST
3261    Tag: leunga-20020219-mlrisc-tools
3262    Description:
3263    
3264       Minor bug fixes to the mlrisc-tools library:
3265    
3266       1.  Fixed up parsing colon suffixed keywords
3267       2.  Added the ability to shut the error messages up
3268       3.  Reimplemented the pretty printer and fixed up/improved
3269           the pretty printing of handle and -> types.
3270       4.  Fixed up generation of literal symbols in the nowhere tool.
3271       5.  Added some SML keywords to to sml.sty
3272    
3273    ----------------------------------------------------------------------
3274    Name: Matthias Blume
3275    Date: 2002/02/19 16:20:00 EST
3276    Tag: blume-20020219-cmffi
3277    Description:
3278    
3279    A wild mix of changes, some minor, some major:
3280    
3281    * All C FFI-related libraries are now anchored under $c:
3282        $/c.cm      --> $c/c.cm
3283        $/c-int.cm  --> $c/internals/c-int.cm
3284        $/memory.cm --> $c/memory/memory.cm
3285    
3286    * "make" tool (in CM) now treats its argument pathname slightly
3287      differently:
3288        1. If the native expansion is an absolute name, then before invoking
3289           the "make" command on it, CM will apply OS.Path.mkRelative
3290           (with relativeTo = OS.FileSys.getDir()) to it.
3291        2. The argument will be passed through to subsequent phases of CM
3292           processing without "going native".  In particular, if the argument
3293           was an anchored path, then "make" will not lose track of that anchor.
3294    
3295    * Compiler backends now "know" their respective C calling conventions
3296      instead of having to be told about it by ml-nlffigen.  This relieves
3297      ml-nlffigen from one of its burdens.
3298    
3299    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3300    
3301    * Export C_DEBUG and C_Debug from $c/c.cm.
3302    
3303    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3304      subtyping relationship between incomplete pointers and their complete
3305      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3306      with the convention of instantiating 'o with ('t, 'c) obj whenever
3307      the pointer target type is complete.  In the incomplete case, 'o
3308      will be instantiated with some "'c iobj" -- a type obtained by
3309      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3310    
3311      Operations that work on both incomplete and complete pointer types are
3312      typed as taking an 'o ptr while operations that require the target to
3313      be known are typed as taking some ('t, 'c) obj ptr.
3314    
3315      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3316      where void is an eqtype without any values.  This makes it possible
3317      to work on voidptr values using functions meant to operate on light
3318      incomplete pointers.
3319    
3320    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3321      been vastly simplified.
3322    
3323    ----------------------------------------------------------------------
3324    Name: Matthias Blume
3325    Date: 2002/02/19 10:48:00 EST
3326    Tag: blume-20020219-pqfix
3327    Description:
3328    
3329    Applied Chris Okasaki's bug fix for priority queues.
3330    
3331    ----------------------------------------------------------------------
3332    Name: Matthias Blume
3333    Date: 2002/02/15 17:05:00
3334    Tag: Release_110_39
3335    Description:
3336    
3337    Last-minute retagging is becoming a tradition... :-(
3338    
3339    This is the working release 110.39.
3340    
3341    ----------------------------------------------------------------------
3342    Name: Matthias Blume
3343    Date: 2002/02/15 16:00:00 EST
3344    Tag: Release_110_39-orig
3345    Description:
3346    
3347    Working release 110.39.  New bootfiles.
3348    
3349    (Update: There was a small bug in the installer so it wouldn't work
3350    with all shells.  So I retagged. -Matthias)
3351    
3352    ----------------------------------------------------------------------
3353    Name: Matthias Blume
3354    Date: 2002/02/15 14:17:00 EST
3355    Tag: blume-20020215-showbindings
3356    Description:
3357    
3358    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3359    the latter can be useful for exploring what bindings are available at
3360    the interactive prompt.  (The first function returns only the list
3361    of symbols that are really bound, the second prints those but also the
3362    ones that CM's autoloading mechanism knows about.)
3363    
3364    ----------------------------------------------------------------------
3365    Name: Matthias Blume
3366    Date: 2002/02/15 12:08:00 EST
3367    Tag: blume-20020215-iptrs
3368    Description:
3369    
3370    Two improvements to ml-nlffigen:
3371    
3372      1. Write files only if they do not exist or if their current contents
3373         do not coincide with what's being written.  (That is, avoid messing
3374         with the time stamps unless absolutely necessary.)
3375    
3376      2. Implement a "repository" mechanism for generated files related
3377         to "incomplete pointer types".   See the README file for details.
3378    
3379    ----------------------------------------------------------------------
3380    Name: Matthias Blume
3381    Date: 2002/02/14 11:50:00 EST
3382    Tag: blume-20020214-quote
3383    Description:
3384    
3385    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3386    because of the new and improved tag generation scheme.  (Thanks to Allen
3387    Leung for pointing it out.)
3388    
3389    ----------------------------------------------------------------------
3390    Name: Lal George
3391    Date: 2002/02/14 09:55:27 EST 2002
3392    Tag: george-20020214-isabelle-bug
3393    Description:
3394    
3395    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3396    of Isabelle on the x86.
3397    
3398    From Allen:
3399    -----------
3400     I've found the problem:
3401    
3402         in ra-core.sml, I use the counter "blocked" to keep track of the
3403         true number of elements in the freeze queue.  When the counter goes
3404         to zero, I skip examining the queue.  But I've messed up the
3405         bookkeeping in combine():
3406    
3407             else ();
3408             case !ucol of
3409               PSEUDO => (if !cntv > 0 then
3410                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3411                                        ^^^^^^^^^^^^^^^^^^^^^^^
3412                      moveu := mergeMoveList(!movev, !moveu)
3413                     )
3414                  else ();
3415    
3416         combine() is called to coalesce two nodes u and v.
3417         I think I was thinking that if the move counts of u and v are both
3418         greater than zero then after they are coalesced then one node is
3419         removed from the freeze queue.  Apparently I was thinking that
3420         both u and v are of low degree, but that's clearly not necessarily true.
3421    
3422    
3423    02/12/2002:
3424        Here's the patch.  HOL now compiles.
3425    
3426        I don't know how this impact on performance (compile
3427        time or runtime).  This bug caused the RA (especially on the x86)
3428        to go thru the potential spill phase when there are still nodes on the
3429        freeze queue.
3430    
3431    
3432    
3433    
3434    ----------------------------------------------------------------------
3435    Name: Matthias Blume
3436    Date: 2002/02/13 22:40:00 EST
3437    Tag: blume-20020213-fptr-rtti
3438    Description:
3439    
3440    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3441    updates.
3442    
3443    ----------------------------------------------------------------------
3444    Name: Matthias Blume
3445    Date: 2002/02/13 16:41:00 EST
3446    Tag: blume-20020213-cmlpq
3447    Description:
3448    
3449    Added new priority queue export symbols (which have just been added to
3450    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3451    not compile and the installer would choke.)
3452    
3453    ----------------------------------------------------------------------
3454    Name: Matthias Blume
3455    Date: 2002/02/13 16:15:00 EST
3456    Tag: blume-20020213-various
3457    Description:
3458    
3459    1. More tweaks to ml-nlffigen:
3460    
3461       - better internal datastructures (resulting in slight speedup)
3462       - "-match" option requires exact match
3463       - "localized" gensym counters (untagged structs/unions nested within
3464         other structs/unions or within typedefs get a fresh counter; their
3465         tag will be prefixed by a concatenation of their parents' tags)
3466       - bug fixes (related to calculation of transitive closure of types
3467         to be included in the output)
3468    
3469    2. Minor Basis updates:
3470    
3471       - added implementations for List.collate and Option.app
3472    
3473    ----------------------------------------------------------------------
3474    Name: Matthias Blume
3475    Date: 2002/02/11 15:55:00 EST
3476    Tag: blume-20020211-gensym
3477    Description:
3478    
3479    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3480    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3481    names (ML structure names that correspond to unnamed C structs, unions,
3482    and enums), so that separate runs of ml-nlffigen do not clash.
3483    
3484    ----------------------------------------------------------------------
3485    Name: Matthias Blume
3486    Date: 2002/02/11 12:05:00 EST
3487    Tag: blume-20020211-gensml
3488    Description:
3489    
3490    A quick fix for a problem with GenSML (in the pgraph-util library):
3491    Make generation of toplevel "local" optional.  (Strictly speaking,
3492    signature definitions within "local" are not legal SML.)
3493    
3494    Other than that: updates to INSTALL and cm/TODO.
3495    
3496    ----------------------------------------------------------------------
3497    Name: Matthias Blume
3498    Date: 2002/02/08 15:00:00 EST
3499    Tag: blume-20020208-uniquepid
3500    Description:
3501    
3502    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3503    
3504    1. The installer (config/install.sh) has gotten smarter:
3505    
3506         - Configuration options are a bit easier to specify now
3507           (in config/targets).
3508         - Bug in recognizing .tar.bz2 files fixed.
3509         - Installer automatically resolves dependencies between
3510           configuration options (e.g., if you ask for eXene, you will
3511           also get cml -- regardless whether you asked for it or not).
3512         - Installer can run in "quieter mode" by setting the environment
3513           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3514           "completely silent", though.
3515         - Build HashCons library as part of smlnj-lib.
3516    
3517    2. A new scheme for assigning persistent identifiers to compilation
3518       units (and, by extension, to types etc.) has been put into place.
3519       This fixes a long-standing bug where types and even dynamic values
3520       can get internally confused, thereby compromising type safety
3521       (abstraction) and dynamic correctness.  See
3522    
3523         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3524    
3525       for an example of how things could go wrong until now.
3526    
3527       The downside of the new scheme is that pids are not quite as
3528       persistent as they used to be: CM will generate a fresh pid
3529       for every compilation unit that it thinks it sees for the first
3530       time.  That means that if you compile starting from a clean, fresh
3531       source tree at two different times, you end up with different
3532       binaries.
3533    
3534       Cutoff recompilation, however, has not been compromised because
3535       CM keeps pid information in special caches between runs.
3536    
3537    ----------------------------------------------------------------------
3538    Name: Lal George
3539    Date: 2002/02/07 15:34:13 EST 2002
3540    Tag: <none>
3541    Description:
3542    
3543    Compilers that generate assembly code may produce  global labels
3544    whose value is resolved at link time. The various peephole optimization
3545    modules did not take this in account.
3546    
3547    TODO. The Labels.addrOf function should really return an option
3548    type so that clients are forced to deal with this issue, rather
3549    than an exception being raised.
3550    
3551    ----------------------------------------------------------------------
3552    Name: Lal George
3553    Date: 2002/02/06 13:55:02 EST
3554    Tag: george-20020206-ra-breakup
3555    Description:
3556    
3557    1. A bug fix from Allen.
3558    
3559        A typo causes extra fstp %st(0)'s to be generated at compensation
3560        edges, which might cause stack underflow traps at runtime.  This
3561        occurs in fft where there are extraneous fstps right before the 'into'
3562        trap instruction (in this case they are harmless since none of the
3563        integers overflow.)
3564    
3565    2. Pulled out various utility modules that were embedded in the modules
3566       of the register allocator. I need these modules for other purposes, but
3567       they are not complete enough to put into a library (just yet).
3568    ----------------------------------------------------------------------
3569    Name: Matthias Blume
3570    Date: 2002/01/31 16:05:00 EST
3571    Tag: blume-20020131-sparc-ccalls
3572    Description:
3573    
3574    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3575       of extra stack space by mistake.  Fixed.
3576    
3577    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3578    
3579    ----------------------------------------------------------------------
3580    Name: Allen Leung
3581    Date: 2002/01/30
3582    Tag: leunga-20020130-nowhere-bug-fix
3583    Description:
3584    
3585       MLRISC bug fixes:
3586       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3587       2. MachineInt.fromString was negating its value.
3588    
3589    ----------------------------------------------------------------------
3590    Name: Matthias Blume
3591    Date: 2002/01/29
3592    Tag: blume-20020129-INSTALL
3593    Description:
3594    
3595    - Added somewhat detailed installation instructions (file INSTALL).
3596    - Fixed curl-detection bug in config/install.sh.
3597    - It is now possible to select the URL getter using the URLGETTER
3598      environment variable:
3599    
3600          not set / "unknown"      --> automatic detection (script tries wget,
3601                                       curl, and lynx)
3602          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3603                                       how to properly invoke them)
3604          other                    --> use $URLGETTER directly, it must take
3605                                       precisely two command-line arguments
3606                                       (source URL and destination file name)
3607    
3608    ----------------------------------------------------------------------
3609    Name: Matthias Blume
3610    Date: 2002/01/28
3611    Tag: blume-20020128-sparc-ccalls
3612    Description:
3613    
3614    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3615    - Make use of the allocParam argument in sparc-c-calls.
3616    
3617    ----------------------------------------------------------------------
3618    Name: Matthias Blume
3619    Date: 2002/01/28
3620    Tag: blume-20020128-allocParam
3621    Description:
3622    
3623    John Reppy:  Changes c-calls API to accept client-callback for
3624    allocating extra stack space.
3625    me: Corresponding changes to mlriscGen (using a dummy argument that
3626        does not change the current behavior).
3627    
3628    ----------------------------------------------------------------------
3629    Name: Matthias Blume
3630    Date: 2002/01/28 12:00:00
3631    Tag: Release_110_38
3632    Description:
3633    
3634    This time for real!!!
3635    
3636    ----------------------------------------------------------------------
3637    Name: Matthias Blume
3638    Date: 2002/01/28 10:56:00 EST
3639    Tag: blume-20020128-retraction
3640    Description:
3641    
3642    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3643       with blume-Release_110_38-retracted.)
3644    
3645    1. Fixed a problem with incorrect rounding modes in real64.sml.
3646       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3647    
3648    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3649       and enums fixed.  The naming of corresponding ML identifiers should
3650       now be consistent again.
3651    
3652    ----------------------------------------------------------------------
3653    Name: Allen Leung
3654    Date: 2002/01/27
3655    Tag: leunga-20020127-nowhere
3656    Description:
3657    
3658       Added a target called nowhere in the configuration scripts.
3659       Enabling this will build the MLRISC 'nowhere' tool (for translating
3660       programs with where-clauses into legal SML code) during installation.
3661    
3662    ----------------------------------------------------------------------
3663    Name: Matthias Blume
3664    Date: 2002/01/25 21:27:00 EST
3665    Tag: blume-Release_110_38-retracted
3666    Description:
3667    
3668    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3669    
3670    README will be added later.
3671    
3672    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3673    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3674    
3675    ----------------------------------------------------------------------
3676    Name: Matthias Blume
3677    Date: 2002/01/25
3678    Tag: blume-20020125-ffi
3679    Description:
3680    
3681    A large number of tweaks and improvements to ml-nlffi-lib and
3682    ml-nlffigen:
3683    
3684       - ML represenation types have been streamlined
3685       - getter and setter functions work with concrete values, not abstract
3686         ones where possible
3687       - ml-nlffigen command line more flexible (see README file there)
3688       - some bugs have been fixed (hopefully)
3689    
3690    ----------------------------------------------------------------------
3691    Name: Lal George
3692    Date: 2002/01/24
3693    Tag: george-20020124-risc-ra-interface
3694    Description:
3695    
3696       There is a dramatic simplification in the interface to the
3697       register allocator for RISC architectures as a result of making
3698       parallel copy instructions explicit.
3699    
3700    ----------------------------------------------------------------------
3701    Name: Matthias Blume
3702    Date: 2002/01/22
3703    Tag: blume-20020122-x86-ccalls
3704    Description:
3705    
3706    Bug fix for c-calls on x86 (having to do with how char- and
3707    short-arguments are being handled).
3708    
3709    ----------------------------------------------------------------------
3710    Name: Matthias Blume
3711    Date: 2002/01/21
3712    Tag: blume-20020121-ff
3713    Description:
3714    
3715    Another day of fiddling with the FFI...
3716    
3717    1. Bug fix/workaround:  CKIT does not complain about negative array
3718       dimensions, so ml-nlffigen has to guard itself against this possibility.
3719       (Otherwise a negative dimension would send it into an infinite loop.)
3720    
3721    2. Some of the abstract types (light objects, light pointers, most "base"
3722       types) in structure C are now eqtypes.
3723    
3724    3. Added constructors and test functions for NULL function pointers.
3725    
3726    ----------------------------------------------------------------------
3727    Name: Matthias Blume
3728    Date: 2002/01/18
3729    Tag: blume-20020118-ready-for-new-release
3730    Description:
3731    
3732    Made config/srcarchiveurl point to a new place.  (Will provide boot
3733    files shortly.)
3734    
3735    Maybe we christen this to be 110.38?
3736    
3737    ----------------------------------------------------------------------
3738    Name: Matthias Blume
3739    Date: 2002/01/18
3740    Tag: blume-20020118-more-ffifiddle
3741    Description:
3742    
3743    Today's FFI fiddling:
3744    
3745      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3746        C.Set.* that use concrete (MLRep.*) arguments and results instead
3747        of abstract ones.
3748    
3749      - Provided word-style bit operations etc. for "int" representation
3750        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3751        and Long.
3752    
3753    ----------------------------------------------------------------------
3754    Name: Matthias Blume
3755    Date: 2002/01/18
3756    Tag: blume-20020118-use-x86-fp
3757    Description:
3758    
3759    Now that x86-fast-fp seems to be working, I turned it back on again
3760    by default.  (Seems to work fine now, even with the FFI.)
3761    
3762    Other than that, I added some documentation about the FFI to
3763    src/ml-nlffigen/README and updated the FFI test examples in
3764    src/ml-nlffi-lib/Tests/*.
3765    
3766    ----------------------------------------------------------------------
3767    Name: Allen Leung
3768    Date: 2002/01/17
3769    Tag: leunga-20020117-x86-fast-fp-call
3770    Description:
3771    
3772       1. Fixed a problem with handling return fp values when x86's fast fp
3773          mode is turned on.
3774    
3775       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3776          of %f32.
3777    
3778       3. Added a constructor INT32lit to the ast of MLRISC tools.
3779    
3780    ----------------------------------------------------------------------
3781    Name: Matthias Blume
3782    Date: 2002/01/16
3783    Tag: blume-20020116-ffifiddle
3784    Description:
3785    
3786    More fiddling with the FFI interface:
3787    
3788     - Make constness 'c instead of rw wherever possible.  This eliminates
3789       the need for certain explicit coercions.  (However, due to ML's
3790       value polymorphism, there will still be many cases where explicit
3791       coercions are necessary.  Phantom types are not the whole answer
3792       to modeling a subtyping relationship in ML.)
3793    
3794     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3795    
3796     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3797       clumsy workaround for varargs problem.)
3798    
3799    ----------------------------------------------------------------------
3800    Name: Lal George
3801    Date: 2002/01/15
3802    Tag: <none>
3803    Description:
3804    
3805    1. Since COPY instructions are no longer native to the architecture,
3806       a generic functor can be used to implement the expandCopies function.
3807    
3808    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3809       TEXT segment.
3810    
3811    ----------------------------------------------------------------------
3812    Name: Matthias Blume
3813    Date: 2002/01/15
3814    Tag: blume-20020115-ffiupdates
3815    Description:
3816    
3817    1. Fix for bug resulting in single-precision float values being returned
3818       incorrectly from FFI calls.
3819    
3820    2. Small modifications to C FFI API:
3821    
3822        - memory-allocation routines return straight objects (no options)
3823          and raise an exception in out-of-memory situations
3824        - unsafe extensions to cast between function pointers and pointers
3825          from/to ints
3826        - added structure C_Debug as an alternative to structure C where
3827          pointer-dereferencing (|*| and |*!) always check for null-pointers
3828        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3829          but also takes a (possibly empty) list of existing library handles
3830          that the current library depends on
3831    
3832    ----------------------------------------------------------------------
3833    Name: Matthias Blume
3834    Date: 2002/01/10
3835    Tag: blume-20020110-newffigen
3836    Description:
3837    
3838    1. Updates to portable graph code.
3839    
3840    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3841       more scalable now so that even huge interfaces such as the one
3842       for GTK compile in finite time and space. :-)
3843       See src/ml-nlffigen/README for details on what's new.
3844    
3845    ----------------------------------------------------------------------
3846    Name: Lal George
3847    Date: 2001/01/09 14:31:35 EST 2002
3848    Tag: george-20011206-rm-native-copy
3849    Description:
3850    
3851            Removed the native COPY and FCOPY instructions
3852            from all the architectures and replaced it with the
3853            explicit COPY instruction from the previous commit.