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 1032, Tue Jan 22 16:59:16 2002 UTC revision 1832, Wed Jul 27 18:49:19 2005 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: John Reppy
17    Date: 2005/07/27
18    Tag:
19    Description:
20    
21    Fixed ml-lex to recognize "\r" as representing carriage return.
22    
23    ----------------------------------------------------------------------
24    Name: John Reppy
25    Date: 2005/07/27
26    Tag:
27    Description:
28    
29    Fixed ml-yacc to work on files with non-native end-of-line
30    encodings (e.g., Windows text file on a Unix system).
31    
32    ----------------------------------------------------------------------
33    Name: John Reppy
34    Date: 2005/07/20
35    Tag:
36    Description:
37    
38    Added changes from Dominic Evans (oldmanuk (at) gmail (dot) com)
39    to support HPUX 11.
40    
41    ----------------------------------------------------------------------
42    Name: John Reppy
43    Date: 2005/07/06
44    Tag:
45    Description:
46    
47    Changes to the SML/NJ library.  See smlnj-lib/CHANGES for details.
48    
49    ----------------------------------------------------------------------
50    Name: Matthias Blume (blume (at) tti - c (dot) org)
51    Date: 2005/07/06 00:45:00 CDT
52    Tag: blume-20050706-slice-copy
53    Description:
54    
55    Fixed reversed logic for deciding whether to "copy up" or "copy down"
56    in *-array-slice.sml.
57    
58    ----------------------------------------------------------------------
59    Name: Allen Leung (leunga (at) reservoir (dot) com)
60    Date: 2005/05/31 17:00:00 EST
61    Tag: leunga-20050531-cygwin-fault-2
62    Description:
63    
64    A typo in the cygwin code fixed.
65    
66    ----------------------------------------------------------------------
67    Name: Allen Leung (leunga (at) reservoir (dot) com)
68    Date: 2005/05/31 16:47:00 EST
69    Tag: leunga-20050531-cygwin-fault
70    Description:
71    
72    Updated Cygwin's fault/signal handling to match the Windows version.
73    Updated the export list.
74    
75    ----------------------------------------------------------------------
76    Name: Matthias Blume (blume (at) tti - c (dot) org)
77    Date: 2005/05/18 13:35:00 CDT
78    Tag: Release_110_54
79    Description:
80    
81    New working version (110.54).  NEW BOOTFILES!
82    
83    ----------------------------------------------------------------------
84    Name: Matthias Blume (blume (at) tti - c (dot) org)
85    Date: 2005/05/18 11:58:00 CDT
86    Tag: blume-20050518-installer
87    Description:
88    
89    Added support scripts for Mac OS X PackageMaker and modified
90    config/install.sh so that it supports re-dumping a heap image
91    after customization.
92    
93    ----------------------------------------------------------------------
94    Name: Matthias Blume (blume (at) tti - c (dot) org)
95    Date: 2005/05/18 10:55:00 CDT
96    Tag: blume-20050518-realdiv-noovld
97    Description:
98    
99    Un-overloaded / to work around bug in overloading resolution code.
100    
101    ----------------------------------------------------------------------
102    Name: Matthias Blume (blume (at) tti - c (dot) org)
103    Date: 2005/05/16 23:50:00 CDT
104    Tag: blume-20050516-redump-heap
105    Description:
106    
107    Added mechanism for re-creating a heap file for the interactive system
108    after configuration variables have been changed.
109    
110       CM.redump_heap : string -> unit
111    
112    This is much like SMLofNJ.exportML, but starting from the resulting
113    heap does not return to the caller of CM.redump_heap but
114    restarts the interactive system from scratch.  The original call of
115    CM.redump_heap does not return but ends the interactive session.
116    Thus, CM.redump_heap is a lot like SMLofNJ.exportFn.
117    
118    Internally, redump_heap winds the dynamic execution context back to
119    the point where the original heap image was created and re-executes
120    the heap image generation code in the boot code.
121    
122    ----------------------------------------------------------------------
123    Name: Matthias Blume (blume (at) tti - c (dot) org)
124    Date: 2005/05/09 21:55:00 CDT
125    Tag: blume-20050509-word64
126    Description:
127    
128    Added a hack to the existing hack known as Word64 to make fromString
129    behave correctly.  I am still not sure whether Word64.scan will work
130    as specified with respect to the interaction of radix and prefix.
131    
132    ----------------------------------------------------------------------
133    Name: Allen Leung (leunga (at) reservoir (dot) com)
134    Date: 2005/05/04 11:50:00 EST
135    Tag: leunga-20050504-checkgc
136    Description:
137    
138    Added a gc protocol checking phase.  This phase is enabled with
139    the flag "check-gc".  "debug-check-gc" turns on the verbose mode.
140    
141    ----------------------------------------------------------------------
142    Name: Matthias Blume (blume (at) tti - c (dot) org)
143    Date: 2005/05/04 10:45:00 CDT
144    Tag: blume-20050504-intinf
145    Description:
146    
147    Fixed a bug in the implementation of div and mod for IntInf.
148    Thanks to Neophytos Michael for reporting the problem.
149    
150    ----------------------------------------------------------------------
151    Name: Matthias Blume (blume (at) tti - c (dot) org)
152    Date: 2005/05/04 10:35:00 CDT
153    Tag: blume-20050504-join
154    Description:
155    
156    Added a "join" combinator to the ParserComb module in smlnj-lib.cm.
157    
158    ----------------------------------------------------------------------
159    Name: Matthias Blume (blume (at) tti - c (dot) org)
160    Date: 2005/02/28 23:40:00 CST
161    Tag: blume-20050228-mVar
162    Description:
163    
164    Fixed serious bug (brown paper bag variety) in new implementation of
165    structure Atom in CML.  (I had accidentally used a mailbox instead of
166    an mvar, leaving the door open for races.)
167    
168    ----------------------------------------------------------------------
169    Name: Matthias Blume (blume (at) tti - c (dot) org)
170    Date: 2005/02/25 15:00:00 CST
171    Tag: Release_110_53
172    Description:
173    
174    New working version (110.53).  NEW BOOTFILES!
175    
176    ----------------------------------------------------------------------
177    Name: Matthias Blume (blume (at) tti - c (dot) org)
178    Date: 2005/02/25 14:50:00 CST
179    Tag: blume-20050225-susp
180    Description:
181    
182    Brought back SMLofNJ.Susp.  The underlying suspension type is the one
183    implemented in Core -- which means that it is the same as the one used
184    by the lazy extension.
185    
186    ----------------------------------------------------------------------
187    Name: Matthias Blume (blume (at) tti - c (dot) org)
188    Date: 2005/02/24 16:50:00 CST
189    Tag: blume-20050224-cml-atom
190    Description:
191    
192    Simpler and at the same time more general implementation of structure
193    Atom in CML.
194    
195    ----------------------------------------------------------------------
196    Name: Matthias Blume (blume (at) tti - c (dot) org)
197    Date: 2005/02/15 17:35:00 CST
198    Tag: blume-20050215-tools
199    Description:
200    
201    Created new "tools" directory under "src" and moved "TraceDebugProf"
202    there.
203    
204    ----------------------------------------------------------------------
205    Name: Matthias Blume (blume (at) tti - c (dot) org)
206    Date: 2005/02/10 17:55:00 CST
207    Tag: blume-20050210-longlong
208    Description:
209    
210    Implemented "long long" arguments and results for NLFFI.  (Only the
211    PPC/MacOS implementation is complete, the other backends still need to
212    be updated.)
213    
214    ----------------------------------------------------------------------
215    Name: Matthias Blume (blume (at) tti - c (dot) org)
216    Date: 2005/01/24 17:40:00 CST
217    Tag: blume-20050124-mlyacc
218    Description:
219    
220    Minor cleanup in ML-Yacc rule printing mechanism.  This should fix a
221    problem with certain "as" patterns which previously got rendered
222    using incorrect syntax.
223    
224    ----------------------------------------------------------------------
225    Name: Matthias Blume (blume (at) tti - c (dot) org)
226    Date: 2005/01/18 12:00:00 CST
227    Tag: blume-20050118-profile
228    Description:
229    
230    Made time profiling code (interrupt handler) in runtime system aware
231    of new array representation.
232    
233    ----------------------------------------------------------------------
234    Name: Matthias Blume (blume (at) tti - c (dot) org)
235    Date: 2005/01/14 18:00:00 CST
236    Tag: blume-20050114-heap2exec
237    Description:
238    
239    Implemented new (but still experimental) heap2exec facility.  This is
240    tested under Mac OS X and should work under Linux (will test shortly).
241    It will probably also work on the Sparc (will test some time later).
242    
243      - removed old "HACKED_STANDALONE" hack from runtime
244    
245    To be able to test this, uncomment the request for "heap2asm" in
246    config/targets prior to installation.  (Notice that this is different
247    from "heap2exec" mentioned below.  Not a typo.)
248    
249    To perform an actual test, run the command
250    
251       $ bin/heap2exec heapfile execfile
252    
253    (You can put heap2exec on your shell's path.)
254    
255    For example, run
256    
257      $ bin/heap2exec bin/.heap/ml-yacc.ppc-darwin mly
258    
259    This will create a standalone executable called "mly" which you can
260    then invoke directly as a command.
261    
262    ----------------------------------------------------------------------
263    Name: Matthias Blume (blume (at) tti - c (dot) org)
264    Date: 2005/01/07 11:44:00 CST
265    Tag: blume-20050107-mlstring
266    Description:
267    
268    fixed off-by-one error in ML_STRING macro (globals.c)
269    
270    ----------------------------------------------------------------------
271    Name: Matthias Blume (blume (at) tti - c (dot) org)
272    Date: 2004/12/23 18:00:00 CST
273    Tag: blume-20041223-santa
274    Description:
275    
276    Made ml-build script "smarter" (but only very little).
277    
278    ----------------------------------------------------------------------
279    Name: Matthias Blume (blume (at) tti - c (dot) org)
280    Date: 2004/12/21 15:05:00 CST
281    Tag: blume-20041221-longlong
282    Description:
283    
284    * Implemented access to signed and unsigned long long data in NLFFI.
285    (The parameter-passing part of the picture has not complete. But
286    data structure access seems to work.)
287    
288    * Fixed CM's incorrect assumption that the PPC is little-endian.
289      (On the Mac, it is big-endian.  And that's currently our only
290      PPC platform.)
291    
292    ----------------------------------------------------------------------
293    Name: Matthias Blume (blume (at) tti - c (dot) org)
294    Date: 2004/12/21 12:50:00 CST
295    Tag: blume-20041221-memory
296    Description:
297    
298    Some cleanup in the $c/memory.cm library: separated some concerns by
299    moving allocation code and memory access code each into their own
300    files.
301    
302    ----------------------------------------------------------------------
303    Name: Allen Leung (leunga (at) reservoir (dot) com)
304    Date: 2004/12/17 16:12:00 EST
305    Tag: leunga-20041217-cygwin-smlnj-home
306    Description:
307    
308        The Unix I/O library of SML/NJ on cygwin does not understand
309    Windows style pathname, so problems arise when SMLNJ_HOME is set
310    to a Windows style pathname.  _run-sml now converts SMLNJ_HOME
311    to a POSIX pathname on cygwin.
312    
313    ----------------------------------------------------------------------
314    Name: Matthias Blume (blume (at) tti - c (dot) org)
315    Date: 2004/12/16 13:35:00 CST
316    Tag: Release_110_52
317    Description:
318    
319    Last-minute changes incorporated into 110.52.  Release tag moved.
320    
321    The changes:
322       - HashString.hashString' -> HashString.hashSubstring
323       - bug fix in UnivariateStats
324    
325    ----------------------------------------------------------------------
326    Name: Matthias Blume (blume (at) tti - c (dot) org)
327    Date: 2004/12/15 23:40:00 CST
328    Tag: blume-20041215-hashSubstring
329    Description:
330    
331    - HashString.hashString' -> HashString.hashSubstring
332    - corresponding changes in atom.sml
333    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
334      added some comments
335    
336    ----------------------------------------------------------------------
337    Name: Matthias Blume (blume (at) tti - c (dot) org)
338    Date: 2004/12/15 15:30:00 CST
339    Tag: (Release_110_52)
340    Description:
341    
342    New working version (110.52).  NEW BOOTFILES!
343    
344    ----------------------------------------------------------------------
345    Name: Matthias Blume (blume (at) tti - c (dot) org)
346    Date: 2004/12/15 12:45:00 CST
347    Tag: blume-20041215-spaces
348    Description:
349    
350    More on the space problem (this time for Win32).
351    
352    ----------------------------------------------------------------------
353    Name: Matthias Blume (blume (at) tti - c (dot) org)
354    Date: 2004/12/14 17:30:00 CST
355    Tag: blume-20041214-spaces
356    Description:
357    
358    Hacked some of the scripts (in particular: the installer) to cope with
359    spaces in filenames a bit better.  But beware: the current "solution"
360    is likely still full of bugs and inherently incomplete.  (We need to
361    do away with those shell scripts for a comprehensive solution.)
362    
363    ----------------------------------------------------------------------
364    Name: Matthias Blume (blume (at) tti - c (dot) org)
365    Date: 2004/12/13 14:45:00 CST
366    Tag: blume-20041213-ml-makedepend
367    Description:
368    
369    Fixed bug in code for ml-makedepend.
370    
371    ----------------------------------------------------------------------
372    Name: Matthias Blume (blume (at) tti - c (dot) org)
373    Date: 2004/12/09 16:30:00 CST
374    Tag: blume-20041209-statistics
375    Description:
376    
377    Added two simple but potentially useful statistics modules to SML/NJ Library.
378    (See CHANGES file there.)
379    
380    ----------------------------------------------------------------------
381    Name: Matthias Blume (blume (at) tti - c (dot) org)
382    Date: 2004/12/01 16:50:00 CST
383    Tag: blume-20041201-atom
384    Description:
385    
386    smlnj-lib:
387    
388      Added function HashString.hashString' for substrings.
389      Hand-inlined CharVector.foldl into HashString (for speed).
390      Modified implementation of structure Atom to avoid extracting
391      strings from substrings unless necessary.
392      (Also see CHANGES file for smlnj-lib.)
393    
394    ----------------------------------------------------------------------
395    Name: Matthias Blume (blume (at) tti - c (dot) org)
396    Date: 2004/11/24 22:15:00 CST
397    Tag: blume-20041124-cml
398    Description:
399    
400    Made sure CML compiles when Position = Int64.
401    
402    ----------------------------------------------------------------------
403    Name: Matthias Blume (blume (at) tti - c (dot) org)
404    Date: 2004/11/24 14:30:00 CST
405    Tag: blume-20041124-position
406    Description:
407    
408    The compiler can now be compiled in a mode that makes structure Position
409    equal to Int64.  The default, however, is unchanged (Position = Int31)
410    for the time being.
411    
412    To enable 64-bit positions, use the following procedure:
413    
414    1. Start sml
415    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
416    3. Type
417          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
418    4. Run CMB.make() as usual.
419    
420    This is barely tested.  The only test so far was a little SML program
421    counting the number of characters in an 8-gigabyte file by
422    reading it character-by-character.  That test was successful.
423    
424    In support of 64-bit positions, a number of new functions have been
425    added to the runtime system.
426    
427    ----------------------------------------------------------------------
428    Name: Matthias Blume (blume (at) tti - c (dot) org)
429    Date: 2004/11/23 14:45:00 CST
430    Tag: blume-20041123-useFile
431    Description:
432    
433    Fixed a problem with unhelpful error messages related to problems with
434    .cm- or .sml files that appear as part of the sml command line.
435    
436    ----------------------------------------------------------------------
437    Name: Matthias Blume (blume (at) tti - c (dot) org)
438    Date: 2004/11/18 15:40:00 CST
439    Tag: Release_110_51
440    Description:
441    
442    New working version (110.51).  NEW BOOTFILES!
443    
444    ----------------------------------------------------------------------
445    Name: Matthias Blume (blume (at) tti - c (dot) org)
446    Date: 2004/11/18 15:35:00 CST
447    Tag: <none>
448    Description:
449    
450    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
451    Joemann).
452    
453    ----------------------------------------------------------------------
454    Name: Allen Leung (leunga (at) reservoir (dot) com)
455    Date: 2004/11/17 16:05:21 EST 2004
456    Tag: leunga-20041117-mlrisc-live-kill
457    Description:
458    
459       Added support for MLTree constructs LIVE and KILL
460    to all the architectures.
461    
462    ----------------------------------------------------------------------
463    Name: Matthias Blume (blume (at) tti - c (dot) org)
464    Date: 2004/11/13 00:20:00 CST
465    Tag: blume-20041113-versiontool
466    Description:
467    
468    - Stripped down the versiontool: It now only handles the version number.
469      The date string is generated at bootstrap time (during makeml).
470    
471    - In a previous commit, fixed a minor issue with how polyequal is being
472      translated.  In particular, the code now "looks through" abstractions.
473      This results in slightly fewer polyEqual warnings and hopefully slightly
474      more efficient code.  Important examples for where this matters are
475      the new int64 and word64 types.
476    
477    ----------------------------------------------------------------------
478    Name: Matthias Blume (blume (at) tti - c (dot) org)
479    Date: 2004/11/12 00:30:00 CST
480    Tag: blume-20041112-int64
481    Description:
482    
483    Structure Int64 fully hooked in.  (The implementation is not very
484    efficient, though.)
485    
486    ----------------------------------------------------------------------
487    Name: Matthias Blume (blume (at) tti - c (dot) org)
488    Date: 2004/11/11 17:30:00 CST
489    Tag: blume-20041111-more64
490    Description:
491    
492    All the pieces of Word64 are now there, with the exception of the
493    conversions from and to LargeWord. (Eventually these need to be identities,
494    but for the time being they don't even make sense because LargeWord is
495    32-bit wide.)
496    
497    Also started to add similar support for Int64, but major pieces of that
498    are still missing.
499    
500    ----------------------------------------------------------------------
501    Name: Matthias Blume (blume (at) tti - c (dot) org)
502    Date: 2004/11/11 00:15:00 CST
503    Tag: blume-20041111-word64
504    Description:
505    
506    Structure Word64 is now (almost) complete, word literals and patterns
507    seem to work.  There are a few odd pieces missing.  In particular,
508    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
509    at the moment.
510    
511    Making Word64 official would mean that LargeWord becomes Word64.  But
512    this requires extreme care because most word-word conversions have to
513    go through LargeWord, so making a mistake means loss of efficiency or
514    worse.  Eventually there will be a solution similar to (but actually
515    simpler than) what I did with IntInf.
516    
517    ----------------------------------------------------------------------
518    Name: Matthias Blume (blume (at) tti - c (dot) org)
519    Date: 2004/11/10 18:12:00 CST
520    Tag: blume-20041110-64bit
521    Description:
522    
523    More 64-bit hacking (but still not even half-way there yet).
524    Also, some assorted improvements to the handling of 8-bit words.
525    
526    ----------------------------------------------------------------------
527    Name: Matthias Blume (blume (at) tti - c (dot) org)
528    Date: 2004/11/09 17:50:00 CST
529    Tag: <none>
530    Description:
531    
532    Started adding some infrastructure for supporting 64-bit int- and
533    word-types.  (Still in its very early stages.)
534    
535    ----------------------------------------------------------------------
536    Name: Matthias Blume (blume (at) tti - c (dot) org)
537    Date: 2004/10/28 10:45:00 CDT
538    Tag: Release_110_50
539    Description:
540    
541    New working version (110.50).  NEW BOOTFILES!
542    
543                 =====================
544    
545    Also:
546    
547      - Changend config/srcarchiveurl from a file just
548        containing the URL string into a file containing
549        shell script code.  The code has access to the $VERSION variable.
550      - Made corresponding changes to config/install.sh and config/unpack.
551      - Default contents of config/srcarchiveurl uses $VERSION and
552        normally does not have to be edited to reflect a version change.
553    
554      (As a result, a version change can be done by just editing
555       config/version, the rest is now automatic.)
556    
557    ----------------------------------------------------------------------
558    Name: Matthias Blume (blume (at) tti - c (dot) org)
559    Date: 2004/10/27 17:50:00 CDT
560    Tag: blume-20041027-btrace-msg
561    Description:
562    
563    BackTrace.monitor now also reports the source of the exception that
564    triggered the trace.
565    
566    ----------------------------------------------------------------------
567    Name: Matthias Blume (blume (at) tti - c (dot) org)
568    Date: 2004/10/27 17:20:00 CDT
569    Tag: blume-20041027-x86-c-calls
570    Description:
571    
572    This is the HISTORY entry for two earlier commits, both concerning
573    the x86 c-calls code in MLRISC:
574    
575      - added a missing LOAD in the code that deals with struct arguments
576      - made sure the caller does not add the wrong number of bytes to the
577        stack pointer after a call of a function returning a struct
578        (the callee already pops the implicit argument which points to
579         the space reserved for the result)
580    
581    ----------------------------------------------------------------------
582    Name: Allen Leung (leunga (at) reservoir (dot) com)
583    Date: 2004/10/24 14:00:00 EST
584    Tag: leunga-20041024-x86-gas-fucomip
585    Description:
586    
587       John discovered a bug in the syntax of fucomip.
588       The opcodes FU?COMIP? have been changed to
589    
590           fu?comip? %st(i), %st
591    
592    ----------------------------------------------------------------------
593    Name: Matthias Blume (blume (at) tti - c (dot) org)
594    Date: 2004/10/20 15:06:00 CDT
595    Tag: blume-20041020-standalone-backtrace
596    Description:
597    
598    Added a mechanism for getting back-trace information from standalone
599    programs.  Here is how it works:
600    
601       1. The part of the program from which you want to get backtrace
602          information (usually the whole program) should be wrapped with
603          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
604          main program could be modified from something like
605    
606             fun main (pgm, args) = ...
607    
608         to
609    
610            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
611    
612       2. To be able to access BackTrace.monitor, you have to add
613          library $smlnj-tdp/plugins.cm to the .cm file that contains your
614          main function.
615    
616       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
617          CM might have created in the past for your project).
618    
619       4. Build the system using this command line:
620    
621             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
622                      myprog.cm MyProg.main myprog
623    
624          instead of the usual
625    
626             ml-build myprog.cm MyProg.main myprog
627    
628    I changed a library name:
629    
630       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
631    
632    New libraries:
633    
634       $smlnj-tdp/back-trace.cm
635          -- when loaded causes the back-trace plugin to be installed
636       $smlnj-tdp/coverage.cm
637          -- when loaded causes the coverage plugin to be installed
638    
639    ----------------------------------------------------------------------
640    Name: Matthias Blume (blume (at) tti - c (dot) org)
641    Date: 2004/10/18 16:45:00 CDT
642    Tag: blume-20041018-groupowner
643    Description:
644    
645    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
646    
647    Eliminated group owner specs from .cm files throughout the source tree.
648    
649    ----------------------------------------------------------------------
650    Name: Matthias Blume (blume (at) tti - c (dot) org)
651    Date: 2004/10/15 15:45:00 CDT
652    Tag: blume-20041015-coverage
653    Description:
654    
655    * Test coverage tool added!
656    * Further reorganization of tracing-, debugging-, and profiling support:
657    
658        - moved original BTImp -- now called BackTrace -- into a separate
659          library called $/trace-debug-profile.cm
660        - eliminated all mentions of BTrace from SMLofNJ.Internals
661        - only the instrumentation mechanism is now left in the compiler proper
662        - BackTrace module is a plugin which is NOT plugged in by default
663        - Coverage module is another such plugin
664    
665    To get the benefits of any of these plugin modules, the code in
666    question must be compiled with tdp instrumentation turned on.  This
667    can be done by setting SMLofNJ.Internals.TDP.mode to true.
668    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
669    
670    Plugins are selected at link time.  (Pre-compiled instrumented code
671    can be re-loaded with different plugins in effect.)  When an
672    instrumented module is linked, whatever plugins are at that time
673    enabled will come into effect for that module.
674    
675    To enable the back-trace plugin, load library $/trace-debug-profile.cm
676    and invoke BackTrace.install() (e.g., from the interactive prompt).
677    To enable the coverage plugin, load the same library and invoke
678    Coverage.install().
679    
680    Back-traces are generated automatically on uncaught exceptions and
681    when the code in question explicitly invokes BackTrace.trigger().
682    
683    Coverage (and execution frequency-) information must be queried
684    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
685    
686    ----------------------------------------------------------------------
687    Name: Matthias Blume (blume (at) tti - c (dot) org)
688    Date: 2004/10/14 17:40:00 CDT
689    Tag: blume-20041014-tdp-core
690    Description:
691    
692    Snapshot of a significant overhaul of how the trace/debug/profile support
693    is hooked into the system (specifically: Core and SMLofNJ.Internals).
694    
695    ----------------------------------------------------------------------
696    Name: Matthias Blume (blume (at) tti - c (dot) org)
697    Date: 2004/10/13 16:34:00 CDT
698    Tag: blume-20041013-tdp
699    Description:
700    
701    Some rationalization of names:
702    
703       structure BTrace -> structure TDPInstrument
704       etc.
705    
706    This is is preparation of using the original back-trace
707    instrumentation for other purposes.  "TDP" stands for
708    Trace/Debug/Profile.
709    
710    The control flag controlling whether instrumentation is on or off is now
711    registered under a different name, so instead of running sml as
712    
713     sml -Cinstrument.btrace-mode=true
714    
715    one has to say
716    
717     sml -Ctdp.instrument=true
718    
719    ----------------------------------------------------------------------
720    Name: Matthias Blume (blume (at) tti - c (dot) org)
721    Date: 2004/10/11 16:37:00 CDT
722    Tag: blume-20041011-regions
723    Description:
724    
725    Made some minor modifications to elabcore.sml to have source regions
726    be propagated more tightly -- resulting in better (i.e., smaller)
727    regions being reported in error- and debug messages.
728    
729    ----------------------------------------------------------------------
730    Name: Matthias Blume (blume (at) tti - c (dot) org)
731    Date: 2004/10/08 22:50:00 CDT
732    Tag: blume-20041008-cmkw
733    Description:
734    
735    Fixed handling of keywords in .cm files: After seeing "is" the lexer
736    treats subsequent occurrences of "group", "library", "source", "is",
737    "*", and "-" as ordinary identifiers rather than keywords.
738    
739    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
740    is supposed to accept a tool argument called "source", but this did
741    not work because of the clash with the keyword.
742    
743    ----------------------------------------------------------------------
744    Name: Matthias Blume (blume (at) tti - c (dot) org)
745    Date: 2004/10/07 16:00:00 CDT
746    Tag: blume-20041007-cleanup
747    Description:
748    
749    Assorted cleanup work:
750    
751      - got rid of intstrmap in favor of using the library's
752        hash table implementation
753      - threw out most of the pathnames stuff, as it was not used anyway
754      - simplified tokentable implementation
755      - fixed some minor spelling errors
756    
757    ----------------------------------------------------------------------
758    Name: Matthias Blume (blume (at) tti - c (dot) org)
759    Date: 2004/10/06 15:15:15 CDT
760    Tag: blume-20041006-handler
761    Description:
762    
763    Cleaned up the absyn to reflect the invariant that HANDLE always
764    carries a FNexp as part of the type definition.  This eliminates some
765    superfluous sanity checks at runtime down the road.
766    
767    Some minor cleanup of the btrace code.
768    
769    ----------------------------------------------------------------------
770    Name: Matthias Blume (blume (at) tti - c (dot) org)
771    Date: 2004/10/01 10:20:30 CDT
772    Tag: blume-20041001-slave
773    Description:
774    
775    Added hack to make slave mode work in the presence of the version
776    tool.  (Still, since the master does two passes over the code for
777    CMB.make, the release number gets bumped twice when slaves are
778    attached.  I don't know if this is worth fixing...)
779    
780    ----------------------------------------------------------------------
781    Name: Matthias Blume (blume (at) tti - c (dot) org)
782    Date: 2004/09/30 10:55:00 CDT
783    Tag: blume-20040930-version
784    Description:
785    
786    * Moved the "version" magic into its own little library under
787      src/system/smlnj/internal.  This avoids expensive reconstruction of
788      a stable src/compiler/core.cm.
789    
790    * At the same time, structure CompilerVersion is now known as
791      structure SMLNJVersion.
792    
793    * Arranged for the version tool to NOT kick in when rebuilding the system
794      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
795      Also, loading the versiontool does not work when rebuilding the system
796      because CM is not properly initialized at that time.
797    
798    ----------------------------------------------------------------------
799    Name: Matthias Blume (blume (at) tti - c (dot) org)
800    Date: 2004/09/29 14:00:00 CDT
801    Tag: blume-20040929-autoversion
802    Description:
803    
804    Implemented some CM magic to have
805        file src/compiler/TopLevel/main/version.sml
806    generated automagically.
807    The version is taken from two files: config/version and config/release.
808    The first is expected to contain a two-part version number such as 110.49.
809    The second should contain a single number, but it may be missing.
810    
811    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
812    time the version tool is loaded (which is the first time you say CMB.make),
813    then the tool will increment the value stored in config/release every
814    time CMB.make is invoked.
815    
816    The binfile format is now insensitive to anything beyond the first
817    two components of a version number, so bumping the release does not render
818    binfiles incompatible.  Auto-bumping can be used to keep track of versions
819    during development without invalidating existing binfiles.
820    
821    In any case, every CMB.make updates the date information in version.sml.
822    (This is the date that is printed in the banner.)
823    
824    ----------------------------------------------------------------------
825    Name: Matthias Blume (blume (at) tti - c (dot) org)
826    Date: 2004/09/28 10:53:00 CDT
827    Tag: blume-20040928-controls
828    Description:
829    
830    Some cleanup of the controls code.
831    
832    ----------------------------------------------------------------------
833    Name: Matthias Blume (blume (at) tti - c (dot) org)
834    Date: 2004/09/27 22:08:00 CDT
835    Tag: blume-20040927-controls
836    Description:
837    
838    Added two pieces of functionality to the Controls interface:
839    
840    1.   val save'restore: 'a control -> unit -> unit
841    
842       grabs the current value of the control in stage 1 and restores it
843       in stage 2.
844    
845    2.   val set' : 'a control * 'a -> unit -> unit
846    
847       stores the given value into the control in stage 2 (i.e., delayed)
848       but does all error checking in stage 1.
849       (This is for string controls that need to do parse their argument --
850        something that might fail.  In some cases, notably in CM, one
851        already knows the intended argument but wants to delay the actual
852        assignment until a time when error recovery would be more difficult.)
853    
854    Changed the handling of controls in tool arguments to classes "sml" and
855    "lazysml":
856       - use Controls.save'restore as a more robust way of restoring the
857         old value (in particular: without having to re-parse the string)
858       - use controls to handle the "overload" keyword in the init group
859         (I believe this change actually fixes a long-standing obscure bug.)
860    
861    ----------------------------------------------------------------------
862    Name: Matthias Blume (blume (at) tti - c (dot) org)
863    Date: 2004/09/27 17:00:00 CDT
864    Tag: blume-20040927-lazysml
865    Description:
866    
867    Added a new tool class called "lazysml" to CM's tool chest.  The only
868    difference to "sml" is that compilation is done with Control.lazysml
869    set to true.  A source of class "lazysml" is automatically recognized
870    by a file name suffix of ".lml".
871    
872    In addition to the above feature, the original class "sml" now also
873    supports a tool argument "lazy" which has the same effect.  As a
874    result, the following three lines are equivalent:
875    
876        foo.sml : lazysml
877        foo.sml : sml (lazy)
878        foo.sml (lazy)
879    
880    The setting goes into effect both during parsing and during
881    compilation.  The original setting is restored right after parsing and
882    after compilation, respectively.
883    
884    In addition to all the above, there is also a general mechanism to set
885    ANY of the "controls" that are available at the command line via
886    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
887    apply as well. (In fact, "lazy" is implemented as a special case of
888    the general mechanism.)
889    
890    The .cm file syntax uses a new keyword tool argument called "with".
891    There are several ways of indicating the desired settings:
892    
893        foo.sml (with:parser.quotations=true)
894        foo.sml (with:(name:parser.quotations value:true))
895        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
896        foo.sml (with:(name1=value1 name2=value2 ...))
897        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
898    
899        etc.
900    
901    Another possible abbreviation is to leave out the =v or value:v part
902    if the name refers to a boolean control (in which case the value is
903    taken to be true).  Thus, one could get lazy sml also by saying:
904    
905        foo.sml (with:parser.lazy-keyword=true)
906        foo.sml (with:parser.lazy-keyword)
907        foo.sml (with:(name:parser.lazy-keyword value:true))
908        foo.sml (with:(name:parser.lazy-keyword))
909    
910    ----------------------------------------------------------------------
911    Name: Matthias Blume (blume (at) tti - c (dot) org)
912    Date: 2004/09/24 16:22:00 CDT
913    Tag: blume-20040924-ppc-long-branch
914    Description:
915    
916    Turned message about "emiting long form of branch" off by default.
917    Added a control flag to turn it back on when desired.
918    
919    ----------------------------------------------------------------------
920    Name: Matthias Blume (blume (at) tti - c (dot) org)
921    Date: 2004/09/24 16:05:00 CDT
922    Tag: blume-20040924-rounding
923    Description:
924    
925    Applied patch for setting rounding modes under Mac OS X.  Thanks to
926    Melissa O'Neill for providing the code!
927    
928    ----------------------------------------------------------------------
929    Name: Matthias Blume (blume (at) tti - c (dot) org)
930    Date: 2004/09/23 17:30:00 CDT
931    Tag: blume-20040923-envvars
932    Description:
933    
934    1. Changed definition of type ControlRegistry.registry_tree to
935       include control_info (i.e., the name of the controlling environment
936       variable).
937    
938    2. Added command-line flags -e and -E to print the names of environment
939       variables that can be used to control internal settings.
940       (This uses the new API mentioned in 1.)
941    
942    ----------------------------------------------------------------------
943    Name: Matthias Blume (blume (at) tti - c (dot) org)
944    Date: 2004/09/13 16:50:00 CDT
945    Tag: Release_110_49
946    Description:
947    
948    New working version (110.49).  NEW BOOTFILES!
949    
950    ----------------------------------------------------------------------
951    Name: Matthias Blume
952    Date: 2004/09/13 16:20:00 CDT
953    Tag: blume-20040913-config-mlrisc
954    Description:
955    
956    Put target "mlrisc" back into the default list.
957    (There is no harm in having it, and some users have expressed their
958    wish to have "mlrisc" included by default.)
959    
960    ----------------------------------------------------------------------
961    Name: John Reppy
962    Date: 2004/09/13
963    Tag: jhr-20040913-signals
964    Description:
965    
966    Fixed the signal masking code to properly nest mask/unmask operations
967    on a per-signal basis.
968    
969    ----------------------------------------------------------------------
970    Name: Matthias Blume (blume (at) tti - c (dot) org)
971    Date: 2004/09/08 13:20:00 CDT
972    Tag: blume-20040908-heap-magic
973    Description:
974    
975    Bumped the heap macig number to 0x09082004 to account for the changed
976    layout of the ML frame under MacOS X.
977    
978    ----------------------------------------------------------------------
979    Name: Allen Leung (leunga (at) reservoir (dot) com)
980    Date: 2004/09/03 11:26:00 EST
981    Tag: leunga-20040903-cygwin-install
982    Description:
983    
984       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
985    The Cygwin runtime is turned on by setting the environment
986    variable SMLNJ_CYGWIN_RUNTIME to 1.
987    
988    ----------------------------------------------------------------------
989    Name: Matthias Blume (blume (at) tti - c (dot) org)
990    Date: 2004/08/31 17:15:00 CDT
991    Tag: blume-20040831-core
992    Description:
993    
994    Added some exports to src/compiler/core.cm upon request by J. Joemann.
995    
996    ----------------------------------------------------------------------
997    Name: Matthias Blume (blume (at) tti - c (dot) org)
998    Date: 2004/08/30 17:55:00 CDT
999    Tag: blume-20040830-installer
1000    Description:
1001    
1002    Upon request by Johannes Joemann:
1003    
1004    - improved ML code of installer to fall back to coping when renaming
1005      fails (i.e., when source and target are on different file systems);
1006      the code compiles but has yet to be tested in anger
1007    - removed mlrisc from list of default targets (config/targets)
1008    
1009    ----------------------------------------------------------------------
1010    Name: Matthias Blume (blume (at) tti - c (dot) org)
1011    Date: 2004/08/27 17:20:00 CDT
1012    Tag: blume-20040827-ptreql
1013    Description:
1014    
1015    Added ptreql primop to structure InlineT (upon request from Larry
1016    Paulson).
1017    
1018    ----------------------------------------------------------------------
1019    Name: Allen Leung (leunga (at) reservoir (dot) com)
1020    Date: 2004/08/15 21:21:00 EST
1021    Tag: leunga-110_48-udgraph
1022    Description:
1023    
1024    Another bug fix from Carl Hauser:
1025    
1026    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
1027    > 48c48
1028    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
1029    > e::L)
1030    > ---
1031    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
1032    > Without this, any deletion of an edge in an undirected graph does severe
1033    > violence to the graph.
1034    
1035    ----------------------------------------------------------------------
1036    Name: Allen Leung (leunga (at) reservoir (dot) com)
1037    Date: 2004/08/10 23:35:00 EST
1038    Tag: leunga-110_48-ppc
1039    Description:
1040    
1041       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
1042    
1043    ----------------------------------------------------------------------
1044    Name: Matthias Blume (blume (at) tti - c (dot) org)
1045    Date: 2004/08/10 12:00:00 CDT
1046    Tag: Release_110_48
1047    Description:
1048    
1049    New working version (110.48).  NEW BOOTFILES!
1050    
1051    ----------------------------------------------------------------------
1052    Name: Allen Leung (leunga (at) reservoir (dot) com)
1053    Date: 2004/08/09 12:21:00 EST
1054    Tag: leunga-110_47-dijsktra
1055    Description:
1056    
1057    Bug fix from Carl Hauser:
1058    
1059    single_source_shortest_paths in dijkstra.sml was observed to get wrong
1060    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
1061    
1062    The problem is that following the expression A.update(dist,s,Num.zero)
1063    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
1064    
1065    ----------------------------------------------------------------------
1066    Name: Matthias Blume (blume (at) tti - c (dot) org)
1067    Date: 2004/08/06 18:10:00 CDT
1068    Tag: blume-20040806-cmdline
1069    Description:
1070    
1071    Fiddled with handling of command-line options:
1072    
1073      * sml now quits after processing the command line
1074        if -H, -S, -h<n>, or -s<n> appears as the last
1075        command-line argument
1076      * a new option -q terminates the session when encountered on
1077        the command line; subsequent arguments will be ignored
1078      * bug fixes: short (erroneous) arguments are no longer ignored
1079        completely
1080    
1081    ----------------------------------------------------------------------
1082    Name: Allen Leung (leunga (at) reservoir (dot) com)
1083    Date: 2004/08/04 18:17:00 EST
1084    Tag: leunga-110_47-ppc-ibm-asm
1085    Description:
1086    
1087      - Added minimal IBM assembly syntax support for PowerPC.
1088    
1089      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
1090        been altered in the runtime.  We need an automatic way to keep the file
1091        in sync.
1092    
1093    ----------------------------------------------------------------------
1094    Name: Matthias Blume (blume (at) tti - c (dot) org)
1095    Date: 2004/08/04 14:00:00 CDT
1096    Tag: Release_110_47
1097    Description:
1098    
1099    New working version (110.47).  NEW BOOTFILES!
1100    
1101    ----------------------------------------------------------------------
1102    Name: Matthias Blume (blume (at) tti - c (dot) org)
1103    Date: 2004/08/03 14:25:00 CDT
1104    Tag: blume-20040803-callingconv
1105    Description:
1106    
1107    Added low-level support for choosing C calling conventions by
1108    twiddling the type of rawccall.  (See
1109    src/compiler/Semant/types/cproto.sml for details.)
1110    
1111    ----------------------------------------------------------------------
1112    Name: Matthias Blume (blume (at) tti - c (dot) org)
1113    Date: 2004/08/02 15:55:00 CDT
1114    Tag: blume-20040802-backout
1115    Description:
1116    
1117    Backed out of change to win32-filesys.c.  The earlier patch to
1118    get_file_time caused CM to produce files with the wrong time stamp.
1119    
1120    ----------------------------------------------------------------------
1121    Name: Matthias Blume (blume (at) tti - c (dot) org)
1122    Date: 2004/08/02 14:45:00 CDT
1123    Tag: blume-20040802-nlffi-win32
1124    Description:
1125    
1126    Added NLFFI support for Win32, adapted from a patch provided by David
1127    Hansel.  This is currently completely untested.  Also, the issue
1128    concerning stdcall vs. ccall is still unresolved.
1129    
1130    ----------------------------------------------------------------------
1131    Name: Matthias Blume (blume (at) tti - c (dot) org)
1132    Date: 2004/07/30 17:55:00 CDT
1133    Tag: blume-20040730-various
1134    Description:
1135    
1136    Gearing up towards 110.47...
1137    
1138    - various minor bugfixes to ml-nlffigen
1139    - a beginning of a manual for nlffi
1140    
1141    - eliminated 'export name=value' in config/install.sh as this does
1142      not work with certain versions of /bin/sh
1143      (Thanks to David King at Motorola for catching this.)
1144    
1145    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
1146      - added a test for tm==NULL to gmtime.c and localtime.c
1147      - applied patch for incorrect GetFileTime under win32
1148      - toSeconds -> toMilliseconds in Win32/win32-process.sml
1149    
1150    ----------------------------------------------------------------------
1151    Name: Matthias Blume (blume (at) tti - c (dot) org)
1152    Date: 2004/07/21 18:20:00 CDT
1153    Tag: blume-20040721-nlffigen
1154    Description:
1155    
1156    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
1157      for a typedef to an incomplete type, but leave out the "typ" member.
1158      (This is just for consistency.)
1159    - Started to produce what is supposed to become better (i.e., comprehensive)
1160      documentation of what ml-nlffigen does and produces.
1161    
1162    ----------------------------------------------------------------------
1163    Name: Matthias Blume (blume (at) tti - c (dot) org)
1164    Date: 2004/07/14 16:25:00 CDT
1165    Tag: blume-20040714-union
1166    Description:
1167    
1168    Added C_UNION to c-calls/c-types.sml and updated the machinery
1169    (ml-nlffigen, cproto.sml) that conveys C function interface
1170    information to the code generator.
1171    
1172    However, the actual architecture-specific implementation of function
1173    arguments and results that are C unions is still not implemented.
1174    
1175    ----------------------------------------------------------------------
1176    Name: Allen Leung (leunga (at) reservoir (dot) com)
1177    Date: 2004/07/14 14:38:00 EST
1178    Tag: leunga-110_46_1-ppc-lwzu
1179    Description:
1180    
1181        Added these instructions to the PowerPC architecture:
1182        LBZU(X), LHZU(X), LWZU(X),
1183        STWU(X), STFDU, STFSU
1184    
1185        etc...
1186    
1187        Note: I haven't added their instruction encoding into the description.
1188    
1189    ----------------------------------------------------------------------
1190    Name: Allen Leung (leunga (at) reservoir (dot) com)
1191    Date: 2004/07/13 15:04:00 EST
1192    Tag: leunga-110_46_1-ppc-lwarx
1193    Description:
1194    
1195       Added the two instructions LWARX and STWCX to the PowerPC
1196    instruction set.
1197    
1198       A (untested) rewrite of loop-structure.sml.  The old version
1199    is completely broken.
1200    
1201    ----------------------------------------------------------------------
1202    Name: Matthias Blume (blume (at) tti - c (dot) org)
1203    Date: 2004/07/13 13:50:00 CDT
1204    Tag: blume-20040713-nlffi
1205    Description:
1206    
1207    - use paramAlloc to report c-calls with too many arguments
1208      (for PPC version where parameter area is pre-allocated)
1209    - added ccall_maxargspace to machspec (to implement the above)
1210    - made "make" commend in CM's "make" tool configurable
1211    - added option (default: on) for passing the name of the SML/NJ's "bin"
1212      directory to "make"; the call looks like this:
1213    
1214         make <options> SMLNJ_BINDIR=<dir> <target>
1215    
1216      This can be used by the Makefile to, e.g., pick the "right" version
1217      of ml-nlffigen.
1218    - minor code tweaks
1219    
1220    ----------------------------------------------------------------------
1221    Name: Matthias Blume (blume (at) tti - c (dot) org)
1222    Date: 2004/07/12 22:50:00 CDT
1223    Tag: blume-110_46_1-macosx-nlffi
1224    Description:
1225    
1226    NLFFI under Mac OS X now working (sort of).  This is largely untested,
1227    though.
1228    
1229    Note:  1. You have to make a new, clean build of the runtime system.
1230           2. There are new BOOTFILES, you have to use them!
1231              (Doing the bootstrap process yourself would be *very* painful!
1232               If you absolutely have to do it, build the system under
1233               a different architecture and then cross-compile.)
1234    
1235    Version bumped to 110.46.1 to account for runtime data format changes.
1236    
1237    ----------------------------------------------------------------------
1238    Name: Matthias Blume (blume (at) tti - c (dot) org)
1239    Date: 2004/06/18 14:30:00 CDT
1240    Tag: blume-20040618-unix
1241    Description:
1242    
1243    Changed the implementation of structure Unix so that the same stream
1244    is returned every time one of the {text,bin}{In,Out}streamOf functions
1245    is invoked on the same proc.  This is not what the spec currently
1246    says -- although IMO it arguably should.  (See discussion below.)
1247    
1248    ----------------------------------------------------------------------
1249    Name: Matthias Blume (blume (at) tti - c (dot) org)
1250    Date: 2004/06/17 18:15:00 CDT
1251    Tag: Release_110_46
1252    Description:
1253    
1254    New working version (110.46).  NEW BOOTFILES!
1255    
1256    ----------------------------------------------------------------------
1257    Name: Matthias Blume (blume (at) tti - c (dot) org)
1258    Date: 2004/06/17 17:20:00 CDT
1259    Tag: blume-20040617-timer-unix
1260    Description:
1261    
1262    Changed the interface of structures Timer and Unix to match the most
1263    recent Basis spec.
1264    
1265    In the case of Unix there still seems to be an open/weird issue:
1266    
1267       The {text,bin}{In,Out}streamOf functions are supposed to create
1268       fresh streams whenever they are called -- as opposed to have them
1269       return the same stream every time.  This design is supposed to
1270       prevent space leaks caused by proc values hanging on to streams.
1271    
1272       The reap function, on the other hand, is supposed to close the
1273       streams.  This cannot be done without having a handle on the
1274       stream in proc after all...
1275    
1276    I took the liberty to implement the following stopgap solution:
1277    
1278      The proc value hangs on to the most recently created stream(s).
1279      Reap closes those.  If either or both of the two streams hadn't
1280      been created at all yet, then reap will close the corresponding
1281      file descriptors directly.
1282    
1283    PS: I don't understand the original space leak argument anymore.  If
1284    a proc hangs on to the imperative stream, then I/O operations on those
1285    will advance the state of the cached stream and avoid the space leak.
1286    
1287    ----------------------------------------------------------------------
1288    Name: Matthias Blume (blume (at) tti - c (dot) org)
1289    Date: 2004/05/28 16:45:00 CDT
1290    Tag: blume-20040528-basis
1291    Description:
1292    
1293    Added signature PACK_REAL and exported functor PrimIO.
1294    
1295    ----------------------------------------------------------------------
1296    Name: Matthias Blume (blume (at) tti - c (dot) org)
1297    Date: 2004/05/25 16:00:00 CDT
1298    Tag: blume-20040525-group-owner
1299    Description:
1300    
1301    CM now ignores (but still accepts) the "owner" information in group
1302    descriptions.  The owner of a group is its next enclosing
1303    library. Each group must have a unique owner.  (There is a virtual
1304    "toplevel" library that owns groups which are not nested within a real
1305    library.)  Previously, each group had to explicitly declare its owner,
1306    and CM would check that such a declaration is correct.  The new scheme
1307    is to have CM check that for each group there is precisely one owning
1308    library.
1309    
1310    The advantage of the new scheme is that the programmer no longer needs
1311    to maintain the somewhat annoying owner information.  The downside is
1312    that CM cannot enforce the ownership rule across multiple runs of
1313    CM.make.  Fortunately, enclosing the same group in two different
1314    libraries A and B which are not part of the same program does not
1315    cause real problems.
1316    
1317    ----------------------------------------------------------------------
1318    Name: Matthias Blume (blume (at) tti - c (dot) org)
1319    Date: 2004/05/20 16:00:00 CDT
1320    Tag: blume-20040520-win32
1321    Description:
1322    
1323    Made the win32 version work again.  (Strangely, a misplaced comma had
1324    slipped into win32-process.c which prevented the runtime from being
1325    compiled correctly.)
1326    
1327    Also, included a minor addition to ml-build.bat analogous to what was
1328    done in blume-20040519-ml-build.
1329    
1330    ----------------------------------------------------------------------
1331    Name: Matthias Blume (blume (at) tti - c (dot) org)
1332    Date: 2004/05/19 22:10:00 CDT
1333    Tag: blume-20040519-ml-build
1334    Description:
1335    
1336    Arranged for ml-build to clean up after itself a little bit better.
1337    The script generates a temporary SML source file and compiles it using
1338    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1339    gets rid of those at the end, so they don't accumulate under .cm.
1340    
1341    This required a minor change to install.sh because the name of the
1342    metadata directory (default: .cm) is actually configurable at
1343    installation time.
1344    
1345    ----------------------------------------------------------------------
1346    Name: Matthias Blume (blume (at) tti - c (dot) org)
1347    Date: 2004/05/18 15:50:00 CDT
1348    Tag: blume-20040518-mkreader
1349    Description:
1350    
1351    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1352    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1353    
1354    ----------------------------------------------------------------------
1355    Name: Matthias Blume (blume (at) tti - c (dot) org)
1356    Date: 2004/05/11 14:35:00 CDT
1357    Tag: blume-20040511-win32sock
1358    Description:
1359    
1360    Added previously missing support for many socket-related functions
1361    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1362    for the voluminous patch!
1363    
1364    (I have not tested this patch under win32 yet.)
1365    
1366    Here is David's e-mail:
1367    
1368    Hi,
1369    
1370    Attached to this email you find a diff against sml/nj 110.45
1371    that will enable socket support under Windows.
1372    
1373    To apply the patch (using unix or cygwin)
1374    1) gunzip runtime.diff.gz
1375    2) "cd" into "src/runtime" in the source tree of a fresh
1376       110.45 installation.
1377    3) patch -p 1 < [your/path/to]runtime.diff
1378    
1379    The code compiles fine but has NOT yet been extensively tested.
1380    I only ran a few tests for basic socket client functionality
1381    (which worked fine).  Especially the functions that use ioctl
1382    are not tested at all and might not work (see below).
1383    
1384    I implemented this since we want to move to a newer version of sml/nj
1385    but need socket support in order to use it.  This is the first time I
1386    even had a look at the sml/nj source,  so please review my changes
1387    before making this part of the distribution!  Here are a few issues
1388    that I think might be better for someone to solve who is more
1389    familiar with the sml/nj source (and socket programming):
1390    
1391    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1392      exception since I could not figure out what the windows equivalent
1393      of these functions is
1394    
1395    - In sockets-osdep.h there are a some #include statements that are
1396      only used in a few files that include sockets-osdep.h
1397    
1398    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1399      does not process its return value since I don't know how to
1400      report an error upwards.
1401    
1402    - It would probably be good to have a call to WSACleanup() when
1403      the library is unloaded (if there is such a possibility).
1404      Otherwise I think Windows will take care of this automatically
1405      when the process finishes.
1406    
1407    - I used ioctlsocket() as a replacement for ioctl() but I have
1408      no idea if that is actually the proper replacement on Windows.
1409    
1410    - All these issues are marked in the code by "FIXME" comments.
1411    
1412    We use sml/nj extensively in our products and are quite happy
1413    with it.  I hope this contribution will help you.
1414    
1415    Keep up the good work!
1416    
1417    David
1418    
1419    ----------------------------------------------------------------------
1420    Name: Matthias Blume (blume (at) tti - c (dot) org)
1421    Date: 2004/05/11 14:20:00 CDT
1422    Tag: blume-20040511-installml
1423    Description:
1424    
1425    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1426    the patch.)
1427    
1428    ----------------------------------------------------------------------
1429    Name: Matthias Blume (blume (at) tti - c (dot) org)
1430    Date: 2004/05/11 14:05:00 CDT
1431    Tag: blume-20040511-nlffi-netbsd
1432    Description:
1433    
1434    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1435    the patch.)
1436    
1437    ----------------------------------------------------------------------
1438    Name: Matthias Blume (blume (at) tti - c (dot) org)
1439    Date: 2004/05/11 12:05:00 CDT
1440    Tag: blume-20040511-exports
1441    Description:
1442    
1443    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1444    export lists in compiler-related .cm-files.
1445    
1446    ----------------------------------------------------------------------
1447    Name: Matthias Blume (blume (at) tti - c (dot) org)
1448    Date: 2004/05/11 11:35:00 CDT
1449    Tag: blume-20040511-allsource
1450    Description:
1451    
1452    The installer now honors the "src-smlnj" target again, although its meaning
1453    has changed from "all sources required for the compiler" to "all sources
1454    the installer knows about".  In other words, if you enable "src-smlnj"
1455    in the "targets" file, then the installer will pull in sources for
1456    everything.  (Notice that this refers to source code only.  Compiled
1457    code is still only installed for modules that were requested explicitly
1458    or which are required for other modules that were requested explicitly.)
1459    
1460    ----------------------------------------------------------------------
1461    Name: Matthias Blume (blume (at) tti - c (dot) org)
1462    Date: 2004/04/23 17:40:00 CDT
1463    Tag: blume-20040423-ieee-scan
1464    Description:
1465    
1466    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1467    in the exponent calculation we get INF or ZERO (depending on the mantissa
1468    and the sign of the exponent).
1469    
1470    ----------------------------------------------------------------------
1471    Name: Matthias Blume (blume (at) tti - c (dot) org)
1472    Date: 2004/04/23 10:40:00 CDT
1473    Tag: blume-20040423-ml-build
1474    Description:
1475    
1476    The ml-build script now terminates with a non-0 status when something
1477    goes wrong.
1478    
1479    ----------------------------------------------------------------------
1480    Name: Matthias Blume (blume (at) tti - c (dot) org)
1481    Date: 2004/04/22 16:35:00 CDT
1482    Tag: blume-20040422-Option
1483    Description:
1484    
1485    Made exception Option to be the same as exception Option.Option
1486    (as it should be).
1487    
1488    ----------------------------------------------------------------------
1489    Name: Allen Leung (leunga (at) reservoir (dot) com)
1490    Date: 2004/03/19 14:40:00 EST
1491    Tag: leunga-20040319-cygwin-nlffi
1492    Description:
1493    
1494    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1495    of SML/NJ.  The problem is that
1496    
1497       lib = dlopen(NULL, ...)
1498       f   = dlsym(lib, "malloc");
1499    
1500    does not work on Windows unless we explicitly export symbols
1501    such as 'malloc' during linking.  We fixed this by explicitly
1502    exporting the required symbols with the magic gcc incantation:
1503    
1504        -Wl,--export-all cygwin.def
1505    
1506    where cygwin.def is a file containing all the symbols that
1507    we wish to export.
1508    
1509    I suspect this is a Windows problem and we'll have to
1510    do the same (somehow with windows compilers) when
1511    we build the native win32 version with the system
1512    calls LoadLibrary/GetProcAddress.
1513    
1514    ----------------------------------------------------------------------
1515    Name: Matthias Blume (blume (at) tti - c (dot) org)
1516    Date: 2004/03/04 16:35:00 CST
1517    Tag: blume-20040304-intinf-fmt
1518    Description:
1519    
1520    Fixed problem with IntInf.fmt (sign would show up on the right instead
1521    of on the left for BIN, OCT, and HEX).
1522    
1523    ----------------------------------------------------------------------
1524    Name: Matthias Blume (blume (at) tti - c (dot) org)
1525    Date: 2004/03/04 11:25:00 CST
1526    Tag: blume-20040304-symlinks
1527    Description:
1528    
1529    Fixed problem with installer script (unix only) where bin/ml-yacc and
1530    friends pointed (via symlinks) to absolute locations instead of just
1531    .run-sml.  This was reported by Vesa A Norrman.
1532    
1533    ----------------------------------------------------------------------
1534    Name: Matthias Blume (blume (at) tti - c (dot) org)
1535    Date: 2004/02/13 14:50:00 CST
1536    Tag: Release_110_45
1537    Description:
1538    
1539    New working version (110.45). New bootfiles.
1540    
1541    ----------------------------------------------------------------------
1542    Name: Matthias Blume (blume (at) tti - c (dot) org)
1543    Date: 2004/01/26 15:15:15 CST
1544    Tag: blume-20040126-toplevel
1545    Description:
1546    
1547    Improved handling of exceptions at the interactive toplevel.
1548    
1549    ----------------------------------------------------------------------
1550    Name: Matthias Blume (blume (at) tti - c (dot) org)
1551    Date: 2004/01/26 11:25:00
1552    Tag: blume-20040126-app
1553    Description:
1554    
1555    Type of top-level "app" corrected.
1556    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1557    
1558    ----------------------------------------------------------------------
1559    Name: Matthias Blume (blume (at) tti - c (dot) org)
1560    Date: 2003/11/18 17:10 CST
1561    Tag: blume-20031118-basis-fiddle
1562    Description:
1563    
1564    - changed Timer interface to what might become the spec
1565    - POSIX_FLAGS -> BIT_FLAGS according to spec
1566    - some other minor discrepancies wrt. spec eliminated
1567    
1568    ----------------------------------------------------------------------
1569    Name: Matthias Blume (blume (at) tti - c (dot) org)
1570    Date: 2003/11/06 12:00:00 CST
1571    Tag: Release_110_44
1572    Description:
1573    
1574    New working version (110.44). New bootfiles.
1575    
1576    ----------------------------------------------------------------------
1577    Name: Matthias Blume (blume (at) tti - c (dot) org)
1578    Date: 2003/11/04 11:50:00 CST
1579    Tag: blume-20031104-move-libraries
1580    Description:
1581    
1582    Eliminated the "dont_move_libraries" directive in config/targets.
1583    (The mechanism was broken and could not be fixed easily.  Moreover,
1584    there does not seem to be any reason not to move all libraries into
1585    lib during installation.  I originally implemented this directive as a
1586    backward-compatibility feature when I first introduced the new CM.
1587    Now that things have been stable for a long time and going back to the
1588    old CM is not an option, there is no reason to keep it around.)
1589    
1590    ----------------------------------------------------------------------
1591    Name: Matthias Blume (blume (at) tti - c (dot) org)
1592    Date: 2003/11/03 16:00:00 CST
1593    Tag: blume-20031103-installdir
1594    Description:
1595    
1596    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1597    Richards for pointing out the problem and providing the solution.)
1598    
1599    ----------------------------------------------------------------------
1600    Name: Matthias Blume (blume (at) tti - c (dot) org)
1601    Date: 2003/10/01 17:05:00 CDT
1602    Tag: blume-20031001-lal-mlrisc
1603    Description:
1604    
1605    MLRISC bug fix from Lal.
1606    
1607    ----------------------------------------------------------------------
1608    Name: Matthias Blume (blume (at) tti - c (dot) org)
1609    Date: 2003/09/30 16:10:00 CDT
1610    Tag: blume-20030930-primio-bat
1611    Description:
1612    
1613    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1614    2. Improved .bat files (for Win32 port) to make things work under Win95.
1615       (thanks to Aaron S. Hawley for this one)
1616    
1617    ----------------------------------------------------------------------
1618    Name: Matthias Blume (blume (at) tti - c (dot) org)
1619    Date: 2003/09/26 16:05:00 CDT
1620    Tag: blume-20030926-wrappriv
1621    Description:
1622    
1623    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1624    
1625    ----------------------------------------------------------------------
1626    Name: Matthias Blume (blume (at) tti - c (dot) org)
1627    Date: 2003/09/26 15:00:00 CDT
1628    Tag: blume-20030926-110_43_3
1629    Description:
1630    
1631    - additional cleanup
1632    - version number bump, NEW BOOTFILES
1633    
1634    ----------------------------------------------------------------------
1635    Name: Matthias Blume (blume (at) tti - c (dot) org)
1636    Date: 2003/09/26 12:00:00 CDT
1637    Tag: blume-20030926-ppautoload
1638    Description:
1639    
1640    I modified the read-eval-print loop so that the autoloader gets
1641    invoked whenever the prettyprinter tries to look up a symbol that
1642    is not currently defined in the toplevel environment but which
1643    appears in CM's autoload registry.  As a result, we see far fewer of
1644    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1645    
1646    In addition to this I tried to clean up some pieces of the Basis
1647    implementation (e.g., Socket, Word8Array) in order to prevent other
1648    instances of these ?.Foo.Bar.xxx names from being printed.
1649    
1650    The mechanism that picks names for types still needs some work, though.
1651    (Right now it seems that if there is a type A.t which is defined to
1652    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1653    "?.B.u" although the perhaps more sensible solution would be to use
1654    "A.t" in this case.  In other words, the prettyprinter should follow
1655    a chain of DEFtycs not farther than there are corresponding toplevel
1656    names in the current environment.)
1657    
1658    ----------------------------------------------------------------------
1659    Name: Matthias Blume (blume (at) tti - c (dot) org)
1660    Date: 2003/09/24 16:31:00 CDT
1661    Tag: blume-20030924-installer
1662    Description:
1663    
1664    Another installer tweak:  All the ML code for the installer is now
1665    compiled during CMB.make and put into a little library called
1666    $smlnj/installer.cm.  The installation then simply invokes
1667    
1668       sml -m $smlnj/installer.cm
1669    
1670    and everything happens automagically.
1671    
1672    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1673    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1674            and BINDIR.
1675    
1676    The new scheme guarantees that the ML code responsible for the installation
1677    is in sync with the APIs of the main system.  Also, the installer is
1678    somewhat faster because the installer script is precompiled.
1679    
1680    ----------------------------------------------------------------------
1681    Name: Matthias Blume (blume (at) tti - c (dot) org)
1682    Date: 2003/09/24 15:35:00 CDT
1683    Tag: blume-20030924-synsock
1684    Description:
1685    
1686    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1687    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1688    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1689    implementations of non-blocking operations from CML's Socket
1690    structure.
1691    
1692    ----------------------------------------------------------------------
1693    Name: Matthias Blume (blume (at) tti - c (dot) org)
1694    Date: 2003/09/24 15:10:05 CDT
1695    Tag: blume-20030924-sockets
1696    Description:
1697    
1698    1. Fixed SOCKET API and implementation to match Basis spec.
1699       This required changing the internal representation of sockets to one
1700       that remembers (for each socket file descriptor) whether it is currently
1701       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1702       call is made only if the state actually needs to change).
1703    
1704    2. OS-specific details of sockets were moved into separate files, thus
1705       making it possible to unify the bulk of the socket implementations
1706       between Unix and Win32.
1707    
1708    3. CML's socket API changed accordingly.
1709       (Note that we need to remove non-blocking functions from this API
1710       since they are redundant in the case of CML!)
1711    
1712    4. CML's socket implementation now makes use of non-blocking functions
1713       provided by Basis, thus removing all OS-dependent code from this part
1714       of CML.
1715    
1716    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1717    
1718    ----------------------------------------------------------------------
1719    Name: Matthias Blume (blume (at) tti - c (dot) org)
1720    Date: 2003/09/22 12:10:00 CDT
1721    Tag: blume-20030922-110_43_2
1722    Description:
1723    
1724    Made a new interim version and bootfiles for developer's bootstrapping
1725    convenience.
1726    
1727       110.43.2 -- NEW BOOTFILES
1728    
1729    ----------------------------------------------------------------------
1730    Name: Matthias Blume (blume (at) tti - c (dot) org)
1731    Date: 2003/09/19 15:55:00 CDT
1732    Tag: blume-20030919-cmdir
1733    Description:
1734    
1735    1. new-install.sh -> install.sh
1736    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1737    3. tweaked installer so that another name instead of .cm can be chosen
1738       at install time (by setting the CM_DIR_ARC environment variable
1739       during installation); once installation is complete, the name is
1740       fixed
1741    
1742    ----------------------------------------------------------------------
1743    Name: Matthias Blume (blume (at) tti - c (dot) org)
1744    Date: 2003/09/18 16:00:00 CDT
1745    Tag: blume-20030918-110_43_1
1746    Description:
1747    
1748    Made a new interim version and bootfiles for developer's bootstrapping
1749    convenience.
1750    
1751       110.43.1 -- NEW BOOTFILES
1752    
1753    ----------------------------------------------------------------------
1754    Name: Matthias Blume (blume (at) tti - c (dot) org)
1755    Date: 2003/09/18 15:20:00 CDT
1756    Tag: blume-20030918-misc
1757    Description:
1758    
1759    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1760       this seems to be controversial at the moment) and used those in
1761       Posix.ProcEnv.times.
1762    
1763    2. Added Time.{from,to}Nanoseconds to Time.
1764    
1765    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1766       For example, fromLargeInt never needs to look at more than 3 "big
1767       digits" to get its 53 bits of precision.
1768    
1769    ----------------------------------------------------------------------
1770    Name: Matthias Blume (blume (at) tti - c (dot) org)
1771    Date: 2003/09/17 16:30:00 CDT
1772    Tag: blume-20030917-real32-slices
1773    Description:
1774    
1775    Added an entry to the primitive environment
1776    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1777    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1778    
1779    Removed some of the "magic" constants in real64.sml and replaced them
1780    with code that generates these values from their corresponding
1781    integer counterparts.
1782    
1783    Made all(?) the slice-related changes to the Basis and made everything
1784    compile again...
1785    
1786    ----------------------------------------------------------------------
1787    Name: Matthias Blume (blume (at) tti - c (dot) org)
1788    Date: 2003/09/15 17:45:00 CDT
1789    Tag: blume-20030915-rbase
1790    Description:
1791    
1792    Fixed bug in Real.fromLargeInt.
1793    
1794    ----------------------------------------------------------------------
1795    Name: Matthias Blume (blume (at) tti - c (dot) org)
1796    Date: 2003/09/13 18:11:00 CDT
1797    Tag: blume-20030913-libinstall
1798    Description:
1799    
1800    Minor bugfix in config/libinstall (set anchor with path to
1801    standalone tool after installing it, otherwise libraries that
1802    need ml-lex or ml-yacc won't compile the first time the installer
1803    runs).
1804    
1805    ----------------------------------------------------------------------
1806    Name: Matthias Blume (blume (at) tti - c (dot) org)
1807    Date: 2003/09/12 11:45:00 CDT
1808    Tag: blume-20030912-various
1809    Description:
1810    
1811     - fixed bug in Real.toLargeInt
1812     - fixed bug in Posix.ProcEnv.times
1813     - changed inputLine functions to return an option
1814     - minor installer improvements / bugfixes
1815     - changed default @SMLalloc parameter for x86/celeron to 64k
1816    
1817    ----------------------------------------------------------------------
1818    Name: Matthias Blume (blume (at) tti - c (dot) org)
1819    Date: 2003/09/09 22:00:00 CDT
1820    Tag: Release_110_43
1821    Description:
1822    
1823        New working release 110.43.  New bootfiles.
1824    
1825    ----------------------------------------------------------------------
1826    Name: Matthias Blume (blume (at) tti - c (dot) org)
1827    Date: 2003/09/09 19:20:00 CDT
1828    Tag: blume-20030909-installer
1829    Description:
1830    
1831    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1832    Modified config/install.bat to take advantage of it.  Also modified
1833    config/install.sh (and called it config/new-install.sh) to take advantage
1834    of it on Unix systems.  (The SML code is (supposed to be) platform-
1835    independent.)
1836    
1837    The installer can now install everything under Win32
1838    as well as under *nix as long as it compiles.
1839    
1840    Other changes:
1841    
1842      - made CML compile again under Win32
1843      - made eXene compile under Win32 (by providing a fake structure UnixSock
1844        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1845      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1846        same as type int;  under Win32 it isn't
1847      - fixed some slice-related problems in the win32-specific parts of CML
1848      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1849        (under Win32, the volume name is case-insensitive, and the
1850        OS.Path code compares volume names for equality)
1851    
1852    ----------------------------------------------------------------------
1853    Name: Matthias Blume (blume (at) tti - c (dot) org)
1854    Date: 2003/09/08 11:55:00 CDT
1855    Tag: blume-20030908-fullpath
1856    Description:
1857    
1858    Made Win32 version of OS.FileSys.fullPath return current directory
1859    when given an empty string.  This is what the spec says, and incidentally,
1860    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1861    cases when presented with the name of a non-existing .cm file.)
1862    
1863    ----------------------------------------------------------------------
1864    Name: Matthias Blume (blume (at) tti - c (dot) org)
1865    Date: 2003/09/04 16:30:00 CDT
1866    Tag: blume-20030905-slices-etc
1867    Description:
1868    
1869      1. Changed interface to vectors and arrays in Basis to match
1870         (draft) Basis spec.
1871      2. Added signatures and implementations of slices according to
1872         Basis spec.
1873      3. Edited source code throughout the system to make it compile again
1874         under 1. and 2.  (In some cases code had to be added to have it
1875         match the new signatures.)
1876      4. MLRISC should be backward-compatible: the copies of the originals
1877         of files that needed to change under 3. were retained, the .cm files
1878         check the compiler version number and use old versions when
1879         appropriate.
1880      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1881         dirstream -> string option (in accordance with Basis spec).
1882      6. When generating code that counts lines, ml-lex used function
1883         CharVector.foldli, taking advantage of its old interface.
1884         This has been replaced with the corresponding code from
1885         CharVectorSlice.  (html-lex must be re-lexed!)
1886      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1887         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1888      8. Removed temporary additions (fromInternal, toInternal) from the
1889         (now obsolete) IntInf in smlnj-lib/Util.
1890      9. Cleaned up structure Byte.
1891     10. Added localOffset, scan, and fromString to Date (according to spec).
1892         Cleaned/corrected implementation of Date.
1893         (Still need to check for correctness; implement better canonicalizeDate.)
1894     11. Added "scan" to signature IEEE_REAL.
1895     12. Some improvements to IntInf [in particular: efficiency-hack for
1896         mod and rem when second operand is 2 (for parity checks).]
1897     13. Changed representation of type Time.time, using a single IntInf.int
1898         value counting microseconds.  This considerably simplified the
1899         implementation of structure Time.  We now support negative time
1900         values; scan and fromString handle signs.
1901     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1902         ArraySlice).
1903    
1904    ----------------------------------------------------------------------
1905    Name: Matthias Blume (blume (at) tti - c (dot) org)
1906    Date: 2003/08/28 17:00:00 CDT
1907    Tag: blume-20030828-intinf
1908    Description:
1909    
1910    This is a major update which comes with a version number bump
1911    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1912    and an implementation of IntInf in the Basis.
1913    
1914    There are a fairly large number of related changes and updates throughout
1915    the system:
1916    
1917    Basis:
1918            - Implemented IntInf.
1919            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1920            - Added some missing Real64 operations, most notably Real.toLargeInt.
1921            - Added FixedInt as a synonym for Int32.
1922    
1923    compiler:
1924            * Added support for a built-in intinf type.
1925              - literals
1926              - pattern matching
1927              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1928              - overloading on literals and operations
1929    
1930            This required adding a primitive type intinf, some additional
1931            primops, and implementations for several non-trivial intinf
1932            operations in Core.  (The intinf type is completely abstract
1933            to the compiler; all operations get delegated back to the Core.)
1934    
1935            * Intinf equality is handled by polyequal.  However, the compiler
1936            does not print its usual warning in this case (since polyequal
1937            is the right thing to do there).
1938    
1939            * Improved the organization of structure InlineT.
1940    
1941            * A word about conversion primops:
1942            If conversions involving intinf do not cancel out during
1943            CPS contract, then the compiler must insert calls to Core functions.
1944            Since all core access must be resolved already during the FLINT
1945            translate phase, it would be too late a the time of CPS contract
1946            to add new Core calls.  For this reason, conversion primops
1947            for intinf carry two arguments:  1. the numeric argument that
1948            they are supposed to convert, and 2. the Core function that
1949            can help with this conversion if necessary.  If CPS contract
1950            eliminates a primop, then the associated Core function becomes
1951            dead and goes away.  Intinf conversion primops that do not get
1952            eliminated by CPS contract get rewritten into calls of their
1953            core functions by a separate, new phase.
1954    
1955    interactive system:
1956            - Control.Print.intinfDepth controls max length of intinf constants
1957            being printed. (Analogous to Control.Print.stringDepth.)
1958            - Cleanup in printutil and pputil: got rid of unused stuff and
1959            duplicates; replaced some of the code with code that makes better
1960            use of library functionality.
1961    
1962    CM:
1963            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1964            are no longer silent.
1965    
1966    CKIT:
1967            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1968            in favor of LargeInt -- which is now the same as IntInf.
1969            CKIT-knowledgable people should check whether this is what's
1970            intended and otherwise change things back to using Int32 or
1971            FixedInt.
1972    
1973    Throughout the code:
1974            Started using IntInf.int literals and built-in operations
1975            (e.g., comparison with 0) where this seems appropriate.
1976    
1977    
1978    ----------------------------------------------------------------------
1979    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1980    Date: 2003/08/13 11:36:00 CDT
1981    Tag: dbm-20030813-mcz-merge1
1982    Description:
1983    
1984    Merging changes from the mcz-branch development branch into trunk.
1985    These changes involve replacement of the emulated old prettyprinter
1986    interface with direct use of the SML/NJ Lib PP library, and fixing
1987    of a couple of bugs (895, 1186) relating to error messages.  A new
1988    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1989    has been added.
1990    
1991    ----------------------------------------------------------------------
1992    Name: Matthias Blume (blume (at) tti - c (dot) org)
1993    Date: 2003/08/11 15:45:00 CDT
1994    Tag: blume-20030811-windows
1995    Description:
1996    
1997    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1998    
1999        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
2000    
2001    This patch restores SML/NJ's ability to run under win32.  There are a
2002    number of changes, including fixes for several bugs that had gone
2003    unnoticed until now:
2004    
2005      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
2006        trying to run the win32 version from within cygwin.)
2007    
2008      - There are a number of simple .bat scripts that substitute for their
2009        corresponding Unix shell-scripts.  (See below.)
2010    
2011      - The internals of ml-build have been modified slightly.  The main
2012        difference is that instead of calling ".link-sml" (or link-sml.bat)
2013        using OS.Process.system, the ML process delegates this task back
2014        to the script.  Otherwise problems arise in mixed environments such
2015        as Cygwin where scripts look and work like Unix scripts, but
2016        where OS.Process.system cannot run them.
2017    
2018      - In CM, the srcpath pickler used native pathname syntax -- which
2019        is incorrect in the case of cross-compilation.  The new pickle format
2020        is independent of platform-specific naming conventions.
2021    
2022      - Path configuration files (such as lib/pathconfig) can now choose
2023        between native and standard syntax.  Placing a line of the form
2024    
2025            standard!
2026    
2027        into the file causes all subsequent paths to be interpreted using
2028        CM standard pathname syntax (= Unix conventions); a line
2029    
2030            native!
2031    
2032        switches back to native style.  This was needed so that
2033        path config files can be written portably, see src/system/pathconfig.
2034    
2035      - Runtime system:
2036    
2037          - win32-filesys.c:  get_file_time and set_file_time now
2038            access modification time, not creation time.
2039    
2040          - I/O code made aware of new array representation.
2041    
2042          - Bug fixes in X86.prim.masm.
2043    
2044       - src/system/makeml made aware of win32. (For use under cygwin
2045            and other Unix-environments for windows.)
2046    
2047       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
2048         which caused BinIO.inputAll to fail consistently.
2049    
2050    .bat scripts:
2051    
2052       Windows .bat scripts assume that SMLNJ_HOME is defined.
2053    
2054       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
2055         applications (sml, ml-yacc, ml-lex).
2056       - ml-build.bat: analogous to ml-build.
2057       - config\install.bat:  Analogous to config/install.sh.  This requires
2058         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
2059         (nmake etc. must be on the path, and vcvars32 must have been run.)
2060         Moreover, sources for ml-lex and ml-yacc need to exist under src,
2061         and the bootfile hierarchy must have been unpacked under
2062         sml.boot.x86-win32.
2063         The script is very primitive and does a poor job at error checking.
2064         It only installs the base system, ml-lex, and ml-yacc.  No other
2065         libraries are being installed (i.e., you get only those that
2066         are part of the compiler.)
2067       - link-sml.bat:  analogous to .link-sml, but not currently used
2068    
2069    Unrelated bug fixes:
2070    
2071      - ml-nlffigen now exports structures ST_* corresponding to incomplete
2072        types.
2073      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
2074        otherwise.)
2075    
2076    ----------------------------------------------------------------------
2077    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
2078    Date: 2003/06/17
2079    Tag: macqueen-20030617-bug895
2080    Description:
2081    
2082    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
2083    Tag will be used for new development branch (mcz-branch) for
2084    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
2085    
2086    ----------------------------------------------------------------------
2087    Name: Matthias Blume (blume (at) tti - c (dot) org)
2088    Date: 2003/05/27 16:55:00 CDT
2089    Tag: blume-20030527-polyeq
2090    Description:
2091    
2092    Tried to eliminated most cases of polymorphic equality.
2093    
2094    ----------------------------------------------------------------------
2095    Name: Matthias Blume (blume (at) tti - c (dot) org)
2096    Date: 2003/05/21 17:45:00 CDT
2097    Tag: blume-20030517-complete
2098    Description:
2099    
2100    Two changes:
2101    
2102      1. Added a flag for controlling whether non-exhaustive bindings will
2103         be treated as errors (default is false).
2104      2. Cleaned up the *entire* source tree so that CMB.make goes through
2105         without a single non-exhaustive match- or bind warning.
2106    
2107    ----------------------------------------------------------------------
2108    Name: Matthias Blume (blume (at) tti - c (dot) org)
2109    Date: 2003/05/17 10:20:00 CDT
2110    Tag: blume-20030517-absyn
2111    Description:
2112    
2113    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
2114    
2115       This mainly affects the quality of error messages.  However, some
2116       of the code is now more straightforward than before. (Treatment of
2117       the above four constructs in translate.sml is much simpler than
2118       the "macro-expansion" that was going on before.  Plus, the mach-
2119       compiler no longer gets invoked just to be able to compile an
2120       if-expression.)
2121    
2122    2. The ErrorMsg.Error exception is now caught and absorbed by the
2123       interactive loop.
2124    
2125    ----------------------------------------------------------------------
2126    Name: Allen Leung
2127    Date: 2003/05/16 13:05:00 CDT
2128    Tag: leunga-20030516-cygwin-runtime
2129    Description:
2130    
2131        Ported the runtime system to cygwin, which uses the unix
2132    x86-unix bin files.  Missing/buggy features:
2133    
2134       o getnetbyname, getnetbyaddr: these functions seem to be missing in
2135         the Cygwin library.
2136       o Ctrl-C handling may be flaky.
2137       o Windows system calls and Windows I/O are not supported.
2138    
2139       A new set of binfiles is located at:
2140    
2141         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
2142    
2143       This is only needed for bootstrapping the cygwin version of smlnj.
2144       Other x86 versions can use the existing binfiles.
2145    
2146    ----------------------------------------------------------------------
2147    Name: Matthias Blume
2148    Date: 2003/04/08 15:42:00 CDT
2149    Tag: blume-20030408-listpair
2150    Description:
2151    
2152    1. Added a target 'mlrisc' to installer.
2153    
2154    2. Added missing elements to structure ListPair.
2155    
2156    ----------------------------------------------------------------------
2157    Name: Allen Leung
2158    Date: 2003/01/07 10:40:00 EST
2159    Tag: leunga-20030107-int-rem
2160    Description:
2161    
2162    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
2163    The arguments to the SUBL instruction were swapped.
2164    
2165    ----------------------------------------------------------------------
2166    Name: Matthias Blume
2167    Date: 2002/12/12 16:25:00 EST
2168    Tag: blume-20021212-risc-ra
2169    Description:
2170    
2171    Fixed a serious bug in the rewrite code for FP spilling/reloading that
2172    sent the RA into an infinite loop when floating point registers get
2173    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
2174    110.37 and 110.38.)
2175    There was another set of potential problems related to the handling of
2176    MLRISC annotations (but those did not yet cause real problems, apparently).
2177    
2178    ----------------------------------------------------------------------
2179    Name: Matthias Blume
2180    Date: 2002/12/06 22:40:00 EST
2181    Tag: blume-20021206-cm-fileid
2182    Description:
2183    
2184    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
2185    This fixes the problem of CM getting confused by files that suddenly
2186    change their identity (e.g., by getting unlinked and recreated by some
2187    text editor such as vi).  There might be a better/cheaper/cleaner way
2188    of doing this, but for now this will have to do.
2189    
2190    ----------------------------------------------------------------------
2191    Name: Matthias Blume
2192    Date: 2002/10/28 09:50:00 EST
2193    Tag: blume-20021028-typecheck
2194    Description:
2195    
2196    Exported structure Typecheck from $smlnj/viscomp/core.cm.
2197    
2198    ----------------------------------------------------------------------
2199    Name: Matthias Blume
2200    Date: 2002/10/17 09:10:00 EDT
2201    Tag: Release_110_42
2202    Description:
2203    
2204    In good old tradition, there has been a slight hiccup so that we have
2205    to patch 110.42 after the fact.   The old release tag has been replaced
2206    (see below).
2207    
2208    The change solves a problem with two competing approaches the
2209    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
2210    each other's way.
2211    
2212    This change only affects the runtime system code and the installer script.
2213    (No new bootfiles.)
2214    
2215    ----------------------------------------------------------------------
2216    Name: Matthias Blume
2217    Date: 2002/10/16 12:00:00 EDT
2218    Tag: Release_110_42_removed
2219    Description:
2220    
2221    New working release.  New bootfiles.
2222    
2223    ----------------------------------------------------------------------
2224    Name: Matthias Blume
2225    Date: 2002/10/10 13:10:00 EDT
2226    Tag: blume-20021010-ppc-divs
2227    Description:
2228    
2229    The mltree operator DIVS must be implemented with an overflow check on
2230    the PPC because the hardware indicates divide-by-zero using "overflow" as
2231    well.
2232    
2233    ----------------------------------------------------------------------
2234    Name: Matthias Blume
2235    Date: 2002/07/23 11:45:00
2236    Tag: blume-20020723-smlnj-home
2237    Description:
2238    
2239    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
2240    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
2241    CM_PATHCONFIG is also set), the path configuration file is assumed
2242    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
2243    the entire tree to some other place and everything will "just work".
2244    
2245    (Companion commands such as ml-build and ml-makedepend also sense this
2246    variable.)
2247    
2248    ----------------------------------------------------------------------
2249    Name: Matthias Blume
2250    Date: 2002/07/12 21:19:00 EDT
2251    Tag: blume-20020712-liveness
2252    Description:
2253    
2254    Exported two useful "step" functions from liveness module (MLRISC).
2255    
2256    ----------------------------------------------------------------------
2257    Name: Matthias Blume
2258    Date: 2002/07/05 16:00 EDT
2259    Tag: Release_110_41
2260    Description:
2261    
2262    New working release.  New bootfiles.
2263    
2264    ----------------------------------------------------------------------
2265    Name: Matthias Blume
2266    Date: 2002/07/05 10:25:00 EDT
2267    Tag: blume-20020705-btimp
2268    Description:
2269    
2270    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
2271    other clients can set up backtracing support.
2272    
2273    ----------------------------------------------------------------------
2274    Name: Matthias Blume
2275    Date: 2002/06/25 17:23:00 EDT
2276    Tag: blume-20020625-fpmax
2277    Description:
2278    
2279    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
2280    case.  (The sense of the isNaN test was reversed -- which made min and
2281    max always return their first argument.)
2282    
2283    ----------------------------------------------------------------------
2284    Name: Matthias Blume
2285    Date: 2002/06/11
2286    Tag: blume-20020611-unixpath
2287    Description:
2288    
2289    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
2290    
2291    ----------------------------------------------------------------------
2292    Name: Matthias Blume
2293    Date: 2002/06/10 16:35:00 EDT
2294    Tag: blume-20020610-ieeereal
2295    Description:
2296    
2297    I back-ported my implementation of IEEEReal.fromString from the
2298    idlbasis-devel branch so that we can test it.
2299    
2300    Another small change is that ppDec tries to give more information
2301    than just "<sig>" in the case of functors.  However, this code is
2302    broken in some mysterious way if the functor's body's signature
2303    has not been declared by ascription but gets inferred from the
2304    implementation.  This needs fixing...
2305    
2306    ----------------------------------------------------------------------
2307    Name: Matthias Blume
2308    Date: 2002/05/31
2309    Tag: blume-20020531-btrace-mode
2310    Description:
2311    
2312    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2313    the wayside when I switched over to using Controls everywhere.)
2314    
2315    ----------------------------------------------------------------------
2316    Name: Lal George
2317    Date: 2002/05/23 12:21:40 EDT
2318    Tag: george-20020523-visual-labels
2319    Description:
2320    
2321    Labels are now displayed in the graphical output to make
2322    the fall-through and target blocks obvious.
2323    
2324    ----------------------------------------------------------------------
2325    Name: Matthias Blume
2326    Date: 2002/05/22 11:03:00 EDT
2327    Tag: blume-20020522-shrink
2328    Description:
2329    
2330    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2331    situation that comes up when allocating huge arrays.
2332    
2333    ----------------------------------------------------------------------
2334    Name: Matthias Blume
2335    Date: 2002/05/21 16:00:00 EDT
2336    Tag: Release_110_40
2337    Description:
2338    
2339    New working release (110.40).  New bootfiles.
2340    
2341    [Also: John Reppy fixed GC bug 1131.]
2342    
2343    ----------------------------------------------------------------------
2344    Name: Matthias Blume
2345    Date: 2002/05/21 12:35:00 EDT
2346    Tag: blume-20020521-cmdoc
2347    Description:
2348    
2349    CM documentation update.
2350    
2351    ----------------------------------------------------------------------
2352    Name: Matthias Blume
2353    Date: 2002/05/21 10:55:00 EDT
2354    Tag: blume-20020521-misc
2355    Description:
2356    
2357    - John tweaked runtime to be silent on heap export (except when
2358      GC messages are on).
2359    - I added a few more things (cross-compiling versions of CMB) to
2360      config/preloads (as suggestions).
2361    
2362    ----------------------------------------------------------------------
2363    Name: Matthias Blume
2364    Date: 2002/05/20 22:25:00 EDT
2365    Tag: blume-20020520-controls
2366    Description:
2367    
2368    - Added ControlUtil structure to control-lib.cm.
2369    - Use it throughout.
2370    - Used Controls facility to define MLRISC controls (as opposed to
2371      registering MLRISC control ref cells with Controls after the fact)
2372    - Fixed messed-up controls priorities.
2373    
2374    * Removed again all the stuff from config/preloads that one wouldn't
2375      be able to preload at the time the initial heap image is built.
2376      (Many libraries, e.g., CML, do not exist yet at this time.  The
2377       only libraries that can be preloaded via config/preloads are those
2378       that come bundled with the bootfiles.)
2379    
2380    ----------------------------------------------------------------------
2381    Name: Matthias Blume
2382    Date: 2002/05/20 10:59:00 EDT
2383    Tag: blume-20020520-preloads
2384    Description:
2385    
2386    Added a lot of commented-out suggestions for things to be included
2387    in config/preloads.
2388    
2389    ----------------------------------------------------------------------
2390    Name: Allen Leung
2391    Date: 2002/05/18 14:20:00 EDT
2392    Tag: leunga-20020518-mdl
2393    Description:
2394    
2395    o Made the mdl tool stuff compile and run again.
2396    o I've disabled all the stuff that depends on RTL specifications; they
2397      are all badly broken anyway.
2398    
2399    ----------------------------------------------------------------------
2400    Name: Matthias Blume
2401    Date: 2002/05/17 16:49:00 EDT
2402    Tag: blume-20020517-controls
2403    Description:
2404    
2405    0. John Reppy made several modifications to the SML/NJ library.
2406       In particular, there is a shiny new controls-lib.cm.
2407    
2408    1. Pushed new controls interface through compiler so that everything
2409       compiles again.
2410    
2411    2. Added FormatComb and FORMAT_COMB to the CML version of the
2412       SML/NJ library (so that CML compiles again).
2413    
2414    3. Modified init scripts because XXX_DEFAULT environment variables
2415       are no longer with us.  (Boot-time initialization is now done
2416       using the same environment variables that are also used for
2417       startup-time initialization of controls.)
2418    
2419    ----------------------------------------------------------------------
2420    Name: Lal George
2421    Date: 2002/05/15 09:20:10 EDT
2422    Tag: george-20020515-pseudo-op-decls
2423    Description:
2424    
2425            All pseudo-ops emitted before the first segment declaration
2426    such as TEXT, DATA, and BSS directives are assumed to be global
2427    declarations and are emitted first in the assembly file. This is
2428    useful in a number of situations where one has pseudo-ops that are not
2429    specific to any segment, and also works around the constraint that one
2430    cannot have client pseudo-ops in the TEXT segment.
2431    
2432            Because no segment is associated with these declarations it is
2433    an error to allocate any space or objects before the first segment
2434    directive and an exception will be raised. However, we cannot make
2435    this check for client pseudo-ops.
2436    
2437            These top level declarations are a field in the CFG graph_info.
2438    In theory you can continue to add to this field after the CFG has been
2439    built -- provided you know what you are doing;-)
2440    
2441    ----------------------------------------------------------------------
2442    Name: Matthias Blume
2443    Date: 2002/05/13 16:40:00 EDT
2444    Tag: blume-20020513-pp-etc
2445    Description:
2446    
2447    A few minor bugfixes:
2448    
2449      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2450        (Bogus printouts for redefined bindings still occur.  Compiler
2451        bug should no longer occur now.  We need to redo the prettyprinter
2452        from scratch.)
2453    
2454      - CM pathname printer now also adds escape sequences for ( and )
2455    
2456      - commend and docu fixes for ml-nlffi
2457    
2458    ----------------------------------------------------------------------
2459    Name: Matthias Blume
2460    Date: 2002/05/10 16:40:00 EDT
2461    Tag: blume-20020510-erg-textio
2462    Description:
2463    
2464    Applied the following bugfix provided by Emden Gansner:
2465    
2466        Output is corrupted when outputSubstr is used rather than output.
2467    
2468        The problem occurs when a substring
2469    
2470            ss = (s, dataStart, dataLen)
2471    
2472        where dataStart > 0, fills a stream buffer with avail bytes left.
2473        avail bytes of s, starting at index dataStart, are copied into the
2474        buffer, the buffer is flushed, and then the remaining dataLen-avail
2475        bytes of ss are copied into the beginning of the buffer. Instead of
2476        starting this copy at index dataStart+avail in s, the current code
2477        starts the copy at index avail.
2478    
2479        Fix:
2480        In text-io-fn.sml, change line 695 from
2481             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2482        to
2483             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2484    
2485    ----------------------------------------------------------------------
2486    Name: Matthias Blume
2487    Date: 2002/04/12 13:55:00 EDT
2488    Tag: blume-20020412-assyntax
2489    Description:
2490    
2491    1. Grabbed newer assyntax.h from the XFree86 project.
2492    2. Fiddled with how to compile X86.prim.asm without warnings.
2493    3. (Very) Minor cleanup in CM.
2494    
2495    ----------------------------------------------------------------------
2496    Name: Matthias Blume
2497    Date: 2002/04/01 (no joke!) 17:07:00 EST
2498    Tag: blume-20020401-x86div
2499    Description:
2500    
2501    Added full support for div/mod/rem/quot on the x86, using the machine
2502    instruction's two results (without clumsily recomputing the remainder)
2503    directly where appropriate.
2504    
2505    Some more extensive power-of-two support was added to the x86 instruction
2506    selector (avoiding expensive divs, mods, and muls where they can be
2507    replaced with cheaper shifts and masks).  However, this sort of thing
2508    ought to be done earlier, e.g., within the CPS optimizer so that
2509    all architectures benefit from it.
2510    
2511    The compiler compiles to a fixed point, but changes might be somewhat
2512    fragile nevertheless.  Please, report any strange things that you might
2513    see wrt. div/mod/quot/rem...
2514    
2515    ----------------------------------------------------------------------
2516    Name: Matthias Blume
2517    Date: 2002/03/29 17:22:00
2518    Tag: blume-20020329-div
2519    Description:
2520    
2521    Fixed my broken div/mod logic.  Unfortunately, this means that the
2522    inline code for div/mod now has one more comparison than before.
2523    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2524    The problem was with quotient = 0, because that alone does not tell
2525    us which way the rounding went.  One then has to look at whether
2526    remainder and divisor have the same sign...  :(
2527    
2528    Anyway, I replaced the bootfiles with fresh ones...
2529    
2530    ----------------------------------------------------------------------
2531    Name: Matthias Blume
2532    Date: 2002/03/29 14:10:00 EST
2533    Tag: blume-20020329-inlprims
2534    Description:
2535    
2536    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2537    
2538    Primops have changed. This means that the bin/boot-file formats have
2539    changed as well.
2540    
2541    To make sure that there is no confusion, I made a new version.
2542    
2543    
2544    CHANGES:
2545    
2546    * removed REMT from mltree (remainder should never overflow).
2547    
2548    * added primops to deal with divisions of all flavors to the frontend
2549    
2550    * handled these primops all the way through so they map to their respective
2551      MLRISC support
2552    
2553    * used these primops in the implementation of Int, Int32, Word, Word32
2554    
2555    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2556    
2557    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2558    
2559    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2560      floating point
2561    
2562    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2563      and Real (but Real.abs maps to a separate floating-point-only primop)
2564    
2565    
2566    TODO items:
2567    
2568    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2569      instructions because the machine instruction encoder cannot handle
2570      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2571      REM in the first place, and REM is supported by the x86 machine coder.)
2572    
2573    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2574      selection phase.  (The two can be streamlined because the hardware
2575      delivers both quotient and remainder at the same time anyway.)
2576    
2577    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2578      (Currently the behavior is inconsistent both across architectures and
2579      wrt. the draft Basis spec.)
2580    
2581    * Word8 should eventually be handled natively, too.
2582    
2583    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2584      as if there currently is no execution path that could trigger it in
2585      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2586      hold for things like multiplication and division.)
2587    
2588    ----------------------------------------------------------------------
2589    Name: Matthias Blume
2590    Date: 2002/03/27 16:27:00 EST
2591    Tag: blume-20020327-mlrisc-divisions
2592    Description:
2593    
2594    Added support for all four division operations (ML's div, mod, quot,
2595    and rem) to MLRISC.  In the course of doing so, I also rationalized
2596    the naming (no more annoying switch-around of DIV and QUOT), by
2597    parameterizing the operation by div_rounding_mode (which can be either
2598    DIV_TO_ZERO or DIV_TO_NEGINF).
2599    
2600    The generic MLTreeGen functor takes care of compiling all four
2601    operations down to only round-to-zero div.
2602    
2603    Missing pieces:
2604    
2605      * Doing something smarter than relying on MLTreeGen on architectures
2606        like, e.g., the x86 where hardware division delivers both quotient and
2607        remainder at the same time.  With this, the implementation of the
2608        round-to-neginf operations could be further streamlined.
2609    
2610      * Remove inlining support for div/mod/rem from the frontend and replace it
2611        with primops that get carried through to the backend.  Do this for all
2612        int and word types.
2613    
2614    ----------------------------------------------------------------------
2615    Name: Matthias Blume
2616    Date: 2002/03/25 17:25:00 EST
2617    Tag: blume-20020325-divmod
2618    Description:
2619    
2620    I improved (hopefully without breaking them) the implementation of Int.div,
2621    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2622    advantage of the following observations:
2623    
2624      Let  q = x quot y      r = x rem y
2625           d = x div  y      m = x mod y
2626    
2627    where "quot" is the round-to-zero version of integer division that
2628    hardware usually provides.  Then we have:
2629    
2630         r = x - q * y        where neither the * nor the - will overflow
2631         d = if q >= 0 orelse x = q * y then q else q - 1
2632                              where neither the * nor the - will overflow
2633         m = if q >= 0 orelse r = 0 then r else r + y
2634                              where the + will not overflow
2635    
2636    This results in substantial simplification of the generated code.
2637    The following table shows the number of CFG nodes and edges generated
2638    for
2639            fun f (x, y) = x OPER y
2640            (* with OPER \in div, mod, quot, rem *)
2641    
2642    
2643        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2644        --------------------------------------------------------
2645         div |         24 |         39 |         12 |         16
2646         mod |         41 |         71 |         12 |         16
2647        quot |          8 |         10 |          8 |         10
2648         rem |         10 |         14 |          8 |         10
2649    
2650    
2651    ----------------------------------------------------------------------
2652    Name: Matthias Blume
2653    Date: 2002/03/25 22:06:00 EST
2654    Tag: blume-20020325-cprotobug
2655    Description:
2656    
2657    Fixed a bug in cproto (c prototype decoder).
2658    
2659    ----------------------------------------------------------------------
2660    Name: Matthias Blume
2661    Date: 2002/03/25 16:00:00 EST
2662    Tag: blume-20020325-raw-primops
2663    Description:
2664    
2665    I did some cleanup to Allen's new primop code and
2666    replaced yesterday's bootfiles with new ones.
2667    (But they are stored in the same place.)
2668    
2669    ----------------------------------------------------------------------
2670    Name: Matthias Blume
2671    Date: 2002/03/24 22:40:00 EST
2672    Tag: blume-20020324-bootfiles
2673    Description:
2674    
2675    Made the bootfiles that Allen asked for.
2676    
2677    ----------------------------------------------------------------------
2678    Name: Allen Leung
2679    Date: 2002/03/23 15:50:00 EST
2680    Tag: leunga-20020323-flint-cps-rcc-primops
2681    Description:
2682    
2683      1. Changes to FLINT primops:
2684    
2685        (* make a call to a C-function;
2686         * The primop carries C function prototype information and specifies
2687         * which of its (ML-) arguments are floating point. C prototype
2688         * information is for use by the backend, ML information is for
2689         * use by the CPS converter. *)
2690      | RAW_CCALL of { c_proto: CTypes.c_proto,
2691                       ml_args: ccall_type list,
2692                       ml_res_opt: ccall_type option,
2693                       reentrant : bool
2694                     } option
2695       (* Allocate uninitialized storage on the heap.
2696        * The record is meant to hold short-lived C objects, i.e., they
2697        * are not ML pointers.  With the tag, the representation is
2698        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2699        *)
2700      | RAW_RECORD of {tag:bool,sz:int}
2701      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2702    
2703      2.  These CPS primops are now overloaded:
2704    
2705           rawload of {kind:numkind}
2706           rawstore of {kind:numkind}
2707    
2708          The one argument form is:
2709    
2710             rawload {kind} address
2711    
2712          The two argument form is:
2713    
2714             rawload {kind} [ml object, byte-offset]
2715    
2716      3. RAW_CCALL/RCC now takes two extra arguments:
2717    
2718         a. The first is whether the C call is reentrant, i.e., whether
2719            ML state should be saved and restored.
2720         b. The second argument is a string argument specifying the name of
2721            library and the C function.
2722    
2723         These things are currently not handled in the code generator, yet.
2724    
2725      4. In CProto,
2726    
2727         An encoding type of "bool" means "ml object" and is mapped into
2728         C prototype of PTR.  Note that "bool" is different than "string",
2729         even though "string" is also mapped into PTR, because "bool"
2730         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2731    
2732      5. Pickler/unpicker
2733    
2734         Changed to handle RAW_RECORD and newest RAW_CCALL
2735    
2736      6. MLRiscGen,
2737    
2738         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2739         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2740            in the file CodeGen/cpscompile/cps-c-calls.sml
2741    
2742      7. Added the conditional move operator
2743    
2744             condmove of branch
2745    
2746         to cps.  Generation of this is still buggy so it is currently
2747         disabled.
2748    
2749    ----------------------------------------------------------------------
2750    Name: Lal George
2751    Date: 2002/03/22 14:18:25 EST
2752    Tag: george-20020322-cps-branch-prob
2753    Description:
2754    
2755    Implemented the Ball-Larus branch prediction-heuristics, and
2756    incorporated graphical viewers for control flow graphs.
2757    
2758    Ball-Larus Heuristics:
2759    ---------------------
2760    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2761    
2762    By design it uses the Dempster-Shafer theory for combining
2763    probabilities.  For example, in the function:
2764    
2765        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2766    
2767    the ball-larus heuristics predicts that the n=0 is unlikely
2768    (OH-heuristic), and the 'then' branch is unlikely because of the
2769    RH-heuristic -- giving the 'then' branch an even lower combined
2770    probability using the Dempster-Shafer theory.
2771    
2772    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2773    probability of the 'then' branch because of the loop in the else
2774    branch.
2775    
2776    
2777    Graphical Viewing:
2778    ------------------
2779    I merely plugged in Allen's graphical viewers into the compiler. The
2780    additional code is not much. At the top level, saying:
2781    
2782            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2783    
2784    will display the graphical view of the control flow graph just before
2785    back-patching.  daVinci must be in your path for this to work. If
2786    daVinci is not available, then the default viewer can be changed
2787    using:
2788    
2789            Control.MLRISC.getString "viewer"
2790    
2791    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2792    course, these viewers must be in your path.
2793    
2794    The above will display the compilation unit at the level of clusters,
2795    many of which are small, boring, and un-interesting. Also setting:
2796    
2797            Control.MLRISC.getInt "cfg-graphical-view_size"
2798    
2799    will display clusters that are larger than the value set by the above.
2800    
2801    
2802    ----------------------------------------------------------------------
2803    Name: Matthias Blume
2804    Date: 2002/03/21 22:20:00 EST
2805    Tag: blume-20020321-kmp-bugfix
2806    Description:
2807    
2808    Changed the interface to the KMP routine in PreString and fixed
2809    a minor bug in one place where it was used.
2810    
2811    ----------------------------------------------------------------------
2812    Name: Allen Leung
2813    Date: 2002/03/21 20:30:00 EST
2814    Tag: leunga-20020321-cfg
2815    Description:
2816    
2817      Fixed a potential problem in cfg edge splitting.
2818    
2819    ----------------------------------------------------------------------
2820    Name: Allen Leung
2821    Date: 2002/03/21 17:15:00 EST
2822    Tag: leunga-20020321-x86-fp-cfg
2823    Description:
2824    
2825      1. Recoded the buggy parts of x86-fp.
2826    
2827         a. All the block reordering code has been removed.
2828            We now depend on the block placement phases to do this work.
2829    
2830         b. Critical edge splitting code has been simplified and moved into the
2831            CFG modules, as where they belong.
2832    
2833         Both of these were quite buggy and complex.  The code is now much, much
2834         simpler.
2835    
2836      2. X86 backend.
2837    
2838         a. Added instructions for 64-bit support.  Instruction selection for
2839            64-bit has not been committed, however, since that
2840            requires changes to MLTREE which haven't been approved by
2841            Lal and John.
2842    
2843         b. Added support for FUCOMI and FUCOMIP when generating code for
2844            PentiumPro and above.  We only generate these instructions in
2845            the fast-fp mode.
2846    
2847         c. Added cases for JP and JNP in X86FreqProps.
2848    
2849      3. CFG
2850    
2851         CFG now has a bunch of methods for edge splitting and merging.
2852    
2853      4. Machine description.
2854    
2855         John's simplification of MLTREE_BASIS.fcond broke a few machine
2856         description things:
2857    
2858         rtl-build.{sig,sml} and hppa.mdl fixed.
2859    
2860         NOTE: the machine description stuff in the repository is still broken.
2861               Again, I can't put my fixes in because that involves
2862               changes to MLTREE.
2863    
2864    ----------------------------------------------------------------------
2865    Name: Matthias Blume
2866    Date: 2002/03/20 15:55:00 EST
2867    Tag: blume-20020320-kmp
2868    Description:
2869    
2870    Implemented Knuth-Morris-Pratt string matching in PreString and used
2871    it for String.isSubstring, Substring.isSubstring, and
2872    Substring.position.
2873    
2874    (Might need some stress-testing.  Simple examples worked fine.)
2875    
2876    ----------------------------------------------------------------------
2877    Name: Matthias Blume
2878    Date: 2002/03/19 16:37:00 EST
2879    Tag: blume-20020319-witnesses
2880    Description:
2881    
2882    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2883    
2884    This implements a generic mechanism for changing constness qualifiers
2885    anywhere within big C types without resorting to outright "casts".
2886    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2887    modify the constness at the outermost level.)
2888    The implementation of "convert" is based on the idea of "witness"
2889    values -- values that are not used by the operation but whose types
2890    "testify" to their applicability.  On the implementation side, "convert"
2891    is simply a projection (returning its second curried argument).  With
2892    cross-module inlining, it should not result in any machine code being
2893    generated.
2894    
2895    ----------------------------------------------------------------------
2896    Name: Matthias Blume
2897    Date: 2002/03/15 16:40:00 EST
2898    Tag: blume-20020315-basis
2899    Description:
2900    
2901    Provided (preliminary?) implementations for
2902    
2903      {String,Substring}.{concatWith,isSuffix,isSubstring}
2904    
2905    and
2906    
2907      Substring.full
2908    
2909    Those are in the Basis spec but they were missing in SML/NJ.
2910    
2911    ----------------------------------------------------------------------
2912    Name: Matthias Blume
2913    Date: 2002/03/14 21:30:00 EST
2914    Tag: blume-20020314-controls
2915    Description:
2916    
2917    Controls:
2918    ---------
2919    
2920    1. Factored out the recently-added Controls : CONTROLS stuff and put
2921       it into its own library $/controls-lib.cm.  The source tree for
2922       this is under src/smlnj-lib/Controls.
2923    
2924    2. Changed the names of types and functions in this interface, so they
2925       make a bit more "sense":
2926    
2927          module -> registry
2928          'a registry -> 'a group
2929    
2930    3. The interface now deals in ref cells only.  The getter/setter interface
2931       is (mostly) gone.
2932    
2933    4. Added a function that lets one register an already-existing ref cell.
2934    
2935    5. Made the corresponding modifications to the rest of the code so that
2936       everything compiles again.
2937    
2938    6. Changed the implementation of Controls.MLRISC back to something closer
2939       to the original.  In particular, this module (and therefore MLRISC)
2940       does not depend on Controls.  There now is some link-time code in
2941       int-sys.sml that registers the MLRISC controls with the Controls
2942       module.
2943    
2944    CM:
2945    ---
2946    
2947      * One can now specify the lambda-split aggressiveness in init.cmi.
2948    
2949    ----------------------------------------------------------------------
2950    Name: Allen Leung
2951    Date: 2002/03/13 17:30:00 EST
2952    Tag: leunga-20020313-x86-fp-unary
2953    Description:
2954    
2955    Bug fix for:
2956    
2957    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2958    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2959    > - fun f(x,(y,z)) = Real.~ y;
2960    > [autoloading]
2961    > [autoloading done]
2962    >       fchsl   (%eax), 184(%esp)
2963    > Error: MLRisc bug: X86MCEmitter.emitInstr
2964    >
2965    > uncaught exception Error
2966    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2967    
2968    The problem was that the code generator did not generate any fp registers
2969    in this case, and the ra didn't know that it needed to run the X86FP phase to
2970    translate the pseudo fp instruction.   This only happened with unary fp
2971    operators in certain situations.
2972    
2973    ----------------------------------------------------------------------
2974    Name: Matthias Blume
2975    Date: 2002/03/13 14:00:00 EST
2976    Tag: blume-20020313-overload-etc
2977    Description:
2978    
2979    1. Added _overload as a synonym for overload for backward compatibility.
2980       (Control.overloadKW must be true for either version to be accepted.)
2981    
2982    2. Fixed bug in install script that caused more things to be installed
2983       than what was requested in config/targets.
2984    
2985    3. Made CM aware of the (_)overload construct so that autoloading
2986       works.
2987    
2988    ----------------------------------------------------------------------
2989    Name: Matthias Blume
2990    Date: 2002/03/12 22:03:00 EST
2991    Tag: blume-20020312-url
2992    Description:
2993    
2994    Forgot to update BOOT and srcarchiveurl.
2995    
2996    ----------------------------------------------------------------------
2997    Name: Matthias Blume
2998    Date: 2002/03/12 17:30:00 EST
2999    Tag: blume-20020312-version110392
3000    Description:
3001    
3002    Yet another version number bump (because of small changes to the
3003    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
3004    
3005    Changes:
3006    
3007      The new pid generation scheme described a few weeks ago was overly
3008      complicated.  I implemented a new mechanism that is simpler and
3009      provides a bit more "stability":  Once CM has seen a compilation
3010      unit, it keeps its identity constant (as long as you do not delete
3011      those crucial CM/GUID/* files).  This means that when you change
3012      an interface, compile, then go back to the old interface, and
3013      compile again, you arrive at the original pid.
3014    
3015      There now also is a mechanism that instructs CM to use the plain
3016      environment hash as a module's pid (effectively making its GUID
3017      the empty string).  For this, "noguid" must be specified as an
3018      option to the .sml file in question within its .cm file.
3019      This is most useful for code that is being generated by tools such
3020      as ml-nlffigen (because during development programmers tend to
3021      erase the tool's entire output directory tree including CM's cached
3022      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
3023      revert to the old, broken behavior of SML/NJ, but in specific cases
3024      where there is no danger of interface confusion, its use is ok
3025      (I think).
3026    
3027      ml-nlffigen by default generates "noguid" annotations.  They can be
3028      turned off by specifying -guid in its command line.
3029    
3030    ----------------------------------------------------------------------
3031    Name: Lal George
3032    Date: 2002/03/12 12 14:42:36 EST
3033    Tag: george-20020312-frequency-computation
3034    Description:
3035    
3036    Integrated jump chaining and static block frequency into the
3037    compiler. More details and numbers later.
3038    
3039    ----------------------------------------------------------------------
3040    Name: Lal George
3041    Date: 2002/03/11 11 22:38:53 EST
3042    Tag: george-20020311-jump-chain-elim
3043    Description:
3044    
3045    Tested the jump chain elimination on all architectures (except the
3046    hppa).  This is on by default right now and is profitable for the
3047    alpha and x86, however, it may not be profitable for the sparc and ppc
3048    when compiling the compiler.
3049    
3050    The gc test will typically jump to a label at the end of the cluster,
3051    where there is another jump to an external cluster containing the actual
3052    code to invoke gc. This is to allow factoring of common gc invocation
3053    sequences. That is to say, we generate:
3054    
3055            f:
3056               testgc
3057               ja   L1      % jump if above to L1
3058    
3059            L1:
3060               jmp L2
3061    
3062    
3063    After jump chain elimination the 'ja L1' instructions is converted to
3064    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
3065    up being implemented in their long form (if L2 is far away) using:
3066    
3067            jbe     L3      % jump if below or equal to L3
3068            jmp     L2
3069         L3:
3070            ...
3071    
3072    
3073    For large compilation units L2  may be far away.
3074    
3075    
3076    ----------------------------------------------------------------------
3077    Name: Matthias Blume
3078    Date: 2002/03/11 13:30:00 EST
3079    Tag: blume-20020311-mltreeeval
3080    Description:
3081    
3082    A functor parameter was missing.
3083    
3084    ----------------------------------------------------------------------
3085    Name: Allen Leung
3086    Date: 2002/03/11 10:30:00 EST
3087    Tag: leunga-20020311-runtime-string0
3088    Description:
3089    
3090       The representation of the empty string now points to a
3091    legal null terminated C string instead of unit.  It is now possible
3092    to convert an ML string into C string with InlineT.CharVector.getData.
3093    This compiles into one single machine instruction.
3094    
3095    ----------------------------------------------------------------------
3096    Name: Allen Leung
3097    Date: 2002/03/10 23:55:00 EST
3098    Tag: leunga-20020310-x86-call
3099    Description:
3100    
3101       Added machine generation for CALL instruction (relative displacement mode)
3102    
3103    ----------------------------------------------------------------------
3104    Name: Matthias Blume
3105    Date: 2002/03/08 16:05:00
3106    Tag: blume-20020308-entrypoints
3107    Description:
3108    
3109    Version number bumped to 110.39.1.  NEW BOOTFILES!
3110    
3111    Entrypoints: non-zero offset into a code object where execution should begin.
3112    
3113    - Added the notion of an entrypoint to CodeObj.
3114    - Added reading/writing of entrypoint info to Binfile.
3115    - Made runtime system bootloader aware of entrypoints.
3116    - Use the address of the label of the first function given to mlriscGen
3117      as the entrypoint.  This address is currently always 0, but it will
3118      not be 0 once we turn on block placement.
3119    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
3120      with entry points) from mlriscGen.
3121    
3122    ----------------------------------------------------------------------
3123    Name: Allen Leung
3124    Date: 2002/03/07 20:45:00 EST
3125    Tag: leunga-20020307-x86-cmov
3126    Description:
3127    
3128       Bug fixes for CMOVcc on x86.
3129    
3130       1. Added machine code generation for CMOVcc
3131       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
3132       3. CMOVcc cannot have an immediate operand as argument.
3133    
3134    ----------------------------------------------------------------------
3135    Name: Matthias Blume
3136    Date: 2002/03/07 16:15:00 EST
3137    Tag: blume-20020307-controls
3138    Description:
3139    
3140    This is a very large but mostly boring patch which makes (almost)
3141    every tuneable compiler knob (i.e., pretty much everything under
3142    Control.* plus a few other things) configurable via both the command
3143    line and environment variables in the style CM did its configuration
3144    until now.
3145    
3146    Try starting sml with '-h' (or, if you are brave, '-H')
3147    
3148    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
3149    implements the underlying generic mechanism.
3150    
3151    The interface to some of the existing such facilities has changed somewhat.
3152    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
3153    (The getFoo interface is still there for backward-compatibility, but its
3154    use is deprecated.)
3155    
3156    The ml-build script passes -Cxxx=yyy command-line arguments through so
3157    that one can now twiddle the compiler settings when using this "batch"
3158    compiler.
3159    
3160    TODO items:
3161    
3162    We should go through and throw out all controls that are no longer
3163    connected to anything.  Moreover, we should go through and provide
3164    meaningful (and correct!) documentation strings for those controls
3165    that still are connected.
3166    
3167    Currently, multiple calls to Controls.new are accepted (only the first
3168    has any effect).  Eventually we should make sure that every control
3169    is being made (via Controls.new) exactly once.  Future access can then
3170    be done using Controls.acc.
3171    
3172    Finally, it would probably be a good idea to use the getter-setter
3173    interface to controls rather than ref cells.  For the time being, both
3174    styles are provided by the Controls module, but getter-setter pairs are
3175    better if thread-safety is of any concern because they can be wrapped.
3176    
3177    *****************************************
3178    
3179    One bug fix: The function blockPlacement in three of the MLRISC
3180    backpatch files used to be hard-wired to one of two possibilities at
3181    link time (according to the value of the placementFlag).  But (I
3182    think) it should rather sense the flag every time.
3183    
3184    *****************************************
3185    
3186    Other assorted changes (by other people who did not supply a HISTORY entry):
3187    
3188    1. the cross-module inliner now works much better (Monnier)
3189    2. representation of weights, frequencies, and probabilities in MLRISC
3190       changed in preparation of using those for weighted block placement
3191       (Reppy, George)
3192    
3193    ----------------------------------------------------------------------
3194    Name: Lal George
3195    Date: 2002/03/07 14:44:24 EST 2002
3196    Tag: george-20020307-weighted-block-placement
3197    
3198    Tested the weighted block placement optimization on all architectures
3199    (except the hppa) using AMPL to generate the block and edge frequencies.
3200    Changes were required in the machine properties to correctly
3201    categorize trap instructions. There is an MLRISC flag
3202    "weighted-block-placement" that can be used to enable weighted block
3203    placement, but this will be ineffective without block/edge
3204    frequencies (coming soon).
3205    
3206    
3207    ----------------------------------------------------------------------
3208    Name: Lal George
3209    Date: 2002/03/05 17:24:48 EST
3210    Tag: george-20020305-linkage-cluster
3211    
3212    In order to support the block placement optimization, a new cluster
3213    is generated as the very first cluster (called the linkage cluster).
3214    It contains a single jump to the 'real' entry point for the compilation
3215    unit. Block placement has no effect on the linkage cluster itself, but
3216    all the other clusters  have full freedom in the manner in which they
3217    reorder blocks or functions.
3218    
3219    On the x86 the typical linkage code that is generated is:
3220       ----------------------
3221            .align 2
3222       L0:
3223            addl    $L1-L0, 72(%esp)
3224            jmp     L1
3225    
3226    
3227            .align  2
3228       L1:
3229       ----------------------
3230    
3231    72(%esp) is the memory location for the stdlink register. This
3232    must contain the address of the CPS function being called. In the
3233    above example, it contains the address of  L0; before
3234    calling L1 (the real entry point for the compilation unit), it
3235    must contain the address for L1, and hence
3236    
3237            addl $L1-L0, 72(%esp)
3238    
3239    I have tested this on all architectures except the hppa.The increase
3240    in code size is of course negligible
3241    
3242    ----------------------------------------------------------------------
3243    Name: Allen Leung
3244    Date: 2002/03/03 13:20:00 EST
3245    Tag: leunga-20020303-mlrisc-tools
3246    
3247      Added #[ ... ] expressions to mlrisc tools
3248    
3249    ----------------------------------------------------------------------
3250    Name: Matthias Blume
3251    Date: 2002/02/27 12:29:00 EST
3252    Tag: blume-20020227-cdebug
3253    Description:
3254    
3255    - made types in structure C and C_Debug to be equal
3256    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
3257    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
3258    
3259    ----------------------------------------------------------------------
3260    Name: Matthias Blume
3261    Date: 2002/02/26 12:00:00 EST
3262    Tag: blume-20020226-ffi
3263    Description:
3264    
3265    1. Fixed a minor bug in CM's "noweb" tool:
3266       If numbering is turned off, then truly don't number (i.e., do not
3267       supply the -L option to noweb).  The previous behavior was to supply
3268       -L'' -- which caused noweb to use the "default" line numbering scheme.
3269       Thanks to Chris Richards for pointing this out (and supplying the fix).
3270    
3271    2. Once again, I reworked some aspects of the FFI:
3272    
3273       A. The incomplete/complete type business:
3274    
3275       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
3276         gone!
3277       - ML types representing an incomplete type are now *equal* to
3278         ML types representing their corresponding complete types (just like
3279         in C).  This is still safe because ml-nlffigen will not generate
3280         RTTI for incomplete types, nor will it generate functions that
3281         require access to such RTTI.   But when ML code generated from both
3282         incomplete and complete versions of the C type meet, the ML types
3283         are trivially interoperable.
3284    
3285         NOTE:  These changes restore the full generality of the translation
3286         (which was previously lost when I eliminated functorization)!
3287    
3288       B. Enum types:
3289    
3290       - Structure C now has a type constructor "enum" that is similar to
3291         how the "su" constructor works.  However, "enum" is not a phantom
3292         type because each "T enum" has values (and is isomorphic to
3293         MLRep.Signed.int).
3294       - There are generic access operations for enum objects (using
3295         MLRep.Signed.int).
3296       - ml-nlffigen will generate a structure E_foo for each "enum foo".
3297         * The structure contains the definition of type "mlrep" (the ML-side
3298         representation type of the enum).  Normally, mlrep is the same
3299         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3300         then mlrep will be defined as a datatype -- thus facilitating
3301         pattern matching on mlrep values.
3302         ("-ec" will be suppressed if there are duplicate values in an
3303          enumeration.)
3304         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3305         will be generated for each C enum constant xxx.
3306         * Conversion functions m2i and i2m convert between mlrep and
3307         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
3308         * Coversion functions c and ml convert between mlrep and "tag enum".
3309         * Access functions (get/set) fetch and store mlrep values.
3310       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3311         enumerations are merged into one single enumeration represented by
3312         structure E_'.
3313    
3314    ----------------------------------------------------------------------
3315    Name: Allen Leung
3316    Date: 2002/02/25 04:45:00 EST
3317    Tag: leunga-20020225-cps-spill
3318    
3319    This is a new implementation of the CPS spill phase.
3320    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3321    In case of problems, replace it with the old file spill.sml
3322    
3323    The current compiler runs into some serious performance problems when
3324    constructing a large record.  This can happen when we try to compile a
3325    structure with many items.  Even a very simple structure like the following
3326    makes the compiler slow down.
3327    
3328        structure Foo = struct
3329           val x_1 = 0w1 : Word32.int
3330           val x_2 = 0w2 : Word32.int
3331           val x_3 = 0w3 : Word32.int
3332           ...
3333           val x_N = 0wN : Word32.int
3334        end
3335    
3336    The following table shows the compile time, from N=1000 to N=4000,
3337    with the old compiler:
3338    
3339    N
3340    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3341           MLRISC ra                               0.06u  0.00s  0.05g
3342              (spills = 0 reloads = 0)
3343           TOTAL                                   0.63u  0.07s  0.21g
3344    
3345    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3346           MLRISC ra                               5.68u  0.59s  3.93g
3347              (spills = 0 reloads = 0)
3348           TOTAL                                   14.71u  0.99s  4.81g
3349    
3350    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3351           MLRISC ra                               5.54u  0.65s  3.91g
3352              (spills = 543 reloads = 1082)
3353           TOTAL                                   65.40u  3.13s  6.00g
3354    
3355    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3356           MLRISC ra                               0.80u  0.10s  0.55g
3357              (spills = 42 reloads = 84)
3358           TOTAL                                   129.42u  5.10s  4.13g
3359    
3360    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3361           MLRISC ra                               2.69u  0.27s  1.38g
3362              (spills = 62 reloads = 124)
3363           TOTAL                                   682.48u  19.61s  13.99g
3364    
3365    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3366           MLRISC ra                               4.96u  0.27s  2.72g
3367              (spills = 85 reloads = 170)
3368           TOTAL                                   2375.26u  57.21s  48.00g
3369    
3370    As you can see the old cps spill module suffers from some serious
3371    performance problem.  But since I cannot decipher the old code fully,
3372    instead of patching the problems up, I'm reimplementing it
3373    with a different algorithm.  The new code is more modular,
3374    smaller when compiled, and substantially faster
3375    (O(n log n) time and O(n) space).  Timing of the new spill module:
3376    
3377    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3378          MLRISC ra                               0.25u  0.02s  0.15g
3379             (spills=1 reloads=3)
3380          TOTAL                                   7.74u  0.34s  1.62g
3381    
3382    Implementation details:
3383    
3384    As far as I can tell, the purpose of the CPS spill module is to make sure the
3385    number of live variables at any program point (the bandwidth)
3386    does not exceed a certain limit, which is determined by the
3387    size of the spill area.
3388    
3389    When the bandwidth is too large, we decrease the register pressure by
3390    packing live variables into spill records.  How we achieve this is
3391    completely different than what we did in the old code.
3392    
3393    First, there is something about the MLRiscGen code generator
3394    that we should be aware of:
3395    
3396    o MLRiscGen performs code motion!
3397    
3398       In particular, it will move floating point computations and
3399       address computations involving only the heap pointer to
3400       their use sites (if there is only a single use).
3401       What this means is that if we have a CPS record construction
3402       statement
3403    
3404           RECORD(k,vl,w,e)
3405    
3406       we should never count the new record address w as live if w
3407       has only one use (which is often the case).
3408    
3409       We should do something similar to floating point, but the transformation
3410       there is much more complex, so I won't deal with that.
3411    
3412    Secondly, there are now two new cps primops at our disposal:
3413    
3414     1. rawrecord of record_kind option
3415        This pure operator allocates some uninitialized storage from the heap.
3416        There are two forms:
3417    
3418         rawrecord NONE [INT n]  allocates a tagless record of length n
3419         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3420                                     and initializes the tag.
3421    
3422     2. rawupdate of cty
3423          rawupdate cty (v,i,x)
3424          Assigns to x to the ith component of record v.
3425          The storelist is not updated.
3426    
3427    We use these new primops for both spilling and increment record construction.
3428    
3429     1. Spilling.
3430    
3431        This is implemented with a linear scan algorithm (but generalized
3432        to trees).  The algorithm will create a single spill record at the
3433        beginning of the cps function and use rawupdate to spill to it,
3434        and SELECT or SELp to reload from it.  So both spills and reloads
3435        are fine-grain operations.  In contrast, in the old algorithm
3436        "spills" have to be bundled together in records.
3437    
3438        Ideally, we should sink the spill record construction to where
3439        it is needed.  We can even split the spill record into multiple ones
3440        at the places where they are needed.  But CPS is not a good
3441        representation for global code motion, so I'll keep it simple and
3442        am not attempting this.
3443    
3444     2. Incremental record construction (aka record splitting).
3445    
3446        Long records with many component values which are simulatenously live
3447        (recall that single use record addresses are not considered to
3448         be live) are constructed with rawrecord and rawupdate.
3449        We allocate space on the heap with rawrecord first, then gradually
3450        fill it in with rawupdate.  This is the technique suggested to me
3451        by Matthias.
3452    
3453        Some restrictions on when this is applicable:
3454        1. It is not a VECTOR record.  The code generator currently does not handle
3455           this case. VECTOR record uses double indirection like arrays.
3456        2. All the record component values are defined in the same "basic block"
3457           as the record constructor.  This is to prevent speculative
3458           record construction.
3459    
3460    ----------------------------------------------------------------------
3461    Name: Allen Leung
3462    Date: 2002/02/22 01:02:00 EST
3463    Tag: leunga-20020222-mlrisc-tools
3464    
3465    Minor bug fixes in the parser and rewriter
3466    
3467    ----------------------------------------------------------------------
3468    Name: Allen Leung
3469    Date: 2002/02/21 20:20:00 EST
3470    Tag: leunga-20020221-peephole
3471    
3472    Regenerated the peephole files.  Some contained typos in the specification
3473    and some didn't compile because of pretty printing bugs in the old version
3474    of 'nowhere'.
3475    
3476    ----------------------------------------------------------------------
3477    Name: Allen Leung
3478    Date: 2002/02/19 20:20:00 EST
3479    Tag: leunga-20020219-mlrisc-tools
3480    Description:
3481    
3482       Minor bug fixes to the mlrisc-tools library:
3483    
3484       1.  Fixed up parsing colon suffixed keywords
3485       2.  Added the ability to shut the error messages up
3486       3.  Reimplemented the pretty printer and fixed up/improved
3487           the pretty printing of handle and -> types.
3488       4.  Fixed up generation of literal symbols in the nowhere tool.
3489       5.  Added some SML keywords to to sml.sty
3490    
3491    ----------------------------------------------------------------------
3492    Name: Matthias Blume
3493    Date: 2002/02/19 16:20:00 EST
3494    Tag: blume-20020219-cmffi
3495    Description:
3496    
3497    A wild mix of changes, some minor, some major:
3498    
3499    * All C FFI-related libraries are now anchored under $c:
3500        $/c.cm      --> $c/c.cm
3501        $/c-int.cm  --> $c/internals/c-int.cm
3502        $/memory.cm --> $c/memory/memory.cm
3503    
3504    * "make" tool (in CM) now treats its argument pathname slightly
3505      differently:
3506        1. If the native expansion is an absolute name, then before invoking
3507           the "make" command on it, CM will apply OS.Path.mkRelative
3508           (with relativeTo = OS.FileSys.getDir()) to it.
3509        2. The argument will be passed through to subsequent phases of CM
3510           processing without "going native".  In particular, if the argument
3511           was an anchored path, then "make" will not lose track of that anchor.
3512    
3513    * Compiler backends now "know" their respective C calling conventions
3514      instead of having to be told about it by ml-nlffigen.  This relieves
3515      ml-nlffigen from one of its burdens.
3516    
3517    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3518    
3519    * Export C_DEBUG and C_Debug from $c/c.cm.
3520    
3521    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3522      subtyping relationship between incomplete pointers and their complete
3523      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3524      with the convention of instantiating 'o with ('t, 'c) obj whenever
3525      the pointer target type is complete.  In the incomplete case, 'o
3526      will be instantiated with some "'c iobj" -- a type obtained by
3527      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3528    
3529      Operations that work on both incomplete and complete pointer types are
3530      typed as taking an 'o ptr while operations that require the target to
3531      be known are typed as taking some ('t, 'c) obj ptr.
3532    
3533      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3534      where void is an eqtype without any values.  This makes it possible
3535      to work on voidptr values using functions meant to operate on light
3536      incomplete pointers.
3537    
3538    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3539      been vastly simplified.
3540    
3541    ----------------------------------------------------------------------
3542    Name: Matthias Blume
3543    Date: 2002/02/19 10:48:00 EST
3544    Tag: blume-20020219-pqfix
3545    Description:
3546    
3547    Applied Chris Okasaki's bug fix for priority queues.
3548    
3549    ----------------------------------------------------------------------
3550    Name: Matthias Blume
3551    Date: 2002/02/15 17:05:00
3552    Tag: Release_110_39
3553    Description:
3554    
3555    Last-minute retagging is becoming a tradition... :-(
3556    
3557    This is the working release 110.39.
3558    
3559    ----------------------------------------------------------------------
3560    Name: Matthias Blume
3561    Date: 2002/02/15 16:00:00 EST
3562    Tag: Release_110_39-orig
3563    Description:
3564    
3565    Working release 110.39.  New bootfiles.
3566    
3567    (Update: There was a small bug in the installer so it wouldn't work
3568    with all shells.  So I retagged. -Matthias)
3569    
3570    ----------------------------------------------------------------------
3571    Name: Matthias Blume
3572    Date: 2002/02/15 14:17:00 EST
3573    Tag: blume-20020215-showbindings
3574    Description:
3575    
3576    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3577    the latter can be useful for exploring what bindings are available at
3578    the interactive prompt.  (The first function returns only the list
3579    of symbols that are really bound, the second prints those but also the
3580    ones that CM's autoloading mechanism knows about.)
3581    
3582    ----------------------------------------------------------------------
3583    Name: Matthias Blume
3584    Date: 2002/02/15 12:08:00 EST
3585    Tag: blume-20020215-iptrs
3586    Description:
3587    
3588    Two improvements to ml-nlffigen:
3589    
3590      1. Write files only if they do not exist or if their current contents
3591         do not coincide with what's being written.  (That is, avoid messing
3592         with the time stamps unless absolutely necessary.)
3593    
3594      2. Implement a "repository" mechanism for generated files related
3595         to "incomplete pointer types".   See the README file for details.
3596    
3597    ----------------------------------------------------------------------
3598    Name: Matthias Blume
3599    Date: 2002/02/14 11:50:00 EST
3600    Tag: blume-20020214-quote
3601    Description:
3602    
3603    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3604    because of the new and improved tag generation scheme.  (Thanks to Allen
3605    Leung for pointing it out.)
3606    
3607    ----------------------------------------------------------------------
3608    Name: Lal George
3609    Date: 2002/02/14 09:55:27 EST 2002
3610    Tag: george-20020214-isabelle-bug
3611    Description:
3612    
3613    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3614    of Isabelle on the x86.
3615    
3616    From Allen:
3617    -----------
3618     I've found the problem:
3619    
3620         in ra-core.sml, I use the counter "blocked" to keep track of the
3621         true number of elements in the freeze queue.  When the counter goes
3622         to zero, I skip examining the queue.  But I've messed up the
3623         bookkeeping in combine():
3624    
3625             else ();
3626             case !ucol of
3627               PSEUDO => (if !cntv > 0 then
3628                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3629                                        ^^^^^^^^^^^^^^^^^^^^^^^
3630                      moveu := mergeMoveList(!movev, !moveu)
3631                     )
3632                  else ();
3633    
3634         combine() is called to coalesce two nodes u and v.
3635         I think I was thinking that if the move counts of u and v are both
3636         greater than zero then after they are coalesced then one node is
3637         removed from the freeze queue.  Apparently I was thinking that
3638         both u and v are of low degree, but that's clearly not necessarily true.
3639    
3640    
3641    02/12/2002:
3642        Here's the patch.  HOL now compiles.
3643    
3644        I don't know how this impact on performance (compile
3645        time or runtime).  This bug caused the RA (especially on the x86)
3646        to go thru the potential spill phase when there are still nodes on the
3647        freeze queue.
3648    
3649    
3650    
3651    
3652    ----------------------------------------------------------------------
3653    Name: Matthias Blume
3654    Date: 2002/02/13 22:40:00 EST
3655    Tag: blume-20020213-fptr-rtti
3656    Description:
3657    
3658    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3659    updates.
3660    
3661    ----------------------------------------------------------------------
3662    Name: Matthias Blume
3663    Date: 2002/02/13 16:41:00 EST
3664    Tag: blume-20020213-cmlpq
3665    Description:
3666    
3667    Added new priority queue export symbols (which have just been added to
3668    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3669    not compile and the installer would choke.)
3670    
3671    ----------------------------------------------------------------------
3672    Name: Matthias Blume
3673    Date: 2002/02/13 16:15:00 EST
3674    Tag: blume-20020213-various
3675    Description:
3676    
3677    1. More tweaks to ml-nlffigen:
3678    
3679       - better internal datastructures (resulting in slight speedup)
3680       - "-match" option requires exact match
3681       - "localized" gensym counters (untagged structs/unions nested within
3682         other structs/unions or within typedefs get a fresh counter; their
3683         tag will be prefixed by a concatenation of their parents' tags)
3684       - bug fixes (related to calculation of transitive closure of types
3685         to be included in the output)
3686    
3687    2. Minor Basis updates:
3688    
3689       - added implementations for List.collate and Option.app
3690    
3691    ----------------------------------------------------------------------
3692    Name: Matthias Blume
3693    Date: 2002/02/11 15:55:00 EST
3694    Tag: blume-20020211-gensym
3695    Description:
3696    
3697    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3698    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3699    names (ML structure names that correspond to unnamed C structs, unions,
3700    and enums), so that separate runs of ml-nlffigen do not clash.
3701    
3702    ----------------------------------------------------------------------
3703    Name: Matthias Blume
3704    Date: 2002/02/11 12:05:00 EST
3705    Tag: blume-20020211-gensml
3706    Description:
3707    
3708    A quick fix for a problem with GenSML (in the pgraph-util library):
3709    Make generation of toplevel "local" optional.  (Strictly speaking,
3710    signature definitions within "local" are not legal SML.)
3711    
3712    Other than that: updates to INSTALL and cm/TODO.
3713    
3714    ----------------------------------------------------------------------
3715    Name: Matthias Blume
3716    Date: 2002/02/08 15:00:00 EST
3717    Tag: blume-20020208-uniquepid
3718    Description:
3719    
3720    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3721    
3722    1. The installer (config/install.sh) has gotten smarter:
3723    
3724         - Configuration options are a bit easier to specify now
3725           (in config/targets).
3726         - Bug in recognizing .tar.bz2 files fixed.
3727         - Installer automatically resolves dependencies between
3728           configuration options (e.g., if you ask for eXene, you will
3729           also get cml -- regardless whether you asked for it or not).
3730         - Installer can run in "quieter mode" by setting the environment
3731           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3732           "completely silent", though.
3733         - Build HashCons library as part of smlnj-lib.
3734    
3735    2. A new scheme for assigning persistent identifiers to compilation
3736       units (and, by extension, to types etc.) has been put into place.
3737       This fixes a long-standing bug where types and even dynamic values
3738       can get internally confused, thereby compromising type safety
3739       (abstraction) and dynamic correctness.  See
3740    
3741         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3742    
3743       for an example of how things could go wrong until now.
3744    
3745       The downside of the new scheme is that pids are not quite as
3746       persistent as they used to be: CM will generate a fresh pid
3747       for every compilation unit that it thinks it sees for the first
3748       time.  That means that if you compile starting from a clean, fresh
3749       source tree at two different times, you end up with different
3750       binaries.
3751    
3752       Cutoff recompilation, however, has not been compromised because
3753       CM keeps pid information in special caches between runs.
3754    
3755    ----------------------------------------------------------------------
3756    Name: Lal George
3757    Date: 2002/02/07 15:34:13 EST 2002
3758    Tag: <none>
3759    Description:
3760    
3761    Compilers that generate assembly code may produce  global labels
3762    whose value is resolved at link time. The various peephole optimization
3763    modules did not take this in account.
3764    
3765    TODO. The Labels.addrOf function should really return an option
3766    type so that clients are forced to deal with this issue, rather
3767    than an exception being raised.
3768    
3769    ----------------------------------------------------------------------
3770    Name: Lal George
3771    Date: 2002/02/06 13:55:02 EST
3772    Tag: george-20020206-ra-breakup
3773    Description:
3774    
3775    1. A bug fix from Allen.
3776    
3777        A typo causes extra fstp %st(0)'s to be generated at compensation
3778        edges, which might cause stack underflow traps at runtime.  This
3779        occurs in fft where there are extraneous fstps right before the 'into'
3780        trap instruction (in this case they are harmless since none of the
3781        integers overflow.)
3782    
3783    2. Pulled out various utility modules that were embedded in the modules
3784       of the register allocator. I need these modules for other purposes, but
3785       they are not complete enough to put into a library (just yet).
3786    ----------------------------------------------------------------------
3787    Name: Matthias Blume
3788    Date: 2002/01/31 16:05:00 EST
3789    Tag: blume-20020131-sparc-ccalls
3790    Description:
3791    
3792    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3793       of extra stack space by mistake.  Fixed.
3794    
3795    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3796    
3797    ----------------------------------------------------------------------
3798    Name: Allen Leung
3799    Date: 2002/01/30
3800    Tag: leunga-20020130-nowhere-bug-fix
3801    Description:
3802    
3803       MLRISC bug fixes:
3804       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3805       2. MachineInt.fromString was negating its value.
3806    
3807    ----------------------------------------------------------------------
3808    Name: Matthias Blume
3809    Date: 2002/01/29
3810    Tag: blume-20020129-INSTALL
3811    Description:
3812    
3813    - Added somewhat detailed installation instructions (file INSTALL).
3814    - Fixed curl-detection bug in config/install.sh.
3815    - It is now possible to select the URL getter using the URLGETTER
3816      environment variable:
3817    
3818          not set / "unknown"      --> automatic detection (script tries wget,
3819                                       curl, and lynx)
3820          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3821                                       how to properly invoke them)
3822          other                    --> use $URLGETTER directly, it must take
3823                                       precisely two command-line arguments
3824                                       (source URL and destination file name)
3825    
3826    ----------------------------------------------------------------------
3827    Name: Matthias Blume
3828    Date: 2002/01/28
3829    Tag: blume-20020128-sparc-ccalls
3830    Description:
3831    
3832    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3833    - Make use of the allocParam argument in sparc-c-calls.
3834    
3835    ----------------------------------------------------------------------
3836    Name: Matthias Blume
3837    Date: 2002/01/28
3838    Tag: blume-20020128-allocParam
3839    Description:
3840    
3841    John Reppy:  Changes c-calls API to accept client-callback for
3842    allocating extra stack space.
3843    me: Corresponding changes to mlriscGen (using a dummy argument that
3844        does not change the current behavior).
3845    
3846    ----------------------------------------------------------------------
3847    Name: Matthias Blume
3848    Date: 2002/01/28 12:00:00
3849    Tag: Release_110_38
3850    Description:
3851    
3852    This time for real!!!
3853    
3854    ----------------------------------------------------------------------
3855    Name: Matthias Blume
3856    Date: 2002/01/28 10:56:00 EST
3857    Tag: blume-20020128-retraction
3858    Description:
3859    
3860    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3861       with blume-Release