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