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