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