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/NOTES/HISTORY
ViewVC logotype

Diff of /sml/trunk/NOTES/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

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