Home My Page Projects Code Snippets Project Openings SML/NJ
Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

SCM Repository

[smlnj] Diff of /sml/trunk/HISTORY
ViewVC logotype

Diff of /sml/trunk/HISTORY

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

revision 903, Mon Aug 20 19:34:06 2001 UTC revision 1686, Thu Nov 11 23:26:27 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    Date: 2004/11/11 17:30:00 CST
17    Tag: blume-20041111-more64
18    Description:
19    
20    All the pieces of Word64 are now there, with the exception of the
21    conversions from and to LargeWord. (Eventually these need to be identities,
22    but for the time being they don't even make sense because LargeWord is
23    32-bit wide.)
24    
25    Also started to add similar support for Int64, but major pieces of that
26    are still missing.
27    
28    ----------------------------------------------------------------------
29    Name: Matthias Blume (blume (at) tti - c (dot) org)
30    Date: 2004/11/11 00:15:00 CST
31    Tag: blume-20041111-word64
32    Description:
33    
34    Structure Word64 is now (almost) complete, word literals and patterns
35    seem to work.  There are a few odd pieces missing.  In particular,
36    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
37    at the moment.
38    
39    Making Word64 official would mean that LargeWord becomes Word64.  But
40    this requires extreme care because most word-word conversions have to
41    go through LargeWord, so making a mistake means loss of efficiency or
42    worse.  Eventually there will be a solution similar to (but actually
43    simpler than) what I did with IntInf.
44    
45    ----------------------------------------------------------------------
46    Name: Matthias Blume (blume (at) tti - c (dot) org)
47    Date: 2004/11/10 18:12:00 CST
48    Tag: blume-20041110-64bit
49    Description:
50    
51    More 64-bit hacking (but still not even half-way there yet).
52    Also, some assorted improvements to the handling of 8-bit words.
53    
54    ----------------------------------------------------------------------
55    Name: Matthias Blume (blume (at) tti - c (dot) org)
56    Date: 2004/11/09 17:50:00 CST
57    Tag: <none>
58    Description:
59    
60    Started adding some infrastructure for supporting 64-bit int- and
61    word-types.  (Still in its very early stages.)
62    
63    ----------------------------------------------------------------------
64    Name: Matthias Blume (blume (at) tti - c (dot) org)
65    Date: 2004/10/28 10:45:00 CDT
66    Tag: Release_110_50
67    Description:
68    
69    New working version (110.50).  NEW BOOTFILES!
70    
71                 =====================
72    
73    Also:
74    
75      - Changend config/srcarchiveurl from a file just
76        containing the URL string into a file containing
77        shell script code.  The code has access to the $VERSION variable.
78      - Made corresponding changes to config/install.sh and config/unpack.
79      - Default contents of config/srcarchiveurl uses $VERSION and
80        normally does not have to be edited to reflect a version change.
81    
82      (As a result, a version change can be done by just editing
83       config/version, the rest is now automatic.)
84    
85    ----------------------------------------------------------------------
86    Name: Matthias Blume (blume (at) tti - c (dot) org)
87    Date: 2004/10/27 17:50:00 CDT
88    Tag: blume-20041027-btrace-msg
89    Description:
90    
91    BackTrace.monitor now also reports the source of the exception that
92    triggered the trace.
93    
94    ----------------------------------------------------------------------
95    Name: Matthias Blume (blume (at) tti - c (dot) org)
96    Date: 2004/10/27 17:20:00 CDT
97    Tag: blume-20041027-x86-c-calls
98    Description:
99    
100    This is the HISTORY entry for two earlier commits, both concerning
101    the x86 c-calls code in MLRISC:
102    
103      - added a missing LOAD in the code that deals with struct arguments
104      - made sure the caller does not add the wrong number of bytes to the
105        stack pointer after a call of a function returning a struct
106        (the callee already pops the implicit argument which points to
107         the space reserved for the result)
108    
109    ----------------------------------------------------------------------
110    Name: Allen Leung (leunga (at) reservoir (dot) com)
111    Date: 2004/10/24 14:00:00 EST
112    Tag: leunga-20041024-x86-gas-fucomip
113    Description:
114    
115       John discovered a bug in the syntax of fucomip.
116       The opcodes FU?COMIP? have been changed to
117    
118           fu?comip? %st(i), %st
119    
120    ----------------------------------------------------------------------
121    Name: Matthias Blume (blume (at) tti - c (dot) org)
122    Date: 2004/10/20 15:06:00 CDT
123    Tag: blume-20041020-standalone-backtrace
124    Description:
125    
126    Added a mechanism for getting back-trace information from standalone
127    programs.  Here is how it works:
128    
129       1. The part of the program from which you want to get backtrace
130          information (usually the whole program) should be wrapped with
131          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
132          main program could be modified from something like
133    
134             fun main (pgm, args) = ...
135    
136         to
137    
138            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
139    
140       2. To be able to access BackTrace.monitor, you have to add
141          library $smlnj-tdp/plugins.cm to the .cm file that contains your
142          main function.
143    
144       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
145          CM might have created in the past for your project).
146    
147       4. Build the system using this command line:
148    
149             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
150                      myprog.cm MyProg.main myprog
151    
152          instead of the usual
153    
154             ml-build myprog.cm MyProg.main myprog
155    
156    I changed a library name:
157    
158       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
159    
160    New libraries:
161    
162       $smlnj-tdp/back-trace.cm
163          -- when loaded causes the back-trace plugin to be installed
164       $smlnj-tdp/coverage.cm
165          -- when loaded causes the coverage plugin to be installed
166    
167    ----------------------------------------------------------------------
168    Name: Matthias Blume (blume (at) tti - c (dot) org)
169    Date: 2004/10/18 16:45:00 CDT
170    Tag: blume-20041018-groupowner
171    Description:
172    
173    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
174    
175    Eliminated group owner specs from .cm files throughout the source tree.
176    
177    ----------------------------------------------------------------------
178    Name: Matthias Blume (blume (at) tti - c (dot) org)
179    Date: 2004/10/15 15:45:00 CDT
180    Tag: blume-20041015-coverage
181    Description:
182    
183    * Test coverage tool added!
184    * Further reorganization of tracing-, debugging-, and profiling support:
185    
186        - moved original BTImp -- now called BackTrace -- into a separate
187          library called $/trace-debug-profile.cm
188        - eliminated all mentions of BTrace from SMLofNJ.Internals
189        - only the instrumentation mechanism is now left in the compiler proper
190        - BackTrace module is a plugin which is NOT plugged in by default
191        - Coverage module is another such plugin
192    
193    To get the benefits of any of these plugin modules, the code in
194    question must be compiled with tdp instrumentation turned on.  This
195    can be done by setting SMLofNJ.Internals.TDP.mode to true.
196    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
197    
198    Plugins are selected at link time.  (Pre-compiled instrumented code
199    can be re-loaded with different plugins in effect.)  When an
200    instrumented module is linked, whatever plugins are at that time
201    enabled will come into effect for that module.
202    
203    To enable the back-trace plugin, load library $/trace-debug-profile.cm
204    and invoke BackTrace.install() (e.g., from the interactive prompt).
205    To enable the coverage plugin, load the same library and invoke
206    Coverage.install().
207    
208    Back-traces are generated automatically on uncaught exceptions and
209    when the code in question explicitly invokes BackTrace.trigger().
210    
211    Coverage (and execution frequency-) information must be queried
212    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
213    
214    ----------------------------------------------------------------------
215    Name: Matthias Blume (blume (at) tti - c (dot) org)
216    Date: 2004/10/14 17:40:00 CDT
217    Tag: blume-20041014-tdp-core
218    Description:
219    
220    Snapshot of a significant overhaul of how the trace/debug/profile support
221    is hooked into the system (specifically: Core and SMLofNJ.Internals).
222    
223    ----------------------------------------------------------------------
224    Name: Matthias Blume (blume (at) tti - c (dot) org)
225    Date: 2004/10/13 16:34:00 CDT
226    Tag: blume-20041013-tdp
227    Description:
228    
229    Some rationalization of names:
230    
231       structure BTrace -> structure TDPInstrument
232       etc.
233    
234    This is is preparation of using the original back-trace
235    instrumentation for other purposes.  "TDP" stands for
236    Trace/Debug/Profile.
237    
238    The control flag controlling whether instrumentation is on or off is now
239    registered under a different name, so instead of running sml as
240    
241     sml -Cinstrument.btrace-mode=true
242    
243    one has to say
244    
245     sml -Ctdp.instrument=true
246    
247    ----------------------------------------------------------------------
248    Name: Matthias Blume (blume (at) tti - c (dot) org)
249    Date: 2004/10/11 16:37:00 CDT
250    Tag: blume-20041011-regions
251    Description:
252    
253    Made some minor modifications to elabcore.sml to have source regions
254    be propagated more tightly -- resulting in better (i.e., smaller)
255    regions being reported in error- and debug messages.
256    
257    ----------------------------------------------------------------------
258    Name: Matthias Blume (blume (at) tti - c (dot) org)
259    Date: 2004/10/08 22:50:00 CDT
260    Tag: blume-20041008-cmkw
261    Description:
262    
263    Fixed handling of keywords in .cm files: After seeing "is" the lexer
264    treats subsequent occurrences of "group", "library", "source", "is",
265    "*", and "-" as ordinary identifiers rather than keywords.
266    
267    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
268    is supposed to accept a tool argument called "source", but this did
269    not work because of the clash with the keyword.
270    
271    ----------------------------------------------------------------------
272    Name: Matthias Blume (blume (at) tti - c (dot) org)
273    Date: 2004/10/07 16:00:00 CDT
274    Tag: blume-20041007-cleanup
275    Description:
276    
277    Assorted cleanup work:
278    
279      - got rid of intstrmap in favor of using the library's
280        hash table implementation
281      - threw out most of the pathnames stuff, as it was not used anyway
282      - simplified tokentable implementation
283      - fixed some minor spelling errors
284    
285    ----------------------------------------------------------------------
286    Name: Matthias Blume (blume (at) tti - c (dot) org)
287    Date: 2004/10/06 15:15:15 CDT
288    Tag: blume-20041006-handler
289    Description:
290    
291    Cleaned up the absyn to reflect the invariant that HANDLE always
292    carries a FNexp as part of the type definition.  This eliminates some
293    superfluous sanity checks at runtime down the road.
294    
295    Some minor cleanup of the btrace code.
296    
297    ----------------------------------------------------------------------
298    Name: Matthias Blume (blume (at) tti - c (dot) org)
299    Date: 2004/10/01 10:20:30 CDT
300    Tag: blume-20041001-slave
301    Description:
302    
303    Added hack to make slave mode work in the presence of the version
304    tool.  (Still, since the master does two passes over the code for
305    CMB.make, the release number gets bumped twice when slaves are
306    attached.  I don't know if this is worth fixing...)
307    
308    ----------------------------------------------------------------------
309    Name: Matthias Blume (blume (at) tti - c (dot) org)
310    Date: 2004/09/30 10:55:00 CDT
311    Tag: blume-20040930-version
312    Description:
313    
314    * Moved the "version" magic into its own little library under
315      src/system/smlnj/internal.  This avoids expensive reconstruction of
316      a stable src/compiler/core.cm.
317    
318    * At the same time, structure CompilerVersion is now known as
319      structure SMLNJVersion.
320    
321    * Arranged for the version tool to NOT kick in when rebuilding the system
322      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
323      Also, loading the versiontool does not work when rebuilding the system
324      because CM is not properly initialized at that time.
325    
326    ----------------------------------------------------------------------
327    Name: Matthias Blume (blume (at) tti - c (dot) org)
328    Date: 2004/09/29 14:00:00 CDT
329    Tag: blume-20040929-autoversion
330    Description:
331    
332    Implemented some CM magic to have
333        file src/compiler/TopLevel/main/version.sml
334    generated automagically.
335    The version is taken from two files: config/version and config/release.
336    The first is expected to contain a two-part version number such as 110.49.
337    The second should contain a single number, but it may be missing.
338    
339    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
340    time the version tool is loaded (which is the first time you say CMB.make),
341    then the tool will increment the value stored in config/release every
342    time CMB.make is invoked.
343    
344    The binfile format is now insensitive to anything beyond the first
345    two components of a version number, so bumping the release does not render
346    binfiles incompatible.  Auto-bumping can be used to keep track of versions
347    during development without invalidating existing binfiles.
348    
349    In any case, every CMB.make updates the date information in version.sml.
350    (This is the date that is printed in the banner.)
351    
352    ----------------------------------------------------------------------
353    Name: Matthias Blume (blume (at) tti - c (dot) org)
354    Date: 2004/09/28 10:53:00 CDT
355    Tag: blume-20040928-controls
356    Description:
357    
358    Some cleanup of the controls code.
359    
360    ----------------------------------------------------------------------
361    Name: Matthias Blume (blume (at) tti - c (dot) org)
362    Date: 2004/09/27 22:08:00 CDT
363    Tag: blume-20040927-controls
364    Description:
365    
366    Added two pieces of functionality to the Controls interface:
367    
368    1.   val save'restore: 'a control -> unit -> unit
369    
370       grabs the current value of the control in stage 1 and restores it
371       in stage 2.
372    
373    2.   val set' : 'a control * 'a -> unit -> unit
374    
375       stores the given value into the control in stage 2 (i.e., delayed)
376       but does all error checking in stage 1.
377       (This is for string controls that need to do parse their argument --
378        something that might fail.  In some cases, notably in CM, one
379        already knows the intended argument but wants to delay the actual
380        assignment until a time when error recovery would be more difficult.)
381    
382    Changed the handling of controls in tool arguments to classes "sml" and
383    "lazysml":
384       - use Controls.save'restore as a more robust way of restoring the
385         old value (in particular: without having to re-parse the string)
386       - use controls to handle the "overload" keyword in the init group
387         (I believe this change actually fixes a long-standing obscure bug.)
388    
389    ----------------------------------------------------------------------
390    Name: Matthias Blume (blume (at) tti - c (dot) org)
391    Date: 2004/09/27 17:00:00 CDT
392    Tag: blume-20040927-lazysml
393    Description:
394    
395    Added a new tool class called "lazysml" to CM's tool chest.  The only
396    difference to "sml" is that compilation is done with Control.lazysml
397    set to true.  A source of class "lazysml" is automatically recognized
398    by a file name suffix of ".lml".
399    
400    In addition to the above feature, the original class "sml" now also
401    supports a tool argument "lazy" which has the same effect.  As a
402    result, the following three lines are equivalent:
403    
404        foo.sml : lazysml
405        foo.sml : sml (lazy)
406        foo.sml (lazy)
407    
408    The setting goes into effect both during parsing and during
409    compilation.  The original setting is restored right after parsing and
410    after compilation, respectively.
411    
412    In addition to all the above, there is also a general mechanism to set
413    ANY of the "controls" that are available at the command line via
414    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
415    apply as well. (In fact, "lazy" is implemented as a special case of
416    the general mechanism.)
417    
418    The .cm file syntax uses a new keyword tool argument called "with".
419    There are several ways of indicating the desired settings:
420    
421        foo.sml (with:parser.quotations=true)
422        foo.sml (with:(name:parser.quotations value:true))
423        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
424        foo.sml (with:(name1=value1 name2=value2 ...))
425        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
426    
427        etc.
428    
429    Another possible abbreviation is to leave out the =v or value:v part
430    if the name refers to a boolean control (in which case the value is
431    taken to be true).  Thus, one could get lazy sml also by saying:
432    
433        foo.sml (with:parser.lazy-keyword=true)
434        foo.sml (with:parser.lazy-keyword)
435        foo.sml (with:(name:parser.lazy-keyword value:true))
436        foo.sml (with:(name:parser.lazy-keyword))
437    
438    ----------------------------------------------------------------------
439    Name: Matthias Blume (blume (at) tti - c (dot) org)
440    Date: 2004/09/24 16:22:00 CDT
441    Tag: blume-20040924-ppc-long-branch
442    Description:
443    
444    Turned message about "emiting long form of branch" off by default.
445    Added a control flag to turn it back on when desired.
446    
447    ----------------------------------------------------------------------
448    Name: Matthias Blume (blume (at) tti - c (dot) org)
449    Date: 2004/09/24 16:05:00 CDT
450    Tag: blume-20040924-rounding
451    Description:
452    
453    Applied patch for setting rounding modes under Mac OS X.  Thanks to
454    Melissa O'Neill for providing the code!
455    
456    ----------------------------------------------------------------------
457    Name: Matthias Blume (blume (at) tti - c (dot) org)
458    Date: 2004/09/23 17:30:00 CDT
459    Tag: blume-20040923-envvars
460    Description:
461    
462    1. Changed definition of type ControlRegistry.registry_tree to
463       include control_info (i.e., the name of the controlling environment
464       variable).
465    
466    2. Added command-line flags -e and -E to print the names of environment
467       variables that can be used to control internal settings.
468       (This uses the new API mentioned in 1.)
469    
470    ----------------------------------------------------------------------
471    Name: Matthias Blume (blume (at) tti - c (dot) org)
472    Date: 2004/09/13 16:50:00 CDT
473    Tag: Release_110_49
474    Description:
475    
476    New working version (110.49).  NEW BOOTFILES!
477    
478    ----------------------------------------------------------------------
479    Name: Matthias Blume
480    Date: 2004/09/13 16:20:00 CDT
481    Tag: blume-20040913-config-mlrisc
482    Description:
483    
484    Put target "mlrisc" back into the default list.
485    (There is no harm in having it, and some users have expressed their
486    wish to have "mlrisc" included by default.)
487    
488    ----------------------------------------------------------------------
489    Name: John Reppy
490    Date: 2004/09/13
491    Tag: jhr-20040913-signals
492    Description:
493    
494    Fixed the signal masking code to properly nest mask/unmask operations
495    on a per-signal basis.
496    
497    ----------------------------------------------------------------------
498    Name: Matthias Blume (blume (at) tti - c (dot) org)
499    Date: 2004/09/08 13:20:00 CDT
500    Tag: blume-20040908-heap-magic
501    Description:
502    
503    Bumped the heap macig number to 0x09082004 to account for the changed
504    layout of the ML frame under MacOS X.
505    
506    ----------------------------------------------------------------------
507    Name: Allen Leung (leunga (at) reservoir (dot) com)
508    Date: 2004/09/03 11:26:00 EST
509    Tag: leunga-20040903-cygwin-install
510    Description:
511    
512       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
513    The Cygwin runtime is turned on by setting the environment
514    variable SMLNJ_CYGWIN_RUNTIME to 1.
515    
516    ----------------------------------------------------------------------
517    Name: Matthias Blume (blume (at) tti - c (dot) org)
518    Date: 2004/08/31 17:15:00 CDT
519    Tag: blume-20040831-core
520    Description:
521    
522    Added some exports to src/compiler/core.cm upon request by J. Joemann.
523    
524    ----------------------------------------------------------------------
525    Name: Matthias Blume (blume (at) tti - c (dot) org)
526    Date: 2004/08/30 17:55:00 CDT
527    Tag: blume-20040830-installer
528    Description:
529    
530    Upon request by Johannes Joemann:
531    
532    - improved ML code of installer to fall back to coping when renaming
533      fails (i.e., when source and target are on different file systems);
534      the code compiles but has yet to be tested in anger
535    - removed mlrisc from list of default targets (config/targets)
536    
537    ----------------------------------------------------------------------
538    Name: Matthias Blume (blume (at) tti - c (dot) org)
539    Date: 2004/08/27 17:20:00 CDT
540    Tag: blume-20040827-ptreql
541    Description:
542    
543    Added ptreql primop to structure InlineT (upon request from Larry
544    Paulson).
545    
546    ----------------------------------------------------------------------
547    Name: Allen Leung (leunga (at) reservoir (dot) com)
548    Date: 2004/08/15 21:21:00 EST
549    Tag: leunga-110_48-udgraph
550    Description:
551    
552    Another bug fix from Carl Hauser:
553    
554    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
555    > 48c48
556    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
557    > e::L)
558    > ---
559    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
560    > Without this, any deletion of an edge in an undirected graph does severe
561    > violence to the graph.
562    
563    ----------------------------------------------------------------------
564    Name: Allen Leung (leunga (at) reservoir (dot) com)
565    Date: 2004/08/10 23:35:00 EST
566    Tag: leunga-110_48-ppc
567    Description:
568    
569       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
570    
571    ----------------------------------------------------------------------
572    Name: Matthias Blume (blume (at) tti - c (dot) org)
573    Date: 2004/08/10 12:00:00 CDT
574    Tag: Release_110_48
575    Description:
576    
577    New working version (110.48).  NEW BOOTFILES!
578    
579    ----------------------------------------------------------------------
580    Name: Allen Leung (leunga (at) reservoir (dot) com)
581    Date: 2004/08/09 12:21:00 EST
582    Tag: leunga-110_47-dijsktra
583    Description:
584    
585    Bug fix from Carl Hauser:
586    
587    single_source_shortest_paths in dijkstra.sml was observed to get wrong
588    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
589    
590    The problem is that following the expression A.update(dist,s,Num.zero)
591    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
592    
593    ----------------------------------------------------------------------
594    Name: Matthias Blume (blume (at) tti - c (dot) org)
595    Date: 2004/08/06 18:10:00 CDT
596    Tag: blume-20040806-cmdline
597    Description:
598    
599    Fiddled with handling of command-line options:
600    
601      * sml now quits after processing the command line
602        if -H, -S, -h<n>, or -s<n> appears as the last
603        command-line argument
604      * a new option -q terminates the session when encountered on
605        the command line; subsequent arguments will be ignored
606      * bug fixes: short (erroneous) arguments are no longer ignored
607        completely
608    
609    ----------------------------------------------------------------------
610    Name: Allen Leung (leunga (at) reservoir (dot) com)
611    Date: 2004/08/04 18:17:00 EST
612    Tag: leunga-110_47-ppc-ibm-asm
613    Description:
614    
615      - Added minimal IBM assembly syntax support for PowerPC.
616    
617      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
618        been altered in the runtime.  We need an automatic way to keep the file
619        in sync.
620    
621    ----------------------------------------------------------------------
622    Name: Matthias Blume (blume (at) tti - c (dot) org)
623    Date: 2004/08/04 14:00:00 CDT
624    Tag: Release_110_47
625    Description:
626    
627    New working version (110.47).  NEW BOOTFILES!
628    
629    ----------------------------------------------------------------------
630    Name: Matthias Blume (blume (at) tti - c (dot) org)
631    Date: 2004/08/03 14:25:00 CDT
632    Tag: blume-20040803-callingconv
633    Description:
634    
635    Added low-level support for choosing C calling conventions by
636    twiddling the type of rawccall.  (See
637    src/compiler/Semant/types/cproto.sml for details.)
638    
639    ----------------------------------------------------------------------
640    Name: Matthias Blume (blume (at) tti - c (dot) org)
641    Date: 2004/08/02 15:55:00 CDT
642    Tag: blume-20040802-backout
643    Description:
644    
645    Backed out of change to win32-filesys.c.  The earlier patch to
646    get_file_time caused CM to produce files with the wrong time stamp.
647    
648    ----------------------------------------------------------------------
649    Name: Matthias Blume (blume (at) tti - c (dot) org)
650    Date: 2004/08/02 14:45:00 CDT
651    Tag: blume-20040802-nlffi-win32
652    Description:
653    
654    Added NLFFI support for Win32, adapted from a patch provided by David
655    Hansel.  This is currently completely untested.  Also, the issue
656    concerning stdcall vs. ccall is still unresolved.
657    
658    ----------------------------------------------------------------------
659    Name: Matthias Blume (blume (at) tti - c (dot) org)
660    Date: 2004/07/30 17:55:00 CDT
661    Tag: blume-20040730-various
662    Description:
663    
664    Gearing up towards 110.47...
665    
666    - various minor bugfixes to ml-nlffigen
667    - a beginning of a manual for nlffi
668    
669    - eliminated 'export name=value' in config/install.sh as this does
670      not work with certain versions of /bin/sh
671      (Thanks to David King at Motorola for catching this.)
672    
673    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
674      - added a test for tm==NULL to gmtime.c and localtime.c
675      - applied patch for incorrect GetFileTime under win32
676      - toSeconds -> toMilliseconds in Win32/win32-process.sml
677    
678    ----------------------------------------------------------------------
679    Name: Matthias Blume (blume (at) tti - c (dot) org)
680    Date: 2004/07/21 18:20:00 CDT
681    Tag: blume-20040721-nlffigen
682    Description:
683    
684    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
685      for a typedef to an incomplete type, but leave out the "typ" member.
686      (This is just for consistency.)
687    - Started to produce what is supposed to become better (i.e., comprehensive)
688      documentation of what ml-nlffigen does and produces.
689    
690    ----------------------------------------------------------------------
691    Name: Matthias Blume (blume (at) tti - c (dot) org)
692    Date: 2004/07/14 16:25:00 CDT
693    Tag: blume-20040714-union
694    Description:
695    
696    Added C_UNION to c-calls/c-types.sml and updated the machinery
697    (ml-nlffigen, cproto.sml) that conveys C function interface
698    information to the code generator.
699    
700    However, the actual architecture-specific implementation of function
701    arguments and results that are C unions is still not implemented.
702    
703    ----------------------------------------------------------------------
704    Name: Allen Leung (leunga (at) reservoir (dot) com)
705    Date: 2004/07/14 14:38:00 EST
706    Tag: leunga-110_46_1-ppc-lwzu
707    Description:
708    
709        Added these instructions to the PowerPC architecture:
710        LBZU(X), LHZU(X), LWZU(X),
711        STWU(X), STFDU, STFSU
712    
713        etc...
714    
715        Note: I haven't added their instruction encoding into the description.
716    
717    ----------------------------------------------------------------------
718    Name: Allen Leung (leunga (at) reservoir (dot) com)
719    Date: 2004/07/13 15:04:00 EST
720    Tag: leunga-110_46_1-ppc-lwarx
721    Description:
722    
723       Added the two instructions LWARX and STWCX to the PowerPC
724    instruction set.
725    
726       A (untested) rewrite of loop-structure.sml.  The old version
727    is completely broken.
728    
729    ----------------------------------------------------------------------
730    Name: Matthias Blume (blume (at) tti - c (dot) org)
731    Date: 2004/07/13 13:50:00 CDT
732    Tag: blume-20040713-nlffi
733    Description:
734    
735    - use paramAlloc to report c-calls with too many arguments
736      (for PPC version where parameter area is pre-allocated)
737    - added ccall_maxargspace to machspec (to implement the above)
738    - made "make" commend in CM's "make" tool configurable
739    - added option (default: on) for passing the name of the SML/NJ's "bin"
740      directory to "make"; the call looks like this:
741    
742         make <options> SMLNJ_BINDIR=<dir> <target>
743    
744      This can be used by the Makefile to, e.g., pick the "right" version
745      of ml-nlffigen.
746    - minor code tweaks
747    
748    ----------------------------------------------------------------------
749    Name: Matthias Blume (blume (at) tti - c (dot) org)
750    Date: 2004/07/12 22:50:00 CDT
751    Tag: blume-110_46_1-macosx-nlffi
752    Description:
753    
754    NLFFI under Mac OS X now working (sort of).  This is largely untested,
755    though.
756    
757    Note:  1. You have to make a new, clean build of the runtime system.
758           2. There are new BOOTFILES, you have to use them!
759              (Doing the bootstrap process yourself would be *very* painful!
760               If you absolutely have to do it, build the system under
761               a different architecture and then cross-compile.)
762    
763    Version bumped to 110.46.1 to account for runtime data format changes.
764    
765    ----------------------------------------------------------------------
766    Name: Matthias Blume (blume (at) tti - c (dot) org)
767    Date: 2004/06/18 14:30:00 CDT
768    Tag: blume-20040618-unix
769    Description:
770    
771    Changed the implementation of structure Unix so that the same stream
772    is returned every time one of the {text,bin}{In,Out}streamOf functions
773    is invoked on the same proc.  This is not what the spec currently
774    says -- although IMO it arguably should.  (See discussion below.)
775    
776    ----------------------------------------------------------------------
777    Name: Matthias Blume (blume (at) tti - c (dot) org)
778    Date: 2004/06/17 18:15:00 CDT
779    Tag: Release_110_46
780    Description:
781    
782    New working version (110.46).  NEW BOOTFILES!
783    
784    ----------------------------------------------------------------------
785    Name: Matthias Blume (blume (at) tti - c (dot) org)
786    Date: 2004/06/17 17:20:00 CDT
787    Tag: blume-20040617-timer-unix
788    Description:
789    
790    Changed the interface of structures Timer and Unix to match the most
791    recent Basis spec.
792    
793    In the case of Unix there still seems to be an open/weird issue:
794    
795       The {text,bin}{In,Out}streamOf functions are supposed to create
796       fresh streams whenever they are called -- as opposed to have them
797       return the same stream every time.  This design is supposed to
798       prevent space leaks caused by proc values hanging on to streams.
799    
800       The reap function, on the other hand, is supposed to close the
801       streams.  This cannot be done without having a handle on the
802       stream in proc after all...
803    
804    I took the liberty to implement the following stopgap solution:
805    
806      The proc value hangs on to the most recently created stream(s).
807      Reap closes those.  If either or both of the two streams hadn't
808      been created at all yet, then reap will close the corresponding
809      file descriptors directly.
810    
811    PS: I don't understand the original space leak argument anymore.  If
812    a proc hangs on to the imperative stream, then I/O operations on those
813    will advance the state of the cached stream and avoid the space leak.
814    
815    ----------------------------------------------------------------------
816    Name: Matthias Blume (blume (at) tti - c (dot) org)
817    Date: 2004/05/28 16:45:00 CDT
818    Tag: blume-20040528-basis
819    Description:
820    
821    Added signature PACK_REAL and exported functor PrimIO.
822    
823    ----------------------------------------------------------------------
824    Name: Matthias Blume (blume (at) tti - c (dot) org)
825    Date: 2004/05/25 16:00:00 CDT
826    Tag: blume-20040525-group-owner
827    Description:
828    
829    CM now ignores (but still accepts) the "owner" information in group
830    descriptions.  The owner of a group is its next enclosing
831    library. Each group must have a unique owner.  (There is a virtual
832    "toplevel" library that owns groups which are not nested within a real
833    library.)  Previously, each group had to explicitly declare its owner,
834    and CM would check that such a declaration is correct.  The new scheme
835    is to have CM check that for each group there is precisely one owning
836    library.
837    
838    The advantage of the new scheme is that the programmer no longer needs
839    to maintain the somewhat annoying owner information.  The downside is
840    that CM cannot enforce the ownership rule across multiple runs of
841    CM.make.  Fortunately, enclosing the same group in two different
842    libraries A and B which are not part of the same program does not
843    cause real problems.
844    
845    ----------------------------------------------------------------------
846    Name: Matthias Blume (blume (at) tti - c (dot) org)
847    Date: 2004/05/20 16:00:00 CDT
848    Tag: blume-20040520-win32
849    Description:
850    
851    Made the win32 version work again.  (Strangely, a misplaced comma had
852    slipped into win32-process.c which prevented the runtime from being
853    compiled correctly.)
854    
855    Also, included a minor addition to ml-build.bat analogous to what was
856    done in blume-20040519-ml-build.
857    
858    ----------------------------------------------------------------------
859    Name: Matthias Blume (blume (at) tti - c (dot) org)
860    Date: 2004/05/19 22:10:00 CDT
861    Tag: blume-20040519-ml-build
862    Description:
863    
864    Arranged for ml-build to clean up after itself a little bit better.
865    The script generates a temporary SML source file and compiles it using
866    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
867    gets rid of those at the end, so they don't accumulate under .cm.
868    
869    This required a minor change to install.sh because the name of the
870    metadata directory (default: .cm) is actually configurable at
871    installation time.
872    
873    ----------------------------------------------------------------------
874    Name: Matthias Blume (blume (at) tti - c (dot) org)
875    Date: 2004/05/18 15:50:00 CDT
876    Tag: blume-20040518-mkreader
877    Description:
878    
879    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
880    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
881    
882    ----------------------------------------------------------------------
883    Name: Matthias Blume (blume (at) tti - c (dot) org)
884    Date: 2004/05/11 14:35:00 CDT
885    Tag: blume-20040511-win32sock
886    Description:
887    
888    Added previously missing support for many socket-related functions
889    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
890    for the voluminous patch!
891    
892    (I have not tested this patch under win32 yet.)
893    
894    Here is David's e-mail:
895    
896    Hi,
897    
898    Attached to this email you find a diff against sml/nj 110.45
899    that will enable socket support under Windows.
900    
901    To apply the patch (using unix or cygwin)
902    1) gunzip runtime.diff.gz
903    2) "cd" into "src/runtime" in the source tree of a fresh
904       110.45 installation.
905    3) patch -p 1 < [your/path/to]runtime.diff
906    
907    The code compiles fine but has NOT yet been extensively tested.
908    I only ran a few tests for basic socket client functionality
909    (which worked fine).  Especially the functions that use ioctl
910    are not tested at all and might not work (see below).
911    
912    I implemented this since we want to move to a newer version of sml/nj
913    but need socket support in order to use it.  This is the first time I
914    even had a look at the sml/nj source,  so please review my changes
915    before making this part of the distribution!  Here are a few issues
916    that I think might be better for someone to solve who is more
917    familiar with the sml/nj source (and socket programming):
918    
919    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
920      exception since I could not figure out what the windows equivalent
921      of these functions is
922    
923    - In sockets-osdep.h there are a some #include statements that are
924      only used in a few files that include sockets-osdep.h
925    
926    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
927      does not process its return value since I don't know how to
928      report an error upwards.
929    
930    - It would probably be good to have a call to WSACleanup() when
931      the library is unloaded (if there is such a possibility).
932      Otherwise I think Windows will take care of this automatically
933      when the process finishes.
934    
935    - I used ioctlsocket() as a replacement for ioctl() but I have
936      no idea if that is actually the proper replacement on Windows.
937    
938    - All these issues are marked in the code by "FIXME" comments.
939    
940    We use sml/nj extensively in our products and are quite happy
941    with it.  I hope this contribution will help you.
942    
943    Keep up the good work!
944    
945    David
946    
947    ----------------------------------------------------------------------
948    Name: Matthias Blume (blume (at) tti - c (dot) org)
949    Date: 2004/05/11 14:20:00 CDT
950    Tag: blume-20040511-installml
951    Description:
952    
953    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
954    the patch.)
955    
956    ----------------------------------------------------------------------
957    Name: Matthias Blume (blume (at) tti - c (dot) org)
958    Date: 2004/05/11 14:05:00 CDT
959    Tag: blume-20040511-nlffi-netbsd
960    Description:
961    
962    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
963    the patch.)
964    
965    ----------------------------------------------------------------------
966    Name: Matthias Blume (blume (at) tti - c (dot) org)
967    Date: 2004/05/11 12:05:00 CDT
968    Tag: blume-20040511-exports
969    Description:
970    
971    As per request by Adam Chlipala <adam@hcoop.net>, extended various
972    export lists in compiler-related .cm-files.
973    
974    ----------------------------------------------------------------------
975    Name: Matthias Blume (blume (at) tti - c (dot) org)
976    Date: 2004/05/11 11:35:00 CDT
977    Tag: blume-20040511-allsource
978    Description:
979    
980    The installer now honors the "src-smlnj" target again, although its meaning
981    has changed from "all sources required for the compiler" to "all sources
982    the installer knows about".  In other words, if you enable "src-smlnj"
983    in the "targets" file, then the installer will pull in sources for
984    everything.  (Notice that this refers to source code only.  Compiled
985    code is still only installed for modules that were requested explicitly
986    or which are required for other modules that were requested explicitly.)
987    
988    ----------------------------------------------------------------------
989    Name: Matthias Blume (blume (at) tti - c (dot) org)
990    Date: 2004/04/23 17:40:00 CDT
991    Tag: blume-20040423-ieee-scan
992    Description:
993    
994    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
995    in the exponent calculation we get INF or ZERO (depending on the mantissa
996    and the sign of the exponent).
997    
998    ----------------------------------------------------------------------
999    Name: Matthias Blume (blume (at) tti - c (dot) org)
1000    Date: 2004/04/23 10:40:00 CDT
1001    Tag: blume-20040423-ml-build
1002    Description:
1003    
1004    The ml-build script now terminates with a non-0 status when something
1005    goes wrong.
1006    
1007    ----------------------------------------------------------------------
1008    Name: Matthias Blume (blume (at) tti - c (dot) org)
1009    Date: 2004/04/22 16:35:00 CDT
1010    Tag: blume-20040422-Option
1011    Description:
1012    
1013    Made exception Option to be the same as exception Option.Option
1014    (as it should be).
1015    
1016    ----------------------------------------------------------------------
1017    Name: Allen Leung (leunga (at) reservoir (dot) com)
1018    Date: 2004/03/19 14:40:00 EST
1019    Tag: leunga-20040319-cygwin-nlffi
1020    Description:
1021    
1022    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1023    of SML/NJ.  The problem is that
1024    
1025       lib = dlopen(NULL, ...)
1026       f   = dlsym(lib, "malloc");
1027    
1028    does not work on Windows unless we explicitly export symbols
1029    such as 'malloc' during linking.  We fixed this by explicitly
1030    exporting the required symbols with the magic gcc incantation:
1031    
1032        -Wl,--export-all cygwin.def
1033    
1034    where cygwin.def is a file containing all the symbols that
1035    we wish to export.
1036    
1037    I suspect this is a Windows problem and we'll have to
1038    do the same (somehow with windows compilers) when
1039    we build the native win32 version with the system
1040    calls LoadLibrary/GetProcAddress.
1041    
1042    ----------------------------------------------------------------------
1043    Name: Matthias Blume (blume (at) tti - c (dot) org)
1044    Date: 2004/03/04 16:35:00 CST
1045    Tag: blume-20040304-intinf-fmt
1046    Description:
1047    
1048    Fixed problem with IntInf.fmt (sign would show up on the right instead
1049    of on the left for BIN, OCT, and HEX).
1050    
1051    ----------------------------------------------------------------------
1052    Name: Matthias Blume (blume (at) tti - c (dot) org)
1053    Date: 2004/03/04 11:25:00 CST
1054    Tag: blume-20040304-symlinks
1055    Description:
1056    
1057    Fixed problem with installer script (unix only) where bin/ml-yacc and
1058    friends pointed (via symlinks) to absolute locations instead of just
1059    .run-sml.  This was reported by Vesa A Norrman.
1060    
1061    ----------------------------------------------------------------------
1062    Name: Matthias Blume (blume (at) tti - c (dot) org)
1063    Date: 2004/02/13 14:50:00 CST
1064    Tag: Release_110_45
1065    Description:
1066    
1067    New working version (110.45). New bootfiles.
1068    
1069    ----------------------------------------------------------------------
1070    Name: Matthias Blume (blume (at) tti - c (dot) org)
1071    Date: 2004/01/26 15:15:15 CST
1072    Tag: blume-20040126-toplevel
1073    Description:
1074    
1075    Improved handling of exceptions at the interactive toplevel.
1076    
1077    ----------------------------------------------------------------------
1078    Name: Matthias Blume (blume (at) tti - c (dot) org)
1079    Date: 2004/01/26 11:25:00
1080    Tag: blume-20040126-app
1081    Description:
1082    
1083    Type of top-level "app" corrected.
1084    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1085    
1086    ----------------------------------------------------------------------
1087    Name: Matthias Blume (blume (at) tti - c (dot) org)
1088    Date: 2003/11/18 17:10 CST
1089    Tag: blume-20031118-basis-fiddle
1090    Description:
1091    
1092    - changed Timer interface to what might become the spec
1093    - POSIX_FLAGS -> BIT_FLAGS according to spec
1094    - some other minor discrepancies wrt. spec eliminated
1095    
1096    ----------------------------------------------------------------------
1097    Name: Matthias Blume (blume (at) tti - c (dot) org)
1098    Date: 2003/11/06 12:00:00 CST
1099    Tag: Release_110_44
1100    Description:
1101    
1102    New working version (110.44). New bootfiles.
1103    
1104    ----------------------------------------------------------------------
1105    Name: Matthias Blume (blume (at) tti - c (dot) org)
1106    Date: 2003/11/04 11:50:00 CST
1107    Tag: blume-20031104-move-libraries
1108    Description:
1109    
1110    Eliminated the "dont_move_libraries" directive in config/targets.
1111    (The mechanism was broken and could not be fixed easily.  Moreover,
1112    there does not seem to be any reason not to move all libraries into
1113    lib during installation.  I originally implemented this directive as a
1114    backward-compatibility feature when I first introduced the new CM.
1115    Now that things have been stable for a long time and going back to the
1116    old CM is not an option, there is no reason to keep it around.)
1117    
1118    ----------------------------------------------------------------------
1119    Name: Matthias Blume (blume (at) tti - c (dot) org)
1120    Date: 2003/11/03 16:00:00 CST
1121    Tag: blume-20031103-installdir
1122    Description:
1123    
1124    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1125    Richards for pointing out the problem and providing the solution.)
1126    
1127    ----------------------------------------------------------------------
1128    Name: Matthias Blume (blume (at) tti - c (dot) org)
1129    Date: 2003/10/01 17:05:00 CDT
1130    Tag: blume-20031001-lal-mlrisc
1131    Description:
1132    
1133    MLRISC bug fix from Lal.
1134    
1135    ----------------------------------------------------------------------
1136    Name: Matthias Blume (blume (at) tti - c (dot) org)
1137    Date: 2003/09/30 16:10:00 CDT
1138    Tag: blume-20030930-primio-bat
1139    Description:
1140    
1141    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1142    2. Improved .bat files (for Win32 port) to make things work under Win95.
1143       (thanks to Aaron S. Hawley for this one)
1144    
1145    ----------------------------------------------------------------------
1146    Name: Matthias Blume (blume (at) tti - c (dot) org)
1147    Date: 2003/09/26 16:05:00 CDT
1148    Tag: blume-20030926-wrappriv
1149    Description:
1150    
1151    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1152    
1153    ----------------------------------------------------------------------
1154    Name: Matthias Blume (blume (at) tti - c (dot) org)
1155    Date: 2003/09/26 15:00:00 CDT
1156    Tag: blume-20030926-110_43_3
1157    Description:
1158    
1159    - additional cleanup
1160    - version number bump, NEW BOOTFILES
1161    
1162    ----------------------------------------------------------------------
1163    Name: Matthias Blume (blume (at) tti - c (dot) org)
1164    Date: 2003/09/26 12:00:00 CDT
1165    Tag: blume-20030926-ppautoload
1166    Description:
1167    
1168    I modified the read-eval-print loop so that the autoloader gets
1169    invoked whenever the prettyprinter tries to look up a symbol that
1170    is not currently defined in the toplevel environment but which
1171    appears in CM's autoload registry.  As a result, we see far fewer of
1172    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1173    
1174    In addition to this I tried to clean up some pieces of the Basis
1175    implementation (e.g., Socket, Word8Array) in order to prevent other
1176    instances of these ?.Foo.Bar.xxx names from being printed.
1177    
1178    The mechanism that picks names for types still needs some work, though.
1179    (Right now it seems that if there is a type A.t which is defined to
1180    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1181    "?.B.u" although the perhaps more sensible solution would be to use
1182    "A.t" in this case.  In other words, the prettyprinter should follow
1183    a chain of DEFtycs not farther than there are corresponding toplevel
1184    names in the current environment.)
1185    
1186    ----------------------------------------------------------------------
1187    Name: Matthias Blume (blume (at) tti - c (dot) org)
1188    Date: 2003/09/24 16:31:00 CDT
1189    Tag: blume-20030924-installer
1190    Description:
1191    
1192    Another installer tweak:  All the ML code for the installer is now
1193    compiled during CMB.make and put into a little library called
1194    $smlnj/installer.cm.  The installation then simply invokes
1195    
1196       sml -m $smlnj/installer.cm
1197    
1198    and everything happens automagically.
1199    
1200    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1201    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1202            and BINDIR.
1203    
1204    The new scheme guarantees that the ML code responsible for the installation
1205    is in sync with the APIs of the main system.  Also, the installer is
1206    somewhat faster because the installer script is precompiled.
1207    
1208    ----------------------------------------------------------------------
1209    Name: Matthias Blume (blume (at) tti - c (dot) org)
1210    Date: 2003/09/24 15:35:00 CDT
1211    Tag: blume-20030924-synsock
1212    Description:
1213    
1214    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1215    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1216    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1217    implementations of non-blocking operations from CML's Socket
1218    structure.
1219    
1220    ----------------------------------------------------------------------
1221    Name: Matthias Blume (blume (at) tti - c (dot) org)
1222    Date: 2003/09/24 15:10:05 CDT
1223    Tag: blume-20030924-sockets
1224    Description:
1225    
1226    1. Fixed SOCKET API and implementation to match Basis spec.
1227       This required changing the internal representation of sockets to one
1228       that remembers (for each socket file descriptor) whether it is currently
1229       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1230       call is made only if the state actually needs to change).
1231    
1232    2. OS-specific details of sockets were moved into separate files, thus
1233       making it possible to unify the bulk of the socket implementations
1234       between Unix and Win32.
1235    
1236    3. CML's socket API changed accordingly.
1237       (Note that we need to remove non-blocking functions from this API
1238       since they are redundant in the case of CML!)
1239    
1240    4. CML's socket implementation now makes use of non-blocking functions
1241       provided by Basis, thus removing all OS-dependent code from this part
1242       of CML.
1243    
1244    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1245    
1246    ----------------------------------------------------------------------
1247    Name: Matthias Blume (blume (at) tti - c (dot) org)
1248    Date: 2003/09/22 12:10:00 CDT
1249    Tag: blume-20030922-110_43_2
1250    Description:
1251    
1252    Made a new interim version and bootfiles for developer's bootstrapping
1253    convenience.
1254    
1255       110.43.2 -- NEW BOOTFILES
1256    
1257    ----------------------------------------------------------------------
1258    Name: Matthias Blume (blume (at) tti - c (dot) org)
1259    Date: 2003/09/19 15:55:00 CDT
1260    Tag: blume-20030919-cmdir
1261    Description:
1262    
1263    1. new-install.sh -> install.sh
1264    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1265    3. tweaked installer so that another name instead of .cm can be chosen
1266       at install time (by setting the CM_DIR_ARC environment variable
1267       during installation); once installation is complete, the name is
1268       fixed
1269    
1270    ----------------------------------------------------------------------
1271    Name: Matthias Blume (blume (at) tti - c (dot) org)
1272    Date: 2003/09/18 16:00:00 CDT
1273    Tag: blume-20030918-110_43_1
1274    Description:
1275    
1276    Made a new interim version and bootfiles for developer's bootstrapping
1277    convenience.
1278    
1279       110.43.1 -- NEW BOOTFILES
1280    
1281    ----------------------------------------------------------------------
1282    Name: Matthias Blume (blume (at) tti - c (dot) org)
1283    Date: 2003/09/18 15:20:00 CDT
1284    Tag: blume-20030918-misc
1285    Description:
1286    
1287    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1288       this seems to be controversial at the moment) and used those in
1289       Posix.ProcEnv.times.
1290    
1291    2. Added Time.{from,to}Nanoseconds to Time.
1292    
1293    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1294       For example, fromLargeInt never needs to look at more than 3 "big
1295       digits" to get its 53 bits of precision.
1296    
1297    ----------------------------------------------------------------------
1298    Name: Matthias Blume (blume (at) tti - c (dot) org)
1299    Date: 2003/09/17 16:30:00 CDT
1300    Tag: blume-20030917-real32-slices
1301    Description:
1302    
1303    Added an entry to the primitive environment
1304    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1305    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1306    
1307    Removed some of the "magic" constants in real64.sml and replaced them
1308    with code that generates these values from their corresponding
1309    integer counterparts.
1310    
1311    Made all(?) the slice-related changes to the Basis and made everything
1312    compile again...
1313    
1314    ----------------------------------------------------------------------
1315    Name: Matthias Blume (blume (at) tti - c (dot) org)
1316    Date: 2003/09/15 17:45:00 CDT
1317    Tag: blume-20030915-rbase
1318    Description:
1319    
1320    Fixed bug in Real.fromLargeInt.
1321    
1322    ----------------------------------------------------------------------
1323    Name: Matthias Blume (blume (at) tti - c (dot) org)
1324    Date: 2003/09/13 18:11:00 CDT
1325    Tag: blume-20030913-libinstall
1326    Description:
1327    
1328    Minor bugfix in config/libinstall (set anchor with path to
1329    standalone tool after installing it, otherwise libraries that
1330    need ml-lex or ml-yacc won't compile the first time the installer
1331    runs).
1332    
1333    ----------------------------------------------------------------------
1334    Name: Matthias Blume (blume (at) tti - c (dot) org)
1335    Date: 2003/09/12 11:45:00 CDT
1336    Tag: blume-20030912-various
1337    Description:
1338    
1339     - fixed bug in Real.toLargeInt
1340     - fixed bug in Posix.ProcEnv.times
1341     - changed inputLine functions to return an option
1342     - minor installer improvements / bugfixes
1343     - changed default @SMLalloc parameter for x86/celeron to 64k
1344    
1345    ----------------------------------------------------------------------
1346    Name: Matthias Blume (blume (at) tti - c (dot) org)
1347    Date: 2003/09/09 22:00:00 CDT
1348    Tag: Release_110_43
1349    Description:
1350    
1351        New working release 110.43.  New bootfiles.
1352    
1353    ----------------------------------------------------------------------
1354    Name: Matthias Blume (blume (at) tti - c (dot) org)
1355    Date: 2003/09/09 19:20:00 CDT
1356    Tag: blume-20030909-installer
1357    Description:
1358    
1359    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1360    Modified config/install.bat to take advantage of it.  Also modified
1361    config/install.sh (and called it config/new-install.sh) to take advantage
1362    of it on Unix systems.  (The SML code is (supposed to be) platform-
1363    independent.)
1364    
1365    The installer can now install everything under Win32
1366    as well as under *nix as long as it compiles.
1367    
1368    Other changes:
1369    
1370      - made CML compile again under Win32
1371      - made eXene compile under Win32 (by providing a fake structure UnixSock
1372        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1373      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1374        same as type int;  under Win32 it isn't
1375      - fixed some slice-related problems in the win32-specific parts of CML
1376      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1377        (under Win32, the volume name is case-insensitive, and the
1378        OS.Path code compares volume names for equality)
1379    
1380    ----------------------------------------------------------------------
1381    Name: Matthias Blume (blume (at) tti - c (dot) org)
1382    Date: 2003/09/08 11:55:00 CDT
1383    Tag: blume-20030908-fullpath
1384    Description:
1385    
1386    Made Win32 version of OS.FileSys.fullPath return current directory
1387    when given an empty string.  This is what the spec says, and incidentally,
1388    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1389    cases when presented with the name of a non-existing .cm file.)
1390    
1391    ----------------------------------------------------------------------
1392    Name: Matthias Blume (blume (at) tti - c (dot) org)
1393    Date: 2003/09/04 16:30:00 CDT
1394    Tag: blume-20030905-slices-etc
1395    Description:
1396    
1397      1. Changed interface to vectors and arrays in Basis to match
1398         (draft) Basis spec.
1399      2. Added signatures and implementations of slices according to
1400         Basis spec.
1401      3. Edited source code throughout the system to make it compile again
1402         under 1. and 2.  (In some cases code had to be added to have it
1403         match the new signatures.)
1404      4. MLRISC should be backward-compatible: the copies of the originals
1405         of files that needed to change under 3. were retained, the .cm files
1406         check the compiler version number and use old versions when
1407         appropriate.
1408      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1409         dirstream -> string option (in accordance with Basis spec).
1410      6. When generating code that counts lines, ml-lex used function
1411         CharVector.foldli, taking advantage of its old interface.
1412         This has been replaced with the corresponding code from
1413         CharVectorSlice.  (html-lex must be re-lexed!)
1414      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1415         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1416      8. Removed temporary additions (fromInternal, toInternal) from the
1417         (now obsolete) IntInf in smlnj-lib/Util.
1418      9. Cleaned up structure Byte.
1419     10. Added localOffset, scan, and fromString to Date (according to spec).
1420         Cleaned/corrected implementation of Date.
1421         (Still need to check for correctness; implement better canonicalizeDate.)
1422     11. Added "scan" to signature IEEE_REAL.
1423     12. Some improvements to IntInf [in particular: efficiency-hack for
1424         mod and rem when second operand is 2 (for parity checks).]
1425     13. Changed representation of type Time.time, using a single IntInf.int
1426         value counting microseconds.  This considerably simplified the
1427         implementation of structure Time.  We now support negative time
1428         values; scan and fromString handle signs.
1429     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1430         ArraySlice).
1431    
1432    ----------------------------------------------------------------------
1433    Name: Matthias Blume (blume (at) tti - c (dot) org)
1434    Date: 2003/08/28 17:00:00 CDT
1435    Tag: blume-20030828-intinf
1436    Description:
1437    
1438    This is a major update which comes with a version number bump
1439    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1440    and an implementation of IntInf in the Basis.
1441    
1442    There are a fairly large number of related changes and updates throughout
1443    the system:
1444    
1445    Basis:
1446            - Implemented IntInf.
1447            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1448            - Added some missing Real64 operations, most notably Real.toLargeInt.
1449            - Added FixedInt as a synonym for Int32.
1450    
1451    compiler:
1452            * Added support for a built-in intinf type.
1453              - literals
1454              - pattern matching
1455              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1456              - overloading on literals and operations
1457    
1458            This required adding a primitive type intinf, some additional
1459            primops, and implementations for several non-trivial intinf
1460            operations in Core.  (The intinf type is completely abstract
1461            to the compiler; all operations get delegated back to the Core.)
1462    
1463            * Intinf equality is handled by polyequal.  However, the compiler
1464            does not print its usual warning in this case (since polyequal
1465            is the right thing to do there).
1466    
1467            * Improved the organization of structure InlineT.
1468    
1469            * A word about conversion primops:
1470            If conversions involving intinf do not cancel out during
1471            CPS contract, then the compiler must insert calls to Core functions.
1472            Since all core access must be resolved already during the FLINT
1473            translate phase, it would be too late a the time of CPS contract
1474            to add new Core calls.  For this reason, conversion primops
1475            for intinf carry two arguments:  1. the numeric argument that
1476            they are supposed to convert, and 2. the Core function that
1477            can help with this conversion if necessary.  If CPS contract
1478            eliminates a primop, then the associated Core function becomes
1479            dead and goes away.  Intinf conversion primops that do not get
1480            eliminated by CPS contract get rewritten into calls of their
1481            core functions by a separate, new phase.
1482    
1483    interactive system:
1484            - Control.Print.intinfDepth controls max length of intinf constants
1485            being printed. (Analogous to Control.Print.stringDepth.)
1486            - Cleanup in printutil and pputil: got rid of unused stuff and
1487            duplicates; replaced some of the code with code that makes better
1488            use of library functionality.
1489    
1490    CM:
1491            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1492            are no longer silent.
1493    
1494    CKIT:
1495            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1496            in favor of LargeInt -- which is now the same as IntInf.
1497            CKIT-knowledgable people should check whether this is what's
1498            intended and otherwise change things back to using Int32 or
1499            FixedInt.
1500    
1501    Throughout the code:
1502            Started using IntInf.int literals and built-in operations
1503            (e.g., comparison with 0) where this seems appropriate.
1504    
1505    
1506    ----------------------------------------------------------------------
1507    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1508    Date: 2003/08/13 11:36:00 CDT
1509    Tag: dbm-20030813-mcz-merge1
1510    Description:
1511    
1512    Merging changes from the mcz-branch development branch into trunk.
1513    These changes involve replacement of the emulated old prettyprinter
1514    interface with direct use of the SML/NJ Lib PP library, and fixing
1515    of a couple of bugs (895, 1186) relating to error messages.  A new
1516    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1517    has been added.
1518    
1519    ----------------------------------------------------------------------
1520    Name: Matthias Blume (blume (at) tti - c (dot) org)
1521    Date: 2003/08/11 15:45:00 CDT
1522    Tag: blume-20030811-windows
1523    Description:
1524    
1525    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1526    
1527        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1528    
1529    This patch restores SML/NJ's ability to run under win32.  There are a
1530    number of changes, including fixes for several bugs that had gone
1531    unnoticed until now:
1532    
1533      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1534        trying to run the win32 version from within cygwin.)
1535    
1536      - There are a number of simple .bat scripts that substitute for their
1537        corresponding Unix shell-scripts.  (See below.)
1538    
1539      - The internals of ml-build have been modified slightly.  The main
1540        difference is that instead of calling ".link-sml" (or link-sml.bat)
1541        using OS.Process.system, the ML process delegates this task back
1542        to the script.  Otherwise problems arise in mixed environments such
1543        as Cygwin where scripts look and work like Unix scripts, but
1544        where OS.Process.system cannot run them.
1545    
1546      - In CM, the srcpath pickler used native pathname syntax -- which
1547        is incorrect in the case of cross-compilation.  The new pickle format
1548        is independent of platform-specific naming conventions.
1549    
1550      - Path configuration files (such as lib/pathconfig) can now choose
1551        between native and standard syntax.  Placing a line of the form
1552    
1553            standard!
1554    
1555        into the file causes all subsequent paths to be interpreted using
1556        CM standard pathname syntax (= Unix conventions); a line
1557    
1558            native!
1559    
1560        switches back to native style.  This was needed so that
1561        path config files can be written portably, see src/system/pathconfig.
1562    
1563      - Runtime system:
1564    
1565          - win32-filesys.c:  get_file_time and set_file_time now
1566            access modification time, not creation time.
1567    
1568          - I/O code made aware of new array representation.
1569    
1570          - Bug fixes in X86.prim.masm.
1571    
1572       - src/system/makeml made aware of win32. (For use under cygwin
1573            and other Unix-environments for windows.)
1574    
1575       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1576         which caused BinIO.inputAll to fail consistently.
1577    
1578    .bat scripts:
1579    
1580       Windows .bat scripts assume that SMLNJ_HOME is defined.
1581    
1582       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1583         applications (sml, ml-yacc, ml-lex).
1584       - ml-build.bat: analogous to ml-build.
1585       - config\install.bat:  Analogous to config/install.sh.  This requires
1586         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1587         (nmake etc. must be on the path, and vcvars32 must have been run.)
1588         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1589         and the bootfile hierarchy must have been unpacked under
1590         sml.boot.x86-win32.
1591         The script is very primitive and does a poor job at error checking.
1592         It only installs the base system, ml-lex, and ml-yacc.  No other
1593         libraries are being installed (i.e., you get only those that
1594         are part of the compiler.)
1595       - link-sml.bat:  analogous to .link-sml, but not currently used
1596    
1597    Unrelated bug fixes:
1598    
1599      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1600        types.
1601      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1602        otherwise.)
1603    
1604    ----------------------------------------------------------------------
1605    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1606    Date: 2003/06/17
1607    Tag: macqueen-20030617-bug895
1608    Description:
1609    
1610    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1611    Tag will be used for new development branch (mcz-branch) for
1612    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1613    
1614    ----------------------------------------------------------------------
1615    Name: Matthias Blume (blume (at) tti - c (dot) org)
1616    Date: 2003/05/27 16:55:00 CDT
1617    Tag: blume-20030527-polyeq
1618    Description:
1619    
1620    Tried to eliminated most cases of polymorphic equality.
1621    
1622    ----------------------------------------------------------------------
1623    Name: Matthias Blume (blume (at) tti - c (dot) org)
1624    Date: 2003/05/21 17:45:00 CDT
1625    Tag: blume-20030517-complete
1626    Description:
1627    
1628    Two changes:
1629    
1630      1. Added a flag for controlling whether non-exhaustive bindings will
1631         be treated as errors (default is false).
1632      2. Cleaned up the *entire* source tree so that CMB.make goes through
1633         without a single non-exhaustive match- or bind warning.
1634    
1635    ----------------------------------------------------------------------
1636    Name: Matthias Blume (blume (at) tti - c (dot) org)
1637    Date: 2003/05/17 10:20:00 CDT
1638    Tag: blume-20030517-absyn
1639    Description:
1640    
1641    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1642    
1643       This mainly affects the quality of error messages.  However, some
1644       of the code is now more straightforward than before. (Treatment of
1645       the above four constructs in translate.sml is much simpler than
1646       the "macro-expansion" that was going on before.  Plus, the mach-
1647       compiler no longer gets invoked just to be able to compile an
1648       if-expression.)
1649    
1650    2. The ErrorMsg.Error exception is now caught and absorbed by the
1651       interactive loop.
1652    
1653    ----------------------------------------------------------------------
1654    Name: Allen Leung
1655    Date: 2003/05/16 13:05:00 CDT
1656    Tag: leunga-20030516-cygwin-runtime
1657    Description:
1658    
1659        Ported the runtime system to cygwin, which uses the unix
1660    x86-unix bin files.  Missing/buggy features:
1661    
1662       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1663         the Cygwin library.
1664       o Ctrl-C handling may be flaky.
1665       o Windows system calls and Windows I/O are not supported.
1666    
1667       A new set of binfiles is located at:
1668    
1669         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1670    
1671       This is only needed for bootstrapping the cygwin version of smlnj.
1672       Other x86 versions can use the existing binfiles.
1673    
1674    ----------------------------------------------------------------------
1675    Name: Matthias Blume
1676    Date: 2003/04/08 15:42:00 CDT
1677    Tag: blume-20030408-listpair
1678    Description:
1679    
1680    1. Added a target 'mlrisc' to installer.
1681    
1682    2. Added missing elements to structure ListPair.
1683    
1684    ----------------------------------------------------------------------
1685    Name: Allen Leung
1686    Date: 2003/01/07 10:40:00 EST
1687    Tag: leunga-20030107-int-rem
1688    Description:
1689    
1690    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1691    The arguments to the SUBL instruction were swapped.
1692    
1693    ----------------------------------------------------------------------
1694    Name: Matthias Blume
1695    Date: 2002/12/12 16:25:00 EST
1696    Tag: blume-20021212-risc-ra
1697    Description:
1698    
1699    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1700    sent the RA into an infinite loop when floating point registers get
1701    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1702    110.37 and 110.38.)
1703    There was another set of potential problems related to the handling of
1704    MLRISC annotations (but those did not yet cause real problems, apparently).
1705    
1706    ----------------------------------------------------------------------
1707    Name: Matthias Blume
1708    Date: 2002/12/06 22:40:00 EST
1709    Tag: blume-20021206-cm-fileid
1710    Description:
1711    
1712    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1713    This fixes the problem of CM getting confused by files that suddenly
1714    change their identity (e.g., by getting unlinked and recreated by some
1715    text editor such as vi).  There might be a better/cheaper/cleaner way
1716    of doing this, but for now this will have to do.
1717    
1718    ----------------------------------------------------------------------
1719    Name: Matthias Blume
1720    Date: 2002/10/28 09:50:00 EST
1721    Tag: blume-20021028-typecheck
1722    Description:
1723    
1724    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1725    
1726    ----------------------------------------------------------------------
1727    Name: Matthias Blume
1728    Date: 2002/10/17 09:10:00 EDT
1729    Tag: Release_110_42
1730    Description:
1731    
1732    In good old tradition, there has been a slight hiccup so that we have
1733    to patch 110.42 after the fact.   The old release tag has been replaced
1734    (see below).
1735    
1736    The change solves a problem with two competing approaches the
1737    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1738    each other's way.
1739    
1740    This change only affects the runtime system code and the installer script.
1741    (No new bootfiles.)
1742    
1743    ----------------------------------------------------------------------
1744    Name: Matthias Blume
1745    Date: 2002/10/16 12:00:00 EDT
1746    Tag: Release_110_42_removed
1747    Description:
1748    
1749    New working release.  New bootfiles.
1750    
1751    ----------------------------------------------------------------------
1752    Name: Matthias Blume
1753    Date: 2002/10/10 13:10:00 EDT
1754    Tag: blume-20021010-ppc-divs
1755    Description:
1756    
1757    The mltree operator DIVS must be implemented with an overflow check on
1758    the PPC because the hardware indicates divide-by-zero using "overflow" as
1759    well.
1760    
1761    ----------------------------------------------------------------------
1762    Name: Matthias Blume
1763    Date: 2002/07/23 11:45:00
1764    Tag: blume-20020723-smlnj-home
1765    Description:
1766    
1767    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1768    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1769    CM_PATHCONFIG is also set), the path configuration file is assumed
1770    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1771    the entire tree to some other place and everything will "just work".
1772    
1773    (Companion commands such as ml-build and ml-makedepend also sense this
1774    variable.)
1775    
1776    ----------------------------------------------------------------------
1777    Name: Matthias Blume
1778    Date: 2002/07/12 21:19:00 EDT
1779    Tag: blume-20020712-liveness
1780    Description:
1781    
1782    Exported two useful "step" functions from liveness module (MLRISC).
1783    
1784    ----------------------------------------------------------------------
1785    Name: Matthias Blume
1786    Date: 2002/07/05 16:00 EDT
1787    Tag: Release_110_41
1788    Description:
1789    
1790    New working release.  New bootfiles.
1791    
1792    ----------------------------------------------------------------------
1793    Name: Matthias Blume
1794    Date: 2002/07/05 10:25:00 EDT
1795    Tag: blume-20020705-btimp
1796    Description:
1797    
1798    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1799    other clients can set up backtracing support.
1800    
1801    ----------------------------------------------------------------------
1802    Name: Matthias Blume
1803    Date: 2002/06/25 17:23:00 EDT
1804    Tag: blume-20020625-fpmax
1805    Description:
1806    
1807    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1808    case.  (The sense of the isNaN test was reversed -- which made min and
1809    max always return their first argument.)
1810    
1811    ----------------------------------------------------------------------
1812    Name: Matthias Blume
1813    Date: 2002/06/11
1814    Tag: blume-20020611-unixpath
1815    Description:
1816    
1817    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1818    
1819    ----------------------------------------------------------------------
1820    Name: Matthias Blume
1821    Date: 2002/06/10 16:35:00 EDT
1822    Tag: blume-20020610-ieeereal
1823    Description:
1824    
1825    I back-ported my implementation of IEEEReal.fromString from the
1826    idlbasis-devel branch so that we can test it.
1827    
1828    Another small change is that ppDec tries to give more information
1829    than just "<sig>" in the case of functors.  However, this code is
1830    broken in some mysterious way if the functor's body's signature
1831    has not been declared by ascription but gets inferred from the
1832    implementation.  This needs fixing...
1833    
1834    ----------------------------------------------------------------------
1835    Name: Matthias Blume
1836    Date: 2002/05/31
1837    Tag: blume-20020531-btrace-mode
1838    Description:
1839    
1840    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1841    the wayside when I switched over to using Controls everywhere.)
1842    
1843    ----------------------------------------------------------------------
1844    Name: Lal George
1845    Date: 2002/05/23 12:21:40 EDT
1846    Tag: george-20020523-visual-labels
1847    Description:
1848    
1849    Labels are now displayed in the graphical output to make
1850    the fall-through and target blocks obvious.
1851    
1852    ----------------------------------------------------------------------
1853    Name: Matthias Blume
1854    Date: 2002/05/22 11:03:00 EDT
1855    Tag: blume-20020522-shrink
1856    Description:
1857    
1858    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1859    situation that comes up when allocating huge arrays.
1860    
1861    ----------------------------------------------------------------------
1862    Name: Matthias Blume
1863    Date: 2002/05/21 16:00:00 EDT
1864    Tag: Release_110_40
1865    Description:
1866    
1867    New working release (110.40).  New bootfiles.
1868    
1869    [Also: John Reppy fixed GC bug 1131.]
1870    
1871    ----------------------------------------------------------------------
1872    Name: Matthias Blume
1873    Date: 2002/05/21 12:35:00 EDT
1874    Tag: blume-20020521-cmdoc
1875    Description:
1876    
1877    CM documentation update.
1878    
1879    ----------------------------------------------------------------------
1880    Name: Matthias Blume
1881    Date: 2002/05/21 10:55:00 EDT
1882    Tag: blume-20020521-misc
1883    Description:
1884    
1885    - John tweaked runtime to be silent on heap export (except when
1886      GC messages are on).
1887    - I added a few more things (cross-compiling versions of CMB) to
1888      config/preloads (as suggestions).
1889    
1890    ----------------------------------------------------------------------
1891    Name: Matthias Blume
1892    Date: 2002/05/20 22:25:00 EDT
1893    Tag: blume-20020520-controls
1894    Description:
1895    
1896    - Added ControlUtil structure to control-lib.cm.
1897    - Use it throughout.
1898    - Used Controls facility to define MLRISC controls (as opposed to
1899      registering MLRISC control ref cells with Controls after the fact)
1900    - Fixed messed-up controls priorities.
1901    
1902    * Removed again all the stuff from config/preloads that one wouldn't
1903      be able to preload at the time the initial heap image is built.
1904      (Many libraries, e.g., CML, do not exist yet at this time.  The
1905       only libraries that can be preloaded via config/preloads are those
1906       that come bundled with the bootfiles.)
1907    
1908    ----------------------------------------------------------------------
1909    Name: Matthias Blume
1910    Date: 2002/05/20 10:59:00 EDT
1911    Tag: blume-20020520-preloads
1912    Description:
1913    
1914    Added a lot of commented-out suggestions for things to be included
1915    in config/preloads.
1916    
1917    ----------------------------------------------------------------------
1918    Name: Allen Leung
1919    Date: 2002/05/18 14:20:00 EDT
1920    Tag: leunga-20020518-mdl
1921    Description:
1922    
1923    o Made the mdl tool stuff compile and run again.
1924    o I've disabled all the stuff that depends on RTL specifications; they
1925      are all badly broken anyway.
1926    
1927    ----------------------------------------------------------------------
1928    Name: Matthias Blume
1929    Date: 2002/05/17 16:49:00 EDT
1930    Tag: blume-20020517-controls
1931    Description:
1932    
1933    0. John Reppy made several modifications to the SML/NJ library.
1934       In particular, there is a shiny new controls-lib.cm.
1935    
1936    1. Pushed new controls interface through compiler so that everything
1937       compiles again.
1938    
1939    2. Added FormatComb and FORMAT_COMB to the CML version of the
1940       SML/NJ library (so that CML compiles again).
1941    
1942    3. Modified init scripts because XXX_DEFAULT environment variables
1943       are no longer with us.  (Boot-time initialization is now done
1944       using the same environment variables that are also used for
1945       startup-time initialization of controls.)
1946    
1947    ----------------------------------------------------------------------
1948    Name: Lal George
1949    Date: 2002/05/15 09:20:10 EDT
1950    Tag: george-20020515-pseudo-op-decls
1951    Description:
1952    
1953            All pseudo-ops emitted before the first segment declaration
1954    such as TEXT, DATA, and BSS directives are assumed to be global
1955    declarations and are emitted first in the assembly file. This is
1956    useful in a number of situations where one has pseudo-ops that are not
1957    specific to any segment, and also works around the constraint that one
1958    cannot have client pseudo-ops in the TEXT segment.
1959    
1960            Because no segment is associated with these declarations it is
1961    an error to allocate any space or objects before the first segment
1962    directive and an exception will be raised. However, we cannot make
1963    this check for client pseudo-ops.
1964    
1965            These top level declarations are a field in the CFG graph_info.
1966    In theory you can continue to add to this field after the CFG has been
1967    built -- provided you know what you are doing;-)
1968    
1969    ----------------------------------------------------------------------
1970    Name: Matthias Blume
1971    Date: 2002/05/13 16:40:00 EDT
1972    Tag: blume-20020513-pp-etc
1973    Description:
1974    
1975    A few minor bugfixes:
1976    
1977      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1978        (Bogus printouts for redefined bindings still occur.  Compiler
1979        bug should no longer occur now.  We need to redo the prettyprinter
1980        from scratch.)
1981    
1982      - CM pathname printer now also adds escape sequences for ( and )
1983    
1984      - commend and docu fixes for ml-nlffi
1985    
1986    ----------------------------------------------------------------------
1987    Name: Matthias Blume
1988    Date: 2002/05/10 16:40:00 EDT
1989    Tag: blume-20020510-erg-textio
1990    Description:
1991    
1992    Applied the following bugfix provided by Emden Gansner:
1993    
1994        Output is corrupted when outputSubstr is used rather than output.
1995    
1996        The problem occurs when a substring
1997    
1998            ss = (s, dataStart, dataLen)
1999    
2000        where dataStart > 0, fills a stream buffer with avail bytes left.
2001        avail bytes of s, starting at index dataStart, are copied into the
2002        buffer, the buffer is flushed, and then the remaining dataLen-avail
2003        bytes of ss are copied into the beginning of the buffer. Instead of
2004        starting this copy at index dataStart+avail in s, the current code
2005        starts the copy at index avail.
2006    
2007        Fix:
2008        In text-io-fn.sml, change line 695 from
2009             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2010        to
2011             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2012    
2013    ----------------------------------------------------------------------
2014    Name: Matthias Blume
2015    Date: 2002/04/12 13:55:00 EDT
2016    Tag: blume-20020412-assyntax
2017    Description:
2018    
2019    1. Grabbed newer assyntax.h from the XFree86 project.
2020    2. Fiddled with how to compile X86.prim.asm without warnings.
2021    3. (Very) Minor cleanup in CM.
2022    
2023    ----------------------------------------------------------------------
2024    Name: Matthias Blume
2025    Date: 2002/04/01 (no joke!) 17:07:00 EST
2026    Tag: blume-20020401-x86div
2027    Description:
2028    
2029    Added full support for div/mod/rem/quot on the x86, using the machine
2030    instruction's two results (without clumsily recomputing the remainder)
2031    directly where appropriate.
2032    
2033    Some more extensive power-of-two support was added to the x86 instruction
2034    selector (avoiding expensive divs, mods, and muls where they can be
2035    replaced with cheaper shifts and masks).  However, this sort of thing
2036    ought to be done earlier, e.g., within the CPS optimizer so that
2037    all architectures benefit from it.
2038    
2039    The compiler compiles to a fixed point, but changes might be somewhat
2040    fragile nevertheless.  Please, report any strange things that you might
2041    see wrt. div/mod/quot/rem...
2042    
2043    ----------------------------------------------------------------------
2044    Name: Matthias Blume
2045    Date: 2002/03/29 17:22:00
2046    Tag: blume-20020329-div
2047    Description:
2048    
2049    Fixed my broken div/mod logic.  Unfortunately, this means that the
2050    inline code for div/mod now has one more comparison than before.
2051    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2052    The problem was with quotient = 0, because that alone does not tell
2053    us which way the rounding went.  One then has to look at whether
2054    remainder and divisor have the same sign...  :(
2055    
2056    Anyway, I replaced the bootfiles with fresh ones...
2057    
2058    ----------------------------------------------------------------------
2059    Name: Matthias Blume
2060    Date: 2002/03/29 14:10:00 EST
2061    Tag: blume-20020329-inlprims
2062    Description:
2063    
2064    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2065    
2066    Primops have changed. This means that the bin/boot-file formats have
2067    changed as well.
2068    
2069    To make sure that there is no confusion, I made a new version.
2070    
2071    
2072    CHANGES:
2073    
2074    * removed REMT from mltree (remainder should never overflow).
2075    
2076    * added primops to deal with divisions of all flavors to the frontend
2077    
2078    * handled these primops all the way through so they map to their respective
2079      MLRISC support
2080    
2081    * used these primops in the implementation of Int, Int32, Word, Word32
2082    
2083    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2084    
2085    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2086    
2087    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2088      floating point
2089    
2090    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2091      and Real (but Real.abs maps to a separate floating-point-only primop)
2092    
2093    
2094    TODO items:
2095    
2096    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2097      instructions because the machine instruction encoder cannot handle
2098      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2099      REM in the first place, and REM is supported by the x86 machine coder.)
2100    
2101    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2102      selection phase.  (The two can be streamlined because the hardware
2103      delivers both quotient and remainder at the same time anyway.)
2104    
2105    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2106      (Currently the behavior is inconsistent both across architectures and
2107      wrt. the draft Basis spec.)
2108    
2109    * Word8 should eventually be handled natively, too.
2110    
2111    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2112      as if there currently is no execution path that could trigger it in
2113      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2114      hold for things like multiplication and division.)
2115    
2116    ----------------------------------------------------------------------
2117    Name: Matthias Blume
2118    Date: 2002/03/27 16:27:00 EST
2119    Tag: blume-20020327-mlrisc-divisions
2120    Description:
2121    
2122    Added support for all four division operations (ML's div, mod, quot,
2123    and rem) to MLRISC.  In the course of doing so, I also rationalized
2124    the naming (no more annoying switch-around of DIV and QUOT), by
2125    parameterizing the operation by div_rounding_mode (which can be either
2126    DIV_TO_ZERO or DIV_TO_NEGINF).
2127    
2128    The generic MLTreeGen functor takes care of compiling all four
2129    operations down to only round-to-zero div.
2130    
2131    Missing pieces:
2132    
2133      * Doing something smarter than relying on MLTreeGen on architectures
2134        like, e.g., the x86 where hardware division delivers both quotient and
2135        remainder at the same time.  With this, the implementation of the
2136        round-to-neginf operations could be further streamlined.
2137    
2138      * Remove inlining support for div/mod/rem from the frontend and replace it
2139        with primops that get carried through to the backend.  Do this for all
2140        int and word types.
2141    
2142    ----------------------------------------------------------------------
2143    Name: Matthias Blume
2144    Date: 2002/03/25 17:25:00 EST
2145    Tag: blume-20020325-divmod
2146    Description:
2147    
2148    I improved (hopefully without breaking them) the implementation of Int.div,
2149    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2150    advantage of the following observations:
2151    
2152      Let  q = x quot y      r = x rem y
2153           d = x div  y      m = x mod y
2154    
2155    where "quot" is the round-to-zero version of integer division that
2156    hardware usually provides.  Then we have:
2157    
2158         r = x - q * y        where neither the * nor the - will overflow
2159         d = if q >= 0 orelse x = q * y then q else q - 1
2160                              where neither the * nor the - will overflow
2161         m = if q >= 0 orelse r = 0 then r else r + y
2162                              where the + will not overflow
2163    
2164    This results in substantial simplification of the generated code.
2165    The following table shows the number of CFG nodes and edges generated
2166    for
2167            fun f (x, y) = x OPER y
2168            (* with OPER \in div, mod, quot, rem *)
2169    
2170    
2171        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2172        --------------------------------------------------------
2173         div |         24 |         39 |         12 |         16
2174         mod |         41 |         71 |         12 |         16
2175        quot |          8 |         10 |          8 |         10
2176         rem |         10 |         14 |          8 |         10
2177    
2178    
2179    ----------------------------------------------------------------------
2180    Name: Matthias Blume
2181    Date: 2002/03/25 22:06:00 EST
2182    Tag: blume-20020325-cprotobug
2183    Description:
2184    
2185    Fixed a bug in cproto (c prototype decoder).
2186    
2187    ----------------------------------------------------------------------
2188    Name: Matthias Blume
2189    Date: 2002/03/25 16:00:00 EST
2190    Tag: blume-20020325-raw-primops
2191    Description:
2192    
2193    I did some cleanup to Allen's new primop code and
2194    replaced yesterday's bootfiles with new ones.
2195    (But they are stored in the same place.)
2196    
2197    ----------------------------------------------------------------------
2198    Name: Matthias Blume
2199    Date: 2002/03/24 22:40:00 EST
2200    Tag: blume-20020324-bootfiles
2201    Description:
2202    
2203    Made the bootfiles that Allen asked for.
2204    
2205    ----------------------------------------------------------------------
2206    Name: Allen Leung
2207    Date: 2002/03/23 15:50:00 EST
2208    Tag: leunga-20020323-flint-cps-rcc-primops
2209    Description:
2210    
2211      1. Changes to FLINT primops:
2212    
2213        (* make a call to a C-function;
2214         * The primop carries C function prototype information and specifies
2215         * which of its (ML-) arguments are floating point. C prototype
2216         * information is for use by the backend, ML information is for
2217         * use by the CPS converter. *)
2218      | RAW_CCALL of { c_proto: CTypes.c_proto,
2219                       ml_args: ccall_type list,
2220                       ml_res_opt: ccall_type option,
2221                       reentrant : bool
2222                     } option
2223       (* Allocate uninitialized storage on the heap.
2224        * The record is meant to hold short-lived C objects, i.e., they
2225        * are not ML pointers.  With the tag, the representation is
2226        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2227        *)
2228      | RAW_RECORD of {tag:bool,sz:int}
2229      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2230    
2231      2.  These CPS primops are now overloaded:
2232    
2233           rawload of {kind:numkind}
2234           rawstore of {kind:numkind}
2235    
2236          The one argument form is:
2237    
2238             rawload {kind} address
2239    
2240          The two argument form is:
2241    
2242             rawload {kind} [ml object, byte-offset]
2243    
2244      3. RAW_CCALL/RCC now takes two extra arguments:
2245    
2246         a. The first is whether the C call is reentrant, i.e., whether
2247            ML state should be saved and restored.
2248         b. The second argument is a string argument specifying the name of
2249            library and the C function.
2250    
2251         These things are currently not handled in the code generator, yet.
2252    
2253      4. In CProto,
2254    
2255         An encoding type of "bool" means "ml object" and is mapped into
2256         C prototype of PTR.  Note that "bool" is different than "string",
2257         even though "string" is also mapped into PTR, because "bool"
2258         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2259    
2260      5. Pickler/unpicker
2261    
2262         Changed to handle RAW_RECORD and newest RAW_CCALL
2263    
2264      6. MLRiscGen,
2265    
2266         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2267         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2268            in the file CodeGen/cpscompile/cps-c-calls.sml
2269    
2270      7. Added the conditional move operator
2271    
2272             condmove of branch
2273    
2274         to cps.  Generation of this is still buggy so it is currently
2275         disabled.
2276    
2277    ----------------------------------------------------------------------
2278    Name: Lal George
2279    Date: 2002/03/22 14:18:25 EST
2280    Tag: george-20020322-cps-branch-prob
2281    Description:
2282    
2283    Implemented the Ball-Larus branch prediction-heuristics, and
2284    incorporated graphical viewers for control flow graphs.
2285    
2286    Ball-Larus Heuristics:
2287    ---------------------
2288    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2289    
2290    By design it uses the Dempster-Shafer theory for combining
2291    probabilities.  For example, in the function:
2292    
2293        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2294    
2295    the ball-larus heuristics predicts that the n=0 is unlikely
2296    (OH-heuristic), and the 'then' branch is unlikely because of the
2297    RH-heuristic -- giving the 'then' branch an even lower combined
2298    probability using the Dempster-Shafer theory.
2299    
2300    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2301    probability of the 'then' branch because of the loop in the else
2302    branch.
2303    
2304    
2305    Graphical Viewing:
2306    ------------------
2307    I merely plugged in Allen's graphical viewers into the compiler. The
2308    additional code is not much. At the top level, saying:
2309    
2310            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2311    
2312    will display the graphical view of the control flow graph just before
2313    back-patching.  daVinci must be in your path for this to work. If
2314    daVinci is not available, then the default viewer can be changed
2315    using:
2316    
2317            Control.MLRISC.getString "viewer"
2318    
2319    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2320    course, these viewers must be in your path.
2321    
2322    The above will display the compilation unit at the level of clusters,
2323    many of which are small, boring, and un-interesting. Also setting:
2324    
2325            Control.MLRISC.getInt "cfg-graphical-view_size"
2326    
2327    will display clusters that are larger than the value set by the above.
2328    
2329    
2330    ----------------------------------------------------------------------
2331    Name: Matthias Blume
2332    Date: 2002/03/21 22:20:00 EST
2333    Tag: blume-20020321-kmp-bugfix
2334    Description:
2335    
2336    Changed the interface to the KMP routine in PreString and fixed
2337    a minor bug in one place where it was used.
2338    
2339    ----------------------------------------------------------------------
2340    Name: Allen Leung
2341    Date: 2002/03/21 20:30:00 EST
2342    Tag: leunga-20020321-cfg
2343    Description:
2344    
2345      Fixed a potential problem in cfg edge splitting.
2346    
2347    ----------------------------------------------------------------------
2348    Name: Allen Leung
2349    Date: 2002/03/21 17:15:00 EST
2350    Tag: leunga-20020321-x86-fp-cfg
2351    Description:
2352    
2353      1. Recoded the buggy parts of x86-fp.
2354    
2355         a. All the block reordering code has been removed.
2356            We now depend on the block placement phases to do this work.
2357    
2358         b. Critical edge splitting code has been simplified and moved into the
2359            CFG modules, as where they belong.
2360    
2361         Both of these were quite buggy and complex.  The code is now much, much
2362         simpler.
2363    
2364      2. X86 backend.
2365    
2366         a. Added instructions for 64-bit support.  Instruction selection for
2367            64-bit has not been committed, however, since that
2368            requires changes to MLTREE which haven't been approved by
2369            Lal and John.
2370    
2371         b. Added support for FUCOMI and FUCOMIP when generating code for
2372            PentiumPro and above.  We only generate these instructions in
2373            the fast-fp mode.
2374    
2375         c. Added cases for JP and JNP in X86FreqProps.
2376    
2377      3. CFG
2378    
2379         CFG now has a bunch of methods for edge splitting and merging.
2380    
2381      4. Machine description.
2382    
2383         John's simplification of MLTREE_BASIS.fcond broke a few machine
2384         description things:
2385    
2386         rtl-build.{sig,sml} and hppa.mdl fixed.
2387    
2388         NOTE: the machine description stuff in the repository is still broken.
2389               Again, I can't put my fixes in because that involves
2390               changes to MLTREE.
2391    
2392    ----------------------------------------------------------------------
2393    Name: Matthias Blume
2394    Date: 2002/03/20 15:55:00 EST
2395    Tag: blume-20020320-kmp
2396    Description:
2397    
2398    Implemented Knuth-Morris-Pratt string matching in PreString and used
2399    it for String.isSubstring, Substring.isSubstring, and
2400    Substring.position.
2401    
2402    (Might need some stress-testing.  Simple examples worked fine.)
2403    
2404    ----------------------------------------------------------------------
2405    Name: Matthias Blume
2406    Date: 2002/03/19 16:37:00 EST
2407    Tag: blume-20020319-witnesses
2408    Description:
2409    
2410    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2411    
2412    This implements a generic mechanism for changing constness qualifiers
2413    anywhere within big C types without resorting to outright "casts".
2414    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2415    modify the constness at the outermost level.)
2416    The implementation of "convert" is based on the idea of "witness"
2417    values -- values that are not used by the operation but whose types
2418    "testify" to their applicability.  On the implementation side, "convert"
2419    is simply a projection (returning its second curried argument).  With
2420    cross-module inlining, it should not result in any machine code being
2421    generated.
2422    
2423    ----------------------------------------------------------------------
2424    Name: Matthias Blume
2425    Date: 2002/03/15 16:40:00 EST
2426    Tag: blume-20020315-basis
2427    Description:
2428    
2429    Provided (preliminary?) implementations for
2430    
2431      {String,Substring}.{concatWith,isSuffix,isSubstring}
2432    
2433    and
2434    
2435      Substring.full
2436    
2437    Those are in the Basis spec but they were missing in SML/NJ.
2438    
2439    ----------------------------------------------------------------------
2440    Name: Matthias Blume
2441    Date: 2002/03/14 21:30:00 EST
2442    Tag: blume-20020314-controls
2443    Description:
2444    
2445    Controls:
2446    ---------
2447    
2448    1. Factored out the recently-added Controls : CONTROLS stuff and put
2449       it into its own library $/controls-lib.cm.  The source tree for
2450       this is under src/smlnj-lib/Controls.
2451    
2452    2. Changed the names of types and functions in this interface, so they
2453       make a bit more "sense":
2454    
2455          module -> registry
2456          'a registry -> 'a group
2457    
2458    3. The interface now deals in ref cells only.  The getter/setter interface
2459       is (mostly) gone.
2460    
2461    4. Added a function that lets one register an already-existing ref cell.
2462    
2463    5. Made the corresponding modifications to the rest of the code so that
2464       everything compiles again.
2465    
2466    6. Changed the implementation of Controls.MLRISC back to something closer
2467       to the original.  In particular, this module (and therefore MLRISC)
2468       does not depend on Controls.  There now is some link-time code in
2469       int-sys.sml that registers the MLRISC controls with the Controls
2470       module.
2471    
2472    CM:
2473    ---
2474    
2475      * One can now specify the lambda-split aggressiveness in init.cmi.
2476    
2477    ----------------------------------------------------------------------
2478    Name: Allen Leung
2479    Date: 2002/03/13 17:30:00 EST
2480    Tag: leunga-20020313-x86-fp-unary
2481    Description:
2482    
2483    Bug fix for:
2484    
2485    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2486    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2487    > - fun f(x,(y,z)) = Real.~ y;
2488    > [autoloading]
2489    > [autoloading done]
2490    >       fchsl   (%eax), 184(%esp)
2491    > Error: MLRisc bug: X86MCEmitter.emitInstr
2492    >
2493    > uncaught exception Error
2494    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2495    
2496    The problem was that the code generator did not generate any fp registers
2497    in this case, and the ra didn't know that it needed to run the X86FP phase to
2498    translate the pseudo fp instruction.   This only happened with unary fp
2499    operators in certain situations.
2500    
2501    ----------------------------------------------------------------------
2502    Name: Matthias Blume
2503    Date: 2002/03/13 14:00:00 EST
2504    Tag: blume-20020313-overload-etc
2505    Description:
2506    
2507    1. Added _overload as a synonym for overload for backward compatibility.
2508       (Control.overloadKW must be true for either version to be accepted.)
2509    
2510    2. Fixed bug in install script that caused more things to be installed
2511       than what was requested in config/targets.
2512    
2513    3. Made CM aware of the (_)overload construct so that autoloading
2514       works.
2515    
2516    ----------------------------------------------------------------------
2517    Name: Matthias Blume
2518    Date: 2002/03/12 22:03:00 EST
2519    Tag: blume-20020312-url
2520    Description:
2521    
2522    Forgot to update BOOT and srcarchiveurl.
2523    
2524    ----------------------------------------------------------------------
2525    Name: Matthias Blume
2526    Date: 2002/03/12 17:30:00 EST
2527    Tag: blume-20020312-version110392
2528    Description:
2529    
2530    Yet another version number bump (because of small changes to the
2531    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2532    
2533    Changes:
2534    
2535      The new pid generation scheme described a few weeks ago was overly
2536      complicated.  I implemented a new mechanism that is simpler and
2537      provides a bit more "stability":  Once CM has seen a compilation
2538      unit, it keeps its identity constant (as long as you do not delete
2539      those crucial CM/GUID/* files).  This means that when you change
2540      an interface, compile, then go back to the old interface, and
2541      compile again, you arrive at the original pid.
2542    
2543      There now also is a mechanism that instructs CM to use the plain
2544      environment hash as a module's pid (effectively making its GUID
2545      the empty string).  For this, "noguid" must be specified as an
2546      option to the .sml file in question within its .cm file.
2547      This is most useful for code that is being generated by tools such
2548      as ml-nlffigen (because during development programmers tend to
2549      erase the tool's entire output directory tree including CM's cached
2550      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2551      revert to the old, broken behavior of SML/NJ, but in specific cases
2552      where there is no danger of interface confusion, its use is ok
2553      (I think).
2554    
2555      ml-nlffigen by default generates "noguid" annotations.  They can be
2556      turned off by specifying -guid in its command line.
2557    
2558    ----------------------------------------------------------------------
2559    Name: Lal George
2560    Date: 2002/03/12 12 14:42:36 EST
2561    Tag: george-20020312-frequency-computation
2562    Description:
2563    
2564    Integrated jump chaining and static block frequency into the
2565    compiler. More details and numbers later.
2566    
2567    ----------------------------------------------------------------------
2568    Name: Lal George
2569    Date: 2002/03/11 11 22:38:53 EST
2570    Tag: george-20020311-jump-chain-elim
2571    Description:
2572    
2573    Tested the jump chain elimination on all architectures (except the
2574    hppa).  This is on by default right now and is profitable for the
2575    alpha and x86, however, it may not be profitable for the sparc and ppc
2576    when compiling the compiler.
2577    
2578    The gc test will typically jump to a label at the end of the cluster,
2579    where there is another jump to an external cluster containing the actual
2580    code to invoke gc. This is to allow factoring of common gc invocation
2581    sequences. That is to say, we generate:
2582    
2583            f:
2584               testgc
2585               ja   L1      % jump if above to L1
2586    
2587            L1:
2588               jmp L2
2589    
2590    
2591    After jump chain elimination the 'ja L1' instructions is converted to
2592    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2593    up being implemented in their long form (if L2 is far away) using:
2594    
2595            jbe     L3      % jump if below or equal to L3
2596            jmp     L2
2597         L3:
2598            ...
2599    
2600    
2601    For large compilation units L2  may be far away.
2602    
2603    
2604    ----------------------------------------------------------------------
2605    Name: Matthias Blume
2606    Date: 2002/03/11 13:30:00 EST
2607    Tag: blume-20020311-mltreeeval
2608    Description:
2609    
2610    A functor parameter was missing.
2611    
2612    ----------------------------------------------------------------------
2613    Name: Allen Leung
2614    Date: 2002/03/11 10:30:00 EST
2615    Tag: leunga-20020311-runtime-string0
2616    Description:
2617    
2618       The representation of the empty string now points to a
2619    legal null terminated C string instead of unit.  It is now possible
2620    to convert an ML string into C string with InlineT.CharVector.getData.
2621    This compiles into one single machine instruction.
2622    
2623    ----------------------------------------------------------------------
2624    Name: Allen Leung
2625    Date: 2002/03/10 23:55:00 EST
2626    Tag: leunga-20020310-x86-call
2627    Description:
2628    
2629       Added machine generation for CALL instruction (relative displacement mode)
2630    
2631    ----------------------------------------------------------------------
2632    Name: Matthias Blume
2633    Date: 2002/03/08 16:05:00
2634    Tag: blume-20020308-entrypoints
2635    Description:
2636    
2637    Version number bumped to 110.39.1.  NEW BOOTFILES!
2638    
2639    Entrypoints: non-zero offset into a code object where execution should begin.
2640    
2641    - Added the notion of an entrypoint to CodeObj.
2642    - Added reading/writing of entrypoint info to Binfile.
2643    - Made runtime system bootloader aware of entrypoints.
2644    - Use the address of the label of the first function given to mlriscGen
2645      as the entrypoint.  This address is currently always 0, but it will
2646      not be 0 once we turn on block placement.
2647    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2648      with entry points) from mlriscGen.
2649    
2650    ----------------------------------------------------------------------
2651    Name: Allen Leung
2652    Date: 2002/03/07 20:45:00 EST
2653    Tag: leunga-20020307-x86-cmov
2654    Description:
2655    
2656       Bug fixes for CMOVcc on x86.
2657    
2658       1. Added machine code generation for CMOVcc
2659       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2660       3. CMOVcc cannot have an immediate operand as argument.
2661    
2662    ----------------------------------------------------------------------
2663    Name: Matthias Blume
2664    Date: 2002/03/07 16:15:00 EST
2665    Tag: blume-20020307-controls
2666    Description:
2667    
2668    This is a very large but mostly boring patch which makes (almost)
2669    every tuneable compiler knob (i.e., pretty much everything under
2670    Control.* plus a few other things) configurable via both the command
2671    line and environment variables in the style CM did its configuration
2672    until now.
2673    
2674    Try starting sml with '-h' (or, if you are brave, '-H')
2675    
2676    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2677    implements the underlying generic mechanism.
2678    
2679    The interface to some of the existing such facilities has changed somewhat.
2680    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2681    (The getFoo interface is still there for backward-compatibility, but its
2682    use is deprecated.)
2683    
2684    The ml-build script passes -Cxxx=yyy command-line arguments through so
2685    that one can now twiddle the compiler settings when using this "batch"
2686    compiler.
2687    
2688    TODO items:
2689    
2690    We should go through and throw out all controls that are no longer
2691    connected to anything.  Moreover, we should go through and provide
2692    meaningful (and correct!) documentation strings for those controls
2693    that still are connected.
2694    
2695    Currently, multiple calls to Controls.new are accepted (only the first
2696    has any effect).  Eventually we should make sure that every control
2697    is being made (via Controls.new) exactly once.  Future access can then
2698    be done using Controls.acc.
2699    
2700    Finally, it would probably be a good idea to use the getter-setter
2701    interface to controls rather than ref cells.  For the time being, both
2702    styles are provided by the Controls module, but getter-setter pairs are
2703    better if thread-safety is of any concern because they can be wrapped.
2704    
2705    *****************************************
2706    
2707    One bug fix: The function blockPlacement in three of the MLRISC
2708    backpatch files used to be hard-wired to one of two possibilities at
2709    link time (according to the value of the placementFlag).  But (I
2710    think) it should rather sense the flag every time.
2711    
2712    *****************************************
2713    
2714    Other assorted changes (by other people who did not supply a HISTORY entry):
2715    
2716    1. the cross-module inliner now works much better (Monnier)
2717    2. representation of weights, frequencies, and probabilities in MLRISC
2718       changed in preparation of using those for weighted block placement
2719       (Reppy, George)
2720    
2721    ----------------------------------------------------------------------
2722    Name: Lal George
2723    Date: 2002/03/07 14:44:24 EST 2002
2724    Tag: george-20020307-weighted-block-placement
2725    
2726    Tested the weighted block placement optimization on all architectures
2727    (except the hppa) using AMPL to generate the block and edge frequencies.
2728    Changes were required in the machine properties to correctly
2729    categorize trap instructions. There is an MLRISC flag
2730    "weighted-block-placement" that can be used to enable weighted block
2731    placement, but this will be ineffective without block/edge
2732    frequencies (coming soon).
2733    
2734    
2735    ----------------------------------------------------------------------
2736    Name: Lal George
2737    Date: 2002/03/05 17:24:48 EST
2738    Tag: george-20020305-linkage-cluster
2739    
2740    In order to support the block placement optimization, a new cluster
2741    is generated as the very first cluster (called the linkage cluster).
2742    It contains a single jump to the 'real' entry point for the compilation
2743    unit. Block placement has no effect on the linkage cluster itself, but
2744    all the other clusters  have full freedom in the manner in which they
2745    reorder blocks or functions.
2746    
2747    On the x86 the typical linkage code that is generated is:
2748       ----------------------
2749            .align 2
2750       L0:
2751            addl    $L1-L0, 72(%esp)
2752            jmp     L1
2753    
2754    
2755            .align  2
2756       L1:
2757       ----------------------
2758    
2759    72(%esp) is the memory location for the stdlink register. This
2760    must contain the address of the CPS function being called. In the
2761    above example, it contains the address of  L0; before
2762    calling L1 (the real entry point for the compilation unit), it
2763    must contain the address for L1, and hence
2764    
2765            addl $L1-L0, 72(%esp)
2766    
2767    I have tested this on all architectures except the hppa.The increase
2768    in code size is of course negligible
2769    
2770    ----------------------------------------------------------------------
2771    Name: Allen Leung
2772    Date: 2002/03/03 13:20:00 EST
2773    Tag: leunga-20020303-mlrisc-tools
2774    
2775      Added #[ ... ] expressions to mlrisc tools
2776    
2777    ----------------------------------------------------------------------
2778    Name: Matthias Blume
2779    Date: 2002/02/27 12:29:00 EST
2780    Tag: blume-20020227-cdebug
2781    Description:
2782    
2783    - made types in structure C and C_Debug to be equal
2784    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2785    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2786    
2787    ----------------------------------------------------------------------
2788    Name: Matthias Blume
2789    Date: 2002/02/26 12:00:00 EST
2790    Tag: blume-20020226-ffi
2791    Description:
2792    
2793    1. Fixed a minor bug in CM's "noweb" tool:
2794       If numbering is turned off, then truly don't number (i.e., do not
2795       supply the -L option to noweb).  The previous behavior was to supply
2796       -L'' -- which caused noweb to use the "default" line numbering scheme.
2797       Thanks to Chris Richards for pointing this out (and supplying the fix).
2798    
2799    2. Once again, I reworked some aspects of the FFI:
2800    
2801       A. The incomplete/complete type business:
2802    
2803       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2804         gone!
2805       - ML types representing an incomplete type are now *equal* to
2806         ML types representing their corresponding complete types (just like
2807         in C).  This is still safe because ml-nlffigen will not generate
2808         RTTI for incomplete types, nor will it generate functions that
2809         require access to such RTTI.   But when ML code generated from both
2810         incomplete and complete versions of the C type meet, the ML types
2811         are trivially interoperable.
2812    
2813         NOTE:  These changes restore the full generality of the translation
2814         (which was previously lost when I eliminated functorization)!
2815    
2816       B. Enum types:
2817    
2818       - Structure C now has a type constructor "enum" that is similar to
2819         how the "su" constructor works.  However, "enum" is not a phantom
2820         type because each "T enum" has values (and is isomorphic to
2821         MLRep.Signed.int).
2822       - There are generic access operations for enum objects (using
2823         MLRep.Signed.int).
2824       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2825         * The structure contains the definition of type "mlrep" (the ML-side
2826         representation type of the enum).  Normally, mlrep is the same
2827         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2828         then mlrep will be defined as a datatype -- thus facilitating
2829         pattern matching on mlrep values.
2830         ("-ec" will be suppressed if there are duplicate values in an
2831          enumeration.)
2832         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2833         will be generated for each C enum constant xxx.
2834         * Conversion functions m2i and i2m convert between mlrep and
2835         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2836         * Coversion functions c and ml convert between mlrep and "tag enum".
2837         * Access functions (get/set) fetch and store mlrep values.
2838       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2839         enumerations are merged into one single enumeration represented by
2840         structure E_'.
2841    
2842    ----------------------------------------------------------------------
2843    Name: Allen Leung
2844    Date: 2002/02/25 04:45:00 EST
2845    Tag: leunga-20020225-cps-spill
2846    
2847    This is a new implementation of the CPS spill phase.
2848    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2849    In case of problems, replace it with the old file spill.sml
2850    
2851    The current compiler runs into some serious performance problems when
2852    constructing a large record.  This can happen when we try to compile a
2853    structure with many items.  Even a very simple structure like the following
2854    makes the compiler slow down.
2855    
2856        structure Foo = struct
2857           val x_1 = 0w1 : Word32.int
2858           val x_2 = 0w2 : Word32.int
2859           val x_3 = 0w3 : Word32.int
2860           ...
2861           val x_N = 0wN : Word32.int
2862        end
2863    
2864    The following table shows the compile time, from N=1000 to N=4000,
2865    with the old compiler:
2866    
2867    N
2868    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2869           MLRISC ra                               0.06u  0.00s  0.05g
2870              (spills = 0 reloads = 0)
2871           TOTAL                                   0.63u  0.07s  0.21g
2872    
2873    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2874           MLRISC ra                               5.68u  0.59s  3.93g
2875              (spills = 0 reloads = 0)
2876           TOTAL                                   14.71u  0.99s  4.81g
2877    
2878    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2879           MLRISC ra                               5.54u  0.65s  3.91g
2880              (spills = 543 reloads = 1082)
2881           TOTAL                                   65.40u  3.13s  6.00g
2882    
2883    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2884           MLRISC ra                               0.80u  0.10s  0.55g
2885              (spills = 42 reloads = 84)
2886           TOTAL                                   129.42u  5.10s  4.13g
2887    
2888    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2889           MLRISC ra                               2.69u  0.27s  1.38g
2890              (spills = 62 reloads = 124)
2891           TOTAL                                   682.48u  19.61s  13.99g
2892    
2893    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2894           MLRISC ra                               4.96u  0.27s  2.72g
2895              (spills = 85 reloads = 170)
2896           TOTAL                                   2375.26u  57.21s  48.00g
2897    
2898    As you can see the old cps spill module suffers from some serious
2899    performance problem.  But since I cannot decipher the old code fully,
2900    instead of patching the problems up, I'm reimplementing it
2901    with a different algorithm.  The new code is more modular,
2902    smaller when compiled, and substantially faster
2903    (O(n log n) time and O(n) space).  Timing of the new spill module:
2904    
2905    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2906          MLRISC ra                               0.25u  0.02s  0.15g
2907             (spills=1 reloads=3)
2908          TOTAL                                   7.74u  0.34s  1.62g
2909    
2910    Implementation details:
2911    
2912    As far as I can tell, the purpose of the CPS spill module is to make sure the
2913    number of live variables at any program point (the bandwidth)
2914    does not exceed a certain limit, which is determined by the
2915    size of the spill area.
2916    
2917    When the bandwidth is too large, we decrease the register pressure by
2918    packing live variables into spill records.  How we achieve this is
2919    completely different than what we did in the old code.
2920    
2921    First, there is something about the MLRiscGen code generator
2922    that we should be aware of:
2923    
2924    o MLRiscGen performs code motion!
2925    
2926       In particular, it will move floating point computations and
2927       address computations involving only the heap pointer to
2928       their use sites (if there is only a single use).
2929       What this means is that if we have a CPS record construction
2930       statement
2931    
2932           RECORD(k,vl,w,e)
2933    
2934       we should never count the new record address w as live if w
2935       has only one use (which is often the case).
2936    
2937       We should do something similar to floating point, but the transformation
2938       there is much more complex, so I won't deal with that.
2939    
2940    Secondly, there are now two new cps primops at our disposal:
2941    
2942     1. rawrecord of record_kind option
2943        This pure operator allocates some uninitialized storage from the heap.
2944        There are two forms:
2945    
2946         rawrecord NONE [INT n]  allocates a tagless record of length n
2947         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2948                                     and initializes the tag.
2949    
2950     2. rawupdate of cty
2951          rawupdate cty (v,i,x)
2952          Assigns to x to the ith component of record v.
2953          The storelist is not updated.
2954    
2955    We use these new primops for both spilling and increment record construction.
2956    
2957     1. Spilling.
2958    
2959        This is implemented with a linear scan algorithm (but generalized
2960        to trees).  The algorithm will create a single spill record at the
2961        beginning of the cps function and use rawupdate to spill to it,
2962        and SELECT or SELp to reload from it.  So both spills and reloads
2963        are fine-grain operations.  In contrast, in the old algorithm
2964        "spills" have to be bundled together in records.
2965    
2966        Ideally, we should sink the spill record construction to where
2967        it is needed.  We can even split the spill record into multiple ones
2968        at the places where they are needed.  But CPS is not a good
2969        representation for global code motion, so I'll keep it simple and
2970        am not attempting this.
2971    
2972     2. Incremental record construction (aka record splitting).
2973    
2974        Long records with many component values which are simulatenously live
2975        (recall that single use record addresses are not considered to
2976         be live) are constructed with rawrecord and rawupdate.
2977        We allocate space on the heap with rawrecord first, then gradually
2978        fill it in with rawupdate.  This is the technique suggested to me
2979        by Matthias.
2980    
2981        Some restrictions on when this is applicable:
2982        1. It is not a VECTOR record.  The code generator currently does not handle
2983           this case. VECTOR record uses double indirection like arrays.
2984        2. All the record component values are defined in the same "basic block"
2985           as the record constructor.  This is to prevent speculative
2986           record construction.
2987    
2988    ----------------------------------------------------------------------
2989    Name: Allen Leung
2990    Date: 2002/02/22 01:02:00 EST
2991    Tag: leunga-20020222-mlrisc-tools
2992    
2993    Minor bug fixes in the parser and rewriter
2994    
2995    ----------------------------------------------------------------------
2996    Name: Allen Leung
2997    Date: 2002/02/21 20:20:00 EST
2998    Tag: leunga-20020221-peephole
2999    
3000    Regenerated the peephole files.  Some contained typos in the specification
3001    and some didn't compile because of pretty printing bugs in the old version
3002    of 'nowhere'.
3003    
3004    ----------------------------------------------------------------------
3005    Name: Allen Leung
3006    Date: 2002/02/19 20:20:00 EST
3007    Tag: leunga-20020219-mlrisc-tools
3008    Description:
3009    
3010       Minor bug fixes to the mlrisc-tools library:
3011    
3012       1.  Fixed up parsing colon suffixed keywords
3013       2.  Added the ability to shut the error messages up
3014       3.  Reimplemented the pretty printer and fixed up/improved
3015           the pretty printing of handle and -> types.
3016       4.  Fixed up generation of literal symbols in the nowhere tool.
3017       5.  Added some SML keywords to to sml.sty
3018    
3019    ----------------------------------------------------------------------
3020    Name: Matthias Blume
3021    Date: 2002/02/19 16:20:00 EST
3022    Tag: blume-20020219-cmffi
3023    Description:
3024    
3025    A wild mix of changes, some minor, some major:
3026    
3027    * All C FFI-related libraries are now anchored under $c:
3028        $/c.cm      --> $c/c.cm
3029        $/c-int.cm  --> $c/internals/c-int.cm
3030        $/memory.cm --> $c/memory/memory.cm
3031    
3032    * "make" tool (in CM) now treats its argument pathname slightly
3033      differently:
3034        1. If the native expansion is an absolute name, then before invoking
3035           the "make" command on it, CM will apply OS.Path.mkRelative
3036           (with relativeTo = OS.FileSys.getDir()) to it.
3037        2. The argument will be passed through to subsequent phases of CM
3038           processing without "going native".  In particular, if the argument
3039           was an anchored path, then "make" will not lose track of that anchor.
3040    
3041    * Compiler backends now "know" their respective C calling conventions
3042      instead of having to be told about it by ml-nlffigen.  This relieves
3043      ml-nlffigen from one of its burdens.
3044    
3045    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3046    
3047    * Export C_DEBUG and C_Debug from $c/c.cm.
3048    
3049    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3050      subtyping relationship between incomplete pointers and their complete
3051      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3052      with the convention of instantiating 'o with ('t, 'c) obj whenever
3053      the pointer target type is complete.  In the incomplete case, 'o
3054      will be instantiated with some "'c iobj" -- a type obtained by
3055      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3056    
3057      Operations that work on both incomplete and complete pointer types are
3058      typed as taking an 'o ptr while operations that require the target to
3059      be known are typed as taking some ('t, 'c) obj ptr.
3060    
3061      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3062      where void is an eqtype without any values.  This makes it possible
3063      to work on voidptr values using functions meant to operate on light
3064      incomplete pointers.
3065    
3066    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3067      been vastly simplified.
3068    
3069    ----------------------------------------------------------------------
3070    Name: Matthias Blume
3071    Date: 2002/02/19 10:48:00 EST
3072    Tag: blume-20020219-pqfix
3073    Description:
3074    
3075    Applied Chris Okasaki's bug fix for priority queues.
3076    
3077    ----------------------------------------------------------------------
3078    Name: Matthias Blume
3079    Date: 2002/02/15 17:05:00
3080    Tag: Release_110_39
3081    Description:
3082    
3083    Last-minute retagging is becoming a tradition... :-(
3084    
3085    This is the working release 110.39.
3086    
3087    ----------------------------------------------------------------------
3088    Name: Matthias Blume
3089    Date: 2002/02/15 16:00:00 EST
3090    Tag: Release_110_39-orig
3091    Description:
3092    
3093    Working release 110.39.  New bootfiles.
3094    
3095    (Update: There was a small bug in the installer so it wouldn't work
3096    with all shells.  So I retagged. -Matthias)
3097    
3098    ----------------------------------------------------------------------
3099    Name: Matthias Blume
3100    Date: 2002/02/15 14:17:00 EST
3101    Tag: blume-20020215-showbindings
3102    Description:
3103    
3104    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3105    the latter can be useful for exploring what bindings are available at
3106    the interactive prompt.  (The first function returns only the list
3107    of symbols that are really bound, the second prints those but also the
3108    ones that CM's autoloading mechanism knows about.)
3109    
3110    ----------------------------------------------------------------------
3111    Name: Matthias Blume
3112    Date: 2002/02/15 12:08:00 EST
3113    Tag: blume-20020215-iptrs
3114    Description:
3115    
3116    Two improvements to ml-nlffigen:
3117    
3118      1. Write files only if they do not exist or if their current contents
3119         do not coincide with what's being written.  (That is, avoid messing
3120         with the time stamps unless absolutely necessary.)
3121    
3122      2. Implement a "repository" mechanism for generated files related
3123         to "incomplete pointer types".   See the README file for details.
3124    
3125    ----------------------------------------------------------------------
3126    Name: Matthias Blume
3127    Date: 2002/02/14 11:50:00 EST
3128    Tag: blume-20020214-quote
3129    Description:
3130    
3131    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3132    because of the new and improved tag generation scheme.  (Thanks to Allen
3133    Leung for pointing it out.)
3134    
3135    ----------------------------------------------------------------------
3136    Name: Lal George
3137    Date: 2002/02/14 09:55:27 EST 2002
3138    Tag: george-20020214-isabelle-bug
3139    Description:
3140    
3141    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3142    of Isabelle on the x86.
3143    
3144    From Allen:
3145    -----------
3146     I've found the problem:
3147    
3148         in ra-core.sml, I use the counter "blocked" to keep track of the
3149         true number of elements in the freeze queue.  When the counter goes
3150         to zero, I skip examining the queue.  But I've messed up the
3151         bookkeeping in combine():
3152    
3153             else ();
3154             case !ucol of
3155               PSEUDO => (if !cntv > 0 then
3156                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3157                                        ^^^^^^^^^^^^^^^^^^^^^^^
3158                      moveu := mergeMoveList(!movev, !moveu)
3159                     )
3160                  else ();
3161    
3162         combine() is called to coalesce two nodes u and v.
3163         I think I was thinking that if the move counts of u and v are both
3164         greater than zero then after they are coalesced then one node is
3165         removed from the freeze queue.  Apparently I was thinking that
3166         both u and v are of low degree, but that's clearly not necessarily true.
3167    
3168    
3169    02/12/2002:
3170        Here's the patch.  HOL now compiles.
3171    
3172        I don't know how this impact on performance (compile
3173        time or runtime).  This bug caused the RA (especially on the x86)
3174        to go thru the potential spill phase when there are still nodes on the
3175        freeze queue.
3176    
3177    
3178    
3179    
3180    ----------------------------------------------------------------------
3181    Name: Matthias Blume
3182    Date: 2002/02/13 22:40:00 EST
3183    Tag: blume-20020213-fptr-rtti
3184    Description:
3185    
3186    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3187    updates.
3188    
3189    ----------------------------------------------------------------------
3190    Name: Matthias Blume
3191    Date: 2002/02/13 16:41:00 EST
3192    Tag: blume-20020213-cmlpq
3193    Description:
3194    
3195    Added new priority queue export symbols (which have just been added to
3196    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3197    not compile and the installer would choke.)
3198    
3199    ----------------------------------------------------------------------
3200    Name: Matthias Blume
3201    Date: 2002/02/13 16:15:00 EST
3202    Tag: blume-20020213-various
3203    Description:
3204    
3205    1. More tweaks to ml-nlffigen:
3206    
3207       - better internal datastructures (resulting in slight speedup)
3208       - "-match" option requires exact match
3209       - "localized" gensym counters (untagged structs/unions nested within
3210         other structs/unions or within typedefs get a fresh counter; their
3211         tag will be prefixed by a concatenation of their parents' tags)
3212       - bug fixes (related to calculation of transitive closure of types
3213         to be included in the output)
3214    
3215    2. Minor Basis updates:
3216    
3217       - added implementations for List.collate and Option.app
3218    
3219    ----------------------------------------------------------------------
3220    Name: Matthias Blume
3221    Date: 2002/02/11 15:55:00 EST
3222    Tag: blume-20020211-gensym
3223    Description:
3224    
3225    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3226    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3227    names (ML structure names that correspond to unnamed C structs, unions,
3228    and enums), so that separate runs of ml-nlffigen do not clash.
3229    
3230    ----------------------------------------------------------------------
3231    Name: Matthias Blume
3232    Date: 2002/02/11 12:05:00 EST
3233    Tag: blume-20020211-gensml
3234    Description:
3235    
3236    A quick fix for a problem with GenSML (in the pgraph-util library):
3237    Make generation of toplevel "local" optional.  (Strictly speaking,
3238    signature definitions within "local" are not legal SML.)
3239    
3240    Other than that: updates to INSTALL and cm/TODO.
3241    
3242    ----------------------------------------------------------------------
3243    Name: Matthias Blume
3244    Date: 2002/02/08 15:00:00 EST
3245    Tag: blume-20020208-uniquepid
3246    Description:
3247    
3248    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3249    
3250    1. The installer (config/install.sh) has gotten smarter:
3251    
3252         - Configuration options are a bit easier to specify now
3253           (in config/targets).
3254         - Bug in recognizing .tar.bz2 files fixed.
3255         - Installer automatically resolves dependencies between
3256           configuration options (e.g., if you ask for eXene, you will
3257           also get cml -- regardless whether you asked for it or not).
3258         - Installer can run in "quieter mode" by setting the environment
3259           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3260           "completely silent", though.
3261         - Build HashCons library as part of smlnj-lib.
3262    
3263    2. A new scheme for assigning persistent identifiers to compilation
3264       units (and, by extension, to types etc.) has been put into place.
3265       This fixes a long-standing bug where types and even dynamic values
3266       can get internally confused, thereby compromising type safety
3267       (abstraction) and dynamic correctness.  See
3268    
3269         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3270    
3271       for an example of how things could go wrong until now.
3272    
3273       The downside of the new scheme is that pids are not quite as
3274       persistent as they used to be: CM will generate a fresh pid
3275       for every compilation unit that it thinks it sees for the first
3276       time.  That means that if you compile starting from a clean, fresh
3277       source tree at two different times, you end up with different
3278       binaries.
3279    
3280       Cutoff recompilation, however, has not been compromised because
3281       CM keeps pid information in special caches between runs.
3282    
3283    ----------------------------------------------------------------------
3284    Name: Lal George
3285    Date: 2002/02/07 15:34:13 EST 2002
3286    Tag: <none>
3287    Description:
3288    
3289    Compilers that generate assembly code may produce  global labels
3290    whose value is resolved at link time. The various peephole optimization
3291    modules did not take this in account.
3292    
3293    TODO. The Labels.addrOf function should really return an option
3294    type so that clients are forced to deal with this issue, rather
3295    than an exception being raised.
3296    
3297    ----------------------------------------------------------------------
3298    Name: Lal George
3299    Date: 2002/02/06 13:55:02 EST
3300    Tag: george-20020206-ra-breakup
3301    Description:
3302    
3303    1. A bug fix from Allen.
3304    
3305        A typo causes extra fstp %st(0)'s to be generated at compensation
3306        edges, which might cause stack underflow traps at runtime.  This
3307        occurs in fft where there are extraneous fstps right before the 'into'
3308        trap instruction (in this case they are harmless since none of the
3309        integers overflow.)
3310    
3311    2. Pulled out various utility modules that were embedded in the modules
3312       of the register allocator. I need these modules for other purposes, but
3313       they are not complete enough to put into a library (just yet).
3314    ----------------------------------------------------------------------
3315    Name: Matthias Blume
3316    Date: 2002/01/31 16:05:00 EST
3317    Tag: blume-20020131-sparc-ccalls
3318    Description:
3319    
3320    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3321       of extra stack space by mistake.  Fixed.
3322    
3323    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3324    
3325    ----------------------------------------------------------------------
3326    Name: Allen Leung
3327    Date: 2002/01/30
3328    Tag: leunga-20020130-nowhere-bug-fix
3329    Description:
3330    
3331       MLRISC bug fixes:
3332       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3333       2. MachineInt.fromString was negating its value.
3334    
3335    ----------------------------------------------------------------------
3336    Name: Matthias Blume
3337    Date: 2002/01/29
3338    Tag: blume-20020129-INSTALL
3339    Description:
3340    
3341    - Added somewhat detailed installation instructions (file INSTALL).
3342    - Fixed curl-detection bug in config/install.sh.
3343    - It is now possible to select the URL getter using the URLGETTER
3344      environment variable:
3345    
3346          not set / "unknown"      --> automatic detection (script tries wget,
3347                                       curl, and lynx)
3348          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3349                                       how to properly invoke them)
3350          other                    --> use $URLGETTER directly, it must take
3351                                       precisely two command-line arguments
3352                                       (source URL and destination file name)
3353    
3354    ----------------------------------------------------------------------
3355    Name: Matthias Blume
3356    Date: 2002/01/28
3357    Tag: blume-20020128-sparc-ccalls
3358    Description:
3359    
3360    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3361    - Make use of the allocParam argument in sparc-c-calls.
3362    
3363    ----------------------------------------------------------------------
3364    Name: Matthias Blume
3365    Date: 2002/01/28
3366    Tag: blume-20020128-allocParam
3367    Description:
3368    
3369    John Reppy:  Changes c-calls API to accept client-callback for
3370    allocating extra stack space.
3371    me: Corresponding changes to mlriscGen (using a dummy argument that
3372        does not change the current behavior).
3373    
3374    ----------------------------------------------------------------------
3375    Name: Matthias Blume
3376    Date: 2002/01/28 12:00:00
3377    Tag: Release_110_38
3378    Description:
3379    
3380    This time for real!!!
3381    
3382    ----------------------------------------------------------------------
3383    Name: Matthias Blume
3384    Date: 2002/01/28 10:56:00 EST
3385    Tag: blume-20020128-retraction
3386    Description:
3387    
3388    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3389       with blume-Release_110_38-retracted.)
3390    
3391    1. Fixed a problem with incorrect rounding modes in real64.sml.
3392       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3393    
3394    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3395       and enums fixed.  The naming of corresponding ML identifiers should
3396       now be consistent again.
3397    
3398    ----------------------------------------------------------------------
3399    Name: Allen Leung
3400    Date: 2002/01/27
3401    Tag: leunga-20020127-nowhere
3402    Description:
3403    
3404       Added a target called nowhere in the configuration scripts.
3405       Enabling this will build the MLRISC 'nowhere' tool (for translating
3406       programs with where-clauses into legal SML code) during installation.
3407    
3408    ----------------------------------------------------------------------
3409    Name: Matthias Blume
3410    Date: 2002/01/25 21:27:00 EST
3411    Tag: blume-Release_110_38-retracted
3412    Description:
3413    
3414    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3415    
3416    README will be added later.
3417    
3418    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3419    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3420    
3421    ----------------------------------------------------------------------
3422    Name: Matthias Blume
3423    Date: 2002/01/25
3424    Tag: blume-20020125-ffi
3425    Description:
3426    
3427    A large number of tweaks and improvements to ml-nlffi-lib and
3428    ml-nlffigen:
3429    
3430       - ML represenation types have been streamlined
3431       - getter and setter functions work with concrete values, not abstract
3432         ones where possible
3433       - ml-nlffigen command line more flexible (see README file there)
3434       - some bugs have been fixed (hopefully)
3435    
3436    ----------------------------------------------------------------------
3437    Name: Lal George
3438    Date: 2002/01/24
3439    Tag: george-20020124-risc-ra-interface
3440    Description:
3441    
3442       There is a dramatic simplification in the interface to the
3443       register allocator for RISC architectures as a result of making
3444       parallel copy instructions explicit.
3445    
3446    ----------------------------------------------------------------------
3447    Name: Matthias Blume
3448    Date: 2002/01/22
3449    Tag: blume-20020122-x86-ccalls
3450    Description:
3451    
3452    Bug fix for c-calls on x86 (having to do with how char- and
3453    short-arguments are being handled).
3454    
3455    ----------------------------------------------------------------------
3456    Name: Matthias Blume
3457    Date: 2002/01/21
3458    Tag: blume-20020121-ff
3459    Description:
3460    
3461    Another day of fiddling with the FFI...
3462    
3463    1. Bug fix/workaround:  CKIT does not complain about negative array
3464       dimensions, so ml-nlffigen has to guard itself against this possibility.
3465       (Otherwise a negative dimension would send it into an infinite loop.)
3466    
3467    2. Some of the abstract types (light objects, light pointers, most "base"
3468       types) in structure C are now eqtypes.
3469    
3470    3. Added constructors and test functions for NULL function pointers.
3471    
3472    ----------------------------------------------------------------------
3473    Name: Matthias Blume
3474    Date: 2002/01/18
3475    Tag: blume-20020118-ready-for-new-release
3476    Description:
3477    
3478    Made config/srcarchiveurl point to a new place.  (Will provide boot
3479    files shortly.)
3480    
3481    Maybe we christen this to be 110.38?
3482    
3483    ----------------------------------------------------------------------
3484    Name: Matthias Blume
3485    Date: 2002/01/18
3486    Tag: blume-20020118-more-ffifiddle
3487    Description:
3488    
3489    Today's FFI fiddling:
3490    
3491      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3492        C.Set.* that use concrete (MLRep.*) arguments and results instead
3493        of abstract ones.
3494    
3495      - Provided word-style bit operations etc. for "int" representation
3496        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3497        and Long.
3498    
3499    ----------------------------------------------------------------------
3500    Name: Matthias Blume
3501    Date: 2002/01/18
3502    Tag: blume-20020118-use-x86-fp
3503    Description:
3504    
3505    Now that x86-fast-fp seems to be working, I turned it back on again
3506    by default.  (Seems to work fine now, even with the FFI.)
3507    
3508    Other than that, I added some documentation about the FFI to
3509    src/ml-nlffigen/README and updated the FFI test examples in
3510    src/ml-nlffi-lib/Tests/*.
3511    
3512    ----------------------------------------------------------------------
3513    Name: Allen Leung
3514    Date: 2002/01/17
3515    Tag: leunga-20020117-x86-fast-fp-call
3516    Description:
3517    
3518       1. Fixed a problem with handling return fp values when x86's fast fp
3519          mode is turned on.
3520    
3521       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3522          of %f32.
3523    
3524       3. Added a constructor INT32lit to the ast of MLRISC tools.
3525    
3526    ----------------------------------------------------------------------
3527    Name: Matthias Blume
3528    Date: 2002/01/16
3529    Tag: blume-20020116-ffifiddle
3530    Description:
3531    
3532    More fiddling with the FFI interface:
3533    
3534     - Make constness 'c instead of rw wherever possible.  This eliminates
3535       the need for certain explicit coercions.  (However, due to ML's
3536       value polymorphism, there will still be many cases where explicit
3537       coercions are necessary.  Phantom types are not the whole answer
3538       to modeling a subtyping relationship in ML.)
3539    
3540     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3541    
3542     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3543       clumsy workaround for varargs problem.)
3544    
3545    ----------------------------------------------------------------------
3546    Name: Lal George
3547    Date: 2002/01/15
3548    Tag: <none>
3549    Description:
3550    
3551    1. Since COPY instructions are no longer native to the architecture,
3552       a generic functor can be used to implement the expandCopies function.
3553    
3554    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3555       TEXT segment.
3556    
3557    ----------------------------------------------------------------------
3558    Name: Matthias Blume
3559    Date: 2002/01/15
3560    Tag: blume-20020115-ffiupdates
3561    Description:
3562    
3563    1. Fix for bug resulting in single-precision float values being returned
3564       incorrectly from FFI calls.
3565    
3566    2. Small modifications to C FFI API:
3567    
3568        - memory-allocation routines return straight objects (no options)
3569          and raise an exception in out-of-memory situations
3570        - unsafe extensions to cast between function pointers and pointers
3571          from/to ints
3572        - added structure C_Debug as an alternative to structure C where
3573          pointer-dereferencing (|*| and |*!) always check for null-pointers
3574        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3575          but also takes a (possibly empty) list of existing library handles
3576          that the current library depends on
3577    
3578    ----------------------------------------------------------------------
3579    Name: Matthias Blume
3580    Date: 2002/01/10
3581    Tag: blume-20020110-newffigen
3582    Description:
3583    
3584    1. Updates to portable graph code.
3585    
3586    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3587       more scalable now so that even huge interfaces such as the one
3588       for GTK compile in finite time and space. :-)
3589       See src/ml-nlffigen/README for details on what's new.
3590    
3591    ----------------------------------------------------------------------
3592    Name: Lal George
3593    Date: 2001/01/09 14:31:35 EST 2002
3594    Tag: george-20011206-rm-native-copy
3595    Description:
3596    
3597            Removed the native COPY and FCOPY instructions
3598            from all the architectures and replaced it with the
3599            explicit COPY instruction from the previous commit.
3600    
3601            It is now possible to simplify many of the optimizations
3602            modules that manipulate copies. This has not been
3603            done in this change.
3604    
3605    ----------------------------------------------------------------------
3606    Name: Lal George
3607    Date: 2001/12/06 16:50:13 EST 2001
3608    Tag: george-20011206-mlrisc-instruction
3609    Description:
3610    
3611    Changed the representation of instructions from being fully abstract
3612    to being partially concrete. That is to say:
3613    
3614      from
3615            type instruction
3616    
3617      to
3618            type instr                              (* machine instruction *)
3619    
3620            datatype instruction =
3621                LIVE of {regs: C.cellset, spilled: C.cellset}
3622              | KILL of {regs: C.cellset, spilled: C.cellset}
3623              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3624              | ANNOTATION of {i: instruction, a: Annotations.annotation}
3625              | INSTR of instr
3626    
3627    This makes the handling of certain special instructions that appear on
3628    all architectures easier and uniform.
3629    
3630    LIVE and KILL say that a list of registers are live or killed at the
3631    program point where they appear. No spill code is generated when an
3632    element of the 'regs' field is spilled, but the register is moved to
3633    the 'spilled' (which is present, more for debugging than anything else).
3634    
3635    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3636    We used to generate:
3637    
3638            DEFFREG f1
3639            f1 := f2 + f3
3640            trapb
3641    
3642    but now generate:
3643    
3644            f1 := f2 + f3
3645            trapb
3646            LIVE {regs=[f1,f2,f3], spilled=[]}
3647    
3648    Furthermore, the DEFFREG (hack) required that all floating point instruction
3649    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3650    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3651    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3652    this architecture).
3653    
3654    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3655    all the architectures. This will result in further simplification of the
3656    register allocator that must be aware of them for coalescing purposes, and
3657    will also simplify certain aspects of the machine description that provides
3658    callbacks related to parallel copies.
3659    
3660    ANNOTATION should be obvious, and now INSTR represents the honest to God
3661    machine instruction set!
3662    
3663    The <arch>/instructions/<arch>Instr.sml files define certain utility
3664    functions for making porting easier -- essentially converting upper case
3665    to lower case. All machine instructions (of type instr) are in upper case,
3666    and the lower case form generates an MLRISC instruction. For example on
3667    the alpha we have:
3668    
3669      datatype instr =
3670         LDA of {r:cell, b:cell, d:operand}
3671       | ...
3672    
3673      val lda : {r:cell, b:cell, d:operand} -> instruction
3674        ...
3675    
3676    where lda is just (INSTR o LDA), etc.
3677    
3678    ----------------------------------------------------------------------
3679    Name: Matthias Blume
3680    Date: 2001/11/22 21:40:00 EST
3681    Tag: Release_110_37
3682    Description:
3683    
3684    Release 110.37.  This time for real.
3685    
3686    ----------------------------------------------------------------------
3687    Name: Matthias Blume
3688    Date: 2001/11/21 16:35:00 EST
3689    Tag: blume-20011121-foot-in-mouth
3690    Description:
3691    
3692    Removed the "Release_110_37" tag because of a serious bug.
3693    This will be re-tagged once the bug is fixed.
3694    
3695    ----------------------------------------------------------------------
3696    Name: Matthias Blume
3697    Date: 2001/11/21 16:14:00 EST
3698    Tag: blume-20011121-forgottenfile
3699    Description:
3700    
3701    Forgot to add a file.  (Just a .tex-file -- part of
3702    the CM manual source.)
3703    
3704    ----------------------------------------------------------------------
3705    Name: Matthias Blume
3706    Date: 2001/11/21 16:10:00 EST
3707    Tag: blume-20011121-invalid_110_37
3708    Description:
3709    
3710    Note: I removed the original tag "Release_110_37" from this commit
3711          because we found a serious bug in all non-x86 backends.
3712          - Matthias
3713    
3714    1. Modifications to the SML/NJ code generator and to the runtime system
3715       so that code object name strings are directly inserted into code
3716       objects at code generation time.  The only business the runtime system
3717       has with this is now to read the name strings on occasions.
3718       (The encoding of the name string has also changed somewhat.)
3719    
3720    2. CM now implements a simple "set calculus" for specifying export lists.
3721       In particular, it is now possible to refer to the export lists of
3722       other libraries/groups/sources and form unions as well as differences.
3723       See the latest CM manual for details.
3724    
3725    3. An separate notion of "proxy" libraries has again be eliminated from
3726       CM's model.  (Proxy libraries are now simply a special case of using
3727       the export list calculus.)
3728    
3729    4. Some of the existing libraries now take advantage of the new set
3730       calculus.
3731       (Notice that not all libraries have been converted because some
3732       of the existing .cm-files are supposed to be backward compatible
3733       with 110.0.x.)
3734    
3735    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3736       "General.exnMessage"!  The former relies on a certain hook to be
3737       initialized, and that often does not happen in the stand-alone case.)
3738    
3739    ----------------------------------------------------------------------
3740    Name: Lal George
3741    Date: 2001/11/21  13:56:18 EST
3742    Tag: george-2001121-pseudo-ops
3743    Description:
3744    
3745      Implemented a complete redesign of MLRISC pseudo-ops. Now there
3746      ought to never be any question of incompatabilities with
3747      pseudo-op syntax expected by host assemblers.
3748    
3749      For now, only modules supporting GAS syntax are implemented
3750      but more should follow, such as MASM, and vendor assembler
3751      syntax, e.g. IBM as, Sun as, etc.
3752    
3753    ----------------------------------------------------------------------
3754    Name: Matthias Blume
3755    Date: 2001/11/14 11:52:00 EST
3756    Tag: blume-20011114-srcname
3757    Description:
3758    
3759    1. Routed the name of the current source file to mlriscgen where it
3760       should be directly emitted into the code object.  (This last part
3761       is yet to be done.)
3762    
3763    2. Some cleanup of the pgraph code to make it match the proposal that
3764       I put out the other day.  (The proposal notwithstanding, things are
3765       still in flux here.)
3766    
3767    ----------------------------------------------------------------------
3768    Name: Lal George
3769    Date: 2001/11/14 09:44:04 EST
3770    Tag:
3771    Description:
3772    
3773      Fix for a backpatching bug reported by Allen.
3774    
3775      Because the boundary between short and long span-dependent
3776      instructions is +/- 128, there are an astounding number of
3777      span-dependent instructions whose size is over estimated.
3778    
3779      Allen came up with the idea of letting the size of span
3780      dependent instructions be non-monotonic, for a maxIter
3781      number of times, after which the size must be monotonically
3782      increasing.
3783    
3784      This table shows the number of span-dependent instructions
3785      whose size was over-estimated as a function of maxIter, for the
3786      file Parse/parse/ml.grm.sml:
3787    
3788         maxIter            # of instructions:
3789            10                      687
3790            20                      438
3791            30                      198
3792            40                        0
3793    
3794      In compiling the compiler, there is no significant difference in
3795      compilation speed between maxIter=10 and maxIter=40. Actually,
3796      my measurements showed that maxIter=40 was a tad faster than
3797      maxIter=10! Also 96% of the  files in the compiler reach a fix
3798      point within 13 iterations, so fixing maxIter at 40, while high,
3799      is okay.
3800    
3801    ----------------------------------------------------------------------
3802    Name: Matthias Blume
3803    Date: 2001/10/31 15:25:00 EST
3804    Tag: blume-20011031-pgraph
3805    Description:
3806    
3807    CKIT:
3808    * Changed the "Function" constructor of type Ast.ctype to carry optional
3809      argument identifiers.
3810    * Changed the return type of TypeUtil.getFunction accordingly.
3811    * Type equality ignores the argument names.
3812    * TypeUtil.composite tries to preserve argument names but gives up quickly
3813      if there is a mismatch.
3814    
3815    installation script:
3816    * attempts to use "curl" if available (unless "wget" is available as well)
3817    
3818    CM:
3819    * has an experimental implementation of "portable graphs" which I will
3820      soon propose as an implementation-independent library format
3821    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
3822    
3823    NLFFI-LIB:
3824    * some cleanup (all cosmetic)
3825    
3826    NLFFIGEN:
3827    * temporarily disabled the mechanism that suppresses ML output for
3828      C definitions whose identifiers start with an underscore character
3829    * generate val bindings for enum constants
3830    * user can request that only one style (light or heavy) is being used;
3831      default is to use both (command-line arguments: -heavy and -light)
3832    * fixed bug in handling of function types involving incomplete pointers
3833    * generate ML entry points that take record arguments (i.e., using
3834      named arguments) for C functions that have a prototype with named
3835      arguments
3836      (see changes to CKIT)
3837    
3838    ----------------------------------------------------------------------
3839    Name: Allen Leung
3840    Date: 2001/10/27 20:34:00 EDT
3841    Tag: leunga-20011027-x86-fast-fp-call
3842    Description:
3843    
3844       Fixed the bug described in blume-20010920-slowfp.
3845    
3846       The fix involves
3847          1. generating FCOPYs in FSTP in ia32-svid
3848