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

SCM Repository

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

Diff of /sml/trunk/NOTES/HISTORY

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

revision 583, Thu Mar 23 21:52:30 2000 UTC revision 1703, Thu Nov 18 21:40:57 2004 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/11/18 15:40:00 CST
18    Tag: Release_110_51
19    Description:
20    
21    New working version (110.51).  NEW BOOTFILES!
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume (blume (at) tti - c (dot) org)
25    Date: 2004/11/18 15:35:00 CST
26    Tag: <none>
27    Description:
28    
29    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
30    Joemann).
31    
32    ----------------------------------------------------------------------
33    Name: Allen Leung (leunga (at) reservoir (dot) com)
34    Date: 2004/11/17 16:05:21 EST 2004
35    Tag: leunga-20041117-mlrisc-live-kill
36    Description:
37    
38       Added support for MLTree constructs LIVE and KILL
39    to all the architectures.
40    
41    ----------------------------------------------------------------------
42    Name: Matthias Blume (blume (at) tti - c (dot) org)
43    Date: 2004/11/13 00:20:00 CST
44    Tag: blume-20041113-versiontool
45    Description:
46    
47    - Stripped down the versiontool: It now only handles the version number.
48      The date string is generated at bootstrap time (during makeml).
49    
50    - In a previous commit, fixed a minor issue with how polyequal is being
51      translated.  In particular, the code now "looks through" abstractions.
52      This results in slightly fewer polyEqual warnings and hopefully slightly
53      more efficient code.  Important examples for where this matters are
54      the new int64 and word64 types.
55    
56    ----------------------------------------------------------------------
57    Name: Matthias Blume (blume (at) tti - c (dot) org)
58    Date: 2004/11/12 00:30:00 CST
59    Tag: blume-20041112-int64
60    Description:
61    
62    Structure Int64 fully hooked in.  (The implementation is not very
63    efficient, though.)
64    
65    ----------------------------------------------------------------------
66    Name: Matthias Blume (blume (at) tti - c (dot) org)
67    Date: 2004/11/11 17:30:00 CST
68    Tag: blume-20041111-more64
69    Description:
70    
71    All the pieces of Word64 are now there, with the exception of the
72    conversions from and to LargeWord. (Eventually these need to be identities,
73    but for the time being they don't even make sense because LargeWord is
74    32-bit wide.)
75    
76    Also started to add similar support for Int64, but major pieces of that
77    are still missing.
78    
79    ----------------------------------------------------------------------
80    Name: Matthias Blume (blume (at) tti - c (dot) org)
81    Date: 2004/11/11 00:15:00 CST
82    Tag: blume-20041111-word64
83    Description:
84    
85    Structure Word64 is now (almost) complete, word literals and patterns
86    seem to work.  There are a few odd pieces missing.  In particular,
87    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
88    at the moment.
89    
90    Making Word64 official would mean that LargeWord becomes Word64.  But
91    this requires extreme care because most word-word conversions have to
92    go through LargeWord, so making a mistake means loss of efficiency or
93    worse.  Eventually there will be a solution similar to (but actually
94    simpler than) what I did with IntInf.
95    
96    ----------------------------------------------------------------------
97    Name: Matthias Blume (blume (at) tti - c (dot) org)
98    Date: 2004/11/10 18:12:00 CST
99    Tag: blume-20041110-64bit
100    Description:
101    
102    More 64-bit hacking (but still not even half-way there yet).
103    Also, some assorted improvements to the handling of 8-bit words.
104    
105    ----------------------------------------------------------------------
106    Name: Matthias Blume (blume (at) tti - c (dot) org)
107    Date: 2004/11/09 17:50:00 CST
108    Tag: <none>
109    Description:
110    
111    Started adding some infrastructure for supporting 64-bit int- and
112    word-types.  (Still in its very early stages.)
113    
114    ----------------------------------------------------------------------
115    Name: Matthias Blume (blume (at) tti - c (dot) org)
116    Date: 2004/10/28 10:45:00 CDT
117    Tag: Release_110_50
118    Description:
119    
120    New working version (110.50).  NEW BOOTFILES!
121    
122                 =====================
123    
124    Also:
125    
126      - Changend config/srcarchiveurl from a file just
127        containing the URL string into a file containing
128        shell script code.  The code has access to the $VERSION variable.
129      - Made corresponding changes to config/install.sh and config/unpack.
130      - Default contents of config/srcarchiveurl uses $VERSION and
131        normally does not have to be edited to reflect a version change.
132    
133      (As a result, a version change can be done by just editing
134       config/version, the rest is now automatic.)
135    
136    ----------------------------------------------------------------------
137    Name: Matthias Blume (blume (at) tti - c (dot) org)
138    Date: 2004/10/27 17:50:00 CDT
139    Tag: blume-20041027-btrace-msg
140    Description:
141    
142    BackTrace.monitor now also reports the source of the exception that
143    triggered the trace.
144    
145    ----------------------------------------------------------------------
146    Name: Matthias Blume (blume (at) tti - c (dot) org)
147    Date: 2004/10/27 17:20:00 CDT
148    Tag: blume-20041027-x86-c-calls
149    Description:
150    
151    This is the HISTORY entry for two earlier commits, both concerning
152    the x86 c-calls code in MLRISC:
153    
154      - added a missing LOAD in the code that deals with struct arguments
155      - made sure the caller does not add the wrong number of bytes to the
156        stack pointer after a call of a function returning a struct
157        (the callee already pops the implicit argument which points to
158         the space reserved for the result)
159    
160    ----------------------------------------------------------------------
161    Name: Allen Leung (leunga (at) reservoir (dot) com)
162    Date: 2004/10/24 14:00:00 EST
163    Tag: leunga-20041024-x86-gas-fucomip
164    Description:
165    
166       John discovered a bug in the syntax of fucomip.
167       The opcodes FU?COMIP? have been changed to
168    
169           fu?comip? %st(i), %st
170    
171    ----------------------------------------------------------------------
172    Name: Matthias Blume (blume (at) tti - c (dot) org)
173    Date: 2004/10/20 15:06:00 CDT
174    Tag: blume-20041020-standalone-backtrace
175    Description:
176    
177    Added a mechanism for getting back-trace information from standalone
178    programs.  Here is how it works:
179    
180       1. The part of the program from which you want to get backtrace
181          information (usually the whole program) should be wrapped with
182          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
183          main program could be modified from something like
184    
185             fun main (pgm, args) = ...
186    
187         to
188    
189            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
190    
191       2. To be able to access BackTrace.monitor, you have to add
192          library $smlnj-tdp/plugins.cm to the .cm file that contains your
193          main function.
194    
195       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
196          CM might have created in the past for your project).
197    
198       4. Build the system using this command line:
199    
200             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
201                      myprog.cm MyProg.main myprog
202    
203          instead of the usual
204    
205             ml-build myprog.cm MyProg.main myprog
206    
207    I changed a library name:
208    
209       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
210    
211    New libraries:
212    
213       $smlnj-tdp/back-trace.cm
214          -- when loaded causes the back-trace plugin to be installed
215       $smlnj-tdp/coverage.cm
216          -- when loaded causes the coverage plugin to be installed
217    
218    ----------------------------------------------------------------------
219    Name: Matthias Blume (blume (at) tti - c (dot) org)
220    Date: 2004/10/18 16:45:00 CDT
221    Tag: blume-20041018-groupowner
222    Description:
223    
224    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
225    
226    Eliminated group owner specs from .cm files throughout the source tree.
227    
228    ----------------------------------------------------------------------
229    Name: Matthias Blume (blume (at) tti - c (dot) org)
230    Date: 2004/10/15 15:45:00 CDT
231    Tag: blume-20041015-coverage
232    Description:
233    
234    * Test coverage tool added!
235    * Further reorganization of tracing-, debugging-, and profiling support:
236    
237        - moved original BTImp -- now called BackTrace -- into a separate
238          library called $/trace-debug-profile.cm
239        - eliminated all mentions of BTrace from SMLofNJ.Internals
240        - only the instrumentation mechanism is now left in the compiler proper
241        - BackTrace module is a plugin which is NOT plugged in by default
242        - Coverage module is another such plugin
243    
244    To get the benefits of any of these plugin modules, the code in
245    question must be compiled with tdp instrumentation turned on.  This
246    can be done by setting SMLofNJ.Internals.TDP.mode to true.
247    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
248    
249    Plugins are selected at link time.  (Pre-compiled instrumented code
250    can be re-loaded with different plugins in effect.)  When an
251    instrumented module is linked, whatever plugins are at that time
252    enabled will come into effect for that module.
253    
254    To enable the back-trace plugin, load library $/trace-debug-profile.cm
255    and invoke BackTrace.install() (e.g., from the interactive prompt).
256    To enable the coverage plugin, load the same library and invoke
257    Coverage.install().
258    
259    Back-traces are generated automatically on uncaught exceptions and
260    when the code in question explicitly invokes BackTrace.trigger().
261    
262    Coverage (and execution frequency-) information must be queried
263    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
264    
265    ----------------------------------------------------------------------
266    Name: Matthias Blume (blume (at) tti - c (dot) org)
267    Date: 2004/10/14 17:40:00 CDT
268    Tag: blume-20041014-tdp-core
269    Description:
270    
271    Snapshot of a significant overhaul of how the trace/debug/profile support
272    is hooked into the system (specifically: Core and SMLofNJ.Internals).
273    
274    ----------------------------------------------------------------------
275    Name: Matthias Blume (blume (at) tti - c (dot) org)
276    Date: 2004/10/13 16:34:00 CDT
277    Tag: blume-20041013-tdp
278    Description:
279    
280    Some rationalization of names:
281    
282       structure BTrace -> structure TDPInstrument
283       etc.
284    
285    This is is preparation of using the original back-trace
286    instrumentation for other purposes.  "TDP" stands for
287    Trace/Debug/Profile.
288    
289    The control flag controlling whether instrumentation is on or off is now
290    registered under a different name, so instead of running sml as
291    
292     sml -Cinstrument.btrace-mode=true
293    
294    one has to say
295    
296     sml -Ctdp.instrument=true
297    
298    ----------------------------------------------------------------------
299    Name: Matthias Blume (blume (at) tti - c (dot) org)
300    Date: 2004/10/11 16:37:00 CDT
301    Tag: blume-20041011-regions
302    Description:
303    
304    Made some minor modifications to elabcore.sml to have source regions
305    be propagated more tightly -- resulting in better (i.e., smaller)
306    regions being reported in error- and debug messages.
307    
308    ----------------------------------------------------------------------
309    Name: Matthias Blume (blume (at) tti - c (dot) org)
310    Date: 2004/10/08 22:50:00 CDT
311    Tag: blume-20041008-cmkw
312    Description:
313    
314    Fixed handling of keywords in .cm files: After seeing "is" the lexer
315    treats subsequent occurrences of "group", "library", "source", "is",
316    "*", and "-" as ordinary identifiers rather than keywords.
317    
318    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
319    is supposed to accept a tool argument called "source", but this did
320    not work because of the clash with the keyword.
321    
322    ----------------------------------------------------------------------
323    Name: Matthias Blume (blume (at) tti - c (dot) org)
324    Date: 2004/10/07 16:00:00 CDT
325    Tag: blume-20041007-cleanup
326    Description:
327    
328    Assorted cleanup work:
329    
330      - got rid of intstrmap in favor of using the library's
331        hash table implementation
332      - threw out most of the pathnames stuff, as it was not used anyway
333      - simplified tokentable implementation
334      - fixed some minor spelling errors
335    
336    ----------------------------------------------------------------------
337    Name: Matthias Blume (blume (at) tti - c (dot) org)
338    Date: 2004/10/06 15:15:15 CDT
339    Tag: blume-20041006-handler
340    Description:
341    
342    Cleaned up the absyn to reflect the invariant that HANDLE always
343    carries a FNexp as part of the type definition.  This eliminates some
344    superfluous sanity checks at runtime down the road.
345    
346    Some minor cleanup of the btrace code.
347    
348    ----------------------------------------------------------------------
349    Name: Matthias Blume (blume (at) tti - c (dot) org)
350    Date: 2004/10/01 10:20:30 CDT
351    Tag: blume-20041001-slave
352    Description:
353    
354    Added hack to make slave mode work in the presence of the version
355    tool.  (Still, since the master does two passes over the code for
356    CMB.make, the release number gets bumped twice when slaves are
357    attached.  I don't know if this is worth fixing...)
358    
359    ----------------------------------------------------------------------
360    Name: Matthias Blume (blume (at) tti - c (dot) org)
361    Date: 2004/09/30 10:55:00 CDT
362    Tag: blume-20040930-version
363    Description:
364    
365    * Moved the "version" magic into its own little library under
366      src/system/smlnj/internal.  This avoids expensive reconstruction of
367      a stable src/compiler/core.cm.
368    
369    * At the same time, structure CompilerVersion is now known as
370      structure SMLNJVersion.
371    
372    * Arranged for the version tool to NOT kick in when rebuilding the system
373      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
374      Also, loading the versiontool does not work when rebuilding the system
375      because CM is not properly initialized at that time.
376    
377    ----------------------------------------------------------------------
378    Name: Matthias Blume (blume (at) tti - c (dot) org)
379    Date: 2004/09/29 14:00:00 CDT
380    Tag: blume-20040929-autoversion
381    Description:
382    
383    Implemented some CM magic to have
384        file src/compiler/TopLevel/main/version.sml
385    generated automagically.
386    The version is taken from two files: config/version and config/release.
387    The first is expected to contain a two-part version number such as 110.49.
388    The second should contain a single number, but it may be missing.
389    
390    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
391    time the version tool is loaded (which is the first time you say CMB.make),
392    then the tool will increment the value stored in config/release every
393    time CMB.make is invoked.
394    
395    The binfile format is now insensitive to anything beyond the first
396    two components of a version number, so bumping the release does not render
397    binfiles incompatible.  Auto-bumping can be used to keep track of versions
398    during development without invalidating existing binfiles.
399    
400    In any case, every CMB.make updates the date information in version.sml.
401    (This is the date that is printed in the banner.)
402    
403    ----------------------------------------------------------------------
404    Name: Matthias Blume (blume (at) tti - c (dot) org)
405    Date: 2004/09/28 10:53:00 CDT
406    Tag: blume-20040928-controls
407    Description:
408    
409    Some cleanup of the controls code.
410    
411    ----------------------------------------------------------------------
412    Name: Matthias Blume (blume (at) tti - c (dot) org)
413    Date: 2004/09/27 22:08:00 CDT
414    Tag: blume-20040927-controls
415    Description:
416    
417    Added two pieces of functionality to the Controls interface:
418    
419    1.   val save'restore: 'a control -> unit -> unit
420    
421       grabs the current value of the control in stage 1 and restores it
422       in stage 2.
423    
424    2.   val set' : 'a control * 'a -> unit -> unit
425    
426       stores the given value into the control in stage 2 (i.e., delayed)
427       but does all error checking in stage 1.
428       (This is for string controls that need to do parse their argument --
429        something that might fail.  In some cases, notably in CM, one
430        already knows the intended argument but wants to delay the actual
431        assignment until a time when error recovery would be more difficult.)
432    
433    Changed the handling of controls in tool arguments to classes "sml" and
434    "lazysml":
435       - use Controls.save'restore as a more robust way of restoring the
436         old value (in particular: without having to re-parse the string)
437       - use controls to handle the "overload" keyword in the init group
438         (I believe this change actually fixes a long-standing obscure bug.)
439    
440    ----------------------------------------------------------------------
441    Name: Matthias Blume (blume (at) tti - c (dot) org)
442    Date: 2004/09/27 17:00:00 CDT
443    Tag: blume-20040927-lazysml
444    Description:
445    
446    Added a new tool class called "lazysml" to CM's tool chest.  The only
447    difference to "sml" is that compilation is done with Control.lazysml
448    set to true.  A source of class "lazysml" is automatically recognized
449    by a file name suffix of ".lml".
450    
451    In addition to the above feature, the original class "sml" now also
452    supports a tool argument "lazy" which has the same effect.  As a
453    result, the following three lines are equivalent:
454    
455        foo.sml : lazysml
456        foo.sml : sml (lazy)
457        foo.sml (lazy)
458    
459    The setting goes into effect both during parsing and during
460    compilation.  The original setting is restored right after parsing and
461    after compilation, respectively.
462    
463    In addition to all the above, there is also a general mechanism to set
464    ANY of the "controls" that are available at the command line via
465    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
466    apply as well. (In fact, "lazy" is implemented as a special case of
467    the general mechanism.)
468    
469    The .cm file syntax uses a new keyword tool argument called "with".
470    There are several ways of indicating the desired settings:
471    
472        foo.sml (with:parser.quotations=true)
473        foo.sml (with:(name:parser.quotations value:true))
474        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
475        foo.sml (with:(name1=value1 name2=value2 ...))
476        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
477    
478        etc.
479    
480    Another possible abbreviation is to leave out the =v or value:v part
481    if the name refers to a boolean control (in which case the value is
482    taken to be true).  Thus, one could get lazy sml also by saying:
483    
484        foo.sml (with:parser.lazy-keyword=true)
485        foo.sml (with:parser.lazy-keyword)
486        foo.sml (with:(name:parser.lazy-keyword value:true))
487        foo.sml (with:(name:parser.lazy-keyword))
488    
489    ----------------------------------------------------------------------
490    Name: Matthias Blume (blume (at) tti - c (dot) org)
491    Date: 2004/09/24 16:22:00 CDT
492    Tag: blume-20040924-ppc-long-branch
493    Description:
494    
495    Turned message about "emiting long form of branch" off by default.
496    Added a control flag to turn it back on when desired.
497    
498    ----------------------------------------------------------------------
499    Name: Matthias Blume (blume (at) tti - c (dot) org)
500    Date: 2004/09/24 16:05:00 CDT
501    Tag: blume-20040924-rounding
502    Description:
503    
504    Applied patch for setting rounding modes under Mac OS X.  Thanks to
505    Melissa O'Neill for providing the code!
506    
507    ----------------------------------------------------------------------
508    Name: Matthias Blume (blume (at) tti - c (dot) org)
509    Date: 2004/09/23 17:30:00 CDT
510    Tag: blume-20040923-envvars
511    Description:
512    
513    1. Changed definition of type ControlRegistry.registry_tree to
514       include control_info (i.e., the name of the controlling environment
515       variable).
516    
517    2. Added command-line flags -e and -E to print the names of environment
518       variables that can be used to control internal settings.
519       (This uses the new API mentioned in 1.)
520    
521    ----------------------------------------------------------------------
522    Name: Matthias Blume (blume (at) tti - c (dot) org)
523    Date: 2004/09/13 16:50:00 CDT
524    Tag: Release_110_49
525    Description:
526    
527    New working version (110.49).  NEW BOOTFILES!
528    
529    ----------------------------------------------------------------------
530    Name: Matthias Blume
531    Date: 2004/09/13 16:20:00 CDT
532    Tag: blume-20040913-config-mlrisc
533    Description:
534    
535    Put target "mlrisc" back into the default list.
536    (There is no harm in having it, and some users have expressed their
537    wish to have "mlrisc" included by default.)
538    
539    ----------------------------------------------------------------------
540    Name: John Reppy
541    Date: 2004/09/13
542    Tag: jhr-20040913-signals
543    Description:
544    
545    Fixed the signal masking code to properly nest mask/unmask operations
546    on a per-signal basis.
547    
548    ----------------------------------------------------------------------
549    Name: Matthias Blume (blume (at) tti - c (dot) org)
550    Date: 2004/09/08 13:20:00 CDT
551    Tag: blume-20040908-heap-magic
552    Description:
553    
554    Bumped the heap macig number to 0x09082004 to account for the changed
555    layout of the ML frame under MacOS X.
556    
557    ----------------------------------------------------------------------
558    Name: Allen Leung (leunga (at) reservoir (dot) com)
559    Date: 2004/09/03 11:26:00 EST
560    Tag: leunga-20040903-cygwin-install
561    Description:
562    
563       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
564    The Cygwin runtime is turned on by setting the environment
565    variable SMLNJ_CYGWIN_RUNTIME to 1.
566    
567    ----------------------------------------------------------------------
568    Name: Matthias Blume (blume (at) tti - c (dot) org)
569    Date: 2004/08/31 17:15:00 CDT
570    Tag: blume-20040831-core
571    Description:
572    
573    Added some exports to src/compiler/core.cm upon request by J. Joemann.
574    
575    ----------------------------------------------------------------------
576    Name: Matthias Blume (blume (at) tti - c (dot) org)
577    Date: 2004/08/30 17:55:00 CDT
578    Tag: blume-20040830-installer
579    Description:
580    
581    Upon request by Johannes Joemann:
582    
583    - improved ML code of installer to fall back to coping when renaming
584      fails (i.e., when source and target are on different file systems);
585      the code compiles but has yet to be tested in anger
586    - removed mlrisc from list of default targets (config/targets)
587    
588    ----------------------------------------------------------------------
589    Name: Matthias Blume (blume (at) tti - c (dot) org)
590    Date: 2004/08/27 17:20:00 CDT
591    Tag: blume-20040827-ptreql
592    Description:
593    
594    Added ptreql primop to structure InlineT (upon request from Larry
595    Paulson).
596    
597    ----------------------------------------------------------------------
598    Name: Allen Leung (leunga (at) reservoir (dot) com)
599    Date: 2004/08/15 21:21:00 EST
600    Tag: leunga-110_48-udgraph
601    Description:
602    
603    Another bug fix from Carl Hauser:
604    
605    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
606    > 48c48
607    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
608    > e::L)
609    > ---
610    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
611    > Without this, any deletion of an edge in an undirected graph does severe
612    > violence to the graph.
613    
614    ----------------------------------------------------------------------
615    Name: Allen Leung (leunga (at) reservoir (dot) com)
616    Date: 2004/08/10 23:35:00 EST
617    Tag: leunga-110_48-ppc
618    Description:
619    
620       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
621    
622    ----------------------------------------------------------------------
623    Name: Matthias Blume (blume (at) tti - c (dot) org)
624    Date: 2004/08/10 12:00:00 CDT
625    Tag: Release_110_48
626    Description:
627    
628    New working version (110.48).  NEW BOOTFILES!
629    
630    ----------------------------------------------------------------------
631    Name: Allen Leung (leunga (at) reservoir (dot) com)
632    Date: 2004/08/09 12:21:00 EST
633    Tag: leunga-110_47-dijsktra
634    Description:
635    
636    Bug fix from Carl Hauser:
637    
638    single_source_shortest_paths in dijkstra.sml was observed to get wrong
639    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
640    
641    The problem is that following the expression A.update(dist,s,Num.zero)
642    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
643    
644    ----------------------------------------------------------------------
645    Name: Matthias Blume (blume (at) tti - c (dot) org)
646    Date: 2004/08/06 18:10:00 CDT
647    Tag: blume-20040806-cmdline
648    Description:
649    
650    Fiddled with handling of command-line options:
651    
652      * sml now quits after processing the command line
653        if -H, -S, -h<n>, or -s<n> appears as the last
654        command-line argument
655      * a new option -q terminates the session when encountered on
656        the command line; subsequent arguments will be ignored
657      * bug fixes: short (erroneous) arguments are no longer ignored
658        completely
659    
660    ----------------------------------------------------------------------
661    Name: Allen Leung (leunga (at) reservoir (dot) com)
662    Date: 2004/08/04 18:17:00 EST
663    Tag: leunga-110_47-ppc-ibm-asm
664    Description:
665    
666      - Added minimal IBM assembly syntax support for PowerPC.
667    
668      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
669        been altered in the runtime.  We need an automatic way to keep the file
670        in sync.
671    
672    ----------------------------------------------------------------------
673    Name: Matthias Blume (blume (at) tti - c (dot) org)
674    Date: 2004/08/04 14:00:00 CDT
675    Tag: Release_110_47
676    Description:
677    
678    New working version (110.47).  NEW BOOTFILES!
679    
680    ----------------------------------------------------------------------
681    Name: Matthias Blume (blume (at) tti - c (dot) org)
682    Date: 2004/08/03 14:25:00 CDT
683    Tag: blume-20040803-callingconv
684    Description:
685    
686    Added low-level support for choosing C calling conventions by
687    twiddling the type of rawccall.  (See
688    src/compiler/Semant/types/cproto.sml for details.)
689    
690    ----------------------------------------------------------------------
691    Name: Matthias Blume (blume (at) tti - c (dot) org)
692    Date: 2004/08/02 15:55:00 CDT
693    Tag: blume-20040802-backout
694    Description:
695    
696    Backed out of change to win32-filesys.c.  The earlier patch to
697    get_file_time caused CM to produce files with the wrong time stamp.
698    
699    ----------------------------------------------------------------------
700    Name: Matthias Blume (blume (at) tti - c (dot) org)
701    Date: 2004/08/02 14:45:00 CDT
702    Tag: blume-20040802-nlffi-win32
703    Description:
704    
705    Added NLFFI support for Win32, adapted from a patch provided by David
706    Hansel.  This is currently completely untested.  Also, the issue
707    concerning stdcall vs. ccall is still unresolved.
708    
709    ----------------------------------------------------------------------
710    Name: Matthias Blume (blume (at) tti - c (dot) org)
711    Date: 2004/07/30 17:55:00 CDT
712    Tag: blume-20040730-various
713    Description:
714    
715    Gearing up towards 110.47...
716    
717    - various minor bugfixes to ml-nlffigen
718    - a beginning of a manual for nlffi
719    
720    - eliminated 'export name=value' in config/install.sh as this does
721      not work with certain versions of /bin/sh
722      (Thanks to David King at Motorola for catching this.)
723    
724    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
725      - added a test for tm==NULL to gmtime.c and localtime.c
726      - applied patch for incorrect GetFileTime under win32
727      - toSeconds -> toMilliseconds in Win32/win32-process.sml
728    
729    ----------------------------------------------------------------------
730    Name: Matthias Blume (blume (at) tti - c (dot) org)
731    Date: 2004/07/21 18:20:00 CDT
732    Tag: blume-20040721-nlffigen
733    Description:
734    
735    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
736      for a typedef to an incomplete type, but leave out the "typ" member.
737      (This is just for consistency.)
738    - Started to produce what is supposed to become better (i.e., comprehensive)
739      documentation of what ml-nlffigen does and produces.
740    
741    ----------------------------------------------------------------------
742    Name: Matthias Blume (blume (at) tti - c (dot) org)
743    Date: 2004/07/14 16:25:00 CDT
744    Tag: blume-20040714-union
745    Description:
746    
747    Added C_UNION to c-calls/c-types.sml and updated the machinery
748    (ml-nlffigen, cproto.sml) that conveys C function interface
749    information to the code generator.
750    
751    However, the actual architecture-specific implementation of function
752    arguments and results that are C unions is still not implemented.
753    
754    ----------------------------------------------------------------------
755    Name: Allen Leung (leunga (at) reservoir (dot) com)
756    Date: 2004/07/14 14:38:00 EST
757    Tag: leunga-110_46_1-ppc-lwzu
758    Description:
759    
760        Added these instructions to the PowerPC architecture:
761        LBZU(X), LHZU(X), LWZU(X),
762        STWU(X), STFDU, STFSU
763    
764        etc...
765    
766        Note: I haven't added their instruction encoding into the description.
767    
768    ----------------------------------------------------------------------
769    Name: Allen Leung (leunga (at) reservoir (dot) com)
770    Date: 2004/07/13 15:04:00 EST
771    Tag: leunga-110_46_1-ppc-lwarx
772    Description:
773    
774       Added the two instructions LWARX and STWCX to the PowerPC
775    instruction set.
776    
777       A (untested) rewrite of loop-structure.sml.  The old version
778    is completely broken.
779    
780    ----------------------------------------------------------------------
781    Name: Matthias Blume (blume (at) tti - c (dot) org)
782    Date: 2004/07/13 13:50:00 CDT
783    Tag: blume-20040713-nlffi
784    Description:
785    
786    - use paramAlloc to report c-calls with too many arguments
787      (for PPC version where parameter area is pre-allocated)
788    - added ccall_maxargspace to machspec (to implement the above)
789    - made "make" commend in CM's "make" tool configurable
790    - added option (default: on) for passing the name of the SML/NJ's "bin"
791      directory to "make"; the call looks like this:
792    
793         make <options> SMLNJ_BINDIR=<dir> <target>
794    
795      This can be used by the Makefile to, e.g., pick the "right" version
796      of ml-nlffigen.
797    - minor code tweaks
798    
799    ----------------------------------------------------------------------
800    Name: Matthias Blume (blume (at) tti - c (dot) org)
801    Date: 2004/07/12 22:50:00 CDT
802    Tag: blume-110_46_1-macosx-nlffi
803    Description:
804    
805    NLFFI under Mac OS X now working (sort of).  This is largely untested,
806    though.
807    
808    Note:  1. You have to make a new, clean build of the runtime system.
809           2. There are new BOOTFILES, you have to use them!
810              (Doing the bootstrap process yourself would be *very* painful!
811               If you absolutely have to do it, build the system under
812               a different architecture and then cross-compile.)
813    
814    Version bumped to 110.46.1 to account for runtime data format changes.
815    
816    ----------------------------------------------------------------------
817    Name: Matthias Blume (blume (at) tti - c (dot) org)
818    Date: 2004/06/18 14:30:00 CDT
819    Tag: blume-20040618-unix
820    Description:
821    
822    Changed the implementation of structure Unix so that the same stream
823    is returned every time one of the {text,bin}{In,Out}streamOf functions
824    is invoked on the same proc.  This is not what the spec currently
825    says -- although IMO it arguably should.  (See discussion below.)
826    
827    ----------------------------------------------------------------------
828    Name: Matthias Blume (blume (at) tti - c (dot) org)
829    Date: 2004/06/17 18:15:00 CDT
830    Tag: Release_110_46
831    Description:
832    
833    New working version (110.46).  NEW BOOTFILES!
834    
835    ----------------------------------------------------------------------
836    Name: Matthias Blume (blume (at) tti - c (dot) org)
837    Date: 2004/06/17 17:20:00 CDT
838    Tag: blume-20040617-timer-unix
839    Description:
840    
841    Changed the interface of structures Timer and Unix to match the most
842    recent Basis spec.
843    
844    In the case of Unix there still seems to be an open/weird issue:
845    
846       The {text,bin}{In,Out}streamOf functions are supposed to create
847       fresh streams whenever they are called -- as opposed to have them
848       return the same stream every time.  This design is supposed to
849       prevent space leaks caused by proc values hanging on to streams.
850    
851       The reap function, on the other hand, is supposed to close the
852       streams.  This cannot be done without having a handle on the
853       stream in proc after all...
854    
855    I took the liberty to implement the following stopgap solution:
856    
857      The proc value hangs on to the most recently created stream(s).
858      Reap closes those.  If either or both of the two streams hadn't
859      been created at all yet, then reap will close the corresponding
860      file descriptors directly.
861    
862    PS: I don't understand the original space leak argument anymore.  If
863    a proc hangs on to the imperative stream, then I/O operations on those
864    will advance the state of the cached stream and avoid the space leak.
865    
866    ----------------------------------------------------------------------
867    Name: Matthias Blume (blume (at) tti - c (dot) org)
868    Date: 2004/05/28 16:45:00 CDT
869    Tag: blume-20040528-basis
870    Description:
871    
872    Added signature PACK_REAL and exported functor PrimIO.
873    
874    ----------------------------------------------------------------------
875    Name: Matthias Blume (blume (at) tti - c (dot) org)
876    Date: 2004/05/25 16:00:00 CDT
877    Tag: blume-20040525-group-owner
878    Description:
879    
880    CM now ignores (but still accepts) the "owner" information in group
881    descriptions.  The owner of a group is its next enclosing
882    library. Each group must have a unique owner.  (There is a virtual
883    "toplevel" library that owns groups which are not nested within a real
884    library.)  Previously, each group had to explicitly declare its owner,
885    and CM would check that such a declaration is correct.  The new scheme
886    is to have CM check that for each group there is precisely one owning
887    library.
888    
889    The advantage of the new scheme is that the programmer no longer needs
890    to maintain the somewhat annoying owner information.  The downside is
891    that CM cannot enforce the ownership rule across multiple runs of
892    CM.make.  Fortunately, enclosing the same group in two different
893    libraries A and B which are not part of the same program does not
894    cause real problems.
895    
896    ----------------------------------------------------------------------
897    Name: Matthias Blume (blume (at) tti - c (dot) org)
898    Date: 2004/05/20 16:00:00 CDT
899    Tag: blume-20040520-win32
900    Description:
901    
902    Made the win32 version work again.  (Strangely, a misplaced comma had
903    slipped into win32-process.c which prevented the runtime from being
904    compiled correctly.)
905    
906    Also, included a minor addition to ml-build.bat analogous to what was
907    done in blume-20040519-ml-build.
908    
909    ----------------------------------------------------------------------
910    Name: Matthias Blume (blume (at) tti - c (dot) org)
911    Date: 2004/05/19 22:10:00 CDT
912    Tag: blume-20040519-ml-build
913    Description:
914    
915    Arranged for ml-build to clean up after itself a little bit better.
916    The script generates a temporary SML source file and compiles it using
917    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
918    gets rid of those at the end, so they don't accumulate under .cm.
919    
920    This required a minor change to install.sh because the name of the
921    metadata directory (default: .cm) is actually configurable at
922    installation time.
923    
924    ----------------------------------------------------------------------
925    Name: Matthias Blume (blume (at) tti - c (dot) org)
926    Date: 2004/05/18 15:50:00 CDT
927    Tag: blume-20040518-mkreader
928    Description:
929    
930    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
931    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
932    
933    ----------------------------------------------------------------------
934    Name: Matthias Blume (blume (at) tti - c (dot) org)
935    Date: 2004/05/11 14:35:00 CDT
936    Tag: blume-20040511-win32sock
937    Description:
938    
939    Added previously missing support for many socket-related functions
940    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
941    for the voluminous patch!
942    
943    (I have not tested this patch under win32 yet.)
944    
945    Here is David's e-mail:
946    
947    Hi,
948    
949    Attached to this email you find a diff against sml/nj 110.45
950    that will enable socket support under Windows.
951    
952    To apply the patch (using unix or cygwin)
953    1) gunzip runtime.diff.gz
954    2) "cd" into "src/runtime" in the source tree of a fresh
955       110.45 installation.
956    3) patch -p 1 < [your/path/to]runtime.diff
957    
958    The code compiles fine but has NOT yet been extensively tested.
959    I only ran a few tests for basic socket client functionality
960    (which worked fine).  Especially the functions that use ioctl
961    are not tested at all and might not work (see below).
962    
963    I implemented this since we want to move to a newer version of sml/nj
964    but need socket support in order to use it.  This is the first time I
965    even had a look at the sml/nj source,  so please review my changes
966    before making this part of the distribution!  Here are a few issues
967    that I think might be better for someone to solve who is more
968    familiar with the sml/nj source (and socket programming):
969    
970    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
971      exception since I could not figure out what the windows equivalent
972      of these functions is
973    
974    - In sockets-osdep.h there are a some #include statements that are
975      only used in a few files that include sockets-osdep.h
976    
977    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
978      does not process its return value since I don't know how to
979      report an error upwards.
980    
981    - It would probably be good to have a call to WSACleanup() when
982      the library is unloaded (if there is such a possibility).
983      Otherwise I think Windows will take care of this automatically
984      when the process finishes.
985    
986    - I used ioctlsocket() as a replacement for ioctl() but I have
987      no idea if that is actually the proper replacement on Windows.
988    
989    - All these issues are marked in the code by "FIXME" comments.
990    
991    We use sml/nj extensively in our products and are quite happy
992    with it.  I hope this contribution will help you.
993    
994    Keep up the good work!
995    
996    David
997    
998    ----------------------------------------------------------------------
999    Name: Matthias Blume (blume (at) tti - c (dot) org)
1000    Date: 2004/05/11 14:20:00 CDT
1001    Tag: blume-20040511-installml
1002    Description:
1003    
1004    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1005    the patch.)
1006    
1007    ----------------------------------------------------------------------
1008    Name: Matthias Blume (blume (at) tti - c (dot) org)
1009    Date: 2004/05/11 14:05:00 CDT
1010    Tag: blume-20040511-nlffi-netbsd
1011    Description:
1012    
1013    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1014    the patch.)
1015    
1016    ----------------------------------------------------------------------
1017    Name: Matthias Blume (blume (at) tti - c (dot) org)
1018    Date: 2004/05/11 12:05:00 CDT
1019    Tag: blume-20040511-exports
1020    Description:
1021    
1022    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1023    export lists in compiler-related .cm-files.
1024    
1025    ----------------------------------------------------------------------
1026    Name: Matthias Blume (blume (at) tti - c (dot) org)
1027    Date: 2004/05/11 11:35:00 CDT
1028    Tag: blume-20040511-allsource
1029    Description:
1030    
1031    The installer now honors the "src-smlnj" target again, although its meaning
1032    has changed from "all sources required for the compiler" to "all sources
1033    the installer knows about".  In other words, if you enable "src-smlnj"
1034    in the "targets" file, then the installer will pull in sources for
1035    everything.  (Notice that this refers to source code only.  Compiled
1036    code is still only installed for modules that were requested explicitly
1037    or which are required for other modules that were requested explicitly.)
1038    
1039    ----------------------------------------------------------------------
1040    Name: Matthias Blume (blume (at) tti - c (dot) org)
1041    Date: 2004/04/23 17:40:00 CDT
1042    Tag: blume-20040423-ieee-scan
1043    Description:
1044    
1045    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1046    in the exponent calculation we get INF or ZERO (depending on the mantissa
1047    and the sign of the exponent).
1048    
1049    ----------------------------------------------------------------------
1050    Name: Matthias Blume (blume (at) tti - c (dot) org)
1051    Date: 2004/04/23 10:40:00 CDT
1052    Tag: blume-20040423-ml-build
1053    Description:
1054    
1055    The ml-build script now terminates with a non-0 status when something
1056    goes wrong.
1057    
1058    ----------------------------------------------------------------------
1059    Name: Matthias Blume (blume (at) tti - c (dot) org)
1060    Date: 2004/04/22 16:35:00 CDT
1061    Tag: blume-20040422-Option
1062    Description:
1063    
1064    Made exception Option to be the same as exception Option.Option
1065    (as it should be).
1066    
1067    ----------------------------------------------------------------------
1068    Name: Allen Leung (leunga (at) reservoir (dot) com)
1069    Date: 2004/03/19 14:40:00 EST
1070    Tag: leunga-20040319-cygwin-nlffi
1071    Description:
1072    
1073    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1074    of SML/NJ.  The problem is that
1075    
1076       lib = dlopen(NULL, ...)
1077       f   = dlsym(lib, "malloc");
1078    
1079    does not work on Windows unless we explicitly export symbols
1080    such as 'malloc' during linking.  We fixed this by explicitly
1081    exporting the required symbols with the magic gcc incantation:
1082    
1083        -Wl,--export-all cygwin.def
1084    
1085    where cygwin.def is a file containing all the symbols that
1086    we wish to export.
1087    
1088    I suspect this is a Windows problem and we'll have to
1089    do the same (somehow with windows compilers) when
1090    we build the native win32 version with the system
1091    calls LoadLibrary/GetProcAddress.
1092    
1093    ----------------------------------------------------------------------
1094    Name: Matthias Blume (blume (at) tti - c (dot) org)
1095    Date: 2004/03/04 16:35:00 CST
1096    Tag: blume-20040304-intinf-fmt
1097    Description:
1098    
1099    Fixed problem with IntInf.fmt (sign would show up on the right instead
1100    of on the left for BIN, OCT, and HEX).
1101    
1102    ----------------------------------------------------------------------
1103    Name: Matthias Blume (blume (at) tti - c (dot) org)
1104    Date: 2004/03/04 11:25:00 CST
1105    Tag: blume-20040304-symlinks
1106    Description:
1107    
1108    Fixed problem with installer script (unix only) where bin/ml-yacc and
1109    friends pointed (via symlinks) to absolute locations instead of just
1110    .run-sml.  This was reported by Vesa A Norrman.
1111    
1112    ----------------------------------------------------------------------
1113    Name: Matthias Blume (blume (at) tti - c (dot) org)
1114    Date: 2004/02/13 14:50:00 CST
1115    Tag: Release_110_45
1116    Description:
1117    
1118    New working version (110.45). New bootfiles.
1119    
1120    ----------------------------------------------------------------------
1121    Name: Matthias Blume (blume (at) tti - c (dot) org)
1122    Date: 2004/01/26 15:15:15 CST
1123    Tag: blume-20040126-toplevel
1124    Description:
1125    
1126    Improved handling of exceptions at the interactive toplevel.
1127    
1128    ----------------------------------------------------------------------
1129    Name: Matthias Blume (blume (at) tti - c (dot) org)
1130    Date: 2004/01/26 11:25:00
1131    Tag: blume-20040126-app
1132    Description:
1133    
1134    Type of top-level "app" corrected.
1135    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1136    
1137    ----------------------------------------------------------------------
1138    Name: Matthias Blume (blume (at) tti - c (dot) org)
1139    Date: 2003/11/18 17:10 CST
1140    Tag: blume-20031118-basis-fiddle
1141    Description:
1142    
1143    - changed Timer interface to what might become the spec
1144    - POSIX_FLAGS -> BIT_FLAGS according to spec
1145    - some other minor discrepancies wrt. spec eliminated
1146    
1147    ----------------------------------------------------------------------
1148    Name: Matthias Blume (blume (at) tti - c (dot) org)
1149    Date: 2003/11/06 12:00:00 CST
1150    Tag: Release_110_44
1151    Description:
1152    
1153    New working version (110.44). New bootfiles.
1154    
1155    ----------------------------------------------------------------------
1156    Name: Matthias Blume (blume (at) tti - c (dot) org)
1157    Date: 2003/11/04 11:50:00 CST
1158    Tag: blume-20031104-move-libraries
1159    Description:
1160    
1161    Eliminated the "dont_move_libraries" directive in config/targets.
1162    (The mechanism was broken and could not be fixed easily.  Moreover,
1163    there does not seem to be any reason not to move all libraries into
1164    lib during installation.  I originally implemented this directive as a
1165    backward-compatibility feature when I first introduced the new CM.
1166    Now that things have been stable for a long time and going back to the
1167    old CM is not an option, there is no reason to keep it around.)
1168    
1169    ----------------------------------------------------------------------
1170    Name: Matthias Blume (blume (at) tti - c (dot) org)
1171    Date: 2003/11/03 16:00:00 CST
1172    Tag: blume-20031103-installdir
1173    Description:
1174    
1175    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1176    Richards for pointing out the problem and providing the solution.)
1177    
1178    ----------------------------------------------------------------------
1179    Name: Matthias Blume (blume (at) tti - c (dot) org)
1180    Date: 2003/10/01 17:05:00 CDT
1181    Tag: blume-20031001-lal-mlrisc
1182    Description:
1183    
1184    MLRISC bug fix from Lal.
1185    
1186    ----------------------------------------------------------------------
1187    Name: Matthias Blume (blume (at) tti - c (dot) org)
1188    Date: 2003/09/30 16:10:00 CDT
1189    Tag: blume-20030930-primio-bat
1190    Description:
1191    
1192    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1193    2. Improved .bat files (for Win32 port) to make things work under Win95.
1194       (thanks to Aaron S. Hawley for this one)
1195    
1196    ----------------------------------------------------------------------
1197    Name: Matthias Blume (blume (at) tti - c (dot) org)
1198    Date: 2003/09/26 16:05:00 CDT
1199    Tag: blume-20030926-wrappriv
1200    Description:
1201    
1202    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1203    
1204    ----------------------------------------------------------------------
1205    Name: Matthias Blume (blume (at) tti - c (dot) org)
1206    Date: 2003/09/26 15:00:00 CDT
1207    Tag: blume-20030926-110_43_3
1208    Description:
1209    
1210    - additional cleanup
1211    - version number bump, NEW BOOTFILES
1212    
1213    ----------------------------------------------------------------------
1214    Name: Matthias Blume (blume (at) tti - c (dot) org)
1215    Date: 2003/09/26 12:00:00 CDT
1216    Tag: blume-20030926-ppautoload
1217    Description:
1218    
1219    I modified the read-eval-print loop so that the autoloader gets
1220    invoked whenever the prettyprinter tries to look up a symbol that
1221    is not currently defined in the toplevel environment but which
1222    appears in CM's autoload registry.  As a result, we see far fewer of
1223    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1224    
1225    In addition to this I tried to clean up some pieces of the Basis
1226    implementation (e.g., Socket, Word8Array) in order to prevent other
1227    instances of these ?.Foo.Bar.xxx names from being printed.
1228    
1229    The mechanism that picks names for types still needs some work, though.
1230    (Right now it seems that if there is a type A.t which is defined to
1231    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1232    "?.B.u" although the perhaps more sensible solution would be to use
1233    "A.t" in this case.  In other words, the prettyprinter should follow
1234    a chain of DEFtycs not farther than there are corresponding toplevel
1235    names in the current environment.)
1236    
1237    ----------------------------------------------------------------------
1238    Name: Matthias Blume (blume (at) tti - c (dot) org)
1239    Date: 2003/09/24 16:31:00 CDT
1240    Tag: blume-20030924-installer
1241    Description:
1242    
1243    Another installer tweak:  All the ML code for the installer is now
1244    compiled during CMB.make and put into a little library called
1245    $smlnj/installer.cm.  The installation then simply invokes
1246    
1247       sml -m $smlnj/installer.cm
1248    
1249    and everything happens automagically.
1250    
1251    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1252    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1253            and BINDIR.
1254    
1255    The new scheme guarantees that the ML code responsible for the installation
1256    is in sync with the APIs of the main system.  Also, the installer is
1257    somewhat faster because the installer script is precompiled.
1258    
1259    ----------------------------------------------------------------------
1260    Name: Matthias Blume (blume (at) tti - c (dot) org)
1261    Date: 2003/09/24 15:35:00 CDT
1262    Tag: blume-20030924-synsock
1263    Description:
1264    
1265    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1266    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1267    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1268    implementations of non-blocking operations from CML's Socket
1269    structure.
1270    
1271    ----------------------------------------------------------------------
1272    Name: Matthias Blume (blume (at) tti - c (dot) org)
1273    Date: 2003/09/24 15:10:05 CDT
1274    Tag: blume-20030924-sockets
1275    Description:
1276    
1277    1. Fixed SOCKET API and implementation to match Basis spec.
1278       This required changing the internal representation of sockets to one
1279       that remembers (for each socket file descriptor) whether it is currently
1280       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1281       call is made only if the state actually needs to change).
1282    
1283    2. OS-specific details of sockets were moved into separate files, thus
1284       making it possible to unify the bulk of the socket implementations
1285       between Unix and Win32.
1286    
1287    3. CML's socket API changed accordingly.
1288       (Note that we need to remove non-blocking functions from this API
1289       since they are redundant in the case of CML!)
1290    
1291    4. CML's socket implementation now makes use of non-blocking functions
1292       provided by Basis, thus removing all OS-dependent code from this part
1293       of CML.
1294    
1295    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1296    
1297    ----------------------------------------------------------------------
1298    Name: Matthias Blume (blume (at) tti - c (dot) org)
1299    Date: 2003/09/22 12:10:00 CDT
1300    Tag: blume-20030922-110_43_2
1301    Description:
1302    
1303    Made a new interim version and bootfiles for developer's bootstrapping
1304    convenience.
1305    
1306       110.43.2 -- NEW BOOTFILES
1307    
1308    ----------------------------------------------------------------------
1309    Name: Matthias Blume (blume (at) tti - c (dot) org)
1310    Date: 2003/09/19 15:55:00 CDT
1311    Tag: blume-20030919-cmdir
1312    Description:
1313    
1314    1. new-install.sh -> install.sh
1315    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1316    3. tweaked installer so that another name instead of .cm can be chosen
1317       at install time (by setting the CM_DIR_ARC environment variable
1318       during installation); once installation is complete, the name is
1319       fixed
1320    
1321    ----------------------------------------------------------------------
1322    Name: Matthias Blume (blume (at) tti - c (dot) org)
1323    Date: 2003/09/18 16:00:00 CDT
1324    Tag: blume-20030918-110_43_1
1325    Description:
1326    
1327    Made a new interim version and bootfiles for developer's bootstrapping
1328    convenience.
1329    
1330       110.43.1 -- NEW BOOTFILES
1331    
1332    ----------------------------------------------------------------------
1333    Name: Matthias Blume (blume (at) tti - c (dot) org)
1334    Date: 2003/09/18 15:20:00 CDT
1335    Tag: blume-20030918-misc
1336    Description:
1337    
1338    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1339       this seems to be controversial at the moment) and used those in
1340       Posix.ProcEnv.times.
1341    
1342    2. Added Time.{from,to}Nanoseconds to Time.
1343    
1344    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1345       For example, fromLargeInt never needs to look at more than 3 "big
1346       digits" to get its 53 bits of precision.
1347    
1348    ----------------------------------------------------------------------
1349    Name: Matthias Blume (blume (at) tti - c (dot) org)
1350    Date: 2003/09/17 16:30:00 CDT
1351    Tag: blume-20030917-real32-slices
1352    Description:
1353    
1354    Added an entry to the primitive environment
1355    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1356    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1357    
1358    Removed some of the "magic" constants in real64.sml and replaced them
1359    with code that generates these values from their corresponding
1360    integer counterparts.
1361    
1362    Made all(?) the slice-related changes to the Basis and made everything
1363    compile again...
1364    
1365    ----------------------------------------------------------------------
1366    Name: Matthias Blume (blume (at) tti - c (dot) org)
1367    Date: 2003/09/15 17:45:00 CDT
1368    Tag: blume-20030915-rbase
1369    Description:
1370    
1371    Fixed bug in Real.fromLargeInt.
1372    
1373    ----------------------------------------------------------------------
1374    Name: Matthias Blume (blume (at) tti - c (dot) org)
1375    Date: 2003/09/13 18:11:00 CDT
1376    Tag: blume-20030913-libinstall
1377    Description:
1378    
1379    Minor bugfix in config/libinstall (set anchor with path to
1380    standalone tool after installing it, otherwise libraries that
1381    need ml-lex or ml-yacc won't compile the first time the installer
1382    runs).
1383    
1384    ----------------------------------------------------------------------
1385    Name: Matthias Blume (blume (at) tti - c (dot) org)
1386    Date: 2003/09/12 11:45:00 CDT
1387    Tag: blume-20030912-various
1388    Description:
1389    
1390     - fixed bug in Real.toLargeInt
1391     - fixed bug in Posix.ProcEnv.times
1392     - changed inputLine functions to return an option
1393     - minor installer improvements / bugfixes
1394     - changed default @SMLalloc parameter for x86/celeron to 64k
1395    
1396    ----------------------------------------------------------------------
1397    Name: Matthias Blume (blume (at) tti - c (dot) org)
1398    Date: 2003/09/09 22:00:00 CDT
1399    Tag: Release_110_43
1400    Description:
1401    
1402        New working release 110.43.  New bootfiles.
1403    
1404    ----------------------------------------------------------------------
1405    Name: Matthias Blume (blume (at) tti - c (dot) org)
1406    Date: 2003/09/09 19:20:00 CDT
1407    Tag: blume-20030909-installer
1408    Description:
1409    
1410    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1411    Modified config/install.bat to take advantage of it.  Also modified
1412    config/install.sh (and called it config/new-install.sh) to take advantage
1413    of it on Unix systems.  (The SML code is (supposed to be) platform-
1414    independent.)
1415    
1416    The installer can now install everything under Win32
1417    as well as under *nix as long as it compiles.
1418    
1419    Other changes:
1420    
1421      - made CML compile again under Win32
1422      - made eXene compile under Win32 (by providing a fake structure UnixSock
1423        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1424      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1425        same as type int;  under Win32 it isn't
1426      - fixed some slice-related problems in the win32-specific parts of CML
1427      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1428        (under Win32, the volume name is case-insensitive, and the
1429        OS.Path code compares volume names for equality)
1430    
1431    ----------------------------------------------------------------------
1432    Name: Matthias Blume (blume (at) tti - c (dot) org)
1433    Date: 2003/09/08 11:55:00 CDT
1434    Tag: blume-20030908-fullpath
1435    Description:
1436    
1437    Made Win32 version of OS.FileSys.fullPath return current directory
1438    when given an empty string.  This is what the spec says, and incidentally,
1439    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1440    cases when presented with the name of a non-existing .cm file.)
1441    
1442    ----------------------------------------------------------------------
1443    Name: Matthias Blume (blume (at) tti - c (dot) org)
1444    Date: 2003/09/04 16:30:00 CDT
1445    Tag: blume-20030905-slices-etc
1446    Description:
1447    
1448      1. Changed interface to vectors and arrays in Basis to match
1449         (draft) Basis spec.
1450      2. Added signatures and implementations of slices according to
1451         Basis spec.
1452      3. Edited source code throughout the system to make it compile again
1453         under 1. and 2.  (In some cases code had to be added to have it
1454         match the new signatures.)
1455      4. MLRISC should be backward-compatible: the copies of the originals
1456         of files that needed to change under 3. were retained, the .cm files
1457         check the compiler version number and use old versions when
1458         appropriate.
1459      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1460         dirstream -> string option (in accordance with Basis spec).
1461      6. When generating code that counts lines, ml-lex used function
1462         CharVector.foldli, taking advantage of its old interface.
1463         This has been replaced with the corresponding code from
1464         CharVectorSlice.  (html-lex must be re-lexed!)
1465      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1466         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1467      8. Removed temporary additions (fromInternal, toInternal) from the
1468         (now obsolete) IntInf in smlnj-lib/Util.
1469      9. Cleaned up structure Byte.
1470     10. Added localOffset, scan, and fromString to Date (according to spec).
1471         Cleaned/corrected implementation of Date.
1472         (Still need to check for correctness; implement better canonicalizeDate.)
1473     11. Added "scan" to signature IEEE_REAL.
1474     12. Some improvements to IntInf [in particular: efficiency-hack for
1475         mod and rem when second operand is 2 (for parity checks).]
1476     13. Changed representation of type Time.time, using a single IntInf.int
1477         value counting microseconds.  This considerably simplified the
1478         implementation of structure Time.  We now support negative time
1479         values; scan and fromString handle signs.
1480     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1481         ArraySlice).
1482    
1483    ----------------------------------------------------------------------
1484    Name: Matthias Blume (blume (at) tti - c (dot) org)
1485    Date: 2003/08/28 17:00:00 CDT
1486    Tag: blume-20030828-intinf
1487    Description:
1488    
1489    This is a major update which comes with a version number bump
1490    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1491    and an implementation of IntInf in the Basis.
1492    
1493    There are a fairly large number of related changes and updates throughout
1494    the system:
1495    
1496    Basis:
1497            - Implemented IntInf.
1498            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1499            - Added some missing Real64 operations, most notably Real.toLargeInt.
1500            - Added FixedInt as a synonym for Int32.
1501    
1502    compiler:
1503            * Added support for a built-in intinf type.
1504              - literals
1505              - pattern matching
1506              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1507              - overloading on literals and operations
1508    
1509            This required adding a primitive type intinf, some additional
1510            primops, and implementations for several non-trivial intinf
1511            operations in Core.  (The intinf type is completely abstract
1512            to the compiler; all operations get delegated back to the Core.)
1513    
1514            * Intinf equality is handled by polyequal.  However, the compiler
1515            does not print its usual warning in this case (since polyequal
1516            is the right thing to do there).
1517    
1518            * Improved the organization of structure InlineT.
1519    
1520            * A word about conversion primops:
1521            If conversions involving intinf do not cancel out during
1522            CPS contract, then the compiler must insert calls to Core functions.
1523            Since all core access must be resolved already during the FLINT
1524            translate phase, it would be too late a the time of CPS contract
1525            to add new Core calls.  For this reason, conversion primops
1526            for intinf carry two arguments:  1. the numeric argument that
1527            they are supposed to convert, and 2. the Core function that
1528            can help with this conversion if necessary.  If CPS contract
1529            eliminates a primop, then the associated Core function becomes
1530            dead and goes away.  Intinf conversion primops that do not get
1531            eliminated by CPS contract get rewritten into calls of their
1532            core functions by a separate, new phase.
1533    
1534    interactive system:
1535            - Control.Print.intinfDepth controls max length of intinf constants
1536            being printed. (Analogous to Control.Print.stringDepth.)
1537            - Cleanup in printutil and pputil: got rid of unused stuff and
1538            duplicates; replaced some of the code with code that makes better
1539            use of library functionality.
1540    
1541    CM:
1542            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1543            are no longer silent.
1544    
1545    CKIT:
1546            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1547            in favor of LargeInt -- which is now the same as IntInf.
1548            CKIT-knowledgable people should check whether this is what's
1549            intended and otherwise change things back to using Int32 or
1550            FixedInt.
1551    
1552    Throughout the code:
1553            Started using IntInf.int literals and built-in operations
1554            (e.g., comparison with 0) where this seems appropriate.
1555    
1556    
1557    ----------------------------------------------------------------------
1558    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1559    Date: 2003/08/13 11:36:00 CDT
1560    Tag: dbm-20030813-mcz-merge1
1561    Description:
1562    
1563    Merging changes from the mcz-branch development branch into trunk.
1564    These changes involve replacement of the emulated old prettyprinter
1565    interface with direct use of the SML/NJ Lib PP library, and fixing
1566    of a couple of bugs (895, 1186) relating to error messages.  A new
1567    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1568    has been added.
1569    
1570    ----------------------------------------------------------------------
1571    Name: Matthias Blume (blume (at) tti - c (dot) org)
1572    Date: 2003/08/11 15:45:00 CDT
1573    Tag: blume-20030811-windows
1574    Description:
1575    
1576    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1577    
1578        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1579    
1580    This patch restores SML/NJ's ability to run under win32.  There are a
1581    number of changes, including fixes for several bugs that had gone
1582    unnoticed until now:
1583    
1584      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1585        trying to run the win32 version from within cygwin.)
1586    
1587      - There are a number of simple .bat scripts that substitute for their
1588        corresponding Unix shell-scripts.  (See below.)
1589    
1590      - The internals of ml-build have been modified slightly.  The main
1591        difference is that instead of calling ".link-sml" (or link-sml.bat)
1592        using OS.Process.system, the ML process delegates this task back
1593        to the script.  Otherwise problems arise in mixed environments such
1594        as Cygwin where scripts look and work like Unix scripts, but
1595        where OS.Process.system cannot run them.
1596    
1597      - In CM, the srcpath pickler used native pathname syntax -- which
1598        is incorrect in the case of cross-compilation.  The new pickle format
1599        is independent of platform-specific naming conventions.
1600    
1601      - Path configuration files (such as lib/pathconfig) can now choose
1602        between native and standard syntax.  Placing a line of the form
1603    
1604            standard!
1605    
1606        into the file causes all subsequent paths to be interpreted using
1607        CM standard pathname syntax (= Unix conventions); a line
1608    
1609            native!
1610    
1611        switches back to native style.  This was needed so that
1612        path config files can be written portably, see src/system/pathconfig.
1613    
1614      - Runtime system:
1615    
1616          - win32-filesys.c:  get_file_time and set_file_time now
1617            access modification time, not creation time.
1618    
1619          - I/O code made aware of new array representation.
1620    
1621          - Bug fixes in X86.prim.masm.
1622    
1623       - src/system/makeml made aware of win32. (For use under cygwin
1624            and other Unix-environments for windows.)
1625    
1626       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1627         which caused BinIO.inputAll to fail consistently.
1628    
1629    .bat scripts:
1630    
1631       Windows .bat scripts assume that SMLNJ_HOME is defined.
1632    
1633       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1634         applications (sml, ml-yacc, ml-lex).
1635       - ml-build.bat: analogous to ml-build.
1636       - config\install.bat:  Analogous to config/install.sh.  This requires
1637         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1638         (nmake etc. must be on the path, and vcvars32 must have been run.)
1639         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1640         and the bootfile hierarchy must have been unpacked under
1641         sml.boot.x86-win32.
1642         The script is very primitive and does a poor job at error checking.
1643         It only installs the base system, ml-lex, and ml-yacc.  No other
1644         libraries are being installed (i.e., you get only those that
1645         are part of the compiler.)
1646       - link-sml.bat:  analogous to .link-sml, but not currently used
1647    
1648    Unrelated bug fixes:
1649    
1650      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1651        types.
1652      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1653        otherwise.)
1654    
1655    ----------------------------------------------------------------------
1656    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1657    Date: 2003/06/17
1658    Tag: macqueen-20030617-bug895
1659    Description:
1660    
1661    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1662    Tag will be used for new development branch (mcz-branch) for
1663    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1664    
1665    ----------------------------------------------------------------------
1666    Name: Matthias Blume (blume (at) tti - c (dot) org)
1667    Date: 2003/05/27 16:55:00 CDT
1668    Tag: blume-20030527-polyeq
1669    Description:
1670    
1671    Tried to eliminated most cases of polymorphic equality.
1672    
1673    ----------------------------------------------------------------------
1674    Name: Matthias Blume (blume (at) tti - c (dot) org)
1675    Date: 2003/05/21 17:45:00 CDT
1676    Tag: blume-20030517-complete
1677    Description:
1678    
1679    Two changes:
1680    
1681      1. Added a flag for controlling whether non-exhaustive bindings will
1682         be treated as errors (default is false).
1683      2. Cleaned up the *entire* source tree so that CMB.make goes through
1684         without a single non-exhaustive match- or bind warning.
1685    
1686    ----------------------------------------------------------------------
1687    Name: Matthias Blume (blume (at) tti - c (dot) org)
1688    Date: 2003/05/17 10:20:00 CDT
1689    Tag: blume-20030517-absyn
1690    Description:
1691    
1692    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1693    
1694       This mainly affects the quality of error messages.  However, some
1695       of the code is now more straightforward than before. (Treatment of
1696       the above four constructs in translate.sml is much simpler than
1697       the "macro-expansion" that was going on before.  Plus, the mach-
1698       compiler no longer gets invoked just to be able to compile an
1699       if-expression.)
1700    
1701    2. The ErrorMsg.Error exception is now caught and absorbed by the
1702       interactive loop.
1703    
1704    ----------------------------------------------------------------------
1705    Name: Allen Leung
1706    Date: 2003/05/16 13:05:00 CDT
1707    Tag: leunga-20030516-cygwin-runtime
1708    Description:
1709    
1710        Ported the runtime system to cygwin, which uses the unix
1711    x86-unix bin files.  Missing/buggy features:
1712    
1713       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1714         the Cygwin library.
1715       o Ctrl-C handling may be flaky.
1716       o Windows system calls and Windows I/O are not supported.
1717    
1718       A new set of binfiles is located at:
1719    
1720         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1721    
1722       This is only needed for bootstrapping the cygwin version of smlnj.
1723       Other x86 versions can use the existing binfiles.
1724    
1725    ----------------------------------------------------------------------
1726    Name: Matthias Blume
1727    Date: 2003/04/08 15:42:00 CDT
1728    Tag: blume-20030408-listpair
1729    Description:
1730    
1731    1. Added a target 'mlrisc' to installer.
1732    
1733    2. Added missing elements to structure ListPair.
1734    
1735    ----------------------------------------------------------------------
1736    Name: Allen Leung
1737    Date: 2003/01/07 10:40:00 EST
1738    Tag: leunga-20030107-int-rem
1739    Description:
1740    
1741    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1742    The arguments to the SUBL instruction were swapped.
1743    
1744    ----------------------------------------------------------------------
1745    Name: Matthias Blume
1746    Date: 2002/12/12 16:25:00 EST
1747    Tag: blume-20021212-risc-ra
1748    Description:
1749    
1750    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1751    sent the RA into an infinite loop when floating point registers get
1752    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1753    110.37 and 110.38.)
1754    There was another set of potential problems related to the handling of
1755    MLRISC annotations (but those did not yet cause real problems, apparently).
1756    
1757    ----------------------------------------------------------------------
1758    Name: Matthias Blume
1759    Date: 2002/12/06 22:40:00 EST
1760    Tag: blume-20021206-cm-fileid
1761    Description:
1762    
1763    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1764    This fixes the problem of CM getting confused by files that suddenly
1765    change their identity (e.g., by getting unlinked and recreated by some
1766    text editor such as vi).  There might be a better/cheaper/cleaner way
1767    of doing this, but for now this will have to do.
1768    
1769    ----------------------------------------------------------------------
1770    Name: Matthias Blume
1771    Date: 2002/10/28 09:50:00 EST
1772    Tag: blume-20021028-typecheck
1773    Description:
1774    
1775    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1776    
1777    ----------------------------------------------------------------------
1778    Name: Matthias Blume
1779    Date: 2002/10/17 09:10:00 EDT
1780    Tag: Release_110_42
1781    Description:
1782    
1783    In good old tradition, there has been a slight hiccup so that we have
1784    to patch 110.42 after the fact.   The old release tag has been replaced
1785    (see below).
1786    
1787    The change solves a problem with two competing approaches the
1788    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1789    each other's way.
1790    
1791    This change only affects the runtime system code and the installer script.
1792    (No new bootfiles.)
1793    
1794    ----------------------------------------------------------------------
1795    Name: Matthias Blume
1796    Date: 2002/10/16 12:00:00 EDT
1797    Tag: Release_110_42_removed
1798    Description:
1799    
1800    New working release.  New bootfiles.
1801    
1802    ----------------------------------------------------------------------
1803    Name: Matthias Blume
1804    Date: 2002/10/10 13:10:00 EDT
1805    Tag: blume-20021010-ppc-divs
1806    Description:
1807    
1808    The mltree operator DIVS must be implemented with an overflow check on
1809    the PPC because the hardware indicates divide-by-zero using "overflow" as
1810    well.
1811    
1812    ----------------------------------------------------------------------
1813    Name: Matthias Blume
1814    Date: 2002/07/23 11:45:00
1815    Tag: blume-20020723-smlnj-home
1816    Description:
1817    
1818    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1819    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1820    CM_PATHCONFIG is also set), the path configuration file is assumed
1821    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1822    the entire tree to some other place and everything will "just work".
1823    
1824    (Companion commands such as ml-build and ml-makedepend also sense this
1825    variable.)
1826    
1827    ----------------------------------------------------------------------
1828    Name: Matthias Blume
1829    Date: 2002/07/12 21:19:00 EDT
1830    Tag: blume-20020712-liveness
1831    Description:
1832    
1833    Exported two useful "step" functions from liveness module (MLRISC).
1834    
1835    ----------------------------------------------------------------------
1836    Name: Matthias Blume
1837    Date: 2002/07/05 16:00 EDT
1838    Tag: Release_110_41
1839    Description:
1840    
1841    New working release.  New bootfiles.
1842    
1843    ----------------------------------------------------------------------
1844    Name: Matthias Blume
1845    Date: 2002/07/05 10:25:00 EDT
1846    Tag: blume-20020705-btimp
1847    Description:
1848    
1849    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1850    other clients can set up backtracing support.
1851    
1852    ----------------------------------------------------------------------
1853    Name: Matthias Blume
1854    Date: 2002/06/25 17:23:00 EDT
1855    Tag: blume-20020625-fpmax
1856    Description:
1857    
1858    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1859    case.  (The sense of the isNaN test was reversed -- which made min and
1860    max always return their first argument.)
1861    
1862    ----------------------------------------------------------------------
1863    Name: Matthias Blume
1864    Date: 2002/06/11
1865    Tag: blume-20020611-unixpath
1866    Description:
1867    
1868    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1869    
1870    ----------------------------------------------------------------------
1871    Name: Matthias Blume
1872    Date: 2002/06/10 16:35:00 EDT
1873    Tag: blume-20020610-ieeereal
1874    Description:
1875    
1876    I back-ported my implementation of IEEEReal.fromString from the
1877    idlbasis-devel branch so that we can test it.
1878    
1879    Another small change is that ppDec tries to give more information
1880    than just "<sig>" in the case of functors.  However, this code is
1881    broken in some mysterious way if the functor's body's signature
1882    has not been declared by ascription but gets inferred from the
1883    implementation.  This needs fixing...
1884    
1885    ----------------------------------------------------------------------
1886    Name: Matthias Blume
1887    Date: 2002/05/31
1888    Tag: blume-20020531-btrace-mode
1889    Description:
1890    
1891    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1892    the wayside when I switched over to using Controls everywhere.)
1893    
1894    ----------------------------------------------------------------------
1895    Name: Lal George
1896    Date: 2002/05/23 12:21:40 EDT
1897    Tag: george-20020523-visual-labels
1898    Description:
1899    
1900    Labels are now displayed in the graphical output to make
1901    the fall-through and target blocks obvious.
1902    
1903    ----------------------------------------------------------------------
1904    Name: Matthias Blume
1905    Date: 2002/05/22 11:03:00 EDT
1906    Tag: blume-20020522-shrink
1907    Description:
1908    
1909    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1910    situation that comes up when allocating huge arrays.
1911    
1912    ----------------------------------------------------------------------
1913    Name: Matthias Blume
1914    Date: 2002/05/21 16:00:00 EDT
1915    Tag: Release_110_40
1916    Description:
1917    
1918    New working release (110.40).  New bootfiles.
1919    
1920    [Also: John Reppy fixed GC bug 1131.]
1921    
1922    ----------------------------------------------------------------------
1923    Name: Matthias Blume
1924    Date: 2002/05/21 12:35:00 EDT
1925    Tag: blume-20020521-cmdoc
1926    Description:
1927    
1928    CM documentation update.
1929    
1930    ----------------------------------------------------------------------
1931    Name: Matthias Blume
1932    Date: 2002/05/21 10:55:00 EDT
1933    Tag: blume-20020521-misc
1934    Description:
1935    
1936    - John tweaked runtime to be silent on heap export (except when
1937      GC messages are on).
1938    - I added a few more things (cross-compiling versions of CMB) to
1939      config/preloads (as suggestions).
1940    
1941    ----------------------------------------------------------------------
1942    Name: Matthias Blume
1943    Date: 2002/05/20 22:25:00 EDT
1944    Tag: blume-20020520-controls
1945    Description:
1946    
1947    - Added ControlUtil structure to control-lib.cm.
1948    - Use it throughout.
1949    - Used Controls facility to define MLRISC controls (as opposed to
1950      registering MLRISC control ref cells with Controls after the fact)
1951    - Fixed messed-up controls priorities.
1952    
1953    * Removed again all the stuff from config/preloads that one wouldn't
1954      be able to preload at the time the initial heap image is built.
1955      (Many libraries, e.g., CML, do not exist yet at this time.  The
1956       only libraries that can be preloaded via config/preloads are those
1957       that come bundled with the bootfiles.)
1958    
1959    ----------------------------------------------------------------------
1960    Name: Matthias Blume
1961    Date: 2002/05/20 10:59:00 EDT
1962    Tag: blume-20020520-preloads
1963    Description:
1964    
1965    Added a lot of commented-out suggestions for things to be included
1966    in config/preloads.
1967    
1968    ----------------------------------------------------------------------
1969    Name: Allen Leung
1970    Date: 2002/05/18 14:20:00 EDT
1971    Tag: leunga-20020518-mdl
1972    Description:
1973    
1974    o Made the mdl tool stuff compile and run again.
1975    o I've disabled all the stuff that depends on RTL specifications; they
1976      are all badly broken anyway.
1977    
1978    ----------------------------------------------------------------------
1979    Name: Matthias Blume
1980    Date: 2002/05/17 16:49:00 EDT
1981    Tag: blume-20020517-controls
1982    Description:
1983    
1984    0. John Reppy made several modifications to the SML/NJ library.
1985       In particular, there is a shiny new controls-lib.cm.
1986    
1987    1. Pushed new controls interface through compiler so that everything
1988       compiles again.
1989    
1990    2. Added FormatComb and FORMAT_COMB to the CML version of the
1991       SML/NJ library (so that CML compiles again).
1992    
1993    3. Modified init scripts because XXX_DEFAULT environment variables
1994       are no longer with us.  (Boot-time initialization is now done
1995       using the same environment variables that are also used for
1996       startup-time initialization of controls.)
1997    
1998    ----------------------------------------------------------------------
1999    Name: Lal George
2000    Date: 2002/05/15 09:20:10 EDT
2001    Tag: george-20020515-pseudo-op-decls
2002    Description:
2003    
2004            All pseudo-ops emitted before the first segment declaration
2005    such as TEXT, DATA, and BSS directives are assumed to be global
2006    declarations and are emitted first in the assembly file. This is
2007    useful in a number of situations where one has pseudo-ops that are not
2008    specific to any segment, and also works around the constraint that one
2009    cannot have client pseudo-ops in the TEXT segment.
2010    
2011            Because no segment is associated with these declarations it is
2012    an error to allocate any space or objects before the first segment
2013    directive and an exception will be raised. However, we cannot make
2014    this check for client pseudo-ops.
2015    
2016            These top level declarations are a field in the CFG graph_info.
2017    In theory you can continue to add to this field after the CFG has been
2018    built -- provided you know what you are doing;-)
2019    
2020    ----------------------------------------------------------------------
2021    Name: Matthias Blume
2022    Date: 2002/05/13 16:40:00 EDT
2023    Tag: blume-20020513-pp-etc
2024    Description:
2025    
2026    A few minor bugfixes:
2027    
2028      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2029        (Bogus printouts for redefined bindings still occur.  Compiler
2030        bug should no longer occur now.  We need to redo the prettyprinter
2031        from scratch.)
2032    
2033      - CM pathname printer now also adds escape sequences for ( and )
2034    
2035      - commend and docu fixes for ml-nlffi
2036    
2037    ----------------------------------------------------------------------
2038    Name: Matthias Blume
2039    Date: 2002/05/10 16:40:00 EDT
2040    Tag: blume-20020510-erg-textio
2041    Description:
2042    
2043    Applied the following bugfix provided by Emden Gansner:
2044    
2045        Output is corrupted when outputSubstr is used rather than output.
2046    
2047        The problem occurs when a substring
2048    
2049            ss = (s, dataStart, dataLen)
2050    
2051        where dataStart > 0, fills a stream buffer with avail bytes left.
2052        avail bytes of s, starting at index dataStart, are copied into the
2053        buffer, the buffer is flushed, and then the remaining dataLen-avail
2054        bytes of ss are copied into the beginning of the buffer. Instead of
2055        starting this copy at index dataStart+avail in s, the current code
2056        starts the copy at index avail.
2057    
2058        Fix:
2059        In text-io-fn.sml, change line 695 from
2060             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2061        to
2062             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2063    
2064    ----------------------------------------------------------------------
2065    Name: Matthias Blume
2066    Date: 2002/04/12 13:55:00 EDT
2067    Tag: blume-20020412-assyntax
2068    Description:
2069    
2070    1. Grabbed newer assyntax.h from the XFree86 project.
2071    2. Fiddled with how to compile X86.prim.asm without warnings.
2072    3. (Very) Minor cleanup in CM.
2073    
2074    ----------------------------------------------------------------------
2075    Name: Matthias Blume
2076    Date: 2002/04/01 (no joke!) 17:07:00 EST
2077    Tag: blume-20020401-x86div
2078    Description:
2079    
2080    Added full support for div/mod/rem/quot on the x86, using the machine
2081    instruction's two results (without clumsily recomputing the remainder)
2082    directly where appropriate.
2083    
2084    Some more extensive power-of-two support was added to the x86 instruction
2085    selector (avoiding expensive divs, mods, and muls where they can be
2086    replaced with cheaper shifts and masks).  However, this sort of thing
2087    ought to be done earlier, e.g., within the CPS optimizer so that
2088    all architectures benefit from it.
2089    
2090    The compiler compiles to a fixed point, but changes might be somewhat
2091    fragile nevertheless.  Please, report any strange things that you might
2092    see wrt. div/mod/quot/rem...
2093    
2094    ----------------------------------------------------------------------
2095    Name: Matthias Blume
2096    Date: 2002/03/29 17:22:00
2097    Tag: blume-20020329-div
2098    Description:
2099    
2100    Fixed my broken div/mod logic.  Unfortunately, this means that the
2101    inline code for div/mod now has one more comparison than before.
2102    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2103    The problem was with quotient = 0, because that alone does not tell
2104    us which way the rounding went.  One then has to look at whether
2105    remainder and divisor have the same sign...  :(
2106    
2107    Anyway, I replaced the bootfiles with fresh ones...
2108    
2109    ----------------------------------------------------------------------
2110    Name: Matthias Blume
2111    Date: 2002/03/29 14:10:00 EST
2112    Tag: blume-20020329-inlprims
2113    Description:
2114    
2115    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2116    
2117    Primops have changed. This means that the bin/boot-file formats have
2118    changed as well.
2119    
2120    To make sure that there is no confusion, I made a new version.
2121    
2122    
2123    CHANGES:
2124    
2125    * removed REMT from mltree (remainder should never overflow).
2126    
2127    * added primops to deal with divisions of all flavors to the frontend
2128    
2129    * handled these primops all the way through so they map to their respective
2130      MLRISC support
2131    
2132    * used these primops in the implementation of Int, Int32, Word, Word32
2133    
2134    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2135    
2136    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2137    
2138    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2139      floating point
2140    
2141    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2142      and Real (but Real.abs maps to a separate floating-point-only primop)
2143    
2144    
2145    TODO items:
2146    
2147    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2148      instructions because the machine instruction encoder cannot handle
2149      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2150      REM in the first place, and REM is supported by the x86 machine coder.)
2151    
2152    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2153      selection phase.  (The two can be streamlined because the hardware
2154      delivers both quotient and remainder at the same time anyway.)
2155    
2156    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2157      (Currently the behavior is inconsistent both across architectures and
2158      wrt. the draft Basis spec.)
2159    
2160    * Word8 should eventually be handled natively, too.
2161    
2162    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2163      as if there currently is no execution path that could trigger it in
2164      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2165      hold for things like multiplication and division.)
2166    
2167    ----------------------------------------------------------------------
2168    Name: Matthias Blume
2169    Date: 2002/03/27 16:27:00 EST
2170    Tag: blume-20020327-mlrisc-divisions
2171    Description:
2172    
2173    Added support for all four division operations (ML's div, mod, quot,
2174    and rem) to MLRISC.  In the course of doing so, I also rationalized
2175    the naming (no more annoying switch-around of DIV and QUOT), by
2176    parameterizing the operation by div_rounding_mode (which can be either
2177    DIV_TO_ZERO or DIV_TO_NEGINF).
2178    
2179    The generic MLTreeGen functor takes care of compiling all four
2180    operations down to only round-to-zero div.
2181    
2182    Missing pieces:
2183    
2184      * Doing something smarter than relying on MLTreeGen on architectures
2185        like, e.g., the x86 where hardware division delivers both quotient and
2186        remainder at the same time.  With this, the implementation of the
2187        round-to-neginf operations could be further streamlined.
2188    
2189      * Remove inlining support for div/mod/rem from the frontend and replace it
2190        with primops that get carried through to the backend.  Do this for all
2191        int and word types.
2192    
2193    ----------------------------------------------------------------------
2194    Name: Matthias Blume
2195    Date: 2002/03/25 17:25:00 EST
2196    Tag: blume-20020325-divmod
2197    Description:
2198    
2199    I improved (hopefully without breaking them) the implementation of Int.div,
2200    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2201    advantage of the following observations:
2202    
2203      Let  q = x quot y      r = x rem y
2204           d = x div  y      m = x mod y
2205    
2206    where "quot" is the round-to-zero version of integer division that
2207    hardware usually provides.  Then we have:
2208    
2209         r = x - q * y        where neither the * nor the - will overflow
2210         d = if q >= 0 orelse x = q * y then q else q - 1
2211                              where neither the * nor the - will overflow
2212         m = if q >= 0 orelse r = 0 then r else r + y
2213                              where the + will not overflow
2214    
2215    This results in substantial simplification of the generated code.
2216    The following table shows the number of CFG nodes and edges generated
2217    for
2218            fun f (x, y) = x OPER y
2219            (* with OPER \in div, mod, quot, rem *)
2220    
2221    
2222        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2223        --------------------------------------------------------
2224         div |         24 |         39 |         12 |         16
2225         mod |         41 |         71 |         12 |         16
2226        quot |          8 |         10 |          8 |         10
2227         rem |         10 |         14 |          8 |         10
2228    
2229    
2230    ----------------------------------------------------------------------
2231    Name: Matthias Blume
2232    Date: 2002/03/25 22:06:00 EST
2233    Tag: blume-20020325-cprotobug
2234    Description:
2235    
2236    Fixed a bug in cproto (c prototype decoder).
2237    
2238    ----------------------------------------------------------------------
2239    Name: Matthias Blume
2240    Date: 2002/03/25 16:00:00 EST
2241    Tag: blume-20020325-raw-primops
2242    Description:
2243    
2244    I did some cleanup to Allen's new primop code and
2245    replaced yesterday's bootfiles with new ones.
2246    (But they are stored in the same place.)
2247    
2248    ----------------------------------------------------------------------
2249    Name: Matthias Blume
2250    Date: 2002/03/24 22:40:00 EST
2251    Tag: blume-20020324-bootfiles
2252    Description:
2253    
2254    Made the bootfiles that Allen asked for.
2255    
2256    ----------------------------------------------------------------------
2257    Name: Allen Leung
2258    Date: 2002/03/23 15:50:00 EST
2259    Tag: leunga-20020323-flint-cps-rcc-primops
2260    Description:
2261    
2262      1. Changes to FLINT primops:
2263    
2264        (* make a call to a C-function;
2265         * The primop carries C function prototype information and specifies
2266         * which of its (ML-) arguments are floating point. C prototype
2267         * information is for use by the backend, ML information is for
2268         * use by the CPS converter. *)
2269      | RAW_CCALL of { c_proto: CTypes.c_proto,
2270                       ml_args: ccall_type list,
2271                       ml_res_opt: ccall_type option,
2272                       reentrant : bool
2273                     } option
2274       (* Allocate uninitialized storage on the heap.
2275        * The record is meant to hold short-lived C objects, i.e., they
2276        * are not ML pointers.  With the tag, the representation is
2277        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2278        *)
2279      | RAW_RECORD of {tag:bool,sz:int}
2280      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2281    
2282      2.  These CPS primops are now overloaded:
2283    
2284           rawload of {kind:numkind}
2285           rawstore of {kind:numkind}
2286    
2287          The one argument form is:
2288    
2289             rawload {kind} address
2290    
2291          The two argument form is:
2292    
2293             rawload {kind} [ml object, byte-offset]
2294    
2295      3. RAW_CCALL/RCC now takes two extra arguments:
2296    
2297         a. The first is whether the C call is reentrant, i.e., whether
2298            ML state should be saved and restored.
2299         b. The second argument is a string argument specifying the name of
2300            library and the C function.
2301    
2302         These things are currently not handled in the code generator, yet.
2303    
2304      4. In CProto,
2305    
2306         An encoding type of "bool" means "ml object" and is mapped into
2307         C prototype of PTR.  Note that "bool" is different than "string",
2308         even though "string" is also mapped into PTR, because "bool"
2309         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2310    
2311      5. Pickler/unpicker
2312    
2313         Changed to handle RAW_RECORD and newest RAW_CCALL
2314    
2315      6. MLRiscGen,
2316    
2317         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2318         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2319            in the file CodeGen/cpscompile/cps-c-calls.sml
2320    
2321      7. Added the conditional move operator
2322    
2323             condmove of branch
2324    
2325         to cps.  Generation of this is still buggy so it is currently
2326         disabled.
2327    
2328    ----------------------------------------------------------------------
2329    Name: Lal George
2330    Date: 2002/03/22 14:18:25 EST
2331    Tag: george-20020322-cps-branch-prob
2332    Description:
2333    
2334    Implemented the Ball-Larus branch prediction-heuristics, and
2335    incorporated graphical viewers for control flow graphs.
2336    
2337    Ball-Larus Heuristics:
2338    ---------------------
2339    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2340    
2341    By design it uses the Dempster-Shafer theory for combining
2342    probabilities.  For example, in the function:
2343    
2344        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2345    
2346    the ball-larus heuristics predicts that the n=0 is unlikely
2347    (OH-heuristic), and the 'then' branch is unlikely because of the
2348    RH-heuristic -- giving the 'then' branch an even lower combined
2349    probability using the Dempster-Shafer theory.
2350    
2351    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2352    probability of the 'then' branch because of the loop in the else
2353    branch.
2354    
2355    
2356    Graphical Viewing:
2357    ------------------
2358    I merely plugged in Allen's graphical viewers into the compiler. The
2359    additional code is not much. At the top level, saying:
2360    
2361            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2362    
2363    will display the graphical view of the control flow graph just before
2364    back-patching.  daVinci must be in your path for this to work. If
2365    daVinci is not available, then the default viewer can be changed
2366    using:
2367    
2368            Control.MLRISC.getString "viewer"
2369    
2370    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2371    course, these viewers must be in your path.
2372    
2373    The above will display the compilation unit at the level of clusters,
2374    many of which are small, boring, and un-interesting. Also setting:
2375    
2376            Control.MLRISC.getInt "cfg-graphical-view_size"
2377    
2378    will display clusters that are larger than the value set by the above.
2379    
2380    
2381    ----------------------------------------------------------------------
2382    Name: Matthias Blume
2383    Date: 2002/03/21 22:20:00 EST
2384    Tag: blume-20020321-kmp-bugfix
2385    Description:
2386    
2387    Changed the interface to the KMP routine in PreString and fixed
2388    a minor bug in one place where it was used.
2389    
2390    ----------------------------------------------------------------------
2391    Name: Allen Leung
2392    Date: 2002/03/21 20:30:00 EST
2393    Tag: leunga-20020321-cfg
2394    Description:
2395    
2396      Fixed a potential problem in cfg edge splitting.
2397    
2398    ----------------------------------------------------------------------
2399    Name: Allen Leung
2400    Date: 2002/03/21 17:15:00 EST
2401    Tag: leunga-20020321-x86-fp-cfg
2402    Description:
2403    
2404      1. Recoded the buggy parts of x86-fp.
2405    
2406         a. All the block reordering code has been removed.
2407            We now depend on the block placement phases to do this work.
2408    
2409         b. Critical edge splitting code has been simplified and moved into the
2410            CFG modules, as where they belong.
2411    
2412         Both of these were quite buggy and complex.  The code is now much, much
2413         simpler.
2414    
2415      2. X86 backend.
2416    
2417         a. Added instructions for 64-bit support.  Instruction selection for
2418            64-bit has not been committed, however, since that
2419            requires changes to MLTREE which haven't been approved by
2420            Lal and John.
2421    
2422         b. Added support for FUCOMI and FUCOMIP when generating code for
2423            PentiumPro and above.  We only generate these instructions in
2424            the fast-fp mode.
2425    
2426         c. Added cases for JP and JNP in X86FreqProps.
2427    
2428      3. CFG
2429    
2430         CFG now has a bunch of methods for edge splitting and merging.
2431    
2432      4. Machine description.
2433    
2434         John's simplification of MLTREE_BASIS.fcond broke a few machine
2435         description things:
2436    
2437         rtl-build.{sig,sml} and hppa.mdl fixed.
2438    
2439         NOTE: the machine description stuff in the repository is still broken.
2440               Again, I can't put my fixes in because that involves
2441               changes to MLTREE.
2442    
2443    ----------------------------------------------------------------------
2444    Name: Matthias Blume
2445    Date: 2002/03/20 15:55:00 EST
2446    Tag: blume-20020320-kmp
2447    Description:
2448    
2449    Implemented Knuth-Morris-Pratt string matching in PreString and used
2450    it for String.isSubstring, Substring.isSubstring, and
2451    Substring.position.
2452    
2453    (Might need some stress-testing.  Simple examples worked fine.)
2454    
2455    ----------------------------------------------------------------------
2456    Name: Matthias Blume
2457    Date: 2002/03/19 16:37:00 EST
2458    Tag: blume-20020319-witnesses
2459    Description:
2460    
2461    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2462    
2463    This implements a generic mechanism for changing constness qualifiers
2464    anywhere within big C types without resorting to outright "casts".
2465    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2466    modify the constness at the outermost level.)
2467    The implementation of "convert" is based on the idea of "witness"
2468    values -- values that are not used by the operation but whose types
2469    "testify" to their applicability.  On the implementation side, "convert"
2470    is simply a projection (returning its second curried argument).  With
2471    cross-module inlining, it should not result in any machine code being
2472    generated.
2473    
2474    ----------------------------------------------------------------------
2475    Name: Matthias Blume
2476    Date: 2002/03/15 16:40:00 EST
2477    Tag: blume-20020315-basis
2478    Description:
2479    
2480    Provided (preliminary?) implementations for
2481    
2482      {String,Substring}.{concatWith,isSuffix,isSubstring}
2483    
2484    and
2485    
2486      Substring.full
2487    
2488    Those are in the Basis spec but they were missing in SML/NJ.
2489    
2490    ----------------------------------------------------------------------
2491    Name: Matthias Blume
2492    Date: 2002/03/14 21:30:00 EST
2493    Tag: blume-20020314-controls
2494    Description:
2495    
2496    Controls:
2497    ---------
2498    
2499    1. Factored out the recently-added Controls : CONTROLS stuff and put
2500       it into its own library $/controls-lib.cm.  The source tree for
2501       this is under src/smlnj-lib/Controls.
2502    
2503    2. Changed the names of types and functions in this interface, so they
2504       make a bit more "sense":
2505    
2506          module -> registry
2507          'a registry -> 'a group
2508    
2509    3. The interface now deals in ref cells only.  The getter/setter interface
2510       is (mostly) gone.
2511    
2512    4. Added a function that lets one register an already-existing ref cell.
2513    
2514    5. Made the corresponding modifications to the rest of the code so that
2515       everything compiles again.
2516    
2517    6. Changed the implementation of Controls.MLRISC back to something closer
2518       to the original.  In particular, this module (and therefore MLRISC)
2519       does not depend on Controls.  There now is some link-time code in
2520       int-sys.sml that registers the MLRISC controls with the Controls
2521       module.
2522    
2523    CM:
2524    ---
2525    
2526      * One can now specify the lambda-split aggressiveness in init.cmi.
2527    
2528    ----------------------------------------------------------------------
2529    Name: Allen Leung
2530    Date: 2002/03/13 17:30:00 EST
2531    Tag: leunga-20020313-x86-fp-unary
2532    Description:
2533    
2534    Bug fix for:
2535    
2536    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2537    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2538    > - fun f(x,(y,z)) = Real.~ y;
2539    > [autoloading]
2540    > [autoloading done]
2541    >       fchsl   (%eax), 184(%esp)
2542    > Error: MLRisc bug: X86MCEmitter.emitInstr
2543    >
2544    > uncaught exception Error
2545    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2546    
2547    The problem was that the code generator did not generate any fp registers
2548    in this case, and the ra didn't know that it needed to run the X86FP phase to
2549    translate the pseudo fp instruction.   This only happened with unary fp
2550    operators in certain situations.
2551    
2552    ----------------------------------------------------------------------
2553    Name: Matthias Blume
2554    Date: 2002/03/13 14:00:00 EST
2555    Tag: blume-20020313-overload-etc
2556    Description:
2557    
2558    1. Added _overload as a synonym for overload for backward compatibility.
2559       (Control.overloadKW must be true for either version to be accepted.)
2560    
2561    2. Fixed bug in install script that caused more things to be installed
2562       than what was requested in config/targets.
2563    
2564    3. Made CM aware of the (_)overload construct so that autoloading
2565       works.
2566    
2567    ----------------------------------------------------------------------
2568    Name: Matthias Blume
2569    Date: 2002/03/12 22:03:00 EST
2570    Tag: blume-20020312-url
2571    Description:
2572    
2573    Forgot to update BOOT and srcarchiveurl.
2574    
2575    ----------------------------------------------------------------------
2576    Name: Matthias Blume
2577    Date: 2002/03/12 17:30:00 EST
2578    Tag: blume-20020312-version110392
2579    Description:
2580    
2581    Yet another version number bump (because of small changes to the
2582    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2583    
2584    Changes:
2585    
2586      The new pid generation scheme described a few weeks ago was overly
2587      complicated.  I implemented a new mechanism that is simpler and
2588      provides a bit more "stability":  Once CM has seen a compilation
2589      unit, it keeps its identity constant (as long as you do not delete
2590      those crucial CM/GUID/* files).  This means that when you change
2591      an interface, compile, then go back to the old interface, and
2592      compile again, you arrive at the original pid.
2593    
2594      There now also is a mechanism that instructs CM to use the plain
2595      environment hash as a module's pid (effectively making its GUID
2596      the empty string).  For this, "noguid" must be specified as an
2597      option to the .sml file in question within its .cm file.
2598      This is most useful for code that is being generated by tools such
2599      as ml-nlffigen (because during development programmers tend to
2600      erase the tool's entire output directory tree including CM's cached
2601      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2602      revert to the old, broken behavior of SML/NJ, but in specific cases
2603      where there is no danger of interface confusion, its use is ok
2604      (I think).
2605    
2606      ml-nlffigen by default generates "noguid" annotations.  They can be
2607      turned off by specifying -guid in its command line.
2608    
2609    ----------------------------------------------------------------------
2610    Name: Lal George
2611    Date: 2002/03/12 12 14:42:36 EST
2612    Tag: george-20020312-frequency-computation
2613    Description:
2614    
2615    Integrated jump chaining and static block frequency into the
2616    compiler. More details and numbers later.
2617    
2618    ----------------------------------------------------------------------
2619    Name: Lal George
2620    Date: 2002/03/11 11 22:38:53 EST
2621    Tag: george-20020311-jump-chain-elim
2622    Description:
2623    
2624    Tested the jump chain elimination on all architectures (except the
2625    hppa).  This is on by default right now and is profitable for the
2626    alpha and x86, however, it may not be profitable for the sparc and ppc
2627    when compiling the compiler.
2628    
2629    The gc test will typically jump to a label at the end of the cluster,
2630    where there is another jump to an external cluster containing the actual
2631    code to invoke gc. This is to allow factoring of common gc invocation
2632    sequences. That is to say, we generate:
2633    
2634            f:
2635               testgc
2636               ja   L1      % jump if above to L1
2637    
2638            L1:
2639               jmp L2
2640    
2641    
2642    After jump chain elimination the 'ja L1' instructions is converted to
2643    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2644    up being implemented in their long form (if L2 is far away) using:
2645    
2646            jbe     L3      % jump if below or equal to L3
2647            jmp     L2
2648         L3:
2649            ...
2650    
2651    
2652    For large compilation units L2  may be far away.
2653    
2654    
2655    ----------------------------------------------------------------------
2656    Name: Matthias Blume
2657    Date: 2002/03/11 13:30:00 EST
2658    Tag: blume-20020311-mltreeeval
2659    Description:
2660    
2661    A functor parameter was missing.
2662    
2663    ----------------------------------------------------------------------
2664    Name: Allen Leung
2665    Date: 2002/03/11 10:30:00 EST
2666    Tag: leunga-20020311-runtime-string0
2667    Description:
2668    
2669       The representation of the empty string now points to a
2670    legal null terminated C string instead of unit.  It is now possible
2671    to convert an ML string into C string with InlineT.CharVector.getData.
2672    This compiles into one single machine instruction.
2673    
2674    ----------------------------------------------------------------------
2675    Name: Allen Leung
2676    Date: 2002/03/10 23:55:00 EST
2677    Tag: leunga-20020310-x86-call
2678    Description:
2679    
2680       Added machine generation for CALL instruction (relative displacement mode)
2681    
2682    ----------------------------------------------------------------------
2683    Name: Matthias Blume
2684    Date: 2002/03/08 16:05:00
2685    Tag: blume-20020308-entrypoints
2686    Description:
2687    
2688    Version number bumped to 110.39.1.  NEW BOOTFILES!
2689    
2690    Entrypoints: non-zero offset into a code object where execution should begin.
2691    
2692    - Added the notion of an entrypoint to CodeObj.
2693    - Added reading/writing of entrypoint info to Binfile.
2694    - Made runtime system bootloader aware of entrypoints.
2695    - Use the address of the label of the first function given to mlriscGen
2696      as the entrypoint.  This address is currently always 0, but it will
2697      not be 0 once we turn on block placement.
2698    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2699      with entry points) from mlriscGen.
2700    
2701    ----------------------------------------------------------------------
2702    Name: Allen Leung
2703    Date: 2002/03/07 20:45:00 EST
2704    Tag: leunga-20020307-x86-cmov
2705    Description:
2706    
2707       Bug fixes for CMOVcc on x86.
2708    
2709       1. Added machine code generation for CMOVcc
2710       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2711       3. CMOVcc cannot have an immediate operand as argument.
2712    
2713    ----------------------------------------------------------------------
2714    Name: Matthias Blume
2715    Date: 2002/03/07 16:15:00 EST
2716    Tag: blume-20020307-controls
2717    Description:
2718    
2719    This is a very large but mostly boring patch which makes (almost)
2720    every tuneable compiler knob (i.e., pretty much everything under
2721    Control.* plus a few other things) configurable via both the command
2722    line and environment variables in the style CM did its configuration
2723    until now.
2724    
2725    Try starting sml with '-h' (or, if you are brave, '-H')
2726    
2727    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2728    implements the underlying generic mechanism.
2729    
2730    The interface to some of the existing such facilities has changed somewhat.
2731    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2732    (The getFoo interface is still there for backward-compatibility, but its
2733    use is deprecated.)
2734    
2735    The ml-build script passes -Cxxx=yyy command-line arguments through so
2736    that one can now twiddle the compiler settings when using this "batch"
2737    compiler.
2738    
2739    TODO items:
2740    
2741    We should go through and throw out all controls that are no longer
2742    connected to anything.  Moreover, we should go through and provide
2743    meaningful (and correct!) documentation strings for those controls
2744    that still are connected.
2745    
2746    Currently, multiple calls to Controls.new are accepted (only the first
2747    has any effect).  Eventually we should make sure that every control
2748    is being made (via Controls.new) exactly once.  Future access can then
2749    be done using Controls.acc.
2750    
2751    Finally, it would probably be a good idea to use the getter-setter
2752    interface to controls rather than ref cells.  For the time being, both
2753    styles are provided by the Controls module, but getter-setter pairs are
2754    better if thread-safety is of any concern because they can be wrapped.
2755    
2756    *****************************************
2757    
2758    One bug fix: The function blockPlacement in three of the MLRISC
2759    backpatch files used to be hard-wired to one of two possibilities at
2760    link time (according to the value of the placementFlag).  But (I
2761    think) it should rather sense the flag every time.
2762    
2763    *****************************************
2764    
2765    Other assorted changes (by other people who did not supply a HISTORY entry):
2766    
2767    1. the cross-module inliner now works much better (Monnier)
2768    2. representation of weights, frequencies, and probabilities in MLRISC
2769       changed in preparation of using those for weighted block placement
2770       (Reppy, George)
2771    
2772    ----------------------------------------------------------------------
2773    Name: Lal George
2774    Date: 2002/03/07 14:44:24 EST 2002
2775    Tag: george-20020307-weighted-block-placement
2776    
2777    Tested the weighted block placement optimization on all architectures
2778    (except the hppa) using AMPL to generate the block and edge frequencies.
2779    Changes were required in the machine properties to correctly
2780    categorize trap instructions. There is an MLRISC flag
2781    "weighted-block-placement" that can be used to enable weighted block
2782    placement, but this will be ineffective without block/edge
2783    frequencies (coming soon).
2784    
2785    
2786    ----------------------------------------------------------------------
2787    Name: Lal George
2788    Date: 2002/03/05 17:24:48 EST
2789    Tag: george-20020305-linkage-cluster
2790    
2791    In order to support the block placement optimization, a new cluster
2792    is generated as the very first cluster (called the linkage cluster).
2793    It contains a single jump to the 'real' entry point for the compilation
2794    unit. Block placement has no effect on the linkage cluster itself, but
2795    all the other clusters  have full freedom in the manner in which they
2796    reorder blocks or functions.
2797    
2798    On the x86 the typical linkage code that is generated is:
2799       ----------------------
2800            .align 2
2801       L0:
2802            addl    $L1-L0, 72(%esp)
2803            jmp     L1
2804    
2805    
2806            .align  2
2807       L1:
2808       ----------------------
2809    
2810    72(%esp) is the memory location for the stdlink register. This
2811    must contain the address of the CPS function being called. In the
2812    above example, it contains the address of  L0; before
2813    calling L1 (the real entry point for the compilation unit), it
2814    must contain the address for L1, and hence
2815    
2816            addl $L1-L0, 72(%esp)
2817    
2818    I have tested this on all architectures except the hppa.The increase
2819    in code size is of course negligible
2820    
2821    ----------------------------------------------------------------------
2822    Name: Allen Leung
2823    Date: 2002/03/03 13:20:00 EST
2824    Tag: leunga-20020303-mlrisc-tools
2825    
2826      Added #[ ... ] expressions to mlrisc tools
2827    
2828    ----------------------------------------------------------------------
2829    Name: Matthias Blume
2830    Date: 2002/02/27 12:29:00 EST
2831    Tag: blume-20020227-cdebug
2832    Description:
2833    
2834    - made types in structure C and C_Debug to be equal
2835    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2836    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2837    
2838    ----------------------------------------------------------------------
2839    Name: Matthias Blume
2840    Date: 2002/02/26 12:00:00 EST
2841    Tag: blume-20020226-ffi
2842    Description:
2843    
2844    1. Fixed a minor bug in CM's "noweb" tool:
2845       If numbering is turned off, then truly don't number (i.e., do not
2846       supply the -L option to noweb).  The previous behavior was to supply
2847       -L'' -- which caused noweb to use the "default" line numbering scheme.
2848       Thanks to Chris Richards for pointing this out (and supplying the fix).
2849    
2850    2. Once again, I reworked some aspects of the FFI:
2851    
2852       A. The incomplete/complete type business:
2853    
2854       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2855         gone!
2856       - ML types representing an incomplete type are now *equal* to
2857         ML types representing their corresponding complete types (just like
2858         in C).  This is still safe because ml-nlffigen will not generate
2859         RTTI for incomplete types, nor will it generate functions that
2860         require access to such RTTI.   But when ML code generated from both
2861         incomplete and complete versions of the C type meet, the ML types
2862         are trivially interoperable.
2863    
2864         NOTE:  These changes restore the full generality of the translation
2865         (which was previously lost when I eliminated functorization)!
2866    
2867       B. Enum types:
2868    
2869       - Structure C now has a type constructor "enum" that is similar to
2870         how the "su" constructor works.  However, "enum" is not a phantom
2871         type because each "T enum" has values (and is isomorphic to
2872         MLRep.Signed.int).
2873       - There are generic access operations for enum objects (using
2874         MLRep.Signed.int).
2875       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2876         * The structure contains the definition of type "mlrep" (the ML-side
2877         representation type of the enum).  Normally, mlrep is the same
2878         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2879         then mlrep will be defined as a datatype -- thus facilitating
2880         pattern matching on mlrep values.
2881         ("-ec" will be suppressed if there are duplicate values in an
2882          enumeration.)
2883         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2884         will be generated for each C enum constant xxx.
2885         * Conversion functions m2i and i2m convert between mlrep and
2886         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2887         * Coversion functions c and ml convert between mlrep and "tag enum".
2888         * Access functions (get/set) fetch and store mlrep values.
2889       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2890         enumerations are merged into one single enumeration represented by
2891         structure E_'.
2892    
2893    ----------------------------------------------------------------------
2894    Name: Allen Leung
2895    Date: 2002/02/25 04:45:00 EST
2896    Tag: leunga-20020225-cps-spill
2897    
2898    This is a new implementation of the CPS spill phase.
2899    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2900    In case of problems, replace it with the old file spill.sml
2901    
2902    The current compiler runs into some serious performance problems when
2903    constructing a large record.  This can happen when we try to compile a
2904    structure with many items.  Even a very simple structure like the following
2905    makes the compiler slow down.
2906    
2907        structure Foo = struct
2908           val x_1 = 0w1 : Word32.int
2909           val x_2 = 0w2 : Word32.int
2910           val x_3 = 0w3 : Word32.int
2911           ...
2912           val x_N = 0wN : Word32.int
2913        end
2914    
2915    The following table shows the compile time, from N=1000 to N=4000,
2916    with the old compiler:
2917    
2918    N
2919    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2920           MLRISC ra                               0.06u  0.00s  0.05g
2921              (spills = 0 reloads = 0)
2922           TOTAL                                   0.63u  0.07s  0.21g
2923    
2924    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2925           MLRISC ra                               5.68u  0.59s  3.93g
2926              (spills = 0 reloads = 0)
2927           TOTAL                                   14.71u  0.99s  4.81g
2928    
2929    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2930           MLRISC ra                               5.54u  0.65s  3.91g
2931              (spills = 543 reloads = 1082)
2932           TOTAL                                   65.40u  3.13s  6.00g
2933    
2934    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2935           MLRISC ra                               0.80u  0.10s  0.55g
2936              (spills = 42 reloads = 84)
2937           TOTAL                                   129.42u  5.10s  4.13g
2938    
2939    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2940           MLRISC ra                               2.69u  0.27s  1.38g
2941              (spills = 62 reloads = 124)
2942           TOTAL                                   682.48u  19.61s  13.99g
2943    
2944    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2945           MLRISC ra                               4.96u  0.27s  2.72g
2946              (spills = 85 reloads = 170)
2947           TOTAL                                   2375.26u  57.21s  48.00g
2948    
2949    As you can see the old cps spill module suffers from some serious
2950    performance problem.  But since I cannot decipher the old code fully,
2951    instead of patching the problems up, I'm reimplementing it
2952    with a different algorithm.  The new code is more modular,
2953    smaller when compiled, and substantially faster
2954    (O(n log n) time and O(n) space).  Timing of the new spill module:
2955    
2956    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2957          MLRISC ra                               0.25u  0.02s  0.15g
2958             (spills=1 reloads=3)
2959          TOTAL                                   7.74u  0.34s  1.62g
2960    
2961    Implementation details:
2962    
2963    As far as I can tell, the purpose of the CPS spill module is to make sure the
2964    number of live variables at any program point (the bandwidth)
2965    does not exceed a certain limit, which is determined by the
2966    size of the spill area.
2967    
2968    When the bandwidth is too large, we decrease the register pressure by
2969    packing live variables into spill records.  How we achieve this is
2970    completely different than what we did in the old code.
2971    
2972    First, there is something about the MLRiscGen code generator
2973    that we should be aware of:
2974    
2975    o MLRiscGen performs code motion!
2976    
2977       In particular, it will move floating point computations and
2978       address computations involving only the heap pointer to
2979       their use sites (if there is only a single use).
2980       What this means is that if we have a CPS record construction
2981       statement
2982    
2983           RECORD(k,vl,w,e)
2984    
2985       we should never count the new record address w as live if w
2986       has only one use (which is often the case).
2987    
2988       We should do something similar to floating point, but the transformation
2989       there is much more complex, so I won't deal with that.
2990    
2991    Secondly, there are now two new cps primops at our disposal:
2992    
2993     1. rawrecord of record_kind option
2994        This pure operator allocates some uninitialized storage from the heap.
2995        There are two forms:
2996    
2997         rawrecord NONE [INT n]  allocates a tagless record of length n
2998         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2999                                     and initializes the tag.
3000    
3001     2. rawupdate of cty
3002          rawupdate cty (v,i,x)
3003          Assigns to x to the ith component of record v.
3004          The storelist is not updated.
3005    
3006    We use these new primops for both spilling and increment record construction.
3007    
3008     1. Spilling.
3009    
3010        This is implemented with a linear scan algorithm (but generalized
3011        to trees).  The algorithm will create a single spill record at the
3012        beginning of the cps function and use rawupdate to spill to it,
3013        and SELECT or SELp to reload from it.  So both spills and reloads
3014        are fine-grain operations.  In contrast, in the old algorithm
3015        "spills" have to be bundled together in records.
3016    
3017        Ideally, we should sink the spill record construction to where
3018        it is needed.  We can even split the spill record into multiple ones
3019        at the places where they are needed.  But CPS is not a good
3020        representation for global code motion, so I'll keep it simple and
3021        am not attempting this.
3022    
3023     2. Incremental record construction (aka record splitting).
3024    
3025        Long records with many component values which are simulatenously live
3026        (recall that single use record addresses are not considered to
3027         be live) are constructed with rawrecord and rawupdate.
3028        We allocate space on the heap with rawrecord first, then gradually
3029        fill it in with rawupdate.  This is the technique suggested to me
3030        by Matthias.
3031    
3032        Some restrictions on when this is applicable:
3033        1. It is not a VECTOR record.  The code generator currently does not handle
3034           this case. VECTOR record uses double indirection like arrays.
3035        2. All the record component values are defined in the same "basic block"
3036           as the record constructor.  This is to prevent speculative
3037           record construction.
3038    
3039    ----------------------------------------------------------------------
3040    Name: Allen Leung
3041    Date: 2002/02/22 01:02:00 EST
3042    Tag: leunga-20020222-mlrisc-tools
3043    
3044    Minor bug fixes in the parser and rewriter
3045    
3046    ----------------------------------------------------------------------
3047    Name: Allen Leung
3048    Date: 2002/02/21 20:20:00 EST
3049    Tag: leunga-20020221-peephole
3050    
3051    Regenerated the peephole files.  Some contained typos in the specification
3052    and some didn't compile because of pretty printing bugs in the old version
3053    of 'nowhere'.
3054    
3055    ----------------------------------------------------------------------
3056    Name: Allen Leung
3057    Date: 2002/02/19 20:20:00 EST
3058    Tag: leunga-20020219-mlrisc-tools
3059    Description:
3060    
3061       Minor bug fixes to the mlrisc-tools library:
3062    
3063       1.  Fixed up parsing colon suffixed keywords
3064       2.  Added the ability to shut the error messages up
3065       3.  Reimplemented the pretty printer and fixed up/improved
3066           the pretty printing of handle and -> types.
3067       4.  Fixed up generation of literal symbols in the nowhere tool.
3068       5.  Added some SML keywords to to sml.sty
3069    
3070    ----------------------------------------------------------------------
3071    Name: Matthias Blume
3072    Date: 2002/02/19 16:20:00 EST
3073    Tag: blume-20020219-cmffi
3074    Description:
3075    
3076    A wild mix of changes, some minor, some major:
3077    
3078    * All C FFI-related libraries are now anchored under $c:
3079        $/c.cm      --> $c/c.cm
3080        $/c-int.cm  --> $c/internals/c-int.cm
3081        $/memory.cm --> $c/memory/memory.cm
3082    
3083    * "make" tool (in CM) now treats its argument pathname slightly
3084      differently:
3085        1. If the native expansion is an absolute name, then before invoking
3086           the "make" command on it, CM will apply OS.Path.mkRelative
3087           (with relativeTo = OS.FileSys.getDir()) to it.
3088        2. The argument will be passed through to subsequent phases of CM
3089           processing without "going native".  In particular, if the argument
3090           was an anchored path, then "make" will not lose track of that anchor.
3091    
3092    * Compiler backends now "know" their respective C calling conventions
3093      instead of having to be told about it by ml-nlffigen.  This relieves
3094      ml-nlffigen from one of its burdens.
3095    
3096    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3097    
3098    * Export C_DEBUG and C_Debug from $c/c.cm.
3099    
3100    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3101      subtyping relationship between incomplete pointers and their complete
3102      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3103      with the convention of instantiating 'o with ('t, 'c) obj whenever
3104      the pointer target type is complete.  In the incomplete case, 'o
3105      will be instantiated with some "'c iobj" -- a type obtained by
3106      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3107    
3108      Operations that work on both incomplete and complete pointer types are
3109      typed as taking an 'o ptr while operations that require the target to
3110      be known are typed as taking some ('t, 'c) obj ptr.
3111    
3112      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3113      where void is an eqtype without any values.  This makes it possible
3114      to work on voidptr values using functions meant to operate on light
3115      incomplete pointers.
3116    
3117    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3118      been vastly simplified.
3119    
3120    ----------------------------------------------------------------------
3121    Name: Matthias Blume
3122    Date: 2002/02/19 10:48:00 EST
3123    Tag: blume-20020219-pqfix
3124    Description:
3125    
3126    Applied Chris Okasaki's bug fix for priority queues.
3127    
3128    ----------------------------------------------------------------------
3129    Name: Matthias Blume
3130    Date: 2002/02/15 17:05:00
3131    Tag: Release_110_39
3132    Description:
3133    
3134    Last-minute retagging is becoming a tradition... :-(
3135    
3136    This is the working release 110.39.
3137    
3138    ----------------------------------------------------------------------
3139    Name: Matthias Blume
3140    Date: 2002/02/15 16:00:00 EST
3141    Tag: Release_110_39-orig
3142    Description:
3143    
3144    Working release 110.39.  New bootfiles.
3145    
3146    (Update: There was a small bug in the installer so it wouldn't work
3147    with all shells.  So I retagged. -Matthias)
3148    
3149    ----------------------------------------------------------------------
3150    Name: Matthias Blume
3151    Date: 2002/02/15 14:17:00 EST
3152    Tag: blume-20020215-showbindings
3153    Description:
3154    
3155    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3156    the latter can be useful for exploring what bindings are available at
3157    the interactive prompt.  (The first function returns only the list
3158    of symbols that are really bound, the second prints those but also the
3159    ones that CM's autoloading mechanism knows about.)
3160    
3161    ----------------------------------------------------------------------
3162    Name: Matthias Blume
3163    Date: 2002/02/15 12:08:00 EST
3164    Tag: blume-20020215-iptrs
3165    Description:
3166    
3167    Two improvements to ml-nlffigen:
3168    
3169      1. Write files only if they do not exist or if their current contents
3170         do not coincide with what's being written.  (That is, avoid messing
3171         with the time stamps unless absolutely necessary.)
3172    
3173      2. Implement a "repository" mechanism for generated files related
3174         to "incomplete pointer types".   See the README file for details.
3175    
3176    ----------------------------------------------------------------------
3177    Name: Matthias Blume
3178    Date: 2002/02/14 11:50:00 EST
3179    Tag: blume-20020214-quote
3180    Description:
3181    
3182    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3183    because of the new and improved tag generation scheme.  (Thanks to Allen
3184    Leung for pointing it out.)
3185    
3186    ----------------------------------------------------------------------
3187    Name: Lal George
3188    Date: 2002/02/14 09:55:27 EST 2002
3189    Tag: george-20020214-isabelle-bug
3190    Description:
3191    
3192    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3193    of Isabelle on the x86.
3194    
3195    From Allen:
3196    -----------
3197     I've found the problem:
3198    
3199         in ra-core.sml, I use the counter "blocked" to keep track of the
3200         true number of elements in the freeze queue.  When the counter goes
3201         to zero, I skip examining the queue.  But I've messed up the
3202         bookkeeping in combine():
3203    
3204             else ();
3205             case !ucol of
3206               PSEUDO => (if !cntv > 0 then
3207                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3208                                        ^^^^^^^^^^^^^^^^^^^^^^^
3209                      moveu := mergeMoveList(!movev, !moveu)
3210                     )
3211                  else ();
3212    
3213         combine() is called to coalesce two nodes u and v.
3214         I think I was thinking that if the move counts of u and v are both
3215         greater than zero then after they are coalesced then one node is
3216         removed from the freeze queue.  Apparently I was thinking that
3217         both u and v are of low degree, but that's clearly not necessarily true.
3218    
3219    
3220    02/12/2002:
3221        Here's the patch.  HOL now compiles.
3222    
3223        I don't know how this impact on performance (compile
3224        time or runtime).  This bug caused the RA (especially on the x86)
3225        to go thru the potential spill phase when there are still nodes on the
3226        freeze queue.
3227    
3228    
3229    
3230    
3231    ----------------------------------------------------------------------
3232    Name: Matthias Blume
3233    Date: 2002/02/13 22:40:00 EST
3234    Tag: blume-20020213-fptr-rtti
3235    Description:
3236    
3237    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3238    updates.
3239    
3240    ----------------------------------------------------------------------
3241    Name: Matthias Blume
3242    Date: 2002/02/13 16:41:00 EST
3243    Tag: blume-20020213-cmlpq
3244    Description:
3245    
3246    Added new priority queue export symbols (which have just been added to
3247    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3248    not compile and the installer would choke.)
3249    
3250    ----------------------------------------------------------------------
3251    Name: Matthias Blume
3252    Date: 2002/02/13 16:15:00 EST
3253    Tag: blume-20020213-various
3254    Description:
3255    
3256    1. More tweaks to ml-nlffigen:
3257    
3258       - better internal datastructures (resulting in slight speedup)
3259       - "-match" option requires exact match
3260       - "localized" gensym counters (untagged structs/unions nested within
3261         other structs/unions or within typedefs get a fresh counter; their
3262         tag will be prefixed by a concatenation of their parents' tags)
3263       - bug fixes (related to calculation of transitive closure of types
3264         to be included in the output)
3265    
3266    2. Minor Basis updates:
3267    
3268       - added implementations for List.collate and Option.app
3269    
3270    ----------------------------------------------------------------------
3271    Name: Matthias Blume
3272    Date: 2002/02/11 15:55:00 EST
3273    Tag: blume-20020211-gensym
3274    Description:
3275    
3276    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3277    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3278    names (ML structure names that correspond to unnamed C structs, unions,
3279    and enums), so that separate runs of ml-nlffigen do not clash.
3280    
3281    ----------------------------------------------------------------------
3282    Name: Matthias Blume
3283    Date: 2002/02/11 12:05:00 EST
3284    Tag: blume-20020211-gensml
3285    Description:
3286    
3287    A quick fix for a problem with GenSML (in the pgraph-util library):
3288    Make generation of toplevel "local" optional.  (Strictly speaking,
3289    signature definitions within "local" are not legal SML.)
3290    
3291    Other than that: updates to INSTALL and cm/TODO.
3292    
3293    ----------------------------------------------------------------------
3294    Name: Matthias Blume
3295    Date: 2002/02/08 15:00:00 EST
3296    Tag: blume-20020208-uniquepid
3297    Description:
3298    
3299    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3300    
3301    1. The installer (config/install.sh) has gotten smarter:
3302    
3303         - Configuration options are a bit easier to specify now
3304           (in config/targets).
3305         - Bug in recognizing .tar.bz2 files fixed.
3306         - Installer automatically resolves dependencies between
3307           configuration options (e.g., if you ask for eXene, you will
3308           also get cml -- regardless whether you asked for it or not).
3309         - Installer can run in "quieter mode" by setting the environment
3310           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3311           "completely silent", though.
3312         - Build HashCons library as part of smlnj-lib.
3313    
3314    2. A new scheme for assigning persistent identifiers to compilation
3315       units (and, by extension, to types etc.) has been put into place.
3316       This fixes a long-standing bug where types and even dynamic values
3317       can get internally confused, thereby compromising type safety
3318       (abstraction) and dynamic correctness.  See
3319    
3320         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3321    
3322       for an example of how things could go wrong until now.
3323    
3324       The downside of the new scheme is that pids are not quite as
3325       persistent as they used to be: CM will generate a fresh pid
3326       for every compilation unit that it thinks it sees for the first
3327       time.  That means that if you compile starting from a clean, fresh
3328       source tree at two different times, you end up with different
3329       binaries.
3330    
3331       Cutoff recompilation, however, has not been compromised because
3332       CM keeps pid information in special caches between runs.
3333    
3334    ----------------------------------------------------------------------
3335    Name: Lal George
3336    Date: 2002/02/07 15:34:13 EST 2002
3337    Tag: <none>
3338    Description:
3339    
3340    Compilers that generate assembly code may produce  global labels
3341    whose value is resolved at link time. The various peephole optimization
3342    modules did not take this in account.
3343    
3344    TODO. The Labels.addrOf function should really return an option
3345    type so that clients are forced to deal with this issue, rather
3346    than an exception being raised.
3347    
3348    ----------------------------------------------------------------------
3349    Name: Lal George
3350    Date: 2002/02/06 13:55:02 EST
3351    Tag: george-20020206-ra-breakup
3352    Description:
3353    
3354    1. A bug fix from Allen.
3355    
3356        A typo causes extra fstp %st(0)'s to be generated at compensation
3357        edges, which might cause stack underflow traps at runtime.  This
3358        occurs in fft where there are extraneous fstps right before the 'into'
3359        trap instruction (in this case they are harmless since none of the
3360        integers overflow.)
3361    
3362    2. Pulled out various utility modules that were embedded in the modules
3363       of the register allocator. I need these modules for other purposes, but
3364       they are not complete enough to put into a library (just yet).
3365    ----------------------------------------------------------------------
3366    Name: Matthias Blume
3367    Date: 2002/01/31 16:05:00 EST
3368    Tag: blume-20020131-sparc-ccalls
3369    Description:
3370    
3371    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3372       of extra stack space by mistake.  Fixed.
3373    
3374    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3375    
3376    ----------------------------------------------------------------------
3377    Name: Allen Leung
3378    Date: 2002/01/30
3379    Tag: leunga-20020130-nowhere-bug-fix
3380    Description:
3381    
3382       MLRISC bug fixes:
3383       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3384       2. MachineInt.fromString was negating its value.
3385    
3386    ----------------------------------------------------------------------
3387    Name: Matthias Blume
3388    Date: 2002/01/29
3389    Tag: blume-20020129-INSTALL
3390    Description:
3391    
3392    - Added somewhat detailed installation instructions (file INSTALL).
3393    - Fixed curl-detection bug in config/install.sh.
3394    - It is now possible to select the URL getter using the URLGETTER
3395      environment variable:
3396    
3397          not set / "unknown"      --> automatic detection (script tries wget,
3398                                       curl, and lynx)
3399          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3400                                       how to properly invoke them)
3401          other                    --> use $URLGETTER directly, it must take
3402                                       precisely two command-line arguments
3403                                       (source URL and destination file name)
3404    
3405    ----------------------------------------------------------------------
3406    Name: Matthias Blume
3407    Date: 2002/01/28
3408    Tag: blume-20020128-sparc-ccalls
3409    Description:
3410    
3411    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3412    - Make use of the allocParam argument in sparc-c-calls.
3413    
3414    ----------------------------------------------------------------------
3415    Name: Matthias Blume
3416    Date: 2002/01/28
3417    Tag: blume-20020128-allocParam
3418    Description:
3419    
3420    John Reppy:  Changes c-calls API to accept client-callback for
3421    allocating extra stack space.
3422    me: Corresponding changes to mlriscGen (using a dummy argument that
3423        does not change the current behavior).
3424    
3425    ----------------------------------------------------------------------
3426    Name: Matthias Blume
3427    Date: 2002/01/28 12:00:00
3428    Tag: Release_110_38
3429    Description:
3430    
3431    This time for real!!!
3432    
3433    ----------------------------------------------------------------------
3434    Name: Matthias Blume
3435    Date: 2002/01/28 10:56:00 EST
3436    Tag: blume-20020128-retraction
3437    Description:
3438    
3439    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3440       with blume-Release_110_38-retracted.)
3441    
3442    1. Fixed a problem with incorrect rounding modes in real64.sml.
3443       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3444    
3445    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3446       and enums fixed.  The naming of corresponding ML identifiers should
3447       now be consistent again.
3448    
3449    ----------------------------------------------------------------------
3450    Name: Allen Leung
3451    Date: 2002/01/27
3452    Tag: leunga-20020127-nowhere
3453    Description:
3454    
3455       Added a target called nowhere in the configuration scripts.
3456       Enabling this will build the MLRISC 'nowhere' tool (for translating
3457       programs with where-clauses into legal SML code) during installation.
3458    
3459    ----------------------------------------------------------------------
3460    Name: Matthias Blume
3461    Date: 2002/01/25 21:27:00 EST
3462    Tag: blume-Release_110_38-retracted
3463    Description:
3464    
3465    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3466    
3467    README will be added later.
3468    
3469    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3470    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3471    
3472    ----------------------------------------------------------------------
3473    Name: Matthias Blume
3474    Date: 2002/01/25
3475    Tag: blume-20020125-ffi
3476    Description:
3477    
3478    A large number of tweaks and improvements to ml-nlffi-lib and
3479    ml-nlffigen:
3480    
3481       - ML represenation types have been streamlined
3482       - getter and setter functions work with concrete values, not abstract
3483         ones where possible
3484       - ml-nlffigen command line more flexible (see README file there)
3485       - some bugs have been fixed (hopefully)
3486    
3487    ----------------------------------------------------------------------
3488    Name: Lal George
3489    Date: 2002/01/24
3490    Tag: george-20020124-risc-ra-interface
3491    Description:
3492    
3493       There is a dramatic simplification in the interface to the
3494       register allocator for RISC architectures as a result of making
3495       parallel copy instructions explicit.
3496    
3497    ----------------------------------------------------------------------
3498    Name: Matthias Blume
3499    Date: 2002/01/22
3500    Tag: blume-20020122-x86-ccalls
3501    Description:
3502    
3503    Bug fix for c-calls on x86 (having to do with how char- and
3504    short-arguments are being handled).
3505    
3506    ----------------------------------------------------------------------
3507    Name: Matthias Blume
3508    Date: 2002/01/21
3509    Tag: blume-20020121-ff
3510    Description:
3511    
3512    Another day of fiddling with the FFI...
3513    
3514    1. Bug fix/workaround:  CKIT does not complain about negative array
3515       dimensions, so ml-nlffigen has to guard itself against this possibility.
3516       (Otherwise a negative dimension would send it into an infinite loop.)
3517    
3518    2. Some of the abstract types (light objects, light pointers, most "base"
3519       types) in structure C are now eqtypes.
3520    
3521    3. Added constructors and test functions for NULL function pointers.
3522    
3523    ----------------------------------------------------------------------
3524    Name: Matthias Blume
3525    Date: 2002/01/18
3526    Tag: blume-20020118-ready-for-new-release
3527    Description:
3528    
3529    Made config/srcarchiveurl point to a new place.  (Will provide boot
3530    files shortly.)
3531    
3532    Maybe we christen this to be 110.38?
3533    
3534    ----------------------------------------------------------------------
3535    Name: Matthias Blume
3536    Date: 2002/01/18
3537    Tag: blume-20020118-more-ffifiddle
3538    Description:
3539    
3540    Today's FFI fiddling:
3541    
3542      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3543        C.Set.* that use concrete (MLRep.*) arguments and results instead
3544        of abstract ones.
3545    
3546      - Provided word-style bit operations etc. for "int" representation
3547        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3548        and Long.
3549    
3550    ----------------------------------------------------------------------
3551    Name: Matthias Blume
3552    Date: 2002/01/18
3553    Tag: blume-20020118-use-x86-fp
3554    Description:
3555    
3556    Now that x86-fast-fp seems to be working, I turned it back on again
3557    by default.  (Seems to work fine now, even with the FFI.)
3558    
3559    Other than that, I added some documentation about the FFI to
3560    src/ml-nlffigen/README and updated the FFI test examples in
3561    src/ml-nlffi-lib/Tests/*.
3562    
3563    ----------------------------------------------------------------------
3564    Name: Allen Leung
3565    Date: 2002/01/17
3566    Tag: leunga-20020117-x86-fast-fp-call
3567    Description:
3568    
3569       1. Fixed a problem with handling return fp values when x86's fast fp
3570          mode is turned on.
3571    
3572       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3573          of %f32.
3574    
3575       3. Added a constructor INT32lit to the ast of MLRISC tools.
3576    
3577    ----------------------------------------------------------------------
3578    Name: Matthias Blume
3579    Date: 2002/01/16
3580    Tag: blume-20020116-ffifiddle
3581    Description:
3582    
3583    More fiddling with the FFI interface:
3584    
3585     - Make constness 'c instead of rw wherever possible.  This eliminates
3586       the need for certain explicit coercions.  (However, due to ML's
3587       value polymorphism, there will still be many cases where explicit
3588       coercions are necessary.  Phantom types are not the whole answer
3589       to modeling a subtyping relationship in ML.)
3590    
3591     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3592    
3593     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3594       clumsy workaround for varargs problem.)
3595    
3596    ----------------------------------------------------------------------
3597    Name: Lal George
3598    Date: 2002/01/15
3599    Tag: <none>
3600    Description:
3601    
3602    1. Since COPY instructions are no longer native to the architecture,
3603       a generic functor can be used to implement the expandCopies function.
3604    
3605    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3606       TEXT segment.
3607    
3608    ----------------------------------------------------------------------
3609    Name: Matthias Blume
3610    Date: 2002/01/15
3611    Tag: blume-20020115-ffiupdates
3612    Description:
3613    
3614    1. Fix for bug resulting in single-precision float values being returned
3615       incorrectly from FFI calls.
3616    
3617    2. Small modifications to C FFI API:
3618    
3619        - memory-allocation routines return straight objects (no options)
3620          and raise an exception in out-of-memory situations
3621        - unsafe extensions to cast between function pointers and pointers