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

SCM Repository

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

Diff of /sml/trunk/HISTORY

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

revision 589, Fri Mar 31 16:14:36 2000 UTC revision 1701, Wed Nov 17 21:07:35 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: Allen Leung (leunga (at) reservoir (dot) com)
17    Date: 2004/11/17 16:05:21 EST 2004
18    Tag: leunga-20041117-mlrisc-live-kill
19    Description:
20    
21       Added support for MLTree constructs LIVE and KILL
22    to all the architectures.
23    
24    ----------------------------------------------------------------------
25    Name: Matthias Blume (blume (at) tti - c (dot) org)
26    Date: 2004/11/13 00:20:00 CST
27    Tag: blume-20041113-versiontool
28    Description:
29    
30    - Stripped down the versiontool: It now only handles the version number.
31      The date string is generated at bootstrap time (during makeml).
32    
33    - In a previous commit, fixed a minor issue with how polyequal is being
34      translated.  In particular, the code now "looks through" abstractions.
35      This results in slightly fewer polyEqual warnings and hopefully slightly
36      more efficient code.  Important examples for where this matters are
37      the new int64 and word64 types.
38    
39    ----------------------------------------------------------------------
40    Name: Matthias Blume (blume (at) tti - c (dot) org)
41    Date: 2004/11/12 00:30:00 CST
42    Tag: blume-20041112-int64
43    Description:
44    
45    Structure Int64 fully hooked in.  (The implementation is not very
46    efficient, though.)
47    
48    ----------------------------------------------------------------------
49    Name: Matthias Blume (blume (at) tti - c (dot) org)
50    Date: 2004/11/11 17:30:00 CST
51    Tag: blume-20041111-more64
52    Description:
53    
54    All the pieces of Word64 are now there, with the exception of the
55    conversions from and to LargeWord. (Eventually these need to be identities,
56    but for the time being they don't even make sense because LargeWord is
57    32-bit wide.)
58    
59    Also started to add similar support for Int64, but major pieces of that
60    are still missing.
61    
62    ----------------------------------------------------------------------
63    Name: Matthias Blume (blume (at) tti - c (dot) org)
64    Date: 2004/11/11 00:15:00 CST
65    Tag: blume-20041111-word64
66    Description:
67    
68    Structure Word64 is now (almost) complete, word literals and patterns
69    seem to work.  There are a few odd pieces missing.  In particular,
70    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
71    at the moment.
72    
73    Making Word64 official would mean that LargeWord becomes Word64.  But
74    this requires extreme care because most word-word conversions have to
75    go through LargeWord, so making a mistake means loss of efficiency or
76    worse.  Eventually there will be a solution similar to (but actually
77    simpler than) what I did with IntInf.
78    
79    ----------------------------------------------------------------------
80    Name: Matthias Blume (blume (at) tti - c (dot) org)
81    Date: 2004/11/10 18:12:00 CST
82    Tag: blume-20041110-64bit
83    Description:
84    
85    More 64-bit hacking (but still not even half-way there yet).
86    Also, some assorted improvements to the handling of 8-bit words.
87    
88    ----------------------------------------------------------------------
89    Name: Matthias Blume (blume (at) tti - c (dot) org)
90    Date: 2004/11/09 17:50:00 CST
91    Tag: <none>
92    Description:
93    
94    Started adding some infrastructure for supporting 64-bit int- and
95    word-types.  (Still in its very early stages.)
96    
97    ----------------------------------------------------------------------
98    Name: Matthias Blume (blume (at) tti - c (dot) org)
99    Date: 2004/10/28 10:45:00 CDT
100    Tag: Release_110_50
101    Description:
102    
103    New working version (110.50).  NEW BOOTFILES!
104    
105                 =====================
106    
107    Also:
108    
109      - Changend config/srcarchiveurl from a file just
110        containing the URL string into a file containing
111        shell script code.  The code has access to the $VERSION variable.
112      - Made corresponding changes to config/install.sh and config/unpack.
113      - Default contents of config/srcarchiveurl uses $VERSION and
114        normally does not have to be edited to reflect a version change.
115    
116      (As a result, a version change can be done by just editing
117       config/version, the rest is now automatic.)
118    
119    ----------------------------------------------------------------------
120    Name: Matthias Blume (blume (at) tti - c (dot) org)
121    Date: 2004/10/27 17:50:00 CDT
122    Tag: blume-20041027-btrace-msg
123    Description:
124    
125    BackTrace.monitor now also reports the source of the exception that
126    triggered the trace.
127    
128    ----------------------------------------------------------------------
129    Name: Matthias Blume (blume (at) tti - c (dot) org)
130    Date: 2004/10/27 17:20:00 CDT
131    Tag: blume-20041027-x86-c-calls
132    Description:
133    
134    This is the HISTORY entry for two earlier commits, both concerning
135    the x86 c-calls code in MLRISC:
136    
137      - added a missing LOAD in the code that deals with struct arguments
138      - made sure the caller does not add the wrong number of bytes to the
139        stack pointer after a call of a function returning a struct
140        (the callee already pops the implicit argument which points to
141         the space reserved for the result)
142    
143    ----------------------------------------------------------------------
144    Name: Allen Leung (leunga (at) reservoir (dot) com)
145    Date: 2004/10/24 14:00:00 EST
146    Tag: leunga-20041024-x86-gas-fucomip
147    Description:
148    
149       John discovered a bug in the syntax of fucomip.
150       The opcodes FU?COMIP? have been changed to
151    
152           fu?comip? %st(i), %st
153    
154    ----------------------------------------------------------------------
155    Name: Matthias Blume (blume (at) tti - c (dot) org)
156    Date: 2004/10/20 15:06:00 CDT
157    Tag: blume-20041020-standalone-backtrace
158    Description:
159    
160    Added a mechanism for getting back-trace information from standalone
161    programs.  Here is how it works:
162    
163       1. The part of the program from which you want to get backtrace
164          information (usually the whole program) should be wrapped with
165          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
166          main program could be modified from something like
167    
168             fun main (pgm, args) = ...
169    
170         to
171    
172            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
173    
174       2. To be able to access BackTrace.monitor, you have to add
175          library $smlnj-tdp/plugins.cm to the .cm file that contains your
176          main function.
177    
178       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
179          CM might have created in the past for your project).
180    
181       4. Build the system using this command line:
182    
183             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
184                      myprog.cm MyProg.main myprog
185    
186          instead of the usual
187    
188             ml-build myprog.cm MyProg.main myprog
189    
190    I changed a library name:
191    
192       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
193    
194    New libraries:
195    
196       $smlnj-tdp/back-trace.cm
197          -- when loaded causes the back-trace plugin to be installed
198       $smlnj-tdp/coverage.cm
199          -- when loaded causes the coverage plugin to be installed
200    
201    ----------------------------------------------------------------------
202    Name: Matthias Blume (blume (at) tti - c (dot) org)
203    Date: 2004/10/18 16:45:00 CDT
204    Tag: blume-20041018-groupowner
205    Description:
206    
207    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
208    
209    Eliminated group owner specs from .cm files throughout the source tree.
210    
211    ----------------------------------------------------------------------
212    Name: Matthias Blume (blume (at) tti - c (dot) org)
213    Date: 2004/10/15 15:45:00 CDT
214    Tag: blume-20041015-coverage
215    Description:
216    
217    * Test coverage tool added!
218    * Further reorganization of tracing-, debugging-, and profiling support:
219    
220        - moved original BTImp -- now called BackTrace -- into a separate
221          library called $/trace-debug-profile.cm
222        - eliminated all mentions of BTrace from SMLofNJ.Internals
223        - only the instrumentation mechanism is now left in the compiler proper
224        - BackTrace module is a plugin which is NOT plugged in by default
225        - Coverage module is another such plugin
226    
227    To get the benefits of any of these plugin modules, the code in
228    question must be compiled with tdp instrumentation turned on.  This
229    can be done by setting SMLofNJ.Internals.TDP.mode to true.
230    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
231    
232    Plugins are selected at link time.  (Pre-compiled instrumented code
233    can be re-loaded with different plugins in effect.)  When an
234    instrumented module is linked, whatever plugins are at that time
235    enabled will come into effect for that module.
236    
237    To enable the back-trace plugin, load library $/trace-debug-profile.cm
238    and invoke BackTrace.install() (e.g., from the interactive prompt).
239    To enable the coverage plugin, load the same library and invoke
240    Coverage.install().
241    
242    Back-traces are generated automatically on uncaught exceptions and
243    when the code in question explicitly invokes BackTrace.trigger().
244    
245    Coverage (and execution frequency-) information must be queried
246    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
247    
248    ----------------------------------------------------------------------
249    Name: Matthias Blume (blume (at) tti - c (dot) org)
250    Date: 2004/10/14 17:40:00 CDT
251    Tag: blume-20041014-tdp-core
252    Description:
253    
254    Snapshot of a significant overhaul of how the trace/debug/profile support
255    is hooked into the system (specifically: Core and SMLofNJ.Internals).
256    
257    ----------------------------------------------------------------------
258    Name: Matthias Blume (blume (at) tti - c (dot) org)
259    Date: 2004/10/13 16:34:00 CDT
260    Tag: blume-20041013-tdp
261    Description:
262    
263    Some rationalization of names:
264    
265       structure BTrace -> structure TDPInstrument
266       etc.
267    
268    This is is preparation of using the original back-trace
269    instrumentation for other purposes.  "TDP" stands for
270    Trace/Debug/Profile.
271    
272    The control flag controlling whether instrumentation is on or off is now
273    registered under a different name, so instead of running sml as
274    
275     sml -Cinstrument.btrace-mode=true
276    
277    one has to say
278    
279     sml -Ctdp.instrument=true
280    
281    ----------------------------------------------------------------------
282    Name: Matthias Blume (blume (at) tti - c (dot) org)
283    Date: 2004/10/11 16:37:00 CDT
284    Tag: blume-20041011-regions
285    Description:
286    
287    Made some minor modifications to elabcore.sml to have source regions
288    be propagated more tightly -- resulting in better (i.e., smaller)
289    regions being reported in error- and debug messages.
290    
291    ----------------------------------------------------------------------
292    Name: Matthias Blume (blume (at) tti - c (dot) org)
293    Date: 2004/10/08 22:50:00 CDT
294    Tag: blume-20041008-cmkw
295    Description:
296    
297    Fixed handling of keywords in .cm files: After seeing "is" the lexer
298    treats subsequent occurrences of "group", "library", "source", "is",
299    "*", and "-" as ordinary identifiers rather than keywords.
300    
301    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
302    is supposed to accept a tool argument called "source", but this did
303    not work because of the clash with the keyword.
304    
305    ----------------------------------------------------------------------
306    Name: Matthias Blume (blume (at) tti - c (dot) org)
307    Date: 2004/10/07 16:00:00 CDT
308    Tag: blume-20041007-cleanup
309    Description:
310    
311    Assorted cleanup work:
312    
313      - got rid of intstrmap in favor of using the library's
314        hash table implementation
315      - threw out most of the pathnames stuff, as it was not used anyway
316      - simplified tokentable implementation
317      - fixed some minor spelling errors
318    
319    ----------------------------------------------------------------------
320    Name: Matthias Blume (blume (at) tti - c (dot) org)
321    Date: 2004/10/06 15:15:15 CDT
322    Tag: blume-20041006-handler
323    Description:
324    
325    Cleaned up the absyn to reflect the invariant that HANDLE always
326    carries a FNexp as part of the type definition.  This eliminates some
327    superfluous sanity checks at runtime down the road.
328    
329    Some minor cleanup of the btrace code.
330    
331    ----------------------------------------------------------------------
332    Name: Matthias Blume (blume (at) tti - c (dot) org)
333    Date: 2004/10/01 10:20:30 CDT
334    Tag: blume-20041001-slave
335    Description:
336    
337    Added hack to make slave mode work in the presence of the version
338    tool.  (Still, since the master does two passes over the code for
339    CMB.make, the release number gets bumped twice when slaves are
340    attached.  I don't know if this is worth fixing...)
341    
342    ----------------------------------------------------------------------
343    Name: Matthias Blume (blume (at) tti - c (dot) org)
344    Date: 2004/09/30 10:55:00 CDT
345    Tag: blume-20040930-version
346    Description:
347    
348    * Moved the "version" magic into its own little library under
349      src/system/smlnj/internal.  This avoids expensive reconstruction of
350      a stable src/compiler/core.cm.
351    
352    * At the same time, structure CompilerVersion is now known as
353      structure SMLNJVersion.
354    
355    * Arranged for the version tool to NOT kick in when rebuilding the system
356      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
357      Also, loading the versiontool does not work when rebuilding the system
358      because CM is not properly initialized at that time.
359    
360    ----------------------------------------------------------------------
361    Name: Matthias Blume (blume (at) tti - c (dot) org)
362    Date: 2004/09/29 14:00:00 CDT
363    Tag: blume-20040929-autoversion
364    Description:
365    
366    Implemented some CM magic to have
367        file src/compiler/TopLevel/main/version.sml
368    generated automagically.
369    The version is taken from two files: config/version and config/release.
370    The first is expected to contain a two-part version number such as 110.49.
371    The second should contain a single number, but it may be missing.
372    
373    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
374    time the version tool is loaded (which is the first time you say CMB.make),
375    then the tool will increment the value stored in config/release every
376    time CMB.make is invoked.
377    
378    The binfile format is now insensitive to anything beyond the first
379    two components of a version number, so bumping the release does not render
380    binfiles incompatible.  Auto-bumping can be used to keep track of versions
381    during development without invalidating existing binfiles.
382    
383    In any case, every CMB.make updates the date information in version.sml.
384    (This is the date that is printed in the banner.)
385    
386    ----------------------------------------------------------------------
387    Name: Matthias Blume (blume (at) tti - c (dot) org)
388    Date: 2004/09/28 10:53:00 CDT
389    Tag: blume-20040928-controls
390    Description:
391    
392    Some cleanup of the controls code.
393    
394    ----------------------------------------------------------------------
395    Name: Matthias Blume (blume (at) tti - c (dot) org)
396    Date: 2004/09/27 22:08:00 CDT
397    Tag: blume-20040927-controls
398    Description:
399    
400    Added two pieces of functionality to the Controls interface:
401    
402    1.   val save'restore: 'a control -> unit -> unit
403    
404       grabs the current value of the control in stage 1 and restores it
405       in stage 2.
406    
407    2.   val set' : 'a control * 'a -> unit -> unit
408    
409       stores the given value into the control in stage 2 (i.e., delayed)
410       but does all error checking in stage 1.
411       (This is for string controls that need to do parse their argument --
412        something that might fail.  In some cases, notably in CM, one
413        already knows the intended argument but wants to delay the actual
414        assignment until a time when error recovery would be more difficult.)
415    
416    Changed the handling of controls in tool arguments to classes "sml" and
417    "lazysml":
418       - use Controls.save'restore as a more robust way of restoring the
419         old value (in particular: without having to re-parse the string)
420       - use controls to handle the "overload" keyword in the init group
421         (I believe this change actually fixes a long-standing obscure bug.)
422    
423    ----------------------------------------------------------------------
424    Name: Matthias Blume (blume (at) tti - c (dot) org)
425    Date: 2004/09/27 17:00:00 CDT
426    Tag: blume-20040927-lazysml
427    Description:
428    
429    Added a new tool class called "lazysml" to CM's tool chest.  The only
430    difference to "sml" is that compilation is done with Control.lazysml
431    set to true.  A source of class "lazysml" is automatically recognized
432    by a file name suffix of ".lml".
433    
434    In addition to the above feature, the original class "sml" now also
435    supports a tool argument "lazy" which has the same effect.  As a
436    result, the following three lines are equivalent:
437    
438        foo.sml : lazysml
439        foo.sml : sml (lazy)
440        foo.sml (lazy)
441    
442    The setting goes into effect both during parsing and during
443    compilation.  The original setting is restored right after parsing and
444    after compilation, respectively.
445    
446    In addition to all the above, there is also a general mechanism to set
447    ANY of the "controls" that are available at the command line via
448    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
449    apply as well. (In fact, "lazy" is implemented as a special case of
450    the general mechanism.)
451    
452    The .cm file syntax uses a new keyword tool argument called "with".
453    There are several ways of indicating the desired settings:
454    
455        foo.sml (with:parser.quotations=true)
456        foo.sml (with:(name:parser.quotations value:true))
457        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
458        foo.sml (with:(name1=value1 name2=value2 ...))
459        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
460    
461        etc.
462    
463    Another possible abbreviation is to leave out the =v or value:v part
464    if the name refers to a boolean control (in which case the value is
465    taken to be true).  Thus, one could get lazy sml also by saying:
466    
467        foo.sml (with:parser.lazy-keyword=true)
468        foo.sml (with:parser.lazy-keyword)
469        foo.sml (with:(name:parser.lazy-keyword value:true))
470        foo.sml (with:(name:parser.lazy-keyword))
471    
472    ----------------------------------------------------------------------
473    Name: Matthias Blume (blume (at) tti - c (dot) org)
474    Date: 2004/09/24 16:22:00 CDT
475    Tag: blume-20040924-ppc-long-branch
476    Description:
477    
478    Turned message about "emiting long form of branch" off by default.
479    Added a control flag to turn it back on when desired.
480    
481    ----------------------------------------------------------------------
482    Name: Matthias Blume (blume (at) tti - c (dot) org)
483    Date: 2004/09/24 16:05:00 CDT
484    Tag: blume-20040924-rounding
485    Description:
486    
487    Applied patch for setting rounding modes under Mac OS X.  Thanks to
488    Melissa O'Neill for providing the code!
489    
490    ----------------------------------------------------------------------
491    Name: Matthias Blume (blume (at) tti - c (dot) org)
492    Date: 2004/09/23 17:30:00 CDT
493    Tag: blume-20040923-envvars
494    Description:
495    
496    1. Changed definition of type ControlRegistry.registry_tree to
497       include control_info (i.e., the name of the controlling environment
498       variable).
499    
500    2. Added command-line flags -e and -E to print the names of environment
501       variables that can be used to control internal settings.
502       (This uses the new API mentioned in 1.)
503    
504    ----------------------------------------------------------------------
505    Name: Matthias Blume (blume (at) tti - c (dot) org)
506    Date: 2004/09/13 16:50:00 CDT
507    Tag: Release_110_49
508    Description:
509    
510    New working version (110.49).  NEW BOOTFILES!
511    
512    ----------------------------------------------------------------------
513    Name: Matthias Blume
514    Date: 2004/09/13 16:20:00 CDT
515    Tag: blume-20040913-config-mlrisc
516    Description:
517    
518    Put target "mlrisc" back into the default list.
519    (There is no harm in having it, and some users have expressed their
520    wish to have "mlrisc" included by default.)
521    
522    ----------------------------------------------------------------------
523    Name: John Reppy
524    Date: 2004/09/13
525    Tag: jhr-20040913-signals
526    Description:
527    
528    Fixed the signal masking code to properly nest mask/unmask operations
529    on a per-signal basis.
530    
531    ----------------------------------------------------------------------
532    Name: Matthias Blume (blume (at) tti - c (dot) org)
533    Date: 2004/09/08 13:20:00 CDT
534    Tag: blume-20040908-heap-magic
535    Description:
536    
537    Bumped the heap macig number to 0x09082004 to account for the changed
538    layout of the ML frame under MacOS X.
539    
540    ----------------------------------------------------------------------
541    Name: Allen Leung (leunga (at) reservoir (dot) com)
542    Date: 2004/09/03 11:26:00 EST
543    Tag: leunga-20040903-cygwin-install
544    Description:
545    
546       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
547    The Cygwin runtime is turned on by setting the environment
548    variable SMLNJ_CYGWIN_RUNTIME to 1.
549    
550    ----------------------------------------------------------------------
551    Name: Matthias Blume (blume (at) tti - c (dot) org)
552    Date: 2004/08/31 17:15:00 CDT
553    Tag: blume-20040831-core
554    Description:
555    
556    Added some exports to src/compiler/core.cm upon request by J. Joemann.
557    
558    ----------------------------------------------------------------------
559    Name: Matthias Blume (blume (at) tti - c (dot) org)
560    Date: 2004/08/30 17:55:00 CDT
561    Tag: blume-20040830-installer
562    Description:
563    
564    Upon request by Johannes Joemann:
565    
566    - improved ML code of installer to fall back to coping when renaming
567      fails (i.e., when source and target are on different file systems);
568      the code compiles but has yet to be tested in anger
569    - removed mlrisc from list of default targets (config/targets)
570    
571    ----------------------------------------------------------------------
572    Name: Matthias Blume (blume (at) tti - c (dot) org)
573    Date: 2004/08/27 17:20:00 CDT
574    Tag: blume-20040827-ptreql
575    Description:
576    
577    Added ptreql primop to structure InlineT (upon request from Larry
578    Paulson).
579    
580    ----------------------------------------------------------------------
581    Name: Allen Leung (leunga (at) reservoir (dot) com)
582    Date: 2004/08/15 21:21:00 EST
583    Tag: leunga-110_48-udgraph
584    Description:
585    
586    Another bug fix from Carl Hauser:
587    
588    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
589    > 48c48
590    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
591    > e::L)
592    > ---
593    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
594    > Without this, any deletion of an edge in an undirected graph does severe
595    > violence to the graph.
596    
597    ----------------------------------------------------------------------
598    Name: Allen Leung (leunga (at) reservoir (dot) com)
599    Date: 2004/08/10 23:35:00 EST
600    Tag: leunga-110_48-ppc
601    Description:
602    
603       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
604    
605    ----------------------------------------------------------------------
606    Name: Matthias Blume (blume (at) tti - c (dot) org)
607    Date: 2004/08/10 12:00:00 CDT
608    Tag: Release_110_48
609    Description:
610    
611    New working version (110.48).  NEW BOOTFILES!
612    
613    ----------------------------------------------------------------------
614    Name: Allen Leung (leunga (at) reservoir (dot) com)
615    Date: 2004/08/09 12:21:00 EST
616    Tag: leunga-110_47-dijsktra
617    Description:
618    
619    Bug fix from Carl Hauser:
620    
621    single_source_shortest_paths in dijkstra.sml was observed to get wrong
622    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
623    
624    The problem is that following the expression A.update(dist,s,Num.zero)
625    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
626    
627    ----------------------------------------------------------------------
628    Name: Matthias Blume (blume (at) tti - c (dot) org)
629    Date: 2004/08/06 18:10:00 CDT
630    Tag: blume-20040806-cmdline
631    Description:
632    
633    Fiddled with handling of command-line options:
634    
635      * sml now quits after processing the command line
636        if -H, -S, -h<n>, or -s<n> appears as the last
637        command-line argument
638      * a new option -q terminates the session when encountered on
639        the command line; subsequent arguments will be ignored
640      * bug fixes: short (erroneous) arguments are no longer ignored
641        completely
642    
643    ----------------------------------------------------------------------
644    Name: Allen Leung (leunga (at) reservoir (dot) com)
645    Date: 2004/08/04 18:17:00 EST
646    Tag: leunga-110_47-ppc-ibm-asm
647    Description:
648    
649      - Added minimal IBM assembly syntax support for PowerPC.
650    
651      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
652        been altered in the runtime.  We need an automatic way to keep the file
653        in sync.
654    
655    ----------------------------------------------------------------------
656    Name: Matthias Blume (blume (at) tti - c (dot) org)
657    Date: 2004/08/04 14:00:00 CDT
658    Tag: Release_110_47
659    Description:
660    
661    New working version (110.47).  NEW BOOTFILES!
662    
663    ----------------------------------------------------------------------
664    Name: Matthias Blume (blume (at) tti - c (dot) org)
665    Date: 2004/08/03 14:25:00 CDT
666    Tag: blume-20040803-callingconv
667    Description:
668    
669    Added low-level support for choosing C calling conventions by
670    twiddling the type of rawccall.  (See
671    src/compiler/Semant/types/cproto.sml for details.)
672    
673    ----------------------------------------------------------------------
674    Name: Matthias Blume (blume (at) tti - c (dot) org)
675    Date: 2004/08/02 15:55:00 CDT
676    Tag: blume-20040802-backout
677    Description:
678    
679    Backed out of change to win32-filesys.c.  The earlier patch to
680    get_file_time caused CM to produce files with the wrong time stamp.
681    
682    ----------------------------------------------------------------------
683    Name: Matthias Blume (blume (at) tti - c (dot) org)
684    Date: 2004/08/02 14:45:00 CDT
685    Tag: blume-20040802-nlffi-win32
686    Description:
687    
688    Added NLFFI support for Win32, adapted from a patch provided by David
689    Hansel.  This is currently completely untested.  Also, the issue
690    concerning stdcall vs. ccall is still unresolved.
691    
692    ----------------------------------------------------------------------
693    Name: Matthias Blume (blume (at) tti - c (dot) org)
694    Date: 2004/07/30 17:55:00 CDT
695    Tag: blume-20040730-various
696    Description:
697    
698    Gearing up towards 110.47...
699    
700    - various minor bugfixes to ml-nlffigen
701    - a beginning of a manual for nlffi
702    
703    - eliminated 'export name=value' in config/install.sh as this does
704      not work with certain versions of /bin/sh
705      (Thanks to David King at Motorola for catching this.)
706    
707    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
708      - added a test for tm==NULL to gmtime.c and localtime.c
709      - applied patch for incorrect GetFileTime under win32
710      - toSeconds -> toMilliseconds in Win32/win32-process.sml
711    
712    ----------------------------------------------------------------------
713    Name: Matthias Blume (blume (at) tti - c (dot) org)
714    Date: 2004/07/21 18:20:00 CDT
715    Tag: blume-20040721-nlffigen
716    Description:
717    
718    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
719      for a typedef to an incomplete type, but leave out the "typ" member.
720      (This is just for consistency.)
721    - Started to produce what is supposed to become better (i.e., comprehensive)
722      documentation of what ml-nlffigen does and produces.
723    
724    ----------------------------------------------------------------------
725    Name: Matthias Blume (blume (at) tti - c (dot) org)
726    Date: 2004/07/14 16:25:00 CDT
727    Tag: blume-20040714-union
728    Description:
729    
730    Added C_UNION to c-calls/c-types.sml and updated the machinery
731    (ml-nlffigen, cproto.sml) that conveys C function interface
732    information to the code generator.
733    
734    However, the actual architecture-specific implementation of function
735    arguments and results that are C unions is still not implemented.
736    
737    ----------------------------------------------------------------------
738    Name: Allen Leung (leunga (at) reservoir (dot) com)
739    Date: 2004/07/14 14:38:00 EST
740    Tag: leunga-110_46_1-ppc-lwzu
741    Description:
742    
743        Added these instructions to the PowerPC architecture:
744        LBZU(X), LHZU(X), LWZU(X),
745        STWU(X), STFDU, STFSU
746    
747        etc...
748    
749        Note: I haven't added their instruction encoding into the description.
750    
751    ----------------------------------------------------------------------
752    Name: Allen Leung (leunga (at) reservoir (dot) com)
753    Date: 2004/07/13 15:04:00 EST
754    Tag: leunga-110_46_1-ppc-lwarx
755    Description:
756    
757       Added the two instructions LWARX and STWCX to the PowerPC
758    instruction set.
759    
760       A (untested) rewrite of loop-structure.sml.  The old version
761    is completely broken.
762    
763    ----------------------------------------------------------------------
764    Name: Matthias Blume (blume (at) tti - c (dot) org)
765    Date: 2004/07/13 13:50:00 CDT
766    Tag: blume-20040713-nlffi
767    Description:
768    
769    - use paramAlloc to report c-calls with too many arguments
770      (for PPC version where parameter area is pre-allocated)
771    - added ccall_maxargspace to machspec (to implement the above)
772    - made "make" commend in CM's "make" tool configurable
773    - added option (default: on) for passing the name of the SML/NJ's "bin"
774      directory to "make"; the call looks like this:
775    
776         make <options> SMLNJ_BINDIR=<dir> <target>
777    
778      This can be used by the Makefile to, e.g., pick the "right" version
779      of ml-nlffigen.
780    - minor code tweaks
781    
782    ----------------------------------------------------------------------
783    Name: Matthias Blume (blume (at) tti - c (dot) org)
784    Date: 2004/07/12 22:50:00 CDT
785    Tag: blume-110_46_1-macosx-nlffi
786    Description:
787    
788    NLFFI under Mac OS X now working (sort of).  This is largely untested,
789    though.
790    
791    Note:  1. You have to make a new, clean build of the runtime system.
792           2. There are new BOOTFILES, you have to use them!
793              (Doing the bootstrap process yourself would be *very* painful!
794               If you absolutely have to do it, build the system under
795               a different architecture and then cross-compile.)
796    
797    Version bumped to 110.46.1 to account for runtime data format changes.
798    
799    ----------------------------------------------------------------------
800    Name: Matthias Blume (blume (at) tti - c (dot) org)
801    Date: 2004/06/18 14:30:00 CDT
802    Tag: blume-20040618-unix
803    Description:
804    
805    Changed the implementation of structure Unix so that the same stream
806    is returned every time one of the {text,bin}{In,Out}streamOf functions
807    is invoked on the same proc.  This is not what the spec currently
808    says -- although IMO it arguably should.  (See discussion below.)
809    
810    ----------------------------------------------------------------------
811    Name: Matthias Blume (blume (at) tti - c (dot) org)
812    Date: 2004/06/17 18:15:00 CDT
813    Tag: Release_110_46
814    Description:
815    
816    New working version (110.46).  NEW BOOTFILES!
817    
818    ----------------------------------------------------------------------
819    Name: Matthias Blume (blume (at) tti - c (dot) org)
820    Date: 2004/06/17 17:20:00 CDT
821    Tag: blume-20040617-timer-unix
822    Description:
823    
824    Changed the interface of structures Timer and Unix to match the most
825    recent Basis spec.
826    
827    In the case of Unix there still seems to be an open/weird issue:
828    
829       The {text,bin}{In,Out}streamOf functions are supposed to create
830       fresh streams whenever they are called -- as opposed to have them
831       return the same stream every time.  This design is supposed to
832       prevent space leaks caused by proc values hanging on to streams.
833    
834       The reap function, on the other hand, is supposed to close the
835       streams.  This cannot be done without having a handle on the
836       stream in proc after all...
837    
838    I took the liberty to implement the following stopgap solution:
839    
840      The proc value hangs on to the most recently created stream(s).
841      Reap closes those.  If either or both of the two streams hadn't
842      been created at all yet, then reap will close the corresponding
843      file descriptors directly.
844    
845    PS: I don't understand the original space leak argument anymore.  If
846    a proc hangs on to the imperative stream, then I/O operations on those
847    will advance the state of the cached stream and avoid the space leak.
848    
849    ----------------------------------------------------------------------
850    Name: Matthias Blume (blume (at) tti - c (dot) org)
851    Date: 2004/05/28 16:45:00 CDT
852    Tag: blume-20040528-basis
853    Description:
854    
855    Added signature PACK_REAL and exported functor PrimIO.
856    
857    ----------------------------------------------------------------------
858    Name: Matthias Blume (blume (at) tti - c (dot) org)
859    Date: 2004/05/25 16:00:00 CDT
860    Tag: blume-20040525-group-owner
861    Description:
862    
863    CM now ignores (but still accepts) the "owner" information in group
864    descriptions.  The owner of a group is its next enclosing
865    library. Each group must have a unique owner.  (There is a virtual
866    "toplevel" library that owns groups which are not nested within a real
867    library.)  Previously, each group had to explicitly declare its owner,
868    and CM would check that such a declaration is correct.  The new scheme
869    is to have CM check that for each group there is precisely one owning
870    library.
871    
872    The advantage of the new scheme is that the programmer no longer needs
873    to maintain the somewhat annoying owner information.  The downside is
874    that CM cannot enforce the ownership rule across multiple runs of
875    CM.make.  Fortunately, enclosing the same group in two different
876    libraries A and B which are not part of the same program does not
877    cause real problems.
878    
879    ----------------------------------------------------------------------
880    Name: Matthias Blume (blume (at) tti - c (dot) org)
881    Date: 2004/05/20 16:00:00 CDT
882    Tag: blume-20040520-win32
883    Description:
884    
885    Made the win32 version work again.  (Strangely, a misplaced comma had
886    slipped into win32-process.c which prevented the runtime from being
887    compiled correctly.)
888    
889    Also, included a minor addition to ml-build.bat analogous to what was
890    done in blume-20040519-ml-build.
891    
892    ----------------------------------------------------------------------
893    Name: Matthias Blume (blume (at) tti - c (dot) org)
894    Date: 2004/05/19 22:10:00 CDT
895    Tag: blume-20040519-ml-build
896    Description:
897    
898    Arranged for ml-build to clean up after itself a little bit better.
899    The script generates a temporary SML source file and compiles it using
900    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
901    gets rid of those at the end, so they don't accumulate under .cm.
902    
903    This required a minor change to install.sh because the name of the
904    metadata directory (default: .cm) is actually configurable at
905    installation time.
906    
907    ----------------------------------------------------------------------
908    Name: Matthias Blume (blume (at) tti - c (dot) org)
909    Date: 2004/05/18 15:50:00 CDT
910    Tag: blume-20040518-mkreader
911    Description:
912    
913    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
914    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
915    
916    ----------------------------------------------------------------------
917    Name: Matthias Blume (blume (at) tti - c (dot) org)
918    Date: 2004/05/11 14:35:00 CDT
919    Tag: blume-20040511-win32sock
920    Description:
921    
922    Added previously missing support for many socket-related functions
923    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
924    for the voluminous patch!
925    
926    (I have not tested this patch under win32 yet.)
927    
928    Here is David's e-mail:
929    
930    Hi,
931    
932    Attached to this email you find a diff against sml/nj 110.45
933    that will enable socket support under Windows.
934    
935    To apply the patch (using unix or cygwin)
936    1) gunzip runtime.diff.gz
937    2) "cd" into "src/runtime" in the source tree of a fresh
938       110.45 installation.
939    3) patch -p 1 < [your/path/to]runtime.diff
940    
941    The code compiles fine but has NOT yet been extensively tested.
942    I only ran a few tests for basic socket client functionality
943    (which worked fine).  Especially the functions that use ioctl
944    are not tested at all and might not work (see below).
945    
946    I implemented this since we want to move to a newer version of sml/nj
947    but need socket support in order to use it.  This is the first time I
948    even had a look at the sml/nj source,  so please review my changes
949    before making this part of the distribution!  Here are a few issues
950    that I think might be better for someone to solve who is more
951    familiar with the sml/nj source (and socket programming):
952    
953    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
954      exception since I could not figure out what the windows equivalent
955      of these functions is
956    
957    - In sockets-osdep.h there are a some #include statements that are
958      only used in a few files that include sockets-osdep.h
959    
960    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
961      does not process its return value since I don't know how to
962      report an error upwards.
963    
964    - It would probably be good to have a call to WSACleanup() when
965      the library is unloaded (if there is such a possibility).
966      Otherwise I think Windows will take care of this automatically
967      when the process finishes.
968    
969    - I used ioctlsocket() as a replacement for ioctl() but I have
970      no idea if that is actually the proper replacement on Windows.
971    
972    - All these issues are marked in the code by "FIXME" comments.
973    
974    We use sml/nj extensively in our products and are quite happy
975    with it.  I hope this contribution will help you.
976    
977    Keep up the good work!
978    
979    David
980    
981    ----------------------------------------------------------------------
982    Name: Matthias Blume (blume (at) tti - c (dot) org)
983    Date: 2004/05/11 14:20:00 CDT
984    Tag: blume-20040511-installml
985    Description:
986    
987    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
988    the patch.)
989    
990    ----------------------------------------------------------------------
991    Name: Matthias Blume (blume (at) tti - c (dot) org)
992    Date: 2004/05/11 14:05:00 CDT
993    Tag: blume-20040511-nlffi-netbsd
994    Description:
995    
996    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
997    the patch.)
998    
999    ----------------------------------------------------------------------
1000    Name: Matthias Blume (blume (at) tti - c (dot) org)
1001    Date: 2004/05/11 12:05:00 CDT
1002    Tag: blume-20040511-exports
1003    Description:
1004    
1005    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1006    export lists in compiler-related .cm-files.
1007    
1008    ----------------------------------------------------------------------
1009    Name: Matthias Blume (blume (at) tti - c (dot) org)
1010    Date: 2004/05/11 11:35:00 CDT
1011    Tag: blume-20040511-allsource
1012    Description:
1013    
1014    The installer now honors the "src-smlnj" target again, although its meaning
1015    has changed from "all sources required for the compiler" to "all sources
1016    the installer knows about".  In other words, if you enable "src-smlnj"
1017    in the "targets" file, then the installer will pull in sources for
1018    everything.  (Notice that this refers to source code only.  Compiled
1019    code is still only installed for modules that were requested explicitly
1020    or which are required for other modules that were requested explicitly.)
1021    
1022    ----------------------------------------------------------------------
1023    Name: Matthias Blume (blume (at) tti - c (dot) org)
1024    Date: 2004/04/23 17:40:00 CDT
1025    Tag: blume-20040423-ieee-scan
1026    Description:
1027    
1028    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1029    in the exponent calculation we get INF or ZERO (depending on the mantissa
1030    and the sign of the exponent).
1031    
1032    ----------------------------------------------------------------------
1033    Name: Matthias Blume (blume (at) tti - c (dot) org)
1034    Date: 2004/04/23 10:40:00 CDT
1035    Tag: blume-20040423-ml-build
1036    Description:
1037    
1038    The ml-build script now terminates with a non-0 status when something
1039    goes wrong.
1040    
1041    ----------------------------------------------------------------------
1042    Name: Matthias Blume (blume (at) tti - c (dot) org)
1043    Date: 2004/04/22 16:35:00 CDT
1044    Tag: blume-20040422-Option
1045    Description:
1046    
1047    Made exception Option to be the same as exception Option.Option
1048    (as it should be).
1049    
1050    ----------------------------------------------------------------------
1051    Name: Allen Leung (leunga (at) reservoir (dot) com)
1052    Date: 2004/03/19 14:40:00 EST
1053    Tag: leunga-20040319-cygwin-nlffi
1054    Description:
1055    
1056    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1057    of SML/NJ.  The problem is that
1058    
1059       lib = dlopen(NULL, ...)
1060       f   = dlsym(lib, "malloc");
1061    
1062    does not work on Windows unless we explicitly export symbols
1063    such as 'malloc' during linking.  We fixed this by explicitly
1064    exporting the required symbols with the magic gcc incantation:
1065    
1066        -Wl,--export-all cygwin.def
1067    
1068    where cygwin.def is a file containing all the symbols that
1069    we wish to export.
1070    
1071    I suspect this is a Windows problem and we'll have to
1072    do the same (somehow with windows compilers) when
1073    we build the native win32 version with the system
1074    calls LoadLibrary/GetProcAddress.
1075    
1076    ----------------------------------------------------------------------
1077    Name: Matthias Blume (blume (at) tti - c (dot) org)
1078    Date: 2004/03/04 16:35:00 CST
1079    Tag: blume-20040304-intinf-fmt
1080    Description:
1081    
1082    Fixed problem with IntInf.fmt (sign would show up on the right instead
1083    of on the left for BIN, OCT, and HEX).
1084    
1085    ----------------------------------------------------------------------
1086    Name: Matthias Blume (blume (at) tti - c (dot) org)
1087    Date: 2004/03/04 11:25:00 CST
1088    Tag: blume-20040304-symlinks
1089    Description:
1090    
1091    Fixed problem with installer script (unix only) where bin/ml-yacc and
1092    friends pointed (via symlinks) to absolute locations instead of just
1093    .run-sml.  This was reported by Vesa A Norrman.
1094    
1095    ----------------------------------------------------------------------
1096    Name: Matthias Blume (blume (at) tti - c (dot) org)
1097    Date: 2004/02/13 14:50:00 CST
1098    Tag: Release_110_45
1099    Description:
1100    
1101    New working version (110.45). New bootfiles.
1102    
1103    ----------------------------------------------------------------------
1104    Name: Matthias Blume (blume (at) tti - c (dot) org)
1105    Date: 2004/01/26 15:15:15 CST
1106    Tag: blume-20040126-toplevel
1107    Description:
1108    
1109    Improved handling of exceptions at the interactive toplevel.
1110    
1111    ----------------------------------------------------------------------
1112    Name: Matthias Blume (blume (at) tti - c (dot) org)
1113    Date: 2004/01/26 11:25:00
1114    Tag: blume-20040126-app
1115    Description:
1116    
1117    Type of top-level "app" corrected.
1118    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1119    
1120    ----------------------------------------------------------------------
1121    Name: Matthias Blume (blume (at) tti - c (dot) org)
1122    Date: 2003/11/18 17:10 CST
1123    Tag: blume-20031118-basis-fiddle
1124    Description:
1125    
1126    - changed Timer interface to what might become the spec
1127    - POSIX_FLAGS -> BIT_FLAGS according to spec
1128    - some other minor discrepancies wrt. spec eliminated
1129    
1130    ----------------------------------------------------------------------
1131    Name: Matthias Blume (blume (at) tti - c (dot) org)
1132    Date: 2003/11/06 12:00:00 CST
1133    Tag: Release_110_44
1134    Description:
1135    
1136    New working version (110.44). New bootfiles.
1137    
1138    ----------------------------------------------------------------------
1139    Name: Matthias Blume (blume (at) tti - c (dot) org)
1140    Date: 2003/11/04 11:50:00 CST
1141    Tag: blume-20031104-move-libraries
1142    Description:
1143    
1144    Eliminated the "dont_move_libraries" directive in config/targets.
1145    (The mechanism was broken and could not be fixed easily.  Moreover,
1146    there does not seem to be any reason not to move all libraries into
1147    lib during installation.  I originally implemented this directive as a
1148    backward-compatibility feature when I first introduced the new CM.
1149    Now that things have been stable for a long time and going back to the
1150    old CM is not an option, there is no reason to keep it around.)
1151    
1152    ----------------------------------------------------------------------
1153    Name: Matthias Blume (blume (at) tti - c (dot) org)
1154    Date: 2003/11/03 16:00:00 CST
1155    Tag: blume-20031103-installdir
1156    Description:
1157    
1158    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1159    Richards for pointing out the problem and providing the solution.)
1160    
1161    ----------------------------------------------------------------------
1162    Name: Matthias Blume (blume (at) tti - c (dot) org)
1163    Date: 2003/10/01 17:05:00 CDT
1164    Tag: blume-20031001-lal-mlrisc
1165    Description:
1166    
1167    MLRISC bug fix from Lal.
1168    
1169    ----------------------------------------------------------------------
1170    Name: Matthias Blume (blume (at) tti - c (dot) org)
1171    Date: 2003/09/30 16:10:00 CDT
1172    Tag: blume-20030930-primio-bat
1173    Description:
1174    
1175    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1176    2. Improved .bat files (for Win32 port) to make things work under Win95.
1177       (thanks to Aaron S. Hawley for this one)
1178    
1179    ----------------------------------------------------------------------
1180    Name: Matthias Blume (blume (at) tti - c (dot) org)
1181    Date: 2003/09/26 16:05:00 CDT
1182    Tag: blume-20030926-wrappriv
1183    Description:
1184    
1185    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1186    
1187    ----------------------------------------------------------------------
1188    Name: Matthias Blume (blume (at) tti - c (dot) org)
1189    Date: 2003/09/26 15:00:00 CDT
1190    Tag: blume-20030926-110_43_3
1191    Description:
1192    
1193    - additional cleanup
1194    - version number bump, NEW BOOTFILES
1195    
1196    ----------------------------------------------------------------------
1197    Name: Matthias Blume (blume (at) tti - c (dot) org)
1198    Date: 2003/09/26 12:00:00 CDT
1199    Tag: blume-20030926-ppautoload
1200    Description:
1201    
1202    I modified the read-eval-print loop so that the autoloader gets
1203    invoked whenever the prettyprinter tries to look up a symbol that
1204    is not currently defined in the toplevel environment but which
1205    appears in CM's autoload registry.  As a result, we see far fewer of
1206    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1207    
1208    In addition to this I tried to clean up some pieces of the Basis
1209    implementation (e.g., Socket, Word8Array) in order to prevent other
1210    instances of these ?.Foo.Bar.xxx names from being printed.
1211    
1212    The mechanism that picks names for types still needs some work, though.
1213    (Right now it seems that if there is a type A.t which is defined to
1214    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1215    "?.B.u" although the perhaps more sensible solution would be to use
1216    "A.t" in this case.  In other words, the prettyprinter should follow
1217    a chain of DEFtycs not farther than there are corresponding toplevel
1218    names in the current environment.)
1219    
1220    ----------------------------------------------------------------------
1221    Name: Matthias Blume (blume (at) tti - c (dot) org)
1222    Date: 2003/09/24 16:31:00 CDT
1223    Tag: blume-20030924-installer
1224    Description:
1225    
1226    Another installer tweak:  All the ML code for the installer is now
1227    compiled during CMB.make and put into a little library called
1228    $smlnj/installer.cm.  The installation then simply invokes
1229    
1230       sml -m $smlnj/installer.cm
1231    
1232    and everything happens automagically.
1233    
1234    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1235    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1236            and BINDIR.
1237    
1238    The new scheme guarantees that the ML code responsible for the installation
1239    is in sync with the APIs of the main system.  Also, the installer is
1240    somewhat faster because the installer script is precompiled.
1241    
1242    ----------------------------------------------------------------------
1243    Name: Matthias Blume (blume (at) tti - c (dot) org)
1244    Date: 2003/09/24 15:35:00 CDT
1245    Tag: blume-20030924-synsock
1246    Description:
1247    
1248    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1249    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1250    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1251    implementations of non-blocking operations from CML's Socket
1252    structure.
1253    
1254    ----------------------------------------------------------------------
1255    Name: Matthias Blume (blume (at) tti - c (dot) org)
1256    Date: 2003/09/24 15:10:05 CDT
1257    Tag: blume-20030924-sockets
1258    Description:
1259    
1260    1. Fixed SOCKET API and implementation to match Basis spec.
1261       This required changing the internal representation of sockets to one
1262       that remembers (for each socket file descriptor) whether it is currently
1263       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1264       call is made only if the state actually needs to change).
1265    
1266    2. OS-specific details of sockets were moved into separate files, thus
1267       making it possible to unify the bulk of the socket implementations
1268       between Unix and Win32.
1269    
1270    3. CML's socket API changed accordingly.
1271       (Note that we need to remove non-blocking functions from this API
1272       since they are redundant in the case of CML!)
1273    
1274    4. CML's socket implementation now makes use of non-blocking functions
1275       provided by Basis, thus removing all OS-dependent code from this part
1276       of CML.
1277    
1278    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1279    
1280    ----------------------------------------------------------------------
1281    Name: Matthias Blume (blume (at) tti - c (dot) org)
1282    Date: 2003/09/22 12:10:00 CDT
1283    Tag: blume-20030922-110_43_2
1284    Description:
1285    
1286    Made a new interim version and bootfiles for developer's bootstrapping
1287    convenience.
1288    
1289       110.43.2 -- NEW BOOTFILES
1290    
1291    ----------------------------------------------------------------------
1292    Name: Matthias Blume (blume (at) tti - c (dot) org)
1293    Date: 2003/09/19 15:55:00 CDT
1294    Tag: blume-20030919-cmdir
1295    Description:
1296    
1297    1. new-install.sh -> install.sh
1298    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1299    3. tweaked installer so that another name instead of .cm can be chosen
1300       at install time (by setting the CM_DIR_ARC environment variable
1301       during installation); once installation is complete, the name is
1302       fixed
1303    
1304    ----------------------------------------------------------------------
1305    Name: Matthias Blume (blume (at) tti - c (dot) org)
1306    Date: 2003/09/18 16:00:00 CDT
1307    Tag: blume-20030918-110_43_1
1308    Description:
1309    
1310    Made a new interim version and bootfiles for developer's bootstrapping
1311    convenience.
1312    
1313       110.43.1 -- NEW BOOTFILES
1314    
1315    ----------------------------------------------------------------------
1316    Name: Matthias Blume (blume (at) tti - c (dot) org)
1317    Date: 2003/09/18 15:20:00 CDT
1318    Tag: blume-20030918-misc
1319    Description:
1320    
1321    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1322       this seems to be controversial at the moment) and used those in
1323       Posix.ProcEnv.times.
1324    
1325    2. Added Time.{from,to}Nanoseconds to Time.
1326    
1327    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1328       For example, fromLargeInt never needs to look at more than 3 "big
1329       digits" to get its 53 bits of precision.
1330    
1331    ----------------------------------------------------------------------
1332    Name: Matthias Blume (blume (at) tti - c (dot) org)
1333    Date: 2003/09/17 16:30:00 CDT
1334    Tag: blume-20030917-real32-slices
1335    Description:
1336    
1337    Added an entry to the primitive environment
1338    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1339    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1340    
1341    Removed some of the "magic" constants in real64.sml and replaced them
1342    with code that generates these values from their corresponding
1343    integer counterparts.
1344    
1345    Made all(?) the slice-related changes to the Basis and made everything
1346    compile again...
1347    
1348    ----------------------------------------------------------------------
1349    Name: Matthias Blume (blume (at) tti - c (dot) org)
1350    Date: 2003/09/15 17:45:00 CDT
1351    Tag: blume-20030915-rbase
1352    Description:
1353    
1354    Fixed bug in Real.fromLargeInt.
1355    
1356    ----------------------------------------------------------------------
1357    Name: Matthias Blume (blume (at) tti - c (dot) org)
1358    Date: 2003/09/13 18:11:00 CDT
1359    Tag: blume-20030913-libinstall
1360    Description:
1361    
1362    Minor bugfix in config/libinstall (set anchor with path to
1363    standalone tool after installing it, otherwise libraries that
1364    need ml-lex or ml-yacc won't compile the first time the installer
1365    runs).
1366    
1367    ----------------------------------------------------------------------
1368    Name: Matthias Blume (blume (at) tti - c (dot) org)
1369    Date: 2003/09/12 11:45:00 CDT
1370    Tag: blume-20030912-various
1371    Description:
1372    
1373     - fixed bug in Real.toLargeInt
1374     - fixed bug in Posix.ProcEnv.times
1375     - changed inputLine functions to return an option
1376     - minor installer improvements / bugfixes
1377     - changed default @SMLalloc parameter for x86/celeron to 64k
1378    
1379    ----------------------------------------------------------------------
1380    Name: Matthias Blume (blume (at) tti - c (dot) org)
1381    Date: 2003/09/09 22:00:00 CDT
1382    Tag: Release_110_43
1383    Description:
1384    
1385        New working release 110.43.  New bootfiles.
1386    
1387    ----------------------------------------------------------------------
1388    Name: Matthias Blume (blume (at) tti - c (dot) org)
1389    Date: 2003/09/09 19:20:00 CDT
1390    Tag: blume-20030909-installer
1391    Description:
1392    
1393    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1394    Modified config/install.bat to take advantage of it.  Also modified
1395    config/install.sh (and called it config/new-install.sh) to take advantage
1396    of it on Unix systems.  (The SML code is (supposed to be) platform-
1397    independent.)
1398    
1399    The installer can now install everything under Win32
1400    as well as under *nix as long as it compiles.
1401    
1402    Other changes:
1403    
1404      - made CML compile again under Win32
1405      - made eXene compile under Win32 (by providing a fake structure UnixSock
1406        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1407      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1408        same as type int;  under Win32 it isn't
1409      - fixed some slice-related problems in the win32-specific parts of CML
1410      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1411        (under Win32, the volume name is case-insensitive, and the
1412        OS.Path code compares volume names for equality)
1413    
1414    ----------------------------------------------------------------------
1415    Name: Matthias Blume (blume (at) tti - c (dot) org)
1416    Date: 2003/09/08 11:55:00 CDT
1417    Tag: blume-20030908-fullpath
1418    Description:
1419    
1420    Made Win32 version of OS.FileSys.fullPath return current directory
1421    when given an empty string.  This is what the spec says, and incidentally,
1422    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1423    cases when presented with the name of a non-existing .cm file.)
1424    
1425    ----------------------------------------------------------------------
1426    Name: Matthias Blume (blume (at) tti - c (dot) org)
1427    Date: 2003/09/04 16:30:00 CDT
1428    Tag: blume-20030905-slices-etc
1429    Description:
1430    
1431      1. Changed interface to vectors and arrays in Basis to match
1432         (draft) Basis spec.
1433      2. Added signatures and implementations of slices according to
1434         Basis spec.
1435      3. Edited source code throughout the system to make it compile again
1436         under 1. and 2.  (In some cases code had to be added to have it
1437         match the new signatures.)
1438      4. MLRISC should be backward-compatible: the copies of the originals
1439         of files that needed to change under 3. were retained, the .cm files
1440         check the compiler version number and use old versions when
1441         appropriate.
1442      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1443         dirstream -> string option (in accordance with Basis spec).
1444      6. When generating code that counts lines, ml-lex used function
1445         CharVector.foldli, taking advantage of its old interface.
1446         This has been replaced with the corresponding code from
1447         CharVectorSlice.  (html-lex must be re-lexed!)
1448      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1449         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1450      8. Removed temporary additions (fromInternal, toInternal) from the
1451         (now obsolete) IntInf in smlnj-lib/Util.
1452      9. Cleaned up structure Byte.
1453     10. Added localOffset, scan, and fromString to Date (according to spec).
1454         Cleaned/corrected implementation of Date.
1455         (Still need to check for correctness; implement better canonicalizeDate.)
1456     11. Added "scan" to signature IEEE_REAL.
1457     12. Some improvements to IntInf [in particular: efficiency-hack for
1458         mod and rem when second operand is 2 (for parity checks).]
1459     13. Changed representation of type Time.time, using a single IntInf.int
1460         value counting microseconds.  This considerably simplified the
1461         implementation of structure Time.  We now support negative time
1462         values; scan and fromString handle signs.
1463     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1464         ArraySlice).
1465    
1466    ----------------------------------------------------------------------
1467    Name: Matthias Blume (blume (at) tti - c (dot) org)
1468    Date: 2003/08/28 17:00:00 CDT
1469    Tag: blume-20030828-intinf
1470    Description:
1471    
1472    This is a major update which comes with a version number bump
1473    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1474    and an implementation of IntInf in the Basis.
1475    
1476    There are a fairly large number of related changes and updates throughout
1477    the system:
1478    
1479    Basis:
1480            - Implemented IntInf.
1481            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1482            - Added some missing Real64 operations, most notably Real.toLargeInt.
1483            - Added FixedInt as a synonym for Int32.
1484    
1485    compiler:
1486            * Added support for a built-in intinf type.
1487              - literals
1488              - pattern matching
1489              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1490              - overloading on literals and operations
1491    
1492            This required adding a primitive type intinf, some additional
1493            primops, and implementations for several non-trivial intinf
1494            operations in Core.  (The intinf type is completely abstract
1495            to the compiler; all operations get delegated back to the Core.)
1496    
1497            * Intinf equality is handled by polyequal.  However, the compiler
1498            does not print its usual warning in this case (since polyequal
1499            is the right thing to do there).
1500    
1501            * Improved the organization of structure InlineT.
1502    
1503            * A word about conversion primops:
1504            If conversions involving intinf do not cancel out during
1505            CPS contract, then the compiler must insert calls to Core functions.
1506            Since all core access must be resolved already during the FLINT
1507            translate phase, it would be too late a the time of CPS contract
1508            to add new Core calls.  For this reason, conversion primops
1509            for intinf carry two arguments:  1. the numeric argument that
1510            they are supposed to convert, and 2. the Core function that
1511            can help with this conversion if necessary.  If CPS contract
1512            eliminates a primop, then the associated Core function becomes
1513            dead and goes away.  Intinf conversion primops that do not get
1514            eliminated by CPS contract get rewritten into calls of their
1515            core functions by a separate, new phase.
1516    
1517    interactive system:
1518            - Control.Print.intinfDepth controls max length of intinf constants
1519            being printed. (Analogous to Control.Print.stringDepth.)
1520            - Cleanup in printutil and pputil: got rid of unused stuff and
1521            duplicates; replaced some of the code with code that makes better
1522            use of library functionality.
1523    
1524    CM:
1525            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1526            are no longer silent.
1527    
1528    CKIT:
1529            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1530            in favor of LargeInt -- which is now the same as IntInf.
1531            CKIT-knowledgable people should check whether this is what's
1532            intended and otherwise change things back to using Int32 or
1533            FixedInt.
1534    
1535    Throughout the code:
1536            Started using IntInf.int literals and built-in operations
1537            (e.g., comparison with 0) where this seems appropriate.
1538    
1539    
1540    ----------------------------------------------------------------------
1541    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1542    Date: 2003/08/13 11:36:00 CDT
1543    Tag: dbm-20030813-mcz-merge1
1544    Description:
1545    
1546    Merging changes from the mcz-branch development branch into trunk.
1547    These changes involve replacement of the emulated old prettyprinter
1548    interface with direct use of the SML/NJ Lib PP library, and fixing
1549    of a couple of bugs (895, 1186) relating to error messages.  A new
1550    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1551    has been added.
1552    
1553    ----------------------------------------------------------------------
1554    Name: Matthias Blume (blume (at) tti - c (dot) org)
1555    Date: 2003/08/11 15:45:00 CDT
1556    Tag: blume-20030811-windows
1557    Description:
1558    
1559    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1560    
1561        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1562    
1563    This patch restores SML/NJ's ability to run under win32.  There are a
1564    number of changes, including fixes for several bugs that had gone
1565    unnoticed until now:
1566    
1567      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1568        trying to run the win32 version from within cygwin.)
1569    
1570      - There are a number of simple .bat scripts that substitute for their
1571        corresponding Unix shell-scripts.  (See below.)
1572    
1573      - The internals of ml-build have been modified slightly.  The main
1574        difference is that instead of calling ".link-sml" (or link-sml.bat)
1575        using OS.Process.system, the ML process delegates this task back
1576        to the script.  Otherwise problems arise in mixed environments such
1577        as Cygwin where scripts look and work like Unix scripts, but
1578        where OS.Process.system cannot run them.
1579    
1580      - In CM, the srcpath pickler used native pathname syntax -- which
1581        is incorrect in the case of cross-compilation.  The new pickle format
1582        is independent of platform-specific naming conventions.
1583    
1584      - Path configuration files (such as lib/pathconfig) can now choose
1585        between native and standard syntax.  Placing a line of the form
1586    
1587            standard!
1588    
1589        into the file causes all subsequent paths to be interpreted using
1590        CM standard pathname syntax (= Unix conventions); a line
1591    
1592            native!
1593    
1594        switches back to native style.  This was needed so that
1595        path config files can be written portably, see src/system/pathconfig.
1596    
1597      - Runtime system:
1598    
1599          - win32-filesys.c:  get_file_time and set_file_time now
1600            access modification time, not creation time.
1601    
1602          - I/O code made aware of new array representation.
1603    
1604          - Bug fixes in X86.prim.masm.
1605    
1606       - src/system/makeml made aware of win32. (For use under cygwin
1607            and other Unix-environments for windows.)
1608    
1609       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1610         which caused BinIO.inputAll to fail consistently.
1611    
1612    .bat scripts:
1613    
1614       Windows .bat scripts assume that SMLNJ_HOME is defined.
1615    
1616       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1617         applications (sml, ml-yacc, ml-lex).
1618       - ml-build.bat: analogous to ml-build.
1619       - config\install.bat:  Analogous to config/install.sh.  This requires
1620         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1621         (nmake etc. must be on the path, and vcvars32 must have been run.)
1622         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1623         and the bootfile hierarchy must have been unpacked under
1624         sml.boot.x86-win32.
1625         The script is very primitive and does a poor job at error checking.
1626         It only installs the base system, ml-lex, and ml-yacc.  No other
1627         libraries are being installed (i.e., you get only those that
1628         are part of the compiler.)
1629       - link-sml.bat:  analogous to .link-sml, but not currently used
1630    
1631    Unrelated bug fixes:
1632    
1633      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1634        types.
1635      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1636        otherwise.)
1637    
1638    ----------------------------------------------------------------------
1639    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1640    Date: 2003/06/17
1641    Tag: macqueen-20030617-bug895
1642    Description:
1643    
1644    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1645    Tag will be used for new development branch (mcz-branch) for
1646    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1647    
1648    ----------------------------------------------------------------------
1649    Name: Matthias Blume (blume (at) tti - c (dot) org)
1650    Date: 2003/05/27 16:55:00 CDT
1651    Tag: blume-20030527-polyeq
1652    Description:
1653    
1654    Tried to eliminated most cases of polymorphic equality.
1655    
1656    ----------------------------------------------------------------------
1657    Name: Matthias Blume (blume (at) tti - c (dot) org)
1658    Date: 2003/05/21 17:45:00 CDT
1659    Tag: blume-20030517-complete
1660    Description:
1661    
1662    Two changes:
1663    
1664      1. Added a flag for controlling whether non-exhaustive bindings will
1665         be treated as errors (default is false).
1666      2. Cleaned up the *entire* source tree so that CMB.make goes through
1667         without a single non-exhaustive match- or bind warning.
1668    
1669    ----------------------------------------------------------------------
1670    Name: Matthias Blume (blume (at) tti - c (dot) org)
1671    Date: 2003/05/17 10:20:00 CDT
1672    Tag: blume-20030517-absyn
1673    Description:
1674    
1675    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1676    
1677       This mainly affects the quality of error messages.  However, some
1678       of the code is now more straightforward than before. (Treatment of
1679       the above four constructs in translate.sml is much simpler than
1680       the "macro-expansion" that was going on before.  Plus, the mach-
1681       compiler no longer gets invoked just to be able to compile an
1682       if-expression.)
1683    
1684    2. The ErrorMsg.Error exception is now caught and absorbed by the
1685       interactive loop.
1686    
1687    ----------------------------------------------------------------------
1688    Name: Allen Leung
1689    Date: 2003/05/16 13:05:00 CDT
1690    Tag: leunga-20030516-cygwin-runtime
1691    Description:
1692    
1693        Ported the runtime system to cygwin, which uses the unix
1694    x86-unix bin files.  Missing/buggy features:
1695    
1696       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1697         the Cygwin library.
1698       o Ctrl-C handling may be flaky.
1699       o Windows system calls and Windows I/O are not supported.
1700    
1701       A new set of binfiles is located at:
1702    
1703         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1704    
1705       This is only needed for bootstrapping the cygwin version of smlnj.
1706       Other x86 versions can use the existing binfiles.
1707    
1708    ----------------------------------------------------------------------
1709    Name: Matthias Blume
1710    Date: 2003/04/08 15:42:00 CDT
1711    Tag: blume-20030408-listpair
1712    Description:
1713    
1714    1. Added a target 'mlrisc' to installer.
1715    
1716    2. Added missing elements to structure ListPair.
1717    
1718    ----------------------------------------------------------------------
1719    Name: Allen Leung
1720    Date: 2003/01/07 10:40:00 EST
1721    Tag: leunga-20030107-int-rem
1722    Description:
1723    
1724    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1725    The arguments to the SUBL instruction were swapped.
1726    
1727    ----------------------------------------------------------------------
1728    Name: Matthias Blume
1729    Date: 2002/12/12 16:25:00 EST
1730    Tag: blume-20021212-risc-ra
1731    Description:
1732    
1733    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1734    sent the RA into an infinite loop when floating point registers get
1735    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1736    110.37 and 110.38.)
1737    There was another set of potential problems related to the handling of
1738    MLRISC annotations (but those did not yet cause real problems, apparently).
1739    
1740    ----------------------------------------------------------------------
1741    Name: Matthias Blume
1742    Date: 2002/12/06 22:40:00 EST
1743    Tag: blume-20021206-cm-fileid
1744    Description:
1745    
1746    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1747    This fixes the problem of CM getting confused by files that suddenly
1748    change their identity (e.g., by getting unlinked and recreated by some
1749    text editor such as vi).  There might be a better/cheaper/cleaner way
1750    of doing this, but for now this will have to do.
1751    
1752    ----------------------------------------------------------------------
1753    Name: Matthias Blume
1754    Date: 2002/10/28 09:50:00 EST
1755    Tag: blume-20021028-typecheck
1756    Description:
1757    
1758    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1759    
1760    ----------------------------------------------------------------------
1761    Name: Matthias Blume
1762    Date: 2002/10/17 09:10:00 EDT
1763    Tag: Release_110_42
1764    Description:
1765    
1766    In good old tradition, there has been a slight hiccup so that we have
1767    to patch 110.42 after the fact.   The old release tag has been replaced
1768    (see below).
1769    
1770    The change solves a problem with two competing approaches the
1771    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1772    each other's way.
1773    
1774    This change only affects the runtime system code and the installer script.
1775    (No new bootfiles.)
1776    
1777    ----------------------------------------------------------------------
1778    Name: Matthias Blume
1779    Date: 2002/10/16 12:00:00 EDT
1780    Tag: Release_110_42_removed
1781    Description:
1782    
1783    New working release.  New bootfiles.
1784    
1785    ----------------------------------------------------------------------
1786    Name: Matthias Blume
1787    Date: 2002/10/10 13:10:00 EDT
1788    Tag: blume-20021010-ppc-divs
1789    Description:
1790    
1791    The mltree operator DIVS must be implemented with an overflow check on
1792    the PPC because the hardware indicates divide-by-zero using "overflow" as
1793    well.
1794    
1795    ----------------------------------------------------------------------
1796    Name: Matthias Blume
1797    Date: 2002/07/23 11:45:00
1798    Tag: blume-20020723-smlnj-home
1799    Description:
1800    
1801    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1802    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1803    CM_PATHCONFIG is also set), the path configuration file is assumed
1804    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1805    the entire tree to some other place and everything will "just work".
1806    
1807    (Companion commands such as ml-build and ml-makedepend also sense this
1808    variable.)
1809    
1810    ----------------------------------------------------------------------
1811    Name: Matthias Blume
1812    Date: 2002/07/12 21:19:00 EDT
1813    Tag: blume-20020712-liveness
1814    Description:
1815    
1816    Exported two useful "step" functions from liveness module (MLRISC).
1817    
1818    ----------------------------------------------------------------------
1819    Name: Matthias Blume
1820    Date: 2002/07/05 16:00 EDT
1821    Tag: Release_110_41
1822    Description:
1823    
1824    New working release.  New bootfiles.
1825    
1826    ----------------------------------------------------------------------
1827    Name: Matthias Blume
1828    Date: 2002/07/05 10:25:00 EDT
1829    Tag: blume-20020705-btimp
1830    Description:
1831    
1832    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1833    other clients can set up backtracing support.
1834    
1835    ----------------------------------------------------------------------
1836    Name: Matthias Blume
1837    Date: 2002/06/25 17:23:00 EDT
1838    Tag: blume-20020625-fpmax
1839    Description:
1840    
1841    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1842    case.  (The sense of the isNaN test was reversed -- which made min and
1843    max always return their first argument.)
1844    
1845    ----------------------------------------------------------------------
1846    Name: Matthias Blume
1847    Date: 2002/06/11
1848    Tag: blume-20020611-unixpath
1849    Description:
1850    
1851    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1852    
1853    ----------------------------------------------------------------------
1854    Name: Matthias Blume
1855    Date: 2002/06/10 16:35:00 EDT
1856    Tag: blume-20020610-ieeereal
1857    Description:
1858    
1859    I back-ported my implementation of IEEEReal.fromString from the
1860    idlbasis-devel branch so that we can test it.
1861    
1862    Another small change is that ppDec tries to give more information
1863    than just "<sig>" in the case of functors.  However, this code is
1864    broken in some mysterious way if the functor's body's signature
1865    has not been declared by ascription but gets inferred from the
1866    implementation.  This needs fixing...
1867    
1868    ----------------------------------------------------------------------
1869    Name: Matthias Blume
1870    Date: 2002/05/31
1871    Tag: blume-20020531-btrace-mode
1872    Description:
1873    
1874    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1875    the wayside when I switched over to using Controls everywhere.)
1876    
1877    ----------------------------------------------------------------------
1878    Name: Lal George
1879    Date: 2002/05/23 12:21:40 EDT
1880    Tag: george-20020523-visual-labels
1881    Description:
1882    
1883    Labels are now displayed in the graphical output to make
1884    the fall-through and target blocks obvious.
1885    
1886    ----------------------------------------------------------------------
1887    Name: Matthias Blume
1888    Date: 2002/05/22 11:03:00 EDT
1889    Tag: blume-20020522-shrink
1890    Description:
1891    
1892    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1893    situation that comes up when allocating huge arrays.
1894    
1895    ----------------------------------------------------------------------
1896    Name: Matthias Blume
1897    Date: 2002/05/21 16:00:00 EDT
1898    Tag: Release_110_40
1899    Description:
1900    
1901    New working release (110.40).  New bootfiles.
1902    
1903    [Also: John Reppy fixed GC bug 1131.]
1904    
1905    ----------------------------------------------------------------------
1906    Name: Matthias Blume
1907    Date: 2002/05/21 12:35:00 EDT
1908    Tag: blume-20020521-cmdoc
1909    Description:
1910    
1911    CM documentation update.
1912    
1913    ----------------------------------------------------------------------
1914    Name: Matthias Blume
1915    Date: 2002/05/21 10:55:00 EDT
1916    Tag: blume-20020521-misc
1917    Description:
1918    
1919    - John tweaked runtime to be silent on heap export (except when
1920      GC messages are on).
1921    - I added a few more things (cross-compiling versions of CMB) to
1922      config/preloads (as suggestions).
1923    
1924    ----------------------------------------------------------------------
1925    Name: Matthias Blume
1926    Date: 2002/05/20 22:25:00 EDT
1927    Tag: blume-20020520-controls
1928    Description:
1929    
1930    - Added ControlUtil structure to control-lib.cm.
1931    - Use it throughout.
1932    - Used Controls facility to define MLRISC controls (as opposed to
1933      registering MLRISC control ref cells with Controls after the fact)
1934    - Fixed messed-up controls priorities.
1935    
1936    * Removed again all the stuff from config/preloads that one wouldn't
1937      be able to preload at the time the initial heap image is built.
1938      (Many libraries, e.g., CML, do not exist yet at this time.  The
1939       only libraries that can be preloaded via config/preloads are those
1940       that come bundled with the bootfiles.)
1941    
1942    ----------------------------------------------------------------------
1943    Name: Matthias Blume
1944    Date: 2002/05/20 10:59:00 EDT
1945    Tag: blume-20020520-preloads
1946    Description:
1947    
1948    Added a lot of commented-out suggestions for things to be included
1949    in config/preloads.
1950    
1951    ----------------------------------------------------------------------
1952    Name: Allen Leung
1953    Date: 2002/05/18 14:20:00 EDT
1954    Tag: leunga-20020518-mdl
1955    Description:
1956    
1957    o Made the mdl tool stuff compile and run again.
1958    o I've disabled all the stuff that depends on RTL specifications; they
1959      are all badly broken anyway.
1960    
1961    ----------------------------------------------------------------------
1962    Name: Matthias Blume
1963    Date: 2002/05/17 16:49:00 EDT
1964    Tag: blume-20020517-controls
1965    Description:
1966    
1967    0. John Reppy made several modifications to the SML/NJ library.
1968       In particular, there is a shiny new controls-lib.cm.
1969    
1970    1. Pushed new controls interface through compiler so that everything
1971       compiles again.
1972    
1973    2. Added FormatComb and FORMAT_COMB to the CML version of the
1974       SML/NJ library (so that CML compiles again).
1975    
1976    3. Modified init scripts because XXX_DEFAULT environment variables
1977       are no longer with us.  (Boot-time initialization is now done
1978       using the same environment variables that are also used for
1979       startup-time initialization of controls.)
1980    
1981    ----------------------------------------------------------------------
1982    Name: Lal George
1983    Date: 2002/05/15 09:20:10 EDT
1984    Tag: george-20020515-pseudo-op-decls
1985    Description:
1986    
1987            All pseudo-ops emitted before the first segment declaration
1988    such as TEXT, DATA, and BSS directives are assumed to be global
1989    declarations and are emitted first in the assembly file. This is
1990    useful in a number of situations where one has pseudo-ops that are not
1991    specific to any segment, and also works around the constraint that one
1992    cannot have client pseudo-ops in the TEXT segment.
1993    
1994            Because no segment is associated with these declarations it is
1995    an error to allocate any space or objects before the first segment
1996    directive and an exception will be raised. However, we cannot make
1997    this check for client pseudo-ops.
1998    
1999            These top level declarations are a field in the CFG graph_info.
2000    In theory you can continue to add to this field after the CFG has been
2001    built -- provided you know what you are doing;-)
2002    
2003    ----------------------------------------------------------------------
2004    Name: Matthias Blume
2005    Date: 2002/05/13 16:40:00 EDT
2006    Tag: blume-20020513-pp-etc
2007    Description:
2008    
2009    A few minor bugfixes:
2010    
2011      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2012        (Bogus printouts for redefined bindings still occur.  Compiler
2013        bug should no longer occur now.  We need to redo the prettyprinter
2014        from scratch.)
2015    
2016      - CM pathname printer now also adds escape sequences for ( and )
2017    
2018      - commend and docu fixes for ml-nlffi
2019    
2020    ----------------------------------------------------------------------
2021    Name: Matthias Blume
2022    Date: 2002/05/10 16:40:00 EDT
2023    Tag: blume-20020510-erg-textio
2024    Description:
2025    
2026    Applied the following bugfix provided by Emden Gansner:
2027    
2028        Output is corrupted when outputSubstr is used rather than output.
2029    
2030        The problem occurs when a substring
2031    
2032            ss = (s, dataStart, dataLen)
2033    
2034        where dataStart > 0, fills a stream buffer with avail bytes left.
2035        avail bytes of s, starting at index dataStart, are copied into the
2036        buffer, the buffer is flushed, and then the remaining dataLen-avail
2037        bytes of ss are copied into the beginning of the buffer. Instead of
2038        starting this copy at index dataStart+avail in s, the current code
2039        starts the copy at index avail.
2040    
2041        Fix:
2042        In text-io-fn.sml, change line 695 from
2043             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2044        to
2045             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2046    
2047    ----------------------------------------------------------------------
2048    Name: Matthias Blume
2049    Date: 2002/04/12 13:55:00 EDT
2050    Tag: blume-20020412-assyntax
2051    Description:
2052    
2053    1. Grabbed newer assyntax.h from the XFree86 project.
2054    2. Fiddled with how to compile X86.prim.asm without warnings.
2055    3. (Very) Minor cleanup in CM.
2056    
2057    ----------------------------------------------------------------------
2058    Name: Matthias Blume
2059    Date: 2002/04/01 (no joke!) 17:07:00 EST
2060    Tag: blume-20020401-x86div
2061    Description:
2062    
2063    Added full support for div/mod/rem/quot on the x86, using the machine
2064    instruction's two results (without clumsily recomputing the remainder)
2065    directly where appropriate.
2066    
2067    Some more extensive power-of-two support was added to the x86 instruction
2068    selector (avoiding expensive divs, mods, and muls where they can be
2069    replaced with cheaper shifts and masks).  However, this sort of thing
2070    ought to be done earlier, e.g., within the CPS optimizer so that
2071    all architectures benefit from it.
2072    
2073    The compiler compiles to a fixed point, but changes might be somewhat
2074    fragile nevertheless.  Please, report any strange things that you might
2075    see wrt. div/mod/quot/rem...
2076    
2077    ----------------------------------------------------------------------
2078    Name: Matthias Blume
2079    Date: 2002/03/29 17:22:00
2080    Tag: blume-20020329-div
2081    Description:
2082    
2083    Fixed my broken div/mod logic.  Unfortunately, this means that the
2084    inline code for div/mod now has one more comparison than before.
2085    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2086    The problem was with quotient = 0, because that alone does not tell
2087    us which way the rounding went.  One then has to look at whether
2088    remainder and divisor have the same sign...  :(
2089    
2090    Anyway, I replaced the bootfiles with fresh ones...
2091    
2092    ----------------------------------------------------------------------
2093    Name: Matthias Blume
2094    Date: 2002/03/29 14:10:00 EST
2095    Tag: blume-20020329-inlprims
2096    Description:
2097    
2098    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2099    
2100    Primops have changed. This means that the bin/boot-file formats have
2101    changed as well.
2102    
2103    To make sure that there is no confusion, I made a new version.
2104    
2105    
2106    CHANGES:
2107    
2108    * removed REMT from mltree (remainder should never overflow).
2109    
2110    * added primops to deal with divisions of all flavors to the frontend
2111    
2112    * handled these primops all the way through so they map to their respective
2113      MLRISC support
2114    
2115    * used these primops in the implementation of Int, Int32, Word, Word32
2116    
2117    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2118    
2119    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2120    
2121    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2122      floating point
2123    
2124    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2125      and Real (but Real.abs maps to a separate floating-point-only primop)
2126    
2127    
2128    TODO items:
2129    
2130    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2131      instructions because the machine instruction encoder cannot handle
2132      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2133      REM in the first place, and REM is supported by the x86 machine coder.)
2134    
2135    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2136      selection phase.  (The two can be streamlined because the hardware
2137      delivers both quotient and remainder at the same time anyway.)
2138    
2139    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2140      (Currently the behavior is inconsistent both across architectures and
2141      wrt. the draft Basis spec.)
2142    
2143    * Word8 should eventually be handled natively, too.
2144    
2145    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2146      as if there currently is no execution path that could trigger it in
2147      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2148      hold for things like multiplication and division.)
2149    
2150    ----------------------------------------------------------------------
2151    Name: Matthias Blume
2152    Date: 2002/03/27 16:27:00 EST
2153    Tag: blume-20020327-mlrisc-divisions
2154    Description:
2155    
2156    Added support for all four division operations (ML's div, mod, quot,
2157    and rem) to MLRISC.  In the course of doing so, I also rationalized
2158    the naming (no more annoying switch-around of DIV and QUOT), by
2159    parameterizing the operation by div_rounding_mode (which can be either
2160    DIV_TO_ZERO or DIV_TO_NEGINF).
2161    
2162    The generic MLTreeGen functor takes care of compiling all four
2163    operations down to only round-to-zero div.
2164    
2165    Missing pieces:
2166    
2167      * Doing something smarter than relying on MLTreeGen on architectures
2168        like, e.g., the x86 where hardware division delivers both quotient and
2169        remainder at the same time.  With this, the implementation of the
2170        round-to-neginf operations could be further streamlined.
2171    
2172      * Remove inlining support for div/mod/rem from the frontend and replace it
2173        with primops that get carried through to the backend.  Do this for all
2174        int and word types.
2175    
2176    ----------------------------------------------------------------------
2177    Name: Matthias Blume
2178    Date: 2002/03/25 17:25:00 EST
2179    Tag: blume-20020325-divmod
2180    Description:
2181    
2182    I improved (hopefully without breaking them) the implementation of Int.div,
2183    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2184    advantage of the following observations:
2185    
2186      Let  q = x quot y      r = x rem y
2187           d = x div  y      m = x mod y
2188    
2189    where "quot" is the round-to-zero version of integer division that
2190    hardware usually provides.  Then we have:
2191    
2192         r = x - q * y        where neither the * nor the - will overflow
2193         d = if q >= 0 orelse x = q * y then q else q - 1
2194                              where neither the * nor the - will overflow
2195         m = if q >= 0 orelse r = 0 then r else r + y
2196                              where the + will not overflow
2197    
2198    This results in substantial simplification of the generated code.
2199    The following table shows the number of CFG nodes and edges generated
2200    for
2201            fun f (x, y) = x OPER y
2202            (* with OPER \in div, mod, quot, rem *)
2203    
2204    
2205        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2206        --------------------------------------------------------
2207         div |         24 |         39 |         12 |         16
2208         mod |         41 |         71 |         12 |         16
2209        quot |          8 |         10 |          8 |         10
2210         rem |         10 |         14 |          8 |         10
2211    
2212    
2213    ----------------------------------------------------------------------
2214    Name: Matthias Blume
2215    Date: 2002/03/25 22:06:00 EST
2216    Tag: blume-20020325-cprotobug
2217    Description:
2218    
2219    Fixed a bug in cproto (c prototype decoder).
2220    
2221    ----------------------------------------------------------------------
2222    Name: Matthias Blume
2223    Date: 2002/03/25 16:00:00 EST
2224    Tag: blume-20020325-raw-primops
2225    Description:
2226    
2227    I did some cleanup to Allen's new primop code and
2228    replaced yesterday's bootfiles with new ones.
2229    (But they are stored in the same place.)
2230    
2231    ----------------------------------------------------------------------
2232    Name: Matthias Blume
2233    Date: 2002/03/24 22:40:00 EST
2234    Tag: blume-20020324-bootfiles
2235    Description:
2236    
2237    Made the bootfiles that Allen asked for.
2238    
2239    ----------------------------------------------------------------------
2240    Name: Allen Leung
2241    Date: 2002/03/23 15:50:00 EST
2242    Tag: leunga-20020323-flint-cps-rcc-primops
2243    Description:
2244    
2245      1. Changes to FLINT primops:
2246    
2247        (* make a call to a C-function;
2248         * The primop carries C function prototype information and specifies
2249         * which of its (ML-) arguments are floating point. C prototype
2250         * information is for use by the backend, ML information is for
2251         * use by the CPS converter. *)
2252      | RAW_CCALL of { c_proto: CTypes.c_proto,
2253                       ml_args: ccall_type list,
2254                       ml_res_opt: ccall_type option,
2255                       reentrant : bool
2256                     } option
2257       (* Allocate uninitialized storage on the heap.
2258        * The record is meant to hold short-lived C objects, i.e., they
2259        * are not ML pointers.  With the tag, the representation is
2260        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2261        *)
2262      | RAW_RECORD of {tag:bool,sz:int}
2263      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2264    
2265      2.  These CPS primops are now overloaded:
2266    
2267           rawload of {kind:numkind}
2268           rawstore of {kind:numkind}
2269    
2270          The one argument form is:
2271    
2272             rawload {kind} address
2273    
2274          The two argument form is:
2275    
2276             rawload {kind} [ml object, byte-offset]
2277    
2278      3. RAW_CCALL/RCC now takes two extra arguments:
2279    
2280         a. The first is whether the C call is reentrant, i.e., whether
2281            ML state should be saved and restored.
2282         b. The second argument is a string argument specifying the name of
2283            library and the C function.
2284    
2285         These things are currently not handled in the code generator, yet.
2286    
2287      4. In CProto,
2288    
2289         An encoding type of "bool" means "ml object" and is mapped into
2290         C prototype of PTR.  Note that "bool" is different than "string",
2291         even though "string" is also mapped into PTR, because "bool"
2292         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2293    
2294      5. Pickler/unpicker
2295    
2296         Changed to handle RAW_RECORD and newest RAW_CCALL
2297    
2298      6. MLRiscGen,
2299    
2300         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2301         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2302            in the file CodeGen/cpscompile/cps-c-calls.sml
2303    
2304      7. Added the conditional move operator
2305    
2306             condmove of branch
2307    
2308         to cps.  Generation of this is still buggy so it is currently
2309         disabled.
2310    
2311    ----------------------------------------------------------------------
2312    Name: Lal George
2313    Date: 2002/03/22 14:18:25 EST
2314    Tag: george-20020322-cps-branch-prob
2315    Description:
2316    
2317    Implemented the Ball-Larus branch prediction-heuristics, and
2318    incorporated graphical viewers for control flow graphs.
2319    
2320    Ball-Larus Heuristics:
2321    ---------------------
2322    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2323    
2324    By design it uses the Dempster-Shafer theory for combining
2325    probabilities.  For example, in the function:
2326    
2327        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2328    
2329    the ball-larus heuristics predicts that the n=0 is unlikely
2330    (OH-heuristic), and the 'then' branch is unlikely because of the
2331    RH-heuristic -- giving the 'then' branch an even lower combined
2332    probability using the Dempster-Shafer theory.
2333    
2334    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2335    probability of the 'then' branch because of the loop in the else
2336    branch.
2337    
2338    
2339    Graphical Viewing:
2340    ------------------
2341    I merely plugged in Allen's graphical viewers into the compiler. The
2342    additional code is not much. At the top level, saying:
2343    
2344            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2345    
2346    will display the graphical view of the control flow graph just before
2347    back-patching.  daVinci must be in your path for this to work. If
2348    daVinci is not available, then the default viewer can be changed
2349    using:
2350    
2351            Control.MLRISC.getString "viewer"
2352    
2353    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2354    course, these viewers must be in your path.
2355    
2356    The above will display the compilation unit at the level of clusters,
2357    many of which are small, boring, and un-interesting. Also setting:
2358    
2359            Control.MLRISC.getInt "cfg-graphical-view_size"
2360    
2361    will display clusters that are larger than the value set by the above.
2362    
2363    
2364    ----------------------------------------------------------------------
2365    Name: Matthias Blume
2366    Date: 2002/03/21 22:20:00 EST
2367    Tag: blume-20020321-kmp-bugfix
2368    Description:
2369    
2370    Changed the interface to the KMP routine in PreString and fixed
2371    a minor bug in one place where it was used.
2372    
2373    ----------------------------------------------------------------------
2374    Name: Allen Leung
2375    Date: 2002/03/21 20:30:00 EST
2376    Tag: leunga-20020321-cfg
2377    Description:
2378    
2379      Fixed a potential problem in cfg edge splitting.
2380    
2381    ----------------------------------------------------------------------
2382    Name: Allen Leung
2383    Date: 2002/03/21 17:15:00 EST
2384    Tag: leunga-20020321-x86-fp-cfg
2385    Description:
2386    
2387      1. Recoded the buggy parts of x86-fp.
2388    
2389         a. All the block reordering code has been removed.
2390            We now depend on the block placement phases to do this work.
2391    
2392         b. Critical edge splitting code has been simplified and moved into the
2393            CFG modules, as where they belong.
2394    
2395         Both of these were quite buggy and complex.  The code is now much, much
2396         simpler.
2397    
2398      2. X86 backend.
2399    
2400         a. Added instructions for 64-bit support.  Instruction selection for
2401            64-bit has not been committed, however, since that
2402            requires changes to MLTREE which haven't been approved by
2403            Lal and John.
2404    
2405         b. Added support for FUCOMI and FUCOMIP when generating code for
2406            PentiumPro and above.  We only generate these instructions in
2407            the fast-fp mode.
2408    
2409         c. Added cases for JP and JNP in X86FreqProps.
2410    
2411      3. CFG
2412    
2413         CFG now has a bunch of methods for edge splitting and merging.
2414    
2415      4. Machine description.
2416    
2417         John's simplification of MLTREE_BASIS.fcond broke a few machine
2418         description things:
2419    
2420         rtl-build.{sig,sml} and hppa.mdl fixed.
2421    
2422         NOTE: the machine description stuff in the repository is still broken.
2423               Again, I can't put my fixes in because that involves
2424               changes to MLTREE.
2425    
2426    ----------------------------------------------------------------------
2427    Name: Matthias Blume
2428    Date: 2002/03/20 15:55:00 EST
2429    Tag: blume-20020320-kmp
2430    Description:
2431    
2432    Implemented Knuth-Morris-Pratt string matching in PreString and used
2433    it for String.isSubstring, Substring.isSubstring, and
2434    Substring.position.
2435    
2436    (Might need some stress-testing.  Simple examples worked fine.)
2437    
2438    ----------------------------------------------------------------------
2439    Name: Matthias Blume
2440    Date: 2002/03/19 16:37:00 EST
2441    Tag: blume-20020319-witnesses
2442    Description:
2443    
2444    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2445    
2446    This implements a generic mechanism for changing constness qualifiers
2447    anywhere within big C types without resorting to outright "casts".
2448    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2449    modify the constness at the outermost level.)
2450    The implementation of "convert" is based on the idea of "witness"
2451    values -- values that are not used by the operation but whose types
2452    "testify" to their applicability.  On the implementation side, "convert"
2453    is simply a projection (returning its second curried argument).  With
2454    cross-module inlining, it should not result in any machine code being
2455    generated.
2456    
2457    ----------------------------------------------------------------------
2458    Name: Matthias Blume
2459    Date: 2002/03/15 16:40:00 EST
2460    Tag: blume-20020315-basis
2461    Description:
2462    
2463    Provided (preliminary?) implementations for
2464    
2465      {String,Substring}.{concatWith,isSuffix,isSubstring}
2466    
2467    and
2468    
2469      Substring.full
2470    
2471    Those are in the Basis spec but they were missing in SML/NJ.
2472    
2473    ----------------------------------------------------------------------
2474    Name: Matthias Blume
2475    Date: 2002/03/14 21:30:00 EST
2476    Tag: blume-20020314-controls
2477    Description:
2478    
2479    Controls:
2480    ---------
2481    
2482    1. Factored out the recently-added Controls : CONTROLS stuff and put
2483       it into its own library $/controls-lib.cm.  The source tree for
2484       this is under src/smlnj-lib/Controls.
2485    
2486    2. Changed the names of types and functions in this interface, so they
2487       make a bit more "sense":
2488    
2489          module -> registry
2490          'a registry -> 'a group
2491    
2492    3. The interface now deals in ref cells only.  The getter/setter interface
2493       is (mostly) gone.
2494    
2495    4. Added a function that lets one register an already-existing ref cell.
2496    
2497    5. Made the corresponding modifications to the rest of the code so that
2498       everything compiles again.
2499    
2500    6. Changed the implementation of Controls.MLRISC back to something closer
2501       to the original.  In particular, this module (and therefore MLRISC)
2502       does not depend on Controls.  There now is some link-time code in
2503       int-sys.sml that registers the MLRISC controls with the Controls
2504       module.
2505    
2506    CM:
2507    ---
2508    
2509      * One can now specify the lambda-split aggressiveness in init.cmi.
2510    
2511    ----------------------------------------------------------------------
2512    Name: Allen Leung
2513    Date: 2002/03/13 17:30:00 EST
2514    Tag: leunga-20020313-x86-fp-unary
2515    Description:
2516    
2517    Bug fix for:
2518    
2519    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2520    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2521    > - fun f(x,(y,z)) = Real.~ y;
2522    > [autoloading]
2523    > [autoloading done]
2524    >       fchsl   (%eax), 184(%esp)
2525    > Error: MLRisc bug: X86MCEmitter.emitInstr
2526    >
2527    > uncaught exception Error
2528    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2529    
2530    The problem was that the code generator did not generate any fp registers
2531    in this case, and the ra didn't know that it needed to run the X86FP phase to
2532    translate the pseudo fp instruction.   This only happened with unary fp
2533    operators in certain situations.
2534    
2535    ----------------------------------------------------------------------
2536    Name: Matthias Blume
2537    Date: 2002/03/13 14:00:00 EST
2538    Tag: blume-20020313-overload-etc
2539    Description:
2540    
2541    1. Added _overload as a synonym for overload for backward compatibility.
2542       (Control.overloadKW must be true for either version to be accepted.)
2543    
2544    2. Fixed bug in install script that caused more things to be installed
2545       than what was requested in config/targets.
2546    
2547    3. Made CM aware of the (_)overload construct so that autoloading
2548       works.
2549    
2550    ----------------------------------------------------------------------
2551    Name: Matthias Blume
2552    Date: 2002/03/12 22:03:00 EST
2553    Tag: blume-20020312-url
2554    Description:
2555    
2556    Forgot to update BOOT and srcarchiveurl.
2557    
2558    ----------------------------------------------------------------------
2559    Name: Matthias Blume
2560    Date: 2002/03/12 17:30:00 EST
2561    Tag: blume-20020312-version110392
2562    Description:
2563    
2564    Yet another version number bump (because of small changes to the
2565    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2566    
2567    Changes:
2568    
2569      The new pid generation scheme described a few weeks ago was overly
2570      complicated.  I implemented a new mechanism that is simpler and
2571      provides a bit more "stability":  Once CM has seen a compilation
2572      unit, it keeps its identity constant (as long as you do not delete
2573      those crucial CM/GUID/* files).  This means that when you change
2574      an interface, compile, then go back to the old interface, and
2575      compile again, you arrive at the original pid.
2576    
2577      There now also is a mechanism that instructs CM to use the plain
2578      environment hash as a module's pid (effectively making its GUID
2579      the empty string).  For this, "noguid" must be specified as an
2580      option to the .sml file in question within its .cm file.
2581      This is most useful for code that is being generated by tools such
2582      as ml-nlffigen (because during development programmers tend to
2583      erase the tool's entire output directory tree including CM's cached
2584      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2585      revert to the old, broken behavior of SML/NJ, but in specific cases
2586      where there is no danger of interface confusion, its use is ok
2587      (I think).
2588    
2589      ml-nlffigen by default generates "noguid" annotations.  They can be
2590      turned off by specifying -guid in its command line.
2591    
2592    ----------------------------------------------------------------------
2593    Name: Lal George
2594    Date: 2002/03/12 12 14:42:36 EST
2595    Tag: george-20020312-frequency-computation
2596    Description:
2597    
2598    Integrated jump chaining and static block frequency into the
2599    compiler. More details and numbers later.
2600    
2601    ----------------------------------------------------------------------
2602    Name: Lal George
2603    Date: 2002/03/11 11 22:38:53 EST
2604    Tag: george-20020311-jump-chain-elim
2605    Description:
2606    
2607    Tested the jump chain elimination on all architectures (except the
2608    hppa).  This is on by default right now and is profitable for the
2609    alpha and x86, however, it may not be profitable for the sparc and ppc
2610    when compiling the compiler.
2611    
2612    The gc test will typically jump to a label at the end of the cluster,
2613    where there is another jump to an external cluster containing the actual
2614    code to invoke gc. This is to allow factoring of common gc invocation
2615    sequences. That is to say, we generate:
2616    
2617            f:
2618               testgc
2619               ja   L1      % jump if above to L1
2620    
2621            L1:
2622               jmp L2
2623    
2624    
2625    After jump chain elimination the 'ja L1' instructions is converted to
2626    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2627    up being implemented in their long form (if L2 is far away) using:
2628    
2629            jbe     L3      % jump if below or equal to L3
2630            jmp     L2
2631         L3:
2632            ...
2633    
2634    
2635    For large compilation units L2  may be far away.
2636    
2637    
2638    ----------------------------------------------------------------------
2639    Name: Matthias Blume
2640    Date: 2002/03/11 13:30:00 EST
2641    Tag: blume-20020311-mltreeeval
2642    Description:
2643    
2644    A functor parameter was missing.
2645    
2646    ----------------------------------------------------------------------
2647    Name: Allen Leung
2648    Date: 2002/03/11 10:30:00 EST
2649    Tag: leunga-20020311-runtime-string0
2650    Description:
2651    
2652       The representation of the empty string now points to a
2653    legal null terminated C string instead of unit.  It is now possible
2654    to convert an ML string into C string with InlineT.CharVector.getData.
2655    This compiles into one single machine instruction.
2656    
2657    ----------------------------------------------------------------------
2658    Name: Allen Leung
2659    Date: 2002/03/10 23:55:00 EST
2660    Tag: leunga-20020310-x86-call
2661    Description:
2662    
2663       Added machine generation for CALL instruction (relative displacement mode)
2664    
2665    ----------------------------------------------------------------------
2666    Name: Matthias Blume
2667    Date: 2002/03/08 16:05:00
2668    Tag: blume-20020308-entrypoints
2669    Description:
2670    
2671    Version number bumped to 110.39.1.  NEW BOOTFILES!
2672    
2673    Entrypoints: non-zero offset into a code object where execution should begin.
2674    
2675    - Added the notion of an entrypoint to CodeObj.
2676    - Added reading/writing of entrypoint info to Binfile.
2677    - Made runtime system bootloader aware of entrypoints.
2678    - Use the address of the label of the first function given to mlriscGen
2679      as the entrypoint.  This address is currently always 0, but it will
2680      not be 0 once we turn on block placement.
2681    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2682      with entry points) from mlriscGen.
2683    
2684    ----------------------------------------------------------------------
2685    Name: Allen Leung
2686    Date: 2002/03/07 20:45:00 EST
2687    Tag: leunga-20020307-x86-cmov
2688    Description:
2689    
2690       Bug fixes for CMOVcc on x86.
2691    
2692       1. Added machine code generation for CMOVcc
2693       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2694       3. CMOVcc cannot have an immediate operand as argument.
2695    
2696    ----------------------------------------------------------------------
2697    Name: Matthias Blume
2698    Date: 2002/03/07 16:15:00 EST
2699    Tag: blume-20020307-controls
2700    Description:
2701    
2702    This is a very large but mostly boring patch which makes (almost)
2703    every tuneable compiler knob (i.e., pretty much everything under
2704    Control.* plus a few other things) configurable via both the command
2705    line and environment variables in the style CM did its configuration
2706    until now.
2707    
2708    Try starting sml with '-h' (or, if you are brave, '-H')
2709    
2710    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2711    implements the underlying generic mechanism.
2712    
2713    The interface to some of the existing such facilities has changed somewhat.
2714    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2715    (The getFoo interface is still there for backward-compatibility, but its
2716    use is deprecated.)
2717    
2718    The ml-build script passes -Cxxx=yyy command-line arguments through so
2719    that one can now twiddle the compiler settings when using this "batch"
2720    compiler.
2721    
2722    TODO items:
2723    
2724    We should go through and throw out all controls that are no longer
2725    connected to anything.  Moreover, we should go through and provide
2726    meaningful (and correct!) documentation strings for those controls
2727    that still are connected.
2728    
2729    Currently, multiple calls to Controls.new are accepted (only the first
2730    has any effect).  Eventually we should make sure that every control
2731    is being made (via Controls.new) exactly once.  Future access can then
2732    be done using Controls.acc.
2733    
2734    Finally, it would probably be a good idea to use the getter-setter
2735    interface to controls rather than ref cells.  For the time being, both
2736    styles are provided by the Controls module, but getter-setter pairs are
2737    better if thread-safety is of any concern because they can be wrapped.
2738    
2739    *****************************************
2740    
2741    One bug fix: The function blockPlacement in three of the MLRISC
2742    backpatch files used to be hard-wired to one of two possibilities at
2743    link time (according to the value of the placementFlag).  But (I
2744    think) it should rather sense the flag every time.
2745    
2746    *****************************************
2747    
2748    Other assorted changes (by other people who did not supply a HISTORY entry):
2749    
2750    1. the cross-module inliner now works much better (Monnier)
2751    2. representation of weights, frequencies, and probabilities in MLRISC
2752       changed in preparation of using those for weighted block placement
2753       (Reppy, George)
2754    
2755    ----------------------------------------------------------------------
2756    Name: Lal George
2757    Date: 2002/03/07 14:44:24 EST 2002
2758    Tag: george-20020307-weighted-block-placement
2759    
2760    Tested the weighted block placement optimization on all architectures
2761    (except the hppa) using AMPL to generate the block and edge frequencies.
2762    Changes were required in the machine properties to correctly
2763    categorize trap instructions. There is an MLRISC flag
2764    "weighted-block-placement" that can be used to enable weighted block
2765    placement, but this will be ineffective without block/edge
2766    frequencies (coming soon).
2767    
2768    
2769    ----------------------------------------------------------------------
2770    Name: Lal George
2771    Date: 2002/03/05 17:24:48 EST
2772    Tag: george-20020305-linkage-cluster
2773    
2774    In order to support the block placement optimization, a new cluster
2775    is generated as the very first cluster (called the linkage cluster).
2776    It contains a single jump to the 'real' entry point for the compilation
2777    unit. Block placement has no effect on the linkage cluster itself, but
2778    all the other clusters  have full freedom in the manner in which they
2779    reorder blocks or functions.
2780    
2781    On the x86 the typical linkage code that is generated is:
2782       ----------------------
2783            .align 2
2784       L0:
2785            addl    $L1-L0, 72(%esp)
2786            jmp     L1
2787    
2788    
2789            .align  2
2790       L1:
2791       ----------------------
2792    
2793    72(%esp) is the memory location for the stdlink register. This
2794    must contain the address of the CPS function being called. In the
2795    above example, it contains the address of  L0; before
2796    calling L1 (the real entry point for the compilation unit), it
2797    must contain the address for L1, and hence
2798    
2799            addl $L1-L0, 72(%esp)
2800    
2801    I have tested this on all architectures except the hppa.The increase
2802    in code size is of course negligible
2803    
2804    ----------------------------------------------------------------------
2805    Name: Allen Leung
2806    Date: 2002/03/03 13:20:00 EST
2807    Tag: leunga-20020303-mlrisc-tools
2808    
2809      Added #[ ... ] expressions to mlrisc tools
2810    
2811    ----------------------------------------------------------------------
2812    Name: Matthias Blume
2813    Date: 2002/02/27 12:29:00 EST
2814    Tag: blume-20020227-cdebug
2815    Description:
2816    
2817    - made types in structure C and C_Debug to be equal
2818    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2819    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2820    
2821    ----------------------------------------------------------------------
2822    Name: Matthias Blume
2823    Date: 2002/02/26 12:00:00 EST
2824    Tag: blume-20020226-ffi
2825    Description:
2826    
2827    1. Fixed a minor bug in CM's "noweb" tool:
2828       If numbering is turned off, then truly don't number (i.e., do not
2829       supply the -L option to noweb).  The previous behavior was to supply
2830       -L'' -- which caused noweb to use the "default" line numbering scheme.
2831       Thanks to Chris Richards for pointing this out (and supplying the fix).
2832    
2833    2. Once again, I reworked some aspects of the FFI:
2834    
2835       A. The incomplete/complete type business:
2836    
2837       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2838         gone!
2839       - ML types representing an incomplete type are now *equal* to
2840         ML types representing their corresponding complete types (just like
2841         in C).  This is still safe because ml-nlffigen will not generate
2842         RTTI for incomplete types, nor will it generate functions that
2843         require access to such RTTI.   But when ML code generated from both
2844         incomplete and complete versions of the C type meet, the ML types
2845         are trivially interoperable.
2846    
2847         NOTE:  These changes restore the full generality of the translation
2848         (which was previously lost when I eliminated functorization)!
2849    
2850       B. Enum types:
2851    
2852       - Structure C now has a type constructor "enum" that is similar to
2853         how the "su" constructor works.  However, "enum" is not a phantom
2854         type because each "T enum" has values (and is isomorphic to
2855         MLRep.Signed.int).
2856       - There are generic access operations for enum objects (using
2857         MLRep.Signed.int).
2858       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2859         * The structure contains the definition of type "mlrep" (the ML-side
2860         representation type of the enum).  Normally, mlrep is the same
2861         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2862         then mlrep will be defined as a datatype -- thus facilitating
2863         pattern matching on mlrep values.
2864         ("-ec" will be suppressed if there are duplicate values in an
2865          enumeration.)
2866         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2867         will be generated for each C enum constant xxx.
2868         * Conversion functions m2i and i2m convert between mlrep and
2869         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2870         * Coversion functions c and ml convert between mlrep and "tag enum".
2871         * Access functions (get/set) fetch and store mlrep values.
2872       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2873         enumerations are merged into one single enumeration represented by
2874         structure E_'.
2875    
2876    ----------------------------------------------------------------------
2877    Name: Allen Leung
2878    Date: 2002/02/25 04:45:00 EST
2879    Tag: leunga-20020225-cps-spill
2880    
2881    This is a new implementation of the CPS spill phase.
2882    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2883    In case of problems, replace it with the old file spill.sml
2884    
2885    The current compiler runs into some serious performance problems when
2886    constructing a large record.  This can happen when we try to compile a
2887    structure with many items.  Even a very simple structure like the following
2888    makes the compiler slow down.
2889    
2890        structure Foo = struct
2891           val x_1 = 0w1 : Word32.int
2892           val x_2 = 0w2 : Word32.int
2893           val x_3 = 0w3 : Word32.int
2894           ...
2895           val x_N = 0wN : Word32.int
2896        end
2897    
2898    The following table shows the compile time, from N=1000 to N=4000,
2899    with the old compiler:
2900    
2901    N
2902    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2903           MLRISC ra                               0.06u  0.00s  0.05g
2904              (spills = 0 reloads = 0)
2905           TOTAL                                   0.63u  0.07s  0.21g
2906    
2907    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2908           MLRISC ra                               5.68u  0.59s  3.93g
2909              (spills = 0 reloads = 0)
2910           TOTAL                                   14.71u  0.99s  4.81g
2911    
2912    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2913           MLRISC ra                               5.54u  0.65s  3.91g
2914              (spills = 543 reloads = 1082)
2915           TOTAL                                   65.40u  3.13s  6.00g
2916    
2917    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2918           MLRISC ra                               0.80u  0.10s  0.55g
2919              (spills = 42 reloads = 84)
2920           TOTAL                                   129.42u  5.10s  4.13g
2921    
2922    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2923           MLRISC ra                               2.69u  0.27s  1.38g
2924              (spills = 62 reloads = 124)
2925           TOTAL                                   682.48u  19.61s  13.99g
2926    
2927    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2928           MLRISC ra                               4.96u  0.27s  2.72g
2929              (spills = 85 reloads = 170)
2930           TOTAL                                   2375.26u  57.21s  48.00g
2931    
2932    As you can see the old cps spill module suffers from some serious
2933    performance problem.  But since I cannot decipher the old code fully,
2934    instead of patching the problems up, I'm reimplementing it
2935    with a different algorithm.  The new code is more modular,
2936    smaller when compiled, and substantially faster
2937    (O(n log n) time and O(n) space).  Timing of the new spill module:
2938    
2939    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2940          MLRISC ra                               0.25u  0.02s  0.15g
2941             (spills=1 reloads=3)
2942          TOTAL                                   7.74u  0.34s  1.62g
2943    
2944    Implementation details:
2945    
2946    As far as I can tell, the purpose of the CPS spill module is to make sure the
2947    number of live variables at any program point (the bandwidth)
2948    does not exceed a certain limit, which is determined by the
2949    size of the spill area.
2950    
2951    When the bandwidth is too large, we decrease the register pressure by
2952    packing live variables into spill records.  How we achieve this is
2953    completely different than what we did in the old code.
2954    
2955    First, there is something about the MLRiscGen code generator
2956    that we should be aware of:
2957    
2958    o MLRiscGen performs code motion!
2959    
2960       In particular, it will move floating point computations and
2961       address computations involving only the heap pointer to
2962       their use sites (if there is only a single use).
2963       What this means is that if we have a CPS record construction
2964       statement
2965    
2966           RECORD(k,vl,w,e)
2967    
2968       we should never count the new record address w as live if w
2969       has only one use (which is often the case).
2970    
2971       We should do something similar to floating point, but the transformation
2972       there is much more complex, so I won't deal with that.
2973    
2974    Secondly, there are now two new cps primops at our disposal:
2975    
2976     1. rawrecord of record_kind option
2977        This pure operator allocates some uninitialized storage from the heap.
2978        There are two forms:
2979    
2980         rawrecord NONE [INT n]  allocates a tagless record of length n
2981         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2982                                     and initializes the tag.
2983    
2984     2. rawupdate of cty
2985          rawupdate cty (v,i,x)
2986          Assigns to x to the ith component of record v.
2987          The storelist is not updated.
2988    
2989    We use these new primops for both spilling and increment record construction.
2990    
2991     1. Spilling.
2992    
2993        This is implemented with a linear scan algorithm (but generalized
2994        to trees).  The algorithm will create a single spill record at the
2995        beginning of the cps function and use rawupdate to spill to it,
2996        and SELECT or SELp to reload from it.  So both spills and reloads
2997        are fine-grain operations.  In contrast, in the old algorithm
2998        "spills" have to be bundled together in records.
2999    
3000        Ideally, we should sink the spill record construction to where
3001        it is needed.  We can even split the spill record into multiple ones
3002        at the places where they are needed.  But CPS is not a good
3003        representation for global code motion, so I'll keep it simple and
3004        am not attempting this.
3005    
3006     2. Incremental record construction (aka record splitting).
3007    
3008        Long records with many component values which are simulatenously live
3009        (recall that single use record addresses are not considered to
3010         be live) are constructed with rawrecord and rawupdate.
3011        We allocate space on the heap with rawrecord first, then gradually
3012        fill it in with rawupdate.  This is the technique suggested to me
3013        by Matthias.
3014    
3015        Some restrictions on when this is applicable:
3016        1. It is not a VECTOR record.  The code generator currently does not handle
3017           this case. VECTOR record uses double indirection like arrays.
3018        2. All the record component values are defined in the same "basic block"
3019           as the record constructor.  This is to prevent speculative
3020           record construction.
3021    
3022    ----------------------------------------------------------------------
3023    Name: Allen Leung
3024    Date: 2002/02/22 01:02:00 EST
3025    Tag: leunga-20020222-mlrisc-tools
3026    
3027    Minor bug fixes in the parser and rewriter
3028    
3029    ----------------------------------------------------------------------
3030    Name: Allen Leung
3031    Date: 2002/02/21 20:20:00 EST
3032    Tag: leunga-20020221-peephole
3033    
3034    Regenerated the peephole files.  Some contained typos in the specification
3035    and some didn't compile because of pretty printing bugs in the old version
3036    of 'nowhere'.
3037    
3038    ----------------------------------------------------------------------
3039    Name: Allen Leung
3040    Date: 2002/02/19 20:20:00 EST
3041    Tag: leunga-20020219-mlrisc-tools
3042    Description:
3043    
3044       Minor bug fixes to the mlrisc-tools library:
3045    
3046       1.  Fixed up parsing colon suffixed keywords
3047       2.  Added the ability to shut the error messages up
3048       3.  Reimplemented the pretty printer and fixed up/improved
3049           the pretty printing of handle and -> types.
3050       4.  Fixed up generation of literal symbols in the nowhere tool.
3051       5.  Added some SML keywords to to sml.sty
3052    
3053    ----------------------------------------------------------------------
3054    Name: Matthias Blume
3055    Date: 2002/02/19 16:20:00 EST
3056    Tag: blume-20020219-cmffi
3057    Description:
3058    
3059    A wild mix of changes, some minor, some major:
3060    
3061    * All C FFI-related libraries are now anchored under $c:
3062        $/c.cm      --> $c/c.cm
3063        $/c-int.cm  --> $c/internals/c-int.cm
3064        $/memory.cm --> $c/memory/memory.cm
3065    
3066    * "make" tool (in CM) now treats its argument pathname slightly
3067      differently:
3068        1. If the native expansion is an absolute name, then before invoking
3069           the "make" command on it, CM will apply OS.Path.mkRelative
3070           (with relativeTo = OS.FileSys.getDir()) to it.
3071        2. The argument will be passed through to subsequent phases of CM
3072           processing without "going native".  In particular, if the argument
3073           was an anchored path, then "make" will not lose track of that anchor.
3074    
3075    * Compiler backends now "know" their respective C calling conventions
3076      instead of having to be told about it by ml-nlffigen.  This relieves
3077      ml-nlffigen from one of its burdens.
3078    
3079    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3080    
3081    * Export C_DEBUG and C_Debug from $c/c.cm.
3082    
3083    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3084      subtyping relationship between incomplete pointers and their complete
3085      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3086      with the convention of instantiating 'o with ('t, 'c) obj whenever
3087      the pointer target type is complete.  In the incomplete case, 'o
3088      will be instantiated with some "'c iobj" -- a type obtained by
3089      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3090    
3091      Operations that work on both incomplete and complete pointer types are
3092      typed as taking an 'o ptr while operations that require the target to
3093      be known are typed as taking some ('t, 'c) obj ptr.
3094    
3095      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3096      where void is an eqtype without any values.  This makes it possible
3097      to work on voidptr values using functions meant to operate on light
3098      incomplete pointers.
3099    
3100    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3101      been vastly simplified.
3102    
3103    ----------------------------------------------------------------------
3104    Name: Matthias Blume
3105    Date: 2002/02/19 10:48:00 EST
3106    Tag: blume-20020219-pqfix
3107    Description:
3108    
3109    Applied Chris Okasaki's bug fix for priority queues.
3110    
3111    ----------------------------------------------------------------------
3112    Name: Matthias Blume
3113    Date: 2002/02/15 17:05:00
3114    Tag: Release_110_39
3115    Description:
3116    
3117    Last-minute retagging is becoming a tradition... :-(
3118    
3119    This is the working release 110.39.
3120    
3121    ----------------------------------------------------------------------
3122    Name: Matthias Blume
3123    Date: 2002/02/15 16:00:00 EST
3124    Tag: Release_110_39-orig
3125    Description:
3126    
3127    Working release 110.39.  New bootfiles.
3128    
3129    (Update: There was a small bug in the installer so it wouldn't work
3130    with all shells.  So I retagged. -Matthias)
3131    
3132    ----------------------------------------------------------------------
3133    Name: Matthias Blume
3134    Date: 2002/02/15 14:17:00 EST
3135    Tag: blume-20020215-showbindings
3136    Description:
3137    
3138    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3139    the latter can be useful for exploring what bindings are available at
3140    the interactive prompt.  (The first function returns only the list
3141    of symbols that are really bound, the second prints those but also the
3142    ones that CM's autoloading mechanism knows about.)
3143    
3144    ----------------------------------------------------------------------
3145    Name: Matthias Blume
3146    Date: 2002/02/15 12:08:00 EST
3147    Tag: blume-20020215-iptrs
3148    Description:
3149    
3150    Two improvements to ml-nlffigen:
3151    
3152      1. Write files only if they do not exist or if their current contents
3153         do not coincide with what's being written.  (That is, avoid messing
3154         with the time stamps unless absolutely necessary.)
3155    
3156      2. Implement a "repository" mechanism for generated files related
3157         to "incomplete pointer types".   See the README file for details.
3158    
3159    ----------------------------------------------------------------------
3160    Name: Matthias Blume
3161    Date: 2002/02/14 11:50:00 EST
3162    Tag: blume-20020214-quote
3163    Description:
3164    
3165    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3166    because of the new and improved tag generation scheme.  (Thanks to Allen
3167    Leung for pointing it out.)
3168    
3169    ----------------------------------------------------------------------
3170    Name: Lal George
3171    Date: 2002/02/14 09:55:27 EST 2002
3172    Tag: george-20020214-isabelle-bug
3173    Description:
3174    
3175    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3176    of Isabelle on the x86.
3177    
3178    From Allen:
3179    -----------
3180     I've found the problem:
3181    
3182         in ra-core.sml, I use the counter "blocked" to keep track of the
3183         true number of elements in the freeze queue.  When the counter goes
3184         to zero, I skip examining the queue.  But I've messed up the
3185         bookkeeping in combine():
3186    
3187             else ();
3188             case !ucol of
3189               PSEUDO => (if !cntv > 0 then
3190                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3191                                        ^^^^^^^^^^^^^^^^^^^^^^^
3192                      moveu := mergeMoveList(!movev, !moveu)
3193                     )
3194                  else ();
3195    
3196         combine() is called to coalesce two nodes u and v.
3197         I think I was thinking that if the move counts of u and v are both
3198         greater than zero then after they are coalesced then one node is
3199         removed from the freeze queue.  Apparently I was thinking that
3200         both u and v are of low degree, but that's clearly not necessarily true.
3201    
3202    
3203    02/12/2002:
3204        Here's the patch.  HOL now compiles.
3205    
3206        I don't know how this impact on performance (compile
3207        time or runtime).  This bug caused the RA (especially on the x86)
3208        to go thru the potential spill phase when there are still nodes on the
3209        freeze queue.
3210    
3211    
3212    
3213    
3214    ----------------------------------------------------------------------
3215    Name: Matthias Blume
3216    Date: 2002/02/13 22:40:00 EST
3217    Tag: blume-20020213-fptr-rtti
3218    Description:
3219    
3220    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3221    updates.
3222    
3223    ----------------------------------------------------------------------
3224    Name: Matthias Blume
3225    Date: 2002/02/13 16:41:00 EST
3226    Tag: blume-20020213-cmlpq
3227    Description:
3228    
3229    Added new priority queue export symbols (which have just been added to
3230    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3231    not compile and the installer would choke.)
3232    
3233    ----------------------------------------------------------------------
3234    Name: Matthias Blume
3235    Date: 2002/02/13 16:15:00 EST
3236    Tag: blume-20020213-various
3237    Description:
3238    
3239    1. More tweaks to ml-nlffigen:
3240    
3241       - better internal datastructures (resulting in slight speedup)
3242       - "-match" option requires exact match
3243       - "localized" gensym counters (untagged structs/unions nested within
3244         other structs/unions or within typedefs get a fresh counter; their
3245         tag will be prefixed by a concatenation of their parents' tags)
3246       - bug fixes (related to calculation of transitive closure of types
3247         to be included in the output)
3248    
3249    2. Minor Basis updates:
3250    
3251       - added implementations for List.collate and Option.app
3252    
3253    ----------------------------------------------------------------------
3254    Name: Matthias Blume
3255    Date: 2002/02/11 15:55:00 EST
3256    Tag: blume-20020211-gensym
3257    Description:
3258    
3259    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3260    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3261    names (ML structure names that correspond to unnamed C structs, unions,
3262    and enums), so that separate runs of ml-nlffigen do not clash.
3263    
3264    ----------------------------------------------------------------------
3265    Name: Matthias Blume
3266    Date: 2002/02/11 12:05:00 EST
3267    Tag: blume-20020211-gensml
3268    Description:
3269    
3270    A quick fix for a problem with GenSML (in the pgraph-util library):
3271    Make generation of toplevel "local" optional.  (Strictly speaking,
3272    signature definitions within "local" are not legal SML.)
3273    
3274    Other than that: updates to INSTALL and cm/TODO.
3275    
3276    ----------------------------------------------------------------------
3277    Name: Matthias Blume
3278    Date: 2002/02/08 15:00:00 EST
3279    Tag: blume-20020208-uniquepid
3280    Description:
3281    
3282    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3283    
3284    1. The installer (config/install.sh) has gotten smarter:
3285    
3286         - Configuration options are a bit easier to specify now
3287           (in config/targets).
3288         - Bug in recognizing .tar.bz2 files fixed.
3289         - Installer automatically resolves dependencies between
3290           configuration options (e.g., if you ask for eXene, you will
3291           also get cml -- regardless whether you asked for it or not).
3292         - Installer can run in "quieter mode" by setting the environment
3293           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3294           "completely silent", though.
3295         - Build HashCons library as part of smlnj-lib.
3296    
3297    2. A new scheme for assigning persistent identifiers to compilation
3298       units (and, by extension, to types etc.) has been put into place.
3299       This fixes a long-standing bug where types and even dynamic values
3300       can get internally confused, thereby compromising type safety
3301       (abstraction) and dynamic correctness.  See
3302    
3303         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3304    
3305       for an example of how things could go wrong until now.
3306    
3307       The downside of the new scheme is that pids are not quite as
3308       persistent as they used to be: CM will generate a fresh pid
3309       for every compilation unit that it thinks it sees for the first
3310       time.  That means that if you compile starting from a clean, fresh
3311       source tree at two different times, you end up with different
3312       binaries.
3313    
3314       Cutoff recompilation, however, has not been compromised because
3315       CM keeps pid information in special caches between runs.
3316    
3317    ----------------------------------------------------------------------
3318    Name: Lal George
3319    Date: 2002/02/07 15:34:13 EST 2002
3320    Tag: <none>
3321    Description:
3322    
3323    Compilers that generate assembly code may produce  global labels
3324    whose value is resolved at link time. The various peephole optimization
3325    modules did not take this in account.
3326    
3327    TODO. The Labels.addrOf function should really return an option
3328    type so that clients are forced to deal with this issue, rather
3329    than an exception being raised.
3330    
3331    ----------------------------------------------------------------------
3332    Name: Lal George
3333    Date: 2002/02/06 13:55:02 EST
3334    Tag: george-20020206-ra-breakup
3335    Description:
3336    
3337    1. A bug fix from Allen.
3338    
3339        A typo causes extra fstp %st(0)'s to be generated at compensation
3340        edges, which might cause stack underflow traps at runtime.  This
3341        occurs in fft where there are extraneous fstps right before the 'into'
3342        trap instruction (in this case they are harmless since none of the
3343        integers overflow.)
3344    
3345    2. Pulled out various utility modules that were embedded in the modules
3346       of the register allocator. I need these modules for other purposes, but
3347       they are not complete enough to put into a library (just yet).
3348    ----------------------------------------------------------------------
3349    Name: Matthias Blume
3350    Date: 2002/01/31 16:05:00 EST
3351    Tag: blume-20020131-sparc-ccalls
3352    Description:
3353    
3354    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3355       of extra stack space by mistake.  Fixed.
3356    
3357    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3358    
3359    ----------------------------------------------------------------------
3360    Name: Allen Leung
3361    Date: 2002/01/30
3362    Tag: leunga-20020130-nowhere-bug-fix
3363    Description:
3364    
3365       MLRISC bug fixes:
3366       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3367       2. MachineInt.fromString was negating its value.
3368    
3369    ----------------------------------------------------------------------
3370    Name: Matthias Blume
3371    Date: 2002/01/29
3372    Tag: blume-20020129-INSTALL
3373    Description:
3374    
3375    - Added somewhat detailed installation instructions (file INSTALL).
3376    - Fixed curl-detection bug in config/install.sh.
3377    - It is now possible to select the URL getter using the URLGETTER
3378      environment variable:
3379    
3380          not set / "unknown"      --> automatic detection (script tries wget,
3381                                       curl, and lynx)
3382          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3383                                       how to properly invoke them)
3384          other                    --> use $URLGETTER directly, it must take
3385                                       precisely two command-line arguments
3386                                       (source URL and destination file name)
3387    
3388    ----------------------------------------------------------------------
3389    Name: Matthias Blume
3390    Date: 2002/01/28
3391    Tag: blume-20020128-sparc-ccalls
3392    Description:
3393    
3394    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3395    - Make use of the allocParam argument in sparc-c-calls.
3396    
3397    ----------------------------------------------------------------------
3398    Name: Matthias Blume
3399    Date: 2002/01/28
3400    Tag: blume-20020128-allocParam
3401    Description:
3402    
3403    John Reppy:  Changes c-calls API to accept client-callback for
3404    allocating extra stack space.
3405    me: Corresponding changes to mlriscGen (using a dummy argument that
3406        does not change the current behavior).
3407    
3408    ----------------------------------------------------------------------
3409    Name: Matthias Blume
3410    Date: 2002/01/28 12:00:00
3411    Tag: Release_110_38
3412    Description:
3413    
3414    This time for real!!!
3415    
3416    ----------------------------------------------------------------------
3417    Name: Matthias Blume
3418    Date: 2002/01/28 10:56:00 EST
3419    Tag: blume-20020128-retraction
3420    Description:
3421    
3422    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3423       with blume-Release_110_38-retracted.)
3424    
3425    1. Fixed a problem with incorrect rounding modes in real64.sml.
3426       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3427    
3428    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3429       and enums fixed.  The naming of corresponding ML identifiers should
3430       now be consistent again.
3431    
3432    ----------------------------------------------------------------------
3433    Name: Allen Leung
3434    Date: 2002/01/27
3435    Tag: leunga-20020127-nowhere
3436    Description:
3437    
3438       Added a target called nowhere in the configuration scripts.
3439       Enabling this will build the MLRISC 'nowhere' tool (for translating
3440       programs with where-clauses into legal SML code) during installation.
3441    
3442    ----------------------------------------------------------------------
3443    Name: Matthias Blume
3444    Date: 2002/01/25 21:27:00 EST
3445    Tag: blume-Release_110_38-retracted
3446    Description:
3447    
3448    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3449    
3450    README will be added later.
3451    
3452    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3453    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3454    
3455    ----------------------------------------------------------------------
3456    Name: Matthias Blume
3457    Date: 2002/01/25
3458    Tag: blume-20020125-ffi
3459    Description:
3460    
3461    A large number of tweaks and improvements to ml-nlffi-lib and
3462    ml-nlffigen:
3463    
3464       - ML represenation types have been streamlined
3465       - getter and setter functions work with concrete values, not abstract
3466         ones where possible
3467       - ml-nlffigen command line more flexible (see README file there)
3468       - some bugs have been fixed (hopefully)
3469    
3470    ----------------------------------------------------------------------
3471    Name: Lal George
3472    Date: 2002/01/24
3473    Tag: george-20020124-risc-ra-interface
3474    Description:
3475    
3476       There is a dramatic simplification in the interface to the
3477       register allocator for RISC architectures as a result of making
3478       parallel copy instructions explicit.
3479    
3480    ----------------------------------------------------------------------
3481    Name: Matthias Blume
3482    Date: 2002/01/22
3483    Tag: blume-20020122-x86-ccalls
3484    Description:
3485    
3486    Bug fix for c-calls on x86 (having to do with how char- and
3487    short-arguments are being handled).
3488    
3489    ----------------------------------------------------------------------
3490    Name: Matthias Blume
3491    Date: 2002/01/21
3492    Tag: blume-20020121-ff
3493    Description:
3494    
3495    Another day of fiddling with the FFI...
3496    
3497    1. Bug fix/workaround:  CKIT does not complain about negative array
3498       dimensions, so ml-nlffigen has to guard itself against this possibility.
3499       (Otherwise a negative dimension would send it into an infinite loop.)
3500    
3501    2. Some of the abstract types (light objects, light pointers, most "base"
3502       types) in structure C are now eqtypes.
3503    
3504    3. Added constructors and test functions for NULL function pointers.
3505    
3506    ----------------------------------------------------------------------
3507    Name: Matthias Blume
3508    Date: 2002/01/18
3509    Tag: blume-20020118-ready-for-new-release
3510    Description:
3511    
3512    Made config/srcarchiveurl point to a new place.  (Will provide boot
3513    files shortly.)
3514    
3515    Maybe we christen this to be 110.38?
3516    
3517    ----------------------------------------------------------------------
3518    Name: Matthias Blume
3519    Date: 2002/01/18
3520    Tag: blume-20020118-more-ffifiddle
3521    Description:
3522    
3523    Today's FFI fiddling:
3524    
3525      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3526        C.Set.* that use concrete (MLRep.*) arguments and results instead
3527        of abstract ones.
3528    
3529      - Provided word-style bit operations etc. for "int" representation
3530        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3531        and Long.
3532    
3533    ----------------------------------------------------------------------
3534    Name: Matthias Blume
3535    Date: 2002/01/18
3536    Tag: blume-20020118-use-x86-fp
3537    Description:
3538    
3539    Now that x86-fast-fp seems to be working, I turned it back on again
3540    by default.  (Seems to work fine now, even with the FFI.)
3541    
3542    Other than that, I added some documentation about the FFI to
3543    src/ml-nlffigen/README and updated the FFI test examples in
3544    src/ml-nlffi-lib/Tests/*.
3545    
3546    ----------------------------------------------------------------------
3547    Name: Allen Leung
3548    Date: 2002/01/17
3549    Tag: leunga-20020117-x86-fast-fp-call
3550    Description:
3551    
3552       1. Fixed a problem with handling return fp values when x86's fast fp
3553          mode is turned on.
3554    
3555       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3556          of %f32.
3557    
3558       3. Added a constructor INT32lit to the ast of MLRISC tools.
3559    
3560    ----------------------------------------------------------------------
3561    Name: Matthias Blume
3562    Date: 2002/01/16
3563    Tag: blume-20020116-ffifiddle
3564    Description:
3565    
3566    More fiddling with the FFI interface:
3567    
3568     - Make constness 'c instead of rw wherever possible.  This eliminates
3569       the need for certain explicit coercions.  (However, due to ML's
3570       value polymorphism, there will still be many cases where explicit
3571       coercions are necessary.  Phantom types are not the whole answer
3572       to modeling a subtyping relationship in ML.)
3573    
3574     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3575    
3576     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3577       clumsy workaround for varargs problem.)
3578    
3579    ----------------------------------------------------------------------
3580    Name: Lal George
3581    Date: 2002/01/15
3582    Tag: <none>
3583    Description:
3584    
3585    1. Since COPY instructions are no longer native to the architecture,
3586       a generic functor can be used to implement the expandCopies function.
3587    
3588    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3589       TEXT segment.
3590    
3591    ----------------------------------------------------------------------
3592    Name: Matthias Blume
3593    Date: 2002/01/15
3594    Tag: blume-20020115-ffiupdates
3595    Description:
3596    
3597    1. Fix for bug resulting in single-precision float values being returned
3598       incorrectly from FFI calls.
3599    
3600    2. Small modifications to C FFI API:
3601    
3602        - memory-allocation routines return straight objects (no options)
3603          and raise an exception in out-of-memory situations
3604        - unsafe extensions to cast between function pointers and pointers
3605          from/to ints
3606        - added structure C_Debug as an alternative to structure C where
3607          pointer-dereferencing (|*| and |*!) always check for null-pointers
3608        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3609          but also takes a (possibly empty) list of existing library handles
3610          that the current library depends on
3611    
3612    ----------------------------------------------------------------------
3613    Name: Matthias Blume
3614    Date: 2002/01/10
3615    Tag: blume-20020110-newffigen
3616    Description:
3617    
3618    1. Updates to portable graph code.
3619    
3620    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3621       more scalable now so that even huge interfaces such as the one
3622       for GTK compile in finite time and space. :-)
3623       See src/ml-nlffigen/README for details on what's new.
3624    
3625    ----------------------------------------------------------------------
3626    Name: Lal George
3627    Date: 2001/01/09 14:31:35 EST 2002
3628    Tag: george-20011206-rm-native-copy
3629    Description:
3630    
3631            Removed the native COPY and FCOPY instructions
3632            from all the architectures and replaced it with the
3633            explicit COPY instruction from the previous commit.
3634    
3635            It is now possible to simplify many of the optimizations
3636            modules that manipulate copies. This has not been
3637            done in this change.
3638    
3639    ----------------------------------------------------------------------
3640    Name: Lal George
3641    Date: 2001/12/06 16:50:13 EST 2001
3642    Tag: george-20011206-mlrisc-instruction
3643    Description:
3644    
3645    Changed the representation of instructions from being fully abstract
3646    to being partially concrete. That is to say:
3647    
3648      from
3649            type instruction
3650    
3651      to
3652            type instr                              (* machine instruction *)
3653    
3654            datatype instruction =
3655                LIVE of {regs: C.cellset, spilled: C.cellset}
3656              | KILL of {regs: C.cellset, spilled: C.cellset}
3657              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3658              | ANNOTATION of {i: instruction, a: Annotations.annotation}
3659              | INSTR of instr
3660    
3661    This makes the handling of certain special instructions that appear on
3662    all architectures easier and uniform.
3663    
3664    LIVE and KILL say that a list of registers are live or killed at the
3665    program point where they appear. No spill code is generated when an
3666    element of the 'regs' field is spilled, but the register is moved to
3667    the 'spilled' (which is present, more for debugging than anything else).
3668    
3669    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3670    We used to generate:
3671    
3672            DEFFREG f1
3673            f1 := f2 + f3
3674            trapb
3675    
3676    but now generate:
3677    
3678            f1 := f2 + f3
3679            trapb
3680            LIVE {regs=[f1,f2,f3], spilled=[]}
3681    
3682    Furthermore, the DEFFREG (hack) required that all floating point instruction
3683    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3684    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3685    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3686    this architecture).
3687    
3688    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3689    all the architectures. This will result in further simplification of the
3690    register allocator that must be aware of them for coalescing purposes, and
3691    will also simplify certain aspects of the machine description that provides
3692    callbacks related to parallel copies.
3693    
3694    ANNOTATION should be obvious, and now INSTR represents the honest to God
3695    machine instruction set!
3696    
3697    The <arch>/instructions/<arch>Instr.sml files define certain utility
3698    functions for making porting easier -- essentially converting upper case
3699    to lower case. All machine instructions (of type instr) are in upper case,
3700    and the lower case form generates an MLRISC instruction. For example on
3701    the alpha we have:
3702    
3703      datatype instr =
3704         LDA of {r:cell, b:cell, d:operand}
3705       | ...
3706    
3707      val lda : {r:cell, b:cell, d:operand} -> instruction
3708        ...
3709    
3710    where lda is just (INSTR o LDA), etc.
3711    
3712    ----------------------------------------------------------------------
3713    Name: Matthias Blume
3714    Date: 2001/11/22 21:40:00 EST
3715    Tag: Release_110_37
3716    Description:
3717    
3718    Release 110.37.  This time for real.
3719    
3720    ----------------------------------------------------------------------
3721    Name: Matthias Blume
3722    Date: 2001/11/21 16:35:00 EST
3723    Tag: blume-20011121-foot-in-mouth
3724    Description:
3725    
3726    Removed the "Release_110_37" tag because of a serious bug.
3727    This will be re-tagged once the bug is fixed.
3728    
3729    ----------------------------------------------------------------------
3730    Name: Matthias Blume
3731    Date: 2001/11/21 16:14:00 EST
3732    Tag: blume-20011121-forgottenfile
3733    Description:
3734    
3735    Forgot to add a file.  (Just a .tex-file -- part of
3736    the CM manual source.)
3737    
3738    ----------------------------------------------------------------------
3739    Name: Matthias Blume
3740    Date: 2001/11/21 16:10:00 EST
3741    Tag: blume-20011121-invalid_110_37
3742    Description:
3743    
3744    Note: I removed the original tag "Release_110_37" from this commit
3745          because we found a serious bug in all non-x86 backends.
3746          - Matthias
3747    
3748    1. Modifications to the SML/NJ code generator and to the runtime system
3749       so that code object name strings are directly inserted into code
3750       objects at code generation time.  The only business the runtime system
3751       has with this is now to read the name strings on occasions.
3752       (The encoding of the name string has also changed somewhat.)
3753    
3754    2. CM now implements a simple "set calculus" for specifying export lists.
3755       In particular, it is now possible to refer to the export lists of
3756       other libraries/groups/sources and form unions as well as differences.
3757       See the latest CM manual for details.
3758    
3759    3. An separate notion of "proxy" libraries has again be eliminated from
3760       CM's model.  (Proxy libraries are now simply a special case of using
3761       the export list calculus.)
3762    
3763    4. Some of the existing libraries now take advantage of the new set
3764       calculus.
3765       (Notice that not all libraries have been converted because some
3766       of the existing .cm-files are supposed to be backward compatible
3767       with 110.0.x.)
3768    
3769    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3770       "General.exnMessage"!  The former relies on a certain hook to be
3771       initialized, and that often does not happen in the stand-alone case.)
3772    
3773    ----------------------------------------------------------------------
3774    Name: Lal George
3775    Date: 2001/11/21  13:56:18 EST
3776    Tag: george-2001121-pseudo-ops
3777    Description:
3778    
3779      Implemented a complete redesign of MLRISC pseudo-ops. Now there
3780      ought to never be any question of incompatabilities with
3781      pseudo-op syntax expected by host assemblers.
3782    
3783      For now, only modules supporting GAS syntax are implemented
3784      but more should follow, such as MASM, and vendor assembler
3785      syntax, e.g. IBM as, Sun as, etc.
3786    
3787    ----------------------------------------------------------------------
3788    Name: Matthias Blume
3789    Date: 2001/11/14 11:52:00 EST
3790    Tag: blume-20011114-srcname
3791    Description:
3792    
3793    1. Routed the name of the current source file to mlriscgen where it
3794       should be directly emitted into the code object.  (This last part
3795       is yet to be done.)
3796    
3797    2. Some cleanup of the pgraph code to make it match the proposal that
3798       I put out the other day.  (The proposal notwithstanding, things are
3799       still in flux here.)
3800    
3801    ----------------------------------------------------------------------
3802    Name: Lal George
3803    Date: 2001/11/14 09:44:04 EST
3804    Tag:
3805    Description:
3806    
3807      Fix for a backpatching bug reported by Allen.
3808    
3809      Because the boundary between short and long span-dependent
3810      instructions is +/- 128, there are an astounding number of
3811      span-dependent instructions whose size is over estimated.
3812    
3813      Allen came up with the idea of letting the size of span
3814      dependent instructions be non-monotonic, for a maxIter
3815      number of times, after which the size must be monotonically
3816      increasing.
3817    
3818      This table shows the number of span-dependent instructions
3819      whose size was over-estimated as a function of maxIter, for the
3820      file Parse/parse/ml.grm.sml:
3821    
3822         maxIter            # of instructions:
3823            10                      687
3824            20                      438
3825            30                      198
3826            40                        0
3827    
3828      In compiling the compiler, there is no significant difference in
3829      compilation speed between maxIter=10 and maxIter=40. Actually,
3830      my measurements showed that maxIter=40 was a tad faster than
3831      maxIter=10! Also 96% of the  files in the compiler reach a fix
3832      point within 13 iterations, so fixing maxIter at 40, while high,
3833      is okay.
3834    
3835    ----------------------------------------------------------------------
3836    Name: Matthias Blume
3837    Date: 2001/10/31 15:25:00 EST
3838    Tag: blume-20011031-pgraph
3839    Description:
3840    
3841    CKIT:
3842    * Changed the "Function" constructor of type Ast.ctype to carry optional
3843      argument identifiers.
3844    * Changed the return type of TypeUtil.getFunction accordingly.
3845    * Type equality ignores the argument names.
3846    * TypeUtil.composite tries to preserve argument names but gives up quickly
3847      if there is a mismatch.