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 694, Thu Jul 27 16:00:25 2000 UTC revision 1715, Wed Nov 24 20:27:30 2004 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/11/24 14:30:00 CST
18    Tag: blume-20041124-position
19    Description:
20    
21    The compiler can now be compiled in a mode that makes structure Position
22    equal to Int64.  The default, however, is unchanged (Position = Int31)
23    for the time being.
24    
25    To enable 64-bit positions, use the following procedure:
26    
27    1. Start sml
28    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
29    3. Type
30          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
31    4. Run CMB.make() as usual.
32    
33    This is barely tested.  The only test so far was a little SML program
34    counting the number of characters in an 8-gigabyte file by
35    reading it character-by-character.  That test was successful.
36    
37    In support of 64-bit positions, a number of new functions have been
38    added to the runtime system.
39    
40    ----------------------------------------------------------------------
41    Name: Matthias Blume (blume (at) tti - c (dot) org)
42    Date: 2004/11/03 14:45:00 CST
43    Tag: blume-20041103-useFile
44    Description:
45    
46    Fixed a problem with unhelpful error messages related to problems with
47    .cm- or .sml files that appear as part of the sml command line.
48    
49    ----------------------------------------------------------------------
50    Name: Matthias Blume (blume (at) tti - c (dot) org)
51    Date: 2004/11/18 15:40:00 CST
52    Tag: Release_110_51
53    Description:
54    
55    New working version (110.51).  NEW BOOTFILES!
56    
57    ----------------------------------------------------------------------
58    Name: Matthias Blume (blume (at) tti - c (dot) org)
59    Date: 2004/11/18 15:35:00 CST
60    Tag: <none>
61    Description:
62    
63    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
64    Joemann).
65    
66    ----------------------------------------------------------------------
67    Name: Allen Leung (leunga (at) reservoir (dot) com)
68    Date: 2004/11/17 16:05:21 EST 2004
69    Tag: leunga-20041117-mlrisc-live-kill
70    Description:
71    
72       Added support for MLTree constructs LIVE and KILL
73    to all the architectures.
74    
75    ----------------------------------------------------------------------
76    Name: Matthias Blume (blume (at) tti - c (dot) org)
77    Date: 2004/11/13 00:20:00 CST
78    Tag: blume-20041113-versiontool
79    Description:
80    
81    - Stripped down the versiontool: It now only handles the version number.
82      The date string is generated at bootstrap time (during makeml).
83    
84    - In a previous commit, fixed a minor issue with how polyequal is being
85      translated.  In particular, the code now "looks through" abstractions.
86      This results in slightly fewer polyEqual warnings and hopefully slightly
87      more efficient code.  Important examples for where this matters are
88      the new int64 and word64 types.
89    
90    ----------------------------------------------------------------------
91    Name: Matthias Blume (blume (at) tti - c (dot) org)
92    Date: 2004/11/12 00:30:00 CST
93    Tag: blume-20041112-int64
94    Description:
95    
96    Structure Int64 fully hooked in.  (The implementation is not very
97    efficient, though.)
98    
99    ----------------------------------------------------------------------
100    Name: Matthias Blume (blume (at) tti - c (dot) org)
101    Date: 2004/11/11 17:30:00 CST
102    Tag: blume-20041111-more64
103    Description:
104    
105    All the pieces of Word64 are now there, with the exception of the
106    conversions from and to LargeWord. (Eventually these need to be identities,
107    but for the time being they don't even make sense because LargeWord is
108    32-bit wide.)
109    
110    Also started to add similar support for Int64, but major pieces of that
111    are still missing.
112    
113    ----------------------------------------------------------------------
114    Name: Matthias Blume (blume (at) tti - c (dot) org)
115    Date: 2004/11/11 00:15:00 CST
116    Tag: blume-20041111-word64
117    Description:
118    
119    Structure Word64 is now (almost) complete, word literals and patterns
120    seem to work.  There are a few odd pieces missing.  In particular,
121    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
122    at the moment.
123    
124    Making Word64 official would mean that LargeWord becomes Word64.  But
125    this requires extreme care because most word-word conversions have to
126    go through LargeWord, so making a mistake means loss of efficiency or
127    worse.  Eventually there will be a solution similar to (but actually
128    simpler than) what I did with IntInf.
129    
130    ----------------------------------------------------------------------
131    Name: Matthias Blume (blume (at) tti - c (dot) org)
132    Date: 2004/11/10 18:12:00 CST
133    Tag: blume-20041110-64bit
134    Description:
135    
136    More 64-bit hacking (but still not even half-way there yet).
137    Also, some assorted improvements to the handling of 8-bit words.
138    
139    ----------------------------------------------------------------------
140    Name: Matthias Blume (blume (at) tti - c (dot) org)
141    Date: 2004/11/09 17:50:00 CST
142    Tag: <none>
143    Description:
144    
145    Started adding some infrastructure for supporting 64-bit int- and
146    word-types.  (Still in its very early stages.)
147    
148    ----------------------------------------------------------------------
149    Name: Matthias Blume (blume (at) tti - c (dot) org)
150    Date: 2004/10/28 10:45:00 CDT
151    Tag: Release_110_50
152    Description:
153    
154    New working version (110.50).  NEW BOOTFILES!
155    
156                 =====================
157    
158    Also:
159    
160      - Changend config/srcarchiveurl from a file just
161        containing the URL string into a file containing
162        shell script code.  The code has access to the $VERSION variable.
163      - Made corresponding changes to config/install.sh and config/unpack.
164      - Default contents of config/srcarchiveurl uses $VERSION and
165        normally does not have to be edited to reflect a version change.
166    
167      (As a result, a version change can be done by just editing
168       config/version, the rest is now automatic.)
169    
170    ----------------------------------------------------------------------
171    Name: Matthias Blume (blume (at) tti - c (dot) org)
172    Date: 2004/10/27 17:50:00 CDT
173    Tag: blume-20041027-btrace-msg
174    Description:
175    
176    BackTrace.monitor now also reports the source of the exception that
177    triggered the trace.
178    
179    ----------------------------------------------------------------------
180    Name: Matthias Blume (blume (at) tti - c (dot) org)
181    Date: 2004/10/27 17:20:00 CDT
182    Tag: blume-20041027-x86-c-calls
183    Description:
184    
185    This is the HISTORY entry for two earlier commits, both concerning
186    the x86 c-calls code in MLRISC:
187    
188      - added a missing LOAD in the code that deals with struct arguments
189      - made sure the caller does not add the wrong number of bytes to the
190        stack pointer after a call of a function returning a struct
191        (the callee already pops the implicit argument which points to
192         the space reserved for the result)
193    
194    ----------------------------------------------------------------------
195    Name: Allen Leung (leunga (at) reservoir (dot) com)
196    Date: 2004/10/24 14:00:00 EST
197    Tag: leunga-20041024-x86-gas-fucomip
198    Description:
199    
200       John discovered a bug in the syntax of fucomip.
201       The opcodes FU?COMIP? have been changed to
202    
203           fu?comip? %st(i), %st
204    
205    ----------------------------------------------------------------------
206    Name: Matthias Blume (blume (at) tti - c (dot) org)
207    Date: 2004/10/20 15:06:00 CDT
208    Tag: blume-20041020-standalone-backtrace
209    Description:
210    
211    Added a mechanism for getting back-trace information from standalone
212    programs.  Here is how it works:
213    
214       1. The part of the program from which you want to get backtrace
215          information (usually the whole program) should be wrapped with
216          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
217          main program could be modified from something like
218    
219             fun main (pgm, args) = ...
220    
221         to
222    
223            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
224    
225       2. To be able to access BackTrace.monitor, you have to add
226          library $smlnj-tdp/plugins.cm to the .cm file that contains your
227          main function.
228    
229       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
230          CM might have created in the past for your project).
231    
232       4. Build the system using this command line:
233    
234             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
235                      myprog.cm MyProg.main myprog
236    
237          instead of the usual
238    
239             ml-build myprog.cm MyProg.main myprog
240    
241    I changed a library name:
242    
243       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
244    
245    New libraries:
246    
247       $smlnj-tdp/back-trace.cm
248          -- when loaded causes the back-trace plugin to be installed
249       $smlnj-tdp/coverage.cm
250          -- when loaded causes the coverage plugin to be installed
251    
252    ----------------------------------------------------------------------
253    Name: Matthias Blume (blume (at) tti - c (dot) org)
254    Date: 2004/10/18 16:45:00 CDT
255    Tag: blume-20041018-groupowner
256    Description:
257    
258    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
259    
260    Eliminated group owner specs from .cm files throughout the source tree.
261    
262    ----------------------------------------------------------------------
263    Name: Matthias Blume (blume (at) tti - c (dot) org)
264    Date: 2004/10/15 15:45:00 CDT
265    Tag: blume-20041015-coverage
266    Description:
267    
268    * Test coverage tool added!
269    * Further reorganization of tracing-, debugging-, and profiling support:
270    
271        - moved original BTImp -- now called BackTrace -- into a separate
272          library called $/trace-debug-profile.cm
273        - eliminated all mentions of BTrace from SMLofNJ.Internals
274        - only the instrumentation mechanism is now left in the compiler proper
275        - BackTrace module is a plugin which is NOT plugged in by default
276        - Coverage module is another such plugin
277    
278    To get the benefits of any of these plugin modules, the code in
279    question must be compiled with tdp instrumentation turned on.  This
280    can be done by setting SMLofNJ.Internals.TDP.mode to true.
281    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
282    
283    Plugins are selected at link time.  (Pre-compiled instrumented code
284    can be re-loaded with different plugins in effect.)  When an
285    instrumented module is linked, whatever plugins are at that time
286    enabled will come into effect for that module.
287    
288    To enable the back-trace plugin, load library $/trace-debug-profile.cm
289    and invoke BackTrace.install() (e.g., from the interactive prompt).
290    To enable the coverage plugin, load the same library and invoke
291    Coverage.install().
292    
293    Back-traces are generated automatically on uncaught exceptions and
294    when the code in question explicitly invokes BackTrace.trigger().
295    
296    Coverage (and execution frequency-) information must be queried
297    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
298    
299    ----------------------------------------------------------------------
300    Name: Matthias Blume (blume (at) tti - c (dot) org)
301    Date: 2004/10/14 17:40:00 CDT
302    Tag: blume-20041014-tdp-core
303    Description:
304    
305    Snapshot of a significant overhaul of how the trace/debug/profile support
306    is hooked into the system (specifically: Core and SMLofNJ.Internals).
307    
308    ----------------------------------------------------------------------
309    Name: Matthias Blume (blume (at) tti - c (dot) org)
310    Date: 2004/10/13 16:34:00 CDT
311    Tag: blume-20041013-tdp
312    Description:
313    
314    Some rationalization of names:
315    
316       structure BTrace -> structure TDPInstrument
317       etc.
318    
319    This is is preparation of using the original back-trace
320    instrumentation for other purposes.  "TDP" stands for
321    Trace/Debug/Profile.
322    
323    The control flag controlling whether instrumentation is on or off is now
324    registered under a different name, so instead of running sml as
325    
326     sml -Cinstrument.btrace-mode=true
327    
328    one has to say
329    
330     sml -Ctdp.instrument=true
331    
332    ----------------------------------------------------------------------
333    Name: Matthias Blume (blume (at) tti - c (dot) org)
334    Date: 2004/10/11 16:37:00 CDT
335    Tag: blume-20041011-regions
336    Description:
337    
338    Made some minor modifications to elabcore.sml to have source regions
339    be propagated more tightly -- resulting in better (i.e., smaller)
340    regions being reported in error- and debug messages.
341    
342    ----------------------------------------------------------------------
343    Name: Matthias Blume (blume (at) tti - c (dot) org)
344    Date: 2004/10/08 22:50:00 CDT
345    Tag: blume-20041008-cmkw
346    Description:
347    
348    Fixed handling of keywords in .cm files: After seeing "is" the lexer
349    treats subsequent occurrences of "group", "library", "source", "is",
350    "*", and "-" as ordinary identifiers rather than keywords.
351    
352    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
353    is supposed to accept a tool argument called "source", but this did
354    not work because of the clash with the keyword.
355    
356    ----------------------------------------------------------------------
357    Name: Matthias Blume (blume (at) tti - c (dot) org)
358    Date: 2004/10/07 16:00:00 CDT
359    Tag: blume-20041007-cleanup
360    Description:
361    
362    Assorted cleanup work:
363    
364      - got rid of intstrmap in favor of using the library's
365        hash table implementation
366      - threw out most of the pathnames stuff, as it was not used anyway
367      - simplified tokentable implementation
368      - fixed some minor spelling errors
369    
370    ----------------------------------------------------------------------
371    Name: Matthias Blume (blume (at) tti - c (dot) org)
372    Date: 2004/10/06 15:15:15 CDT
373    Tag: blume-20041006-handler
374    Description:
375    
376    Cleaned up the absyn to reflect the invariant that HANDLE always
377    carries a FNexp as part of the type definition.  This eliminates some
378    superfluous sanity checks at runtime down the road.
379    
380    Some minor cleanup of the btrace code.
381    
382    ----------------------------------------------------------------------
383    Name: Matthias Blume (blume (at) tti - c (dot) org)
384    Date: 2004/10/01 10:20:30 CDT
385    Tag: blume-20041001-slave
386    Description:
387    
388    Added hack to make slave mode work in the presence of the version
389    tool.  (Still, since the master does two passes over the code for
390    CMB.make, the release number gets bumped twice when slaves are
391    attached.  I don't know if this is worth fixing...)
392    
393    ----------------------------------------------------------------------
394    Name: Matthias Blume (blume (at) tti - c (dot) org)
395    Date: 2004/09/30 10:55:00 CDT
396    Tag: blume-20040930-version
397    Description:
398    
399    * Moved the "version" magic into its own little library under
400      src/system/smlnj/internal.  This avoids expensive reconstruction of
401      a stable src/compiler/core.cm.
402    
403    * At the same time, structure CompilerVersion is now known as
404      structure SMLNJVersion.
405    
406    * Arranged for the version tool to NOT kick in when rebuilding the system
407      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
408      Also, loading the versiontool does not work when rebuilding the system
409      because CM is not properly initialized at that time.
410    
411    ----------------------------------------------------------------------
412    Name: Matthias Blume (blume (at) tti - c (dot) org)
413    Date: 2004/09/29 14:00:00 CDT
414    Tag: blume-20040929-autoversion
415    Description:
416    
417    Implemented some CM magic to have
418        file src/compiler/TopLevel/main/version.sml
419    generated automagically.
420    The version is taken from two files: config/version and config/release.
421    The first is expected to contain a two-part version number such as 110.49.
422    The second should contain a single number, but it may be missing.
423    
424    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
425    time the version tool is loaded (which is the first time you say CMB.make),
426    then the tool will increment the value stored in config/release every
427    time CMB.make is invoked.
428    
429    The binfile format is now insensitive to anything beyond the first
430    two components of a version number, so bumping the release does not render
431    binfiles incompatible.  Auto-bumping can be used to keep track of versions
432    during development without invalidating existing binfiles.
433    
434    In any case, every CMB.make updates the date information in version.sml.
435    (This is the date that is printed in the banner.)
436    
437    ----------------------------------------------------------------------
438    Name: Matthias Blume (blume (at) tti - c (dot) org)
439    Date: 2004/09/28 10:53:00 CDT
440    Tag: blume-20040928-controls
441    Description:
442    
443    Some cleanup of the controls code.
444    
445    ----------------------------------------------------------------------
446    Name: Matthias Blume (blume (at) tti - c (dot) org)
447    Date: 2004/09/27 22:08:00 CDT
448    Tag: blume-20040927-controls
449    Description:
450    
451    Added two pieces of functionality to the Controls interface:
452    
453    1.   val save'restore: 'a control -> unit -> unit
454    
455       grabs the current value of the control in stage 1 and restores it
456       in stage 2.
457    
458    2.   val set' : 'a control * 'a -> unit -> unit
459    
460       stores the given value into the control in stage 2 (i.e., delayed)
461       but does all error checking in stage 1.
462       (This is for string controls that need to do parse their argument --
463        something that might fail.  In some cases, notably in CM, one
464        already knows the intended argument but wants to delay the actual
465        assignment until a time when error recovery would be more difficult.)
466    
467    Changed the handling of controls in tool arguments to classes "sml" and
468    "lazysml":
469       - use Controls.save'restore as a more robust way of restoring the
470         old value (in particular: without having to re-parse the string)
471       - use controls to handle the "overload" keyword in the init group
472         (I believe this change actually fixes a long-standing obscure bug.)
473    
474    ----------------------------------------------------------------------
475    Name: Matthias Blume (blume (at) tti - c (dot) org)
476    Date: 2004/09/27 17:00:00 CDT
477    Tag: blume-20040927-lazysml
478    Description:
479    
480    Added a new tool class called "lazysml" to CM's tool chest.  The only
481    difference to "sml" is that compilation is done with Control.lazysml
482    set to true.  A source of class "lazysml" is automatically recognized
483    by a file name suffix of ".lml".
484    
485    In addition to the above feature, the original class "sml" now also
486    supports a tool argument "lazy" which has the same effect.  As a
487    result, the following three lines are equivalent:
488    
489        foo.sml : lazysml
490        foo.sml : sml (lazy)
491        foo.sml (lazy)
492    
493    The setting goes into effect both during parsing and during
494    compilation.  The original setting is restored right after parsing and
495    after compilation, respectively.
496    
497    In addition to all the above, there is also a general mechanism to set
498    ANY of the "controls" that are available at the command line via
499    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
500    apply as well. (In fact, "lazy" is implemented as a special case of
501    the general mechanism.)
502    
503    The .cm file syntax uses a new keyword tool argument called "with".
504    There are several ways of indicating the desired settings:
505    
506        foo.sml (with:parser.quotations=true)
507        foo.sml (with:(name:parser.quotations value:true))
508        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
509        foo.sml (with:(name1=value1 name2=value2 ...))
510        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
511    
512        etc.
513    
514    Another possible abbreviation is to leave out the =v or value:v part
515    if the name refers to a boolean control (in which case the value is
516    taken to be true).  Thus, one could get lazy sml also by saying:
517    
518        foo.sml (with:parser.lazy-keyword=true)
519        foo.sml (with:parser.lazy-keyword)
520        foo.sml (with:(name:parser.lazy-keyword value:true))
521        foo.sml (with:(name:parser.lazy-keyword))
522    
523    ----------------------------------------------------------------------
524    Name: Matthias Blume (blume (at) tti - c (dot) org)
525    Date: 2004/09/24 16:22:00 CDT
526    Tag: blume-20040924-ppc-long-branch
527    Description:
528    
529    Turned message about "emiting long form of branch" off by default.
530    Added a control flag to turn it back on when desired.
531    
532    ----------------------------------------------------------------------
533    Name: Matthias Blume (blume (at) tti - c (dot) org)
534    Date: 2004/09/24 16:05:00 CDT
535    Tag: blume-20040924-rounding
536    Description:
537    
538    Applied patch for setting rounding modes under Mac OS X.  Thanks to
539    Melissa O'Neill for providing the code!
540    
541    ----------------------------------------------------------------------
542    Name: Matthias Blume (blume (at) tti - c (dot) org)
543    Date: 2004/09/23 17:30:00 CDT
544    Tag: blume-20040923-envvars
545    Description:
546    
547    1. Changed definition of type ControlRegistry.registry_tree to
548       include control_info (i.e., the name of the controlling environment
549       variable).
550    
551    2. Added command-line flags -e and -E to print the names of environment
552       variables that can be used to control internal settings.
553       (This uses the new API mentioned in 1.)
554    
555    ----------------------------------------------------------------------
556    Name: Matthias Blume (blume (at) tti - c (dot) org)
557    Date: 2004/09/13 16:50:00 CDT
558    Tag: Release_110_49
559    Description:
560    
561    New working version (110.49).  NEW BOOTFILES!
562    
563    ----------------------------------------------------------------------
564    Name: Matthias Blume
565    Date: 2004/09/13 16:20:00 CDT
566    Tag: blume-20040913-config-mlrisc
567    Description:
568    
569    Put target "mlrisc" back into the default list.
570    (There is no harm in having it, and some users have expressed their
571    wish to have "mlrisc" included by default.)
572    
573    ----------------------------------------------------------------------
574    Name: John Reppy
575    Date: 2004/09/13
576    Tag: jhr-20040913-signals
577    Description:
578    
579    Fixed the signal masking code to properly nest mask/unmask operations
580    on a per-signal basis.
581    
582    ----------------------------------------------------------------------
583    Name: Matthias Blume (blume (at) tti - c (dot) org)
584    Date: 2004/09/08 13:20:00 CDT
585    Tag: blume-20040908-heap-magic
586    Description:
587    
588    Bumped the heap macig number to 0x09082004 to account for the changed
589    layout of the ML frame under MacOS X.
590    
591    ----------------------------------------------------------------------
592    Name: Allen Leung (leunga (at) reservoir (dot) com)
593    Date: 2004/09/03 11:26:00 EST
594    Tag: leunga-20040903-cygwin-install
595    Description:
596    
597       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
598    The Cygwin runtime is turned on by setting the environment
599    variable SMLNJ_CYGWIN_RUNTIME to 1.
600    
601    ----------------------------------------------------------------------
602    Name: Matthias Blume (blume (at) tti - c (dot) org)
603    Date: 2004/08/31 17:15:00 CDT
604    Tag: blume-20040831-core
605    Description:
606    
607    Added some exports to src/compiler/core.cm upon request by J. Joemann.
608    
609    ----------------------------------------------------------------------
610    Name: Matthias Blume (blume (at) tti - c (dot) org)
611    Date: 2004/08/30 17:55:00 CDT
612    Tag: blume-20040830-installer
613    Description:
614    
615    Upon request by Johannes Joemann:
616    
617    - improved ML code of installer to fall back to coping when renaming
618      fails (i.e., when source and target are on different file systems);
619      the code compiles but has yet to be tested in anger
620    - removed mlrisc from list of default targets (config/targets)
621    
622    ----------------------------------------------------------------------
623    Name: Matthias Blume (blume (at) tti - c (dot) org)
624    Date: 2004/08/27 17:20:00 CDT
625    Tag: blume-20040827-ptreql
626    Description:
627    
628    Added ptreql primop to structure InlineT (upon request from Larry
629    Paulson).
630    
631    ----------------------------------------------------------------------
632    Name: Allen Leung (leunga (at) reservoir (dot) com)
633    Date: 2004/08/15 21:21:00 EST
634    Tag: leunga-110_48-udgraph
635    Description:
636    
637    Another bug fix from Carl Hauser:
638    
639    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
640    > 48c48
641    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
642    > e::L)
643    > ---
644    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
645    > Without this, any deletion of an edge in an undirected graph does severe
646    > violence to the graph.
647    
648    ----------------------------------------------------------------------
649    Name: Allen Leung (leunga (at) reservoir (dot) com)
650    Date: 2004/08/10 23:35:00 EST
651    Tag: leunga-110_48-ppc
652    Description:
653    
654       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
655    
656    ----------------------------------------------------------------------
657    Name: Matthias Blume (blume (at) tti - c (dot) org)
658    Date: 2004/08/10 12:00:00 CDT
659    Tag: Release_110_48
660    Description:
661    
662    New working version (110.48).  NEW BOOTFILES!
663    
664    ----------------------------------------------------------------------
665    Name: Allen Leung (leunga (at) reservoir (dot) com)
666    Date: 2004/08/09 12:21:00 EST
667    Tag: leunga-110_47-dijsktra
668    Description:
669    
670    Bug fix from Carl Hauser:
671    
672    single_source_shortest_paths in dijkstra.sml was observed to get wrong
673    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
674    
675    The problem is that following the expression A.update(dist,s,Num.zero)
676    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
677    
678    ----------------------------------------------------------------------
679    Name: Matthias Blume (blume (at) tti - c (dot) org)
680    Date: 2004/08/06 18:10:00 CDT
681    Tag: blume-20040806-cmdline
682    Description:
683    
684    Fiddled with handling of command-line options:
685    
686      * sml now quits after processing the command line
687        if -H, -S, -h<n>, or -s<n> appears as the last
688        command-line argument
689      * a new option -q terminates the session when encountered on
690        the command line; subsequent arguments will be ignored
691      * bug fixes: short (erroneous) arguments are no longer ignored
692        completely
693    
694    ----------------------------------------------------------------------
695    Name: Allen Leung (leunga (at) reservoir (dot) com)
696    Date: 2004/08/04 18:17:00 EST
697    Tag: leunga-110_47-ppc-ibm-asm
698    Description:
699    
700      - Added minimal IBM assembly syntax support for PowerPC.
701    
702      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
703        been altered in the runtime.  We need an automatic way to keep the file
704        in sync.
705    
706    ----------------------------------------------------------------------
707    Name: Matthias Blume (blume (at) tti - c (dot) org)
708    Date: 2004/08/04 14:00:00 CDT
709    Tag: Release_110_47
710    Description:
711    
712    New working version (110.47).  NEW BOOTFILES!
713    
714    ----------------------------------------------------------------------
715    Name: Matthias Blume (blume (at) tti - c (dot) org)
716    Date: 2004/08/03 14:25:00 CDT
717    Tag: blume-20040803-callingconv
718    Description:
719    
720    Added low-level support for choosing C calling conventions by
721    twiddling the type of rawccall.  (See
722    src/compiler/Semant/types/cproto.sml for details.)
723    
724    ----------------------------------------------------------------------
725    Name: Matthias Blume (blume (at) tti - c (dot) org)
726    Date: 2004/08/02 15:55:00 CDT
727    Tag: blume-20040802-backout
728    Description:
729    
730    Backed out of change to win32-filesys.c.  The earlier patch to
731    get_file_time caused CM to produce files with the wrong time stamp.
732    
733    ----------------------------------------------------------------------
734    Name: Matthias Blume (blume (at) tti - c (dot) org)
735    Date: 2004/08/02 14:45:00 CDT
736    Tag: blume-20040802-nlffi-win32
737    Description:
738    
739    Added NLFFI support for Win32, adapted from a patch provided by David
740    Hansel.  This is currently completely untested.  Also, the issue
741    concerning stdcall vs. ccall is still unresolved.
742    
743    ----------------------------------------------------------------------
744    Name: Matthias Blume (blume (at) tti - c (dot) org)
745    Date: 2004/07/30 17:55:00 CDT
746    Tag: blume-20040730-various
747    Description:
748    
749    Gearing up towards 110.47...
750    
751    - various minor bugfixes to ml-nlffigen
752    - a beginning of a manual for nlffi
753    
754    - eliminated 'export name=value' in config/install.sh as this does
755      not work with certain versions of /bin/sh
756      (Thanks to David King at Motorola for catching this.)
757    
758    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
759      - added a test for tm==NULL to gmtime.c and localtime.c
760      - applied patch for incorrect GetFileTime under win32
761      - toSeconds -> toMilliseconds in Win32/win32-process.sml
762    
763    ----------------------------------------------------------------------
764    Name: Matthias Blume (blume (at) tti - c (dot) org)
765    Date: 2004/07/21 18:20:00 CDT
766    Tag: blume-20040721-nlffigen
767    Description:
768    
769    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
770      for a typedef to an incomplete type, but leave out the "typ" member.
771      (This is just for consistency.)
772    - Started to produce what is supposed to become better (i.e., comprehensive)
773      documentation of what ml-nlffigen does and produces.
774    
775    ----------------------------------------------------------------------
776    Name: Matthias Blume (blume (at) tti - c (dot) org)
777    Date: 2004/07/14 16:25:00 CDT
778    Tag: blume-20040714-union
779    Description:
780    
781    Added C_UNION to c-calls/c-types.sml and updated the machinery
782    (ml-nlffigen, cproto.sml) that conveys C function interface
783    information to the code generator.
784    
785    However, the actual architecture-specific implementation of function
786    arguments and results that are C unions is still not implemented.
787    
788    ----------------------------------------------------------------------
789    Name: Allen Leung (leunga (at) reservoir (dot) com)
790    Date: 2004/07/14 14:38:00 EST
791    Tag: leunga-110_46_1-ppc-lwzu
792    Description:
793    
794        Added these instructions to the PowerPC architecture:
795        LBZU(X), LHZU(X), LWZU(X),
796        STWU(X), STFDU, STFSU
797    
798        etc...
799    
800        Note: I haven't added their instruction encoding into the description.
801    
802    ----------------------------------------------------------------------
803    Name: Allen Leung (leunga (at) reservoir (dot) com)
804    Date: 2004/07/13 15:04:00 EST
805    Tag: leunga-110_46_1-ppc-lwarx
806    Description:
807    
808       Added the two instructions LWARX and STWCX to the PowerPC
809    instruction set.
810    
811       A (untested) rewrite of loop-structure.sml.  The old version
812    is completely broken.
813    
814    ----------------------------------------------------------------------
815    Name: Matthias Blume (blume (at) tti - c (dot) org)
816    Date: 2004/07/13 13:50:00 CDT
817    Tag: blume-20040713-nlffi
818    Description:
819    
820    - use paramAlloc to report c-calls with too many arguments
821      (for PPC version where parameter area is pre-allocated)
822    - added ccall_maxargspace to machspec (to implement the above)
823    - made "make" commend in CM's "make" tool configurable
824    - added option (default: on) for passing the name of the SML/NJ's "bin"
825      directory to "make"; the call looks like this:
826    
827         make <options> SMLNJ_BINDIR=<dir> <target>
828    
829      This can be used by the Makefile to, e.g., pick the "right" version
830      of ml-nlffigen.
831    - minor code tweaks
832    
833    ----------------------------------------------------------------------
834    Name: Matthias Blume (blume (at) tti - c (dot) org)
835    Date: 2004/07/12 22:50:00 CDT
836    Tag: blume-110_46_1-macosx-nlffi
837    Description:
838    
839    NLFFI under Mac OS X now working (sort of).  This is largely untested,
840    though.
841    
842    Note:  1. You have to make a new, clean build of the runtime system.
843           2. There are new BOOTFILES, you have to use them!
844              (Doing the bootstrap process yourself would be *very* painful!
845               If you absolutely have to do it, build the system under
846               a different architecture and then cross-compile.)
847    
848    Version bumped to 110.46.1 to account for runtime data format changes.
849    
850    ----------------------------------------------------------------------
851    Name: Matthias Blume (blume (at) tti - c (dot) org)
852    Date: 2004/06/18 14:30:00 CDT
853    Tag: blume-20040618-unix
854    Description:
855    
856    Changed the implementation of structure Unix so that the same stream
857    is returned every time one of the {text,bin}{In,Out}streamOf functions
858    is invoked on the same proc.  This is not what the spec currently
859    says -- although IMO it arguably should.  (See discussion below.)
860    
861    ----------------------------------------------------------------------
862    Name: Matthias Blume (blume (at) tti - c (dot) org)
863    Date: 2004/06/17 18:15:00 CDT
864    Tag: Release_110_46
865    Description:
866    
867    New working version (110.46).  NEW BOOTFILES!
868    
869    ----------------------------------------------------------------------
870    Name: Matthias Blume (blume (at) tti - c (dot) org)
871    Date: 2004/06/17 17:20:00 CDT
872    Tag: blume-20040617-timer-unix
873    Description:
874    
875    Changed the interface of structures Timer and Unix to match the most
876    recent Basis spec.
877    
878    In the case of Unix there still seems to be an open/weird issue:
879    
880       The {text,bin}{In,Out}streamOf functions are supposed to create
881       fresh streams whenever they are called -- as opposed to have them
882       return the same stream every time.  This design is supposed to
883       prevent space leaks caused by proc values hanging on to streams.
884    
885       The reap function, on the other hand, is supposed to close the
886       streams.  This cannot be done without having a handle on the
887       stream in proc after all...
888    
889    I took the liberty to implement the following stopgap solution:
890    
891      The proc value hangs on to the most recently created stream(s).
892      Reap closes those.  If either or both of the two streams hadn't
893      been created at all yet, then reap will close the corresponding
894      file descriptors directly.
895    
896    PS: I don't understand the original space leak argument anymore.  If
897    a proc hangs on to the imperative stream, then I/O operations on those
898    will advance the state of the cached stream and avoid the space leak.
899    
900    ----------------------------------------------------------------------
901    Name: Matthias Blume (blume (at) tti - c (dot) org)
902    Date: 2004/05/28 16:45:00 CDT
903    Tag: blume-20040528-basis
904    Description:
905    
906    Added signature PACK_REAL and exported functor PrimIO.
907    
908    ----------------------------------------------------------------------
909    Name: Matthias Blume (blume (at) tti - c (dot) org)
910    Date: 2004/05/25 16:00:00 CDT
911    Tag: blume-20040525-group-owner
912    Description:
913    
914    CM now ignores (but still accepts) the "owner" information in group
915    descriptions.  The owner of a group is its next enclosing
916    library. Each group must have a unique owner.  (There is a virtual
917    "toplevel" library that owns groups which are not nested within a real
918    library.)  Previously, each group had to explicitly declare its owner,
919    and CM would check that such a declaration is correct.  The new scheme
920    is to have CM check that for each group there is precisely one owning
921    library.
922    
923    The advantage of the new scheme is that the programmer no longer needs
924    to maintain the somewhat annoying owner information.  The downside is
925    that CM cannot enforce the ownership rule across multiple runs of
926    CM.make.  Fortunately, enclosing the same group in two different
927    libraries A and B which are not part of the same program does not
928    cause real problems.
929    
930    ----------------------------------------------------------------------
931    Name: Matthias Blume (blume (at) tti - c (dot) org)
932    Date: 2004/05/20 16:00:00 CDT
933    Tag: blume-20040520-win32
934    Description:
935    
936    Made the win32 version work again.  (Strangely, a misplaced comma had
937    slipped into win32-process.c which prevented the runtime from being
938    compiled correctly.)
939    
940    Also, included a minor addition to ml-build.bat analogous to what was
941    done in blume-20040519-ml-build.
942    
943    ----------------------------------------------------------------------
944    Name: Matthias Blume (blume (at) tti - c (dot) org)
945    Date: 2004/05/19 22:10:00 CDT
946    Tag: blume-20040519-ml-build
947    Description:
948    
949    Arranged for ml-build to clean up after itself a little bit better.
950    The script generates a temporary SML source file and compiles it using
951    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
952    gets rid of those at the end, so they don't accumulate under .cm.
953    
954    This required a minor change to install.sh because the name of the
955    metadata directory (default: .cm) is actually configurable at
956    installation time.
957    
958    ----------------------------------------------------------------------
959    Name: Matthias Blume (blume (at) tti - c (dot) org)
960    Date: 2004/05/18 15:50:00 CDT
961    Tag: blume-20040518-mkreader
962    Description:
963    
964    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
965    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
966    
967    ----------------------------------------------------------------------
968    Name: Matthias Blume (blume (at) tti - c (dot) org)
969    Date: 2004/05/11 14:35:00 CDT
970    Tag: blume-20040511-win32sock
971    Description:
972    
973    Added previously missing support for many socket-related functions
974    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
975    for the voluminous patch!
976    
977    (I have not tested this patch under win32 yet.)
978    
979    Here is David's e-mail:
980    
981    Hi,
982    
983    Attached to this email you find a diff against sml/nj 110.45
984    that will enable socket support under Windows.
985    
986    To apply the patch (using unix or cygwin)
987    1) gunzip runtime.diff.gz
988    2) "cd" into "src/runtime" in the source tree of a fresh
989       110.45 installation.
990    3) patch -p 1 < [your/path/to]runtime.diff
991    
992    The code compiles fine but has NOT yet been extensively tested.
993    I only ran a few tests for basic socket client functionality
994    (which worked fine).  Especially the functions that use ioctl
995    are not tested at all and might not work (see below).
996    
997    I implemented this since we want to move to a newer version of sml/nj
998    but need socket support in order to use it.  This is the first time I
999    even had a look at the sml/nj source,  so please review my changes
1000    before making this part of the distribution!  Here are a few issues
1001    that I think might be better for someone to solve who is more
1002    familiar with the sml/nj source (and socket programming):
1003    
1004    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1005      exception since I could not figure out what the windows equivalent
1006      of these functions is
1007    
1008    - In sockets-osdep.h there are a some #include statements that are
1009      only used in a few files that include sockets-osdep.h
1010    
1011    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1012      does not process its return value since I don't know how to
1013      report an error upwards.
1014    
1015    - It would probably be good to have a call to WSACleanup() when
1016      the library is unloaded (if there is such a possibility).
1017      Otherwise I think Windows will take care of this automatically
1018      when the process finishes.
1019    
1020    - I used ioctlsocket() as a replacement for ioctl() but I have
1021      no idea if that is actually the proper replacement on Windows.
1022    
1023    - All these issues are marked in the code by "FIXME" comments.
1024    
1025    We use sml/nj extensively in our products and are quite happy
1026    with it.  I hope this contribution will help you.
1027    
1028    Keep up the good work!
1029    
1030    David
1031    
1032    ----------------------------------------------------------------------
1033    Name: Matthias Blume (blume (at) tti - c (dot) org)
1034    Date: 2004/05/11 14:20:00 CDT
1035    Tag: blume-20040511-installml
1036    Description:
1037    
1038    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1039    the patch.)
1040    
1041    ----------------------------------------------------------------------
1042    Name: Matthias Blume (blume (at) tti - c (dot) org)
1043    Date: 2004/05/11 14:05:00 CDT
1044    Tag: blume-20040511-nlffi-netbsd
1045    Description:
1046    
1047    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1048    the patch.)
1049    
1050    ----------------------------------------------------------------------
1051    Name: Matthias Blume (blume (at) tti - c (dot) org)
1052    Date: 2004/05/11 12:05:00 CDT
1053    Tag: blume-20040511-exports
1054    Description:
1055    
1056    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1057    export lists in compiler-related .cm-files.
1058    
1059    ----------------------------------------------------------------------
1060    Name: Matthias Blume (blume (at) tti - c (dot) org)
1061    Date: 2004/05/11 11:35:00 CDT
1062    Tag: blume-20040511-allsource
1063    Description:
1064    
1065    The installer now honors the "src-smlnj" target again, although its meaning
1066    has changed from "all sources required for the compiler" to "all sources
1067    the installer knows about".  In other words, if you enable "src-smlnj"
1068    in the "targets" file, then the installer will pull in sources for
1069    everything.  (Notice that this refers to source code only.  Compiled
1070    code is still only installed for modules that were requested explicitly
1071    or which are required for other modules that were requested explicitly.)
1072    
1073    ----------------------------------------------------------------------
1074    Name: Matthias Blume (blume (at) tti - c (dot) org)
1075    Date: 2004/04/23 17:40:00 CDT
1076    Tag: blume-20040423-ieee-scan
1077    Description:
1078    
1079    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1080    in the exponent calculation we get INF or ZERO (depending on the mantissa
1081    and the sign of the exponent).
1082    
1083    ----------------------------------------------------------------------
1084    Name: Matthias Blume (blume (at) tti - c (dot) org)
1085    Date: 2004/04/23 10:40:00 CDT
1086    Tag: blume-20040423-ml-build
1087    Description:
1088    
1089    The ml-build script now terminates with a non-0 status when something
1090    goes wrong.
1091    
1092    ----------------------------------------------------------------------
1093    Name: Matthias Blume (blume (at) tti - c (dot) org)
1094    Date: 2004/04/22 16:35:00 CDT
1095    Tag: blume-20040422-Option
1096    Description:
1097    
1098    Made exception Option to be the same as exception Option.Option
1099    (as it should be).
1100    
1101    ----------------------------------------------------------------------
1102    Name: Allen Leung (leunga (at) reservoir (dot) com)
1103    Date: 2004/03/19 14:40:00 EST
1104    Tag: leunga-20040319-cygwin-nlffi
1105    Description:
1106    
1107    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1108    of SML/NJ.  The problem is that
1109    
1110       lib = dlopen(NULL, ...)
1111       f   = dlsym(lib, "malloc");
1112    
1113    does not work on Windows unless we explicitly export symbols
1114    such as 'malloc' during linking.  We fixed this by explicitly
1115    exporting the required symbols with the magic gcc incantation:
1116    
1117        -Wl,--export-all cygwin.def
1118    
1119    where cygwin.def is a file containing all the symbols that
1120    we wish to export.
1121    
1122    I suspect this is a Windows problem and we'll have to
1123    do the same (somehow with windows compilers) when
1124    we build the native win32 version with the system
1125    calls LoadLibrary/GetProcAddress.
1126    
1127    ----------------------------------------------------------------------
1128    Name: Matthias Blume (blume (at) tti - c (dot) org)
1129    Date: 2004/03/04 16:35:00 CST
1130    Tag: blume-20040304-intinf-fmt
1131    Description:
1132    
1133    Fixed problem with IntInf.fmt (sign would show up on the right instead
1134    of on the left for BIN, OCT, and HEX).
1135    
1136    ----------------------------------------------------------------------
1137    Name: Matthias Blume (blume (at) tti - c (dot) org)
1138    Date: 2004/03/04 11:25:00 CST
1139    Tag: blume-20040304-symlinks
1140    Description:
1141    
1142    Fixed problem with installer script (unix only) where bin/ml-yacc and
1143    friends pointed (via symlinks) to absolute locations instead of just
1144    .run-sml.  This was reported by Vesa A Norrman.
1145    
1146    ----------------------------------------------------------------------
1147    Name: Matthias Blume (blume (at) tti - c (dot) org)
1148    Date: 2004/02/13 14:50:00 CST
1149    Tag: Release_110_45
1150    Description:
1151    
1152    New working version (110.45). New bootfiles.
1153    
1154    ----------------------------------------------------------------------
1155    Name: Matthias Blume (blume (at) tti - c (dot) org)
1156    Date: 2004/01/26 15:15:15 CST
1157    Tag: blume-20040126-toplevel
1158    Description:
1159    
1160    Improved handling of exceptions at the interactive toplevel.
1161    
1162    ----------------------------------------------------------------------
1163    Name: Matthias Blume (blume (at) tti - c (dot) org)
1164    Date: 2004/01/26 11:25:00
1165    Tag: blume-20040126-app
1166    Description:
1167    
1168    Type of top-level "app" corrected.
1169    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1170    
1171    ----------------------------------------------------------------------
1172    Name: Matthias Blume (blume (at) tti - c (dot) org)
1173    Date: 2003/11/18 17:10 CST
1174    Tag: blume-20031118-basis-fiddle
1175    Description:
1176    
1177    - changed Timer interface to what might become the spec
1178    - POSIX_FLAGS -> BIT_FLAGS according to spec
1179    - some other minor discrepancies wrt. spec eliminated
1180    
1181    ----------------------------------------------------------------------
1182    Name: Matthias Blume (blume (at) tti - c (dot) org)
1183    Date: 2003/11/06 12:00:00 CST
1184    Tag: Release_110_44
1185    Description:
1186    
1187    New working version (110.44). New bootfiles.
1188    
1189    ----------------------------------------------------------------------
1190    Name: Matthias Blume (blume (at) tti - c (dot) org)
1191    Date: 2003/11/04 11:50:00 CST
1192    Tag: blume-20031104-move-libraries
1193    Description:
1194    
1195    Eliminated the "dont_move_libraries" directive in config/targets.
1196    (The mechanism was broken and could not be fixed easily.  Moreover,
1197    there does not seem to be any reason not to move all libraries into
1198    lib during installation.  I originally implemented this directive as a
1199    backward-compatibility feature when I first introduced the new CM.
1200    Now that things have been stable for a long time and going back to the
1201    old CM is not an option, there is no reason to keep it around.)
1202    
1203    ----------------------------------------------------------------------
1204    Name: Matthias Blume (blume (at) tti - c (dot) org)
1205    Date: 2003/11/03 16:00:00 CST
1206    Tag: blume-20031103-installdir
1207    Description:
1208    
1209    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1210    Richards for pointing out the problem and providing the solution.)
1211    
1212    ----------------------------------------------------------------------
1213    Name: Matthias Blume (blume (at) tti - c (dot) org)
1214    Date: 2003/10/01 17:05:00 CDT
1215    Tag: blume-20031001-lal-mlrisc
1216    Description:
1217    
1218    MLRISC bug fix from Lal.
1219    
1220    ----------------------------------------------------------------------
1221    Name: Matthias Blume (blume (at) tti - c (dot) org)
1222    Date: 2003/09/30 16:10:00 CDT
1223    Tag: blume-20030930-primio-bat
1224    Description:
1225    
1226    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1227    2. Improved .bat files (for Win32 port) to make things work under Win95.
1228       (thanks to Aaron S. Hawley for this one)
1229    
1230    ----------------------------------------------------------------------
1231    Name: Matthias Blume (blume (at) tti - c (dot) org)
1232    Date: 2003/09/26 16:05:00 CDT
1233    Tag: blume-20030926-wrappriv
1234    Description:
1235    
1236    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1237    
1238    ----------------------------------------------------------------------
1239    Name: Matthias Blume (blume (at) tti - c (dot) org)
1240    Date: 2003/09/26 15:00:00 CDT
1241    Tag: blume-20030926-110_43_3
1242    Description:
1243    
1244    - additional cleanup
1245    - version number bump, NEW BOOTFILES
1246    
1247    ----------------------------------------------------------------------
1248    Name: Matthias Blume (blume (at) tti - c (dot) org)
1249    Date: 2003/09/26 12:00:00 CDT
1250    Tag: blume-20030926-ppautoload
1251    Description:
1252    
1253    I modified the read-eval-print loop so that the autoloader gets
1254    invoked whenever the prettyprinter tries to look up a symbol that
1255    is not currently defined in the toplevel environment but which
1256    appears in CM's autoload registry.  As a result, we see far fewer of
1257    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1258    
1259    In addition to this I tried to clean up some pieces of the Basis
1260    implementation (e.g., Socket, Word8Array) in order to prevent other
1261    instances of these ?.Foo.Bar.xxx names from being printed.
1262    
1263    The mechanism that picks names for types still needs some work, though.
1264    (Right now it seems that if there is a type A.t which is defined to
1265    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1266    "?.B.u" although the perhaps more sensible solution would be to use
1267    "A.t" in this case.  In other words, the prettyprinter should follow
1268    a chain of DEFtycs not farther than there are corresponding toplevel
1269    names in the current environment.)
1270    
1271    ----------------------------------------------------------------------
1272    Name: Matthias Blume (blume (at) tti - c (dot) org)
1273    Date: 2003/09/24 16:31:00 CDT
1274    Tag: blume-20030924-installer
1275    Description:
1276    
1277    Another installer tweak:  All the ML code for the installer is now
1278    compiled during CMB.make and put into a little library called
1279    $smlnj/installer.cm.  The installation then simply invokes
1280    
1281       sml -m $smlnj/installer.cm
1282    
1283    and everything happens automagically.
1284    
1285    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1286    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1287            and BINDIR.
1288    
1289    The new scheme guarantees that the ML code responsible for the installation
1290    is in sync with the APIs of the main system.  Also, the installer is
1291    somewhat faster because the installer script is precompiled.
1292    
1293    ----------------------------------------------------------------------
1294    Name: Matthias Blume (blume (at) tti - c (dot) org)
1295    Date: 2003/09/24 15:35:00 CDT
1296    Tag: blume-20030924-synsock
1297    Description:
1298    
1299    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1300    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1301    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1302    implementations of non-blocking operations from CML's Socket
1303    structure.
1304    
1305    ----------------------------------------------------------------------
1306    Name: Matthias Blume (blume (at) tti - c (dot) org)
1307    Date: 2003/09/24 15:10:05 CDT
1308    Tag: blume-20030924-sockets
1309    Description:
1310    
1311    1. Fixed SOCKET API and implementation to match Basis spec.
1312       This required changing the internal representation of sockets to one
1313       that remembers (for each socket file descriptor) whether it is currently
1314       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1315       call is made only if the state actually needs to change).
1316    
1317    2. OS-specific details of sockets were moved into separate files, thus
1318       making it possible to unify the bulk of the socket implementations
1319       between Unix and Win32.
1320    
1321    3. CML's socket API changed accordingly.
1322       (Note that we need to remove non-blocking functions from this API
1323       since they are redundant in the case of CML!)
1324    
1325    4. CML's socket implementation now makes use of non-blocking functions
1326       provided by Basis, thus removing all OS-dependent code from this part
1327       of CML.
1328    
1329    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1330    
1331    ----------------------------------------------------------------------
1332    Name: Matthias Blume (blume (at) tti - c (dot) org)
1333    Date: 2003/09/22 12:10:00 CDT
1334    Tag: blume-20030922-110_43_2
1335    Description:
1336    
1337    Made a new interim version and bootfiles for developer's bootstrapping
1338    convenience.
1339    
1340       110.43.2 -- NEW BOOTFILES
1341    
1342    ----------------------------------------------------------------------
1343    Name: Matthias Blume (blume (at) tti - c (dot) org)
1344    Date: 2003/09/19 15:55:00 CDT
1345    Tag: blume-20030919-cmdir
1346    Description:
1347    
1348    1. new-install.sh -> install.sh
1349    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1350    3. tweaked installer so that another name instead of .cm can be chosen
1351       at install time (by setting the CM_DIR_ARC environment variable
1352       during installation); once installation is complete, the name is
1353       fixed
1354    
1355    ----------------------------------------------------------------------
1356    Name: Matthias Blume (blume (at) tti - c (dot) org)
1357    Date: 2003/09/18 16:00:00 CDT
1358    Tag: blume-20030918-110_43_1
1359    Description:
1360    
1361    Made a new interim version and bootfiles for developer's bootstrapping
1362    convenience.
1363    
1364       110.43.1 -- NEW BOOTFILES
1365    
1366    ----------------------------------------------------------------------
1367    Name: Matthias Blume (blume (at) tti - c (dot) org)
1368    Date: 2003/09/18 15:20:00 CDT
1369    Tag: blume-20030918-misc
1370    Description:
1371    
1372    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1373       this seems to be controversial at the moment) and used those in
1374       Posix.ProcEnv.times.
1375    
1376    2. Added Time.{from,to}Nanoseconds to Time.
1377    
1378    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1379       For example, fromLargeInt never needs to look at more than 3 "big
1380       digits" to get its 53 bits of precision.
1381    
1382    ----------------------------------------------------------------------
1383    Name: Matthias Blume (blume (at) tti - c (dot) org)
1384    Date: 2003/09/17 16:30:00 CDT
1385    Tag: blume-20030917-real32-slices
1386    Description:
1387    
1388    Added an entry to the primitive environment
1389    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1390    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1391    
1392    Removed some of the "magic" constants in real64.sml and replaced them
1393    with code that generates these values from their corresponding
1394    integer counterparts.
1395    
1396    Made all(?) the slice-related changes to the Basis and made everything
1397    compile again...
1398    
1399    ----------------------------------------------------------------------
1400    Name: Matthias Blume (blume (at) tti - c (dot) org)
1401    Date: 2003/09/15 17:45:00 CDT
1402    Tag: blume-20030915-rbase
1403    Description:
1404    
1405    Fixed bug in Real.fromLargeInt.
1406    
1407    ----------------------------------------------------------------------
1408    Name: Matthias Blume (blume (at) tti - c (dot) org)
1409    Date: 2003/09/13 18:11:00 CDT
1410    Tag: blume-20030913-libinstall
1411    Description:
1412    
1413    Minor bugfix in config/libinstall (set anchor with path to
1414    standalone tool after installing it, otherwise libraries that
1415    need ml-lex or ml-yacc won't compile the first time the installer
1416    runs).
1417    
1418    ----------------------------------------------------------------------
1419    Name: Matthias Blume (blume (at) tti - c (dot) org)
1420    Date: 2003/09/12 11:45:00 CDT
1421    Tag: blume-20030912-various
1422    Description:
1423    
1424     - fixed bug in Real.toLargeInt
1425     - fixed bug in Posix.ProcEnv.times
1426     - changed inputLine functions to return an option
1427     - minor installer improvements / bugfixes
1428     - changed default @SMLalloc parameter for x86/celeron to 64k
1429    
1430    ----------------------------------------------------------------------
1431    Name: Matthias Blume (blume (at) tti - c (dot) org)
1432    Date: 2003/09/09 22:00:00 CDT
1433    Tag: Release_110_43
1434    Description:
1435    
1436        New working release 110.43.  New bootfiles.
1437    
1438    ----------------------------------------------------------------------
1439    Name: Matthias Blume (blume (at) tti - c (dot) org)
1440    Date: 2003/09/09 19:20:00 CDT
1441    Tag: blume-20030909-installer
1442    Description:
1443    
1444    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1445    Modified config/install.bat to take advantage of it.  Also modified
1446    config/install.sh (and called it config/new-install.sh) to take advantage
1447    of it on Unix systems.  (The SML code is (supposed to be) platform-
1448    independent.)
1449    
1450    The installer can now install everything under Win32
1451    as well as under *nix as long as it compiles.
1452    
1453    Other changes:
1454    
1455      - made CML compile again under Win32
1456      - made eXene compile under Win32 (by providing a fake structure UnixSock
1457        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1458      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1459        same as type int;  under Win32 it isn't
1460      - fixed some slice-related problems in the win32-specific parts of CML
1461      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1462        (under Win32, the volume name is case-insensitive, and the
1463        OS.Path code compares volume names for equality)
1464    
1465    ----------------------------------------------------------------------
1466    Name: Matthias Blume (blume (at) tti - c (dot) org)
1467    Date: 2003/09/08 11:55:00 CDT
1468    Tag: blume-20030908-fullpath
1469    Description:
1470    
1471    Made Win32 version of OS.FileSys.fullPath return current directory
1472    when given an empty string.  This is what the spec says, and incidentally,
1473    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1474    cases when presented with the name of a non-existing .cm file.)
1475    
1476    ----------------------------------------------------------------------
1477    Name: Matthias Blume (blume (at) tti - c (dot) org)
1478    Date: 2003/09/04 16:30:00 CDT
1479    Tag: blume-20030905-slices-etc
1480    Description:
1481    
1482      1. Changed interface to vectors and arrays in Basis to match
1483         (draft) Basis spec.
1484      2. Added signatures and implementations of slices according to
1485         Basis spec.
1486      3. Edited source code throughout the system to make it compile again
1487         under 1. and 2.  (In some cases code had to be added to have it
1488         match the new signatures.)
1489      4. MLRISC should be backward-compatible: the copies of the originals
1490         of files that needed to change under 3. were retained, the .cm files
1491         check the compiler version number and use old versions when
1492         appropriate.
1493      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1494         dirstream -> string option (in accordance with Basis spec).
1495      6. When generating code that counts lines, ml-lex used function
1496         CharVector.foldli, taking advantage of its old interface.
1497         This has been replaced with the corresponding code from
1498         CharVectorSlice.  (html-lex must be re-lexed!)
1499      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1500         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1501      8. Removed temporary additions (fromInternal, toInternal) from the
1502         (now obsolete) IntInf in smlnj-lib/Util.
1503      9. Cleaned up structure Byte.
1504     10. Added localOffset, scan, and fromString to Date (according to spec).
1505         Cleaned/corrected implementation of Date.
1506         (Still need to check for correctness; implement better canonicalizeDate.)
1507     11. Added "scan" to signature IEEE_REAL.
1508     12. Some improvements to IntInf [in particular: efficiency-hack for
1509         mod and rem when second operand is 2 (for parity checks).]
1510     13. Changed representation of type Time.time, using a single IntInf.int
1511         value counting microseconds.  This considerably simplified the
1512         implementation of structure Time.  We now support negative time
1513         values; scan and fromString handle signs.
1514     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1515         ArraySlice).
1516    
1517    ----------------------------------------------------------------------
1518    Name: Matthias Blume (blume (at) tti - c (dot) org)
1519    Date: 2003/08/28 17:00:00 CDT
1520    Tag: blume-20030828-intinf
1521    Description:
1522    
1523    This is a major update which comes with a version number bump
1524    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1525    and an implementation of IntInf in the Basis.
1526    
1527    There are a fairly large number of related changes and updates throughout
1528    the system:
1529    
1530    Basis:
1531            - Implemented IntInf.
1532            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1533            - Added some missing Real64 operations, most notably Real.toLargeInt.
1534            - Added FixedInt as a synonym for Int32.
1535    
1536    compiler:
1537            * Added support for a built-in intinf type.
1538              - literals
1539              - pattern matching
1540              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1541              - overloading on literals and operations
1542    
1543            This required adding a primitive type intinf, some additional
1544            primops, and implementations for several non-trivial intinf
1545            operations in Core.  (The intinf type is completely abstract
1546            to the compiler; all operations get delegated back to the Core.)
1547    
1548            * Intinf equality is handled by polyequal.  However, the compiler
1549            does not print its usual warning in this case (since polyequal
1550            is the right thing to do there).
1551    
1552            * Improved the organization of structure InlineT.
1553    
1554            * A word about conversion primops:
1555            If conversions involving intinf do not cancel out during
1556            CPS contract, then the compiler must insert calls to Core functions.
1557            Since all core access must be resolved already during the FLINT
1558            translate phase, it would be too late a the time of CPS contract
1559            to add new Core calls.  For this reason, conversion primops
1560            for intinf carry two arguments:  1. the numeric argument that
1561            they are supposed to convert, and 2. the Core function that
1562            can help with this conversion if necessary.  If CPS contract
1563            eliminates a primop, then the associated Core function becomes
1564            dead and goes away.  Intinf conversion primops that do not get
1565            eliminated by CPS contract get rewritten into calls of their
1566            core functions by a separate, new phase.
1567    
1568    interactive system:
1569            - Control.Print.intinfDepth controls max length of intinf constants
1570            being printed. (Analogous to Control.Print.stringDepth.)
1571            - Cleanup in printutil and pputil: got rid of unused stuff and
1572            duplicates; replaced some of the code with code that makes better
1573            use of library functionality.
1574    
1575    CM:
1576            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1577            are no longer silent.
1578    
1579    CKIT:
1580            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1581            in favor of LargeInt -- which is now the same as IntInf.
1582            CKIT-knowledgable people should check whether this is what's
1583            intended and otherwise change things back to using Int32 or
1584            FixedInt.
1585    
1586    Throughout the code:
1587            Started using IntInf.int literals and built-in operations
1588            (e.g., comparison with 0) where this seems appropriate.
1589    
1590    
1591    ----------------------------------------------------------------------
1592    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1593    Date: 2003/08/13 11:36:00 CDT
1594    Tag: dbm-20030813-mcz-merge1
1595    Description:
1596    
1597    Merging changes from the mcz-branch development branch into trunk.
1598    These changes involve replacement of the emulated old prettyprinter
1599    interface with direct use of the SML/NJ Lib PP library, and fixing
1600    of a couple of bugs (895, 1186) relating to error messages.  A new
1601    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1602    has been added.
1603    
1604    ----------------------------------------------------------------------
1605    Name: Matthias Blume (blume (at) tti - c (dot) org)
1606    Date: 2003/08/11 15:45:00 CDT
1607    Tag: blume-20030811-windows
1608    Description:
1609    
1610    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1611    
1612        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1613    
1614    This patch restores SML/NJ's ability to run under win32.  There are a
1615    number of changes, including fixes for several bugs that had gone
1616    unnoticed until now:
1617    
1618      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1619        trying to run the win32 version from within cygwin.)
1620    
1621      - There are a number of simple .bat scripts that substitute for their
1622        corresponding Unix shell-scripts.  (See below.)
1623    
1624      - The internals of ml-build have been modified slightly.  The main
1625        difference is that instead of calling ".link-sml" (or link-sml.bat)
1626        using OS.Process.system, the ML process delegates this task back
1627        to the script.  Otherwise problems arise in mixed environments such
1628        as Cygwin where scripts look and work like Unix scripts, but
1629        where OS.Process.system cannot run them.
1630    
1631      - In CM, the srcpath pickler used native pathname syntax -- which
1632        is incorrect in the case of cross-compilation.  The new pickle format
1633        is independent of platform-specific naming conventions.
1634    
1635      - Path configuration files (such as lib/pathconfig) can now choose
1636        between native and standard syntax.  Placing a line of the form
1637    
1638            standard!
1639    
1640        into the file causes all subsequent paths to be interpreted using
1641        CM standard pathname syntax (= Unix conventions); a line
1642    
1643            native!
1644    
1645        switches back to native style.  This was needed so that
1646        path config files can be written portably, see src/system/pathconfig.
1647    
1648      - Runtime system:
1649    
1650          - win32-filesys.c:  get_file_time and set_file_time now
1651            access modification time, not creation time.
1652    
1653          - I/O code made aware of new array representation.
1654    
1655          - Bug fixes in X86.prim.masm.
1656    
1657       - src/system/makeml made aware of win32. (For use under cygwin
1658            and other Unix-environments for windows.)
1659    
1660       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1661         which caused BinIO.inputAll to fail consistently.
1662    
1663    .bat scripts:
1664    
1665       Windows .bat scripts assume that SMLNJ_HOME is defined.
1666    
1667       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1668         applications (sml, ml-yacc, ml-lex).
1669       - ml-build.bat: analogous to ml-build.
1670       - config\install.bat:  Analogous to config/install.sh.  This requires
1671         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1672         (nmake etc. must be on the path, and vcvars32 must have been run.)
1673         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1674         and the bootfile hierarchy must have been unpacked under
1675         sml.boot.x86-win32.
1676         The script is very primitive and does a poor job at error checking.
1677         It only installs the base system, ml-lex, and ml-yacc.  No other
1678         libraries are being installed (i.e., you get only those that
1679         are part of the compiler.)
1680       - link-sml.bat:  analogous to .link-sml, but not currently used
1681    
1682    Unrelated bug fixes:
1683    
1684      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1685        types.
1686      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1687        otherwise.)
1688    
1689    ----------------------------------------------------------------------
1690    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1691    Date: 2003/06/17
1692    Tag: macqueen-20030617-bug895
1693    Description:
1694    
1695    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1696    Tag will be used for new development branch (mcz-branch) for
1697    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1698    
1699    ----------------------------------------------------------------------
1700    Name: Matthias Blume (blume (at) tti - c (dot) org)
1701    Date: 2003/05/27 16:55:00 CDT
1702    Tag: blume-20030527-polyeq
1703    Description:
1704    
1705    Tried to eliminated most cases of polymorphic equality.
1706    
1707    ----------------------------------------------------------------------
1708    Name: Matthias Blume (blume (at) tti - c (dot) org)
1709    Date: 2003/05/21 17:45:00 CDT
1710    Tag: blume-20030517-complete
1711    Description:
1712    
1713    Two changes:
1714    
1715      1. Added a flag for controlling whether non-exhaustive bindings will
1716         be treated as errors (default is false).
1717      2. Cleaned up the *entire* source tree so that CMB.make goes through
1718         without a single non-exhaustive match- or bind warning.
1719    
1720    ----------------------------------------------------------------------
1721    Name: Matthias Blume (blume (at) tti - c (dot) org)
1722    Date: 2003/05/17 10:20:00 CDT
1723    Tag: blume-20030517-absyn
1724    Description:
1725    
1726    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1727    
1728       This mainly affects the quality of error messages.  However, some
1729       of the code is now more straightforward than before. (Treatment of
1730       the above four constructs in translate.sml is much simpler than
1731       the "macro-expansion" that was going on before.  Plus, the mach-
1732       compiler no longer gets invoked just to be able to compile an
1733       if-expression.)
1734    
1735    2. The ErrorMsg.Error exception is now caught and absorbed by the
1736       interactive loop.
1737    
1738    ----------------------------------------------------------------------
1739    Name: Allen Leung
1740    Date: 2003/05/16 13:05:00 CDT
1741    Tag: leunga-20030516-cygwin-runtime
1742    Description:
1743    
1744        Ported the runtime system to cygwin, which uses the unix
1745    x86-unix bin files.  Missing/buggy features:
1746    
1747       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1748         the Cygwin library.
1749       o Ctrl-C handling may be flaky.
1750       o Windows system calls and Windows I/O are not supported.
1751    
1752       A new set of binfiles is located at:
1753    
1754         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1755    
1756       This is only needed for bootstrapping the cygwin version of smlnj.
1757       Other x86 versions can use the existing binfiles.
1758    
1759    ----------------------------------------------------------------------
1760    Name: Matthias Blume
1761    Date: 2003/04/08 15:42:00 CDT
1762    Tag: blume-20030408-listpair
1763    Description:
1764    
1765    1. Added a target 'mlrisc' to installer.
1766    
1767    2. Added missing elements to structure ListPair.
1768    
1769    ----------------------------------------------------------------------
1770    Name: Allen Leung
1771    Date: 2003/01/07 10:40:00 EST
1772    Tag: leunga-20030107-int-rem
1773    Description:
1774    
1775    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1776    The arguments to the SUBL instruction were swapped.
1777    
1778    ----------------------------------------------------------------------
1779    Name: Matthias Blume
1780    Date: 2002/12/12 16:25:00 EST
1781    Tag: blume-20021212-risc-ra
1782    Description:
1783    
1784    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1785    sent the RA into an infinite loop when floating point registers get
1786    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1787    110.37 and 110.38.)
1788    There was another set of potential problems related to the handling of
1789    MLRISC annotations (but those did not yet cause real problems, apparently).
1790    
1791    ----------------------------------------------------------------------
1792    Name: Matthias Blume
1793    Date: 2002/12/06 22:40:00 EST
1794    Tag: blume-20021206-cm-fileid
1795    Description:
1796    
1797    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1798    This fixes the problem of CM getting confused by files that suddenly
1799    change their identity (e.g., by getting unlinked and recreated by some
1800    text editor such as vi).  There might be a better/cheaper/cleaner way
1801    of doing this, but for now this will have to do.
1802    
1803    ----------------------------------------------------------------------
1804    Name: Matthias Blume
1805    Date: 2002/10/28 09:50:00 EST
1806    Tag: blume-20021028-typecheck
1807    Description:
1808    
1809    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1810    
1811    ----------------------------------------------------------------------
1812    Name: Matthias Blume
1813    Date: 2002/10/17 09:10:00 EDT
1814    Tag: Release_110_42
1815    Description:
1816    
1817    In good old tradition, there has been a slight hiccup so that we have
1818    to patch 110.42 after the fact.   The old release tag has been replaced
1819    (see below).
1820    
1821    The change solves a problem with two competing approaches the
1822    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1823    each other's way.
1824    
1825    This change only affects the runtime system code and the installer script.
1826    (No new bootfiles.)
1827    
1828    ----------------------------------------------------------------------
1829    Name: Matthias Blume
1830    Date: 2002/10/16 12:00:00 EDT
1831    Tag: Release_110_42_removed
1832    Description:
1833    
1834    New working release.  New bootfiles.
1835    
1836    ----------------------------------------------------------------------
1837    Name: Matthias Blume
1838    Date: 2002/10/10 13:10:00 EDT
1839    Tag: blume-20021010-ppc-divs
1840    Description:
1841    
1842    The mltree operator DIVS must be implemented with an overflow check on
1843    the PPC because the hardware indicates divide-by-zero using "overflow" as
1844    well.
1845    
1846    ----------------------------------------------------------------------
1847    Name: Matthias Blume
1848    Date: 2002/07/23 11:45:00
1849    Tag: blume-20020723-smlnj-home
1850    Description:
1851    
1852    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1853    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1854    CM_PATHCONFIG is also set), the path configuration file is assumed
1855    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1856    the entire tree to some other place and everything will "just work".
1857    
1858    (Companion commands such as ml-build and ml-makedepend also sense this
1859    variable.)
1860    
1861    ----------------------------------------------------------------------
1862    Name: Matthias Blume
1863    Date: 2002/07/12 21:19:00 EDT
1864    Tag: blume-20020712-liveness
1865    Description:
1866    
1867    Exported two useful "step" functions from liveness module (MLRISC).
1868    
1869    ----------------------------------------------------------------------
1870    Name: Matthias Blume
1871    Date: 2002/07/05 16:00 EDT
1872    Tag: Release_110_41
1873    Description:
1874    
1875    New working release.  New bootfiles.
1876    
1877    ----------------------------------------------------------------------
1878    Name: Matthias Blume
1879    Date: 2002/07/05 10:25:00 EDT
1880    Tag: blume-20020705-btimp
1881    Description:
1882    
1883    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1884    other clients can set up backtracing support.
1885    
1886    ----------------------------------------------------------------------
1887    Name: Matthias Blume
1888    Date: 2002/06/25 17:23:00 EDT
1889    Tag: blume-20020625-fpmax
1890    Description:
1891    
1892    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1893    case.  (The sense of the isNaN test was reversed -- which made min and
1894    max always return their first argument.)
1895    
1896    ----------------------------------------------------------------------
1897    Name: Matthias Blume
1898    Date: 2002/06/11
1899    Tag: blume-20020611-unixpath
1900    Description:
1901    
1902    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1903    
1904    ----------------------------------------------------------------------
1905    Name: Matthias Blume
1906    Date: 2002/06/10 16:35:00 EDT
1907    Tag: blume-20020610-ieeereal
1908    Description:
1909    
1910    I back-ported my implementation of IEEEReal.fromString from the
1911    idlbasis-devel branch so that we can test it.
1912    
1913    Another small change is that ppDec tries to give more information
1914    than just "<sig>" in the case of functors.  However, this code is
1915    broken in some mysterious way if the functor's body's signature
1916    has not been declared by ascription but gets inferred from the
1917    implementation.  This needs fixing...
1918    
1919    ----------------------------------------------------------------------
1920    Name: Matthias Blume
1921    Date: 2002/05/31
1922    Tag: blume-20020531-btrace-mode
1923    Description:
1924    
1925    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1926    the wayside when I switched over to using Controls everywhere.)
1927    
1928    ----------------------------------------------------------------------
1929    Name: Lal George
1930    Date: 2002/05/23 12:21:40 EDT
1931    Tag: george-20020523-visual-labels
1932    Description:
1933    
1934    Labels are now displayed in the graphical output to make
1935    the fall-through and target blocks obvious.
1936    
1937    ----------------------------------------------------------------------
1938    Name: Matthias Blume
1939    Date: 2002/05/22 11:03:00 EDT
1940    Tag: blume-20020522-shrink
1941    Description:
1942    
1943    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1944    situation that comes up when allocating huge arrays.
1945    
1946    ----------------------------------------------------------------------
1947    Name: Matthias Blume
1948    Date: 2002/05/21 16:00:00 EDT
1949    Tag: Release_110_40
1950    Description:
1951    
1952    New working release (110.40).  New bootfiles.
1953    
1954    [Also: John Reppy fixed GC bug 1131.]
1955    
1956    ----------------------------------------------------------------------
1957    Name: Matthias Blume
1958    Date: 2002/05/21 12:35:00 EDT
1959    Tag: blume-20020521-cmdoc
1960    Description:
1961    
1962    CM documentation update.
1963    
1964    ----------------------------------------------------------------------
1965    Name: Matthias Blume
1966    Date: 2002/05/21 10:55:00 EDT
1967    Tag: blume-20020521-misc
1968    Description:
1969    
1970    - John tweaked runtime to be silent on heap export (except when
1971      GC messages are on).
1972    - I added a few more things (cross-compiling versions of CMB) to
1973      config/preloads (as suggestions).
1974    
1975    ----------------------------------------------------------------------
1976    Name: Matthias Blume
1977    Date: 2002/05/20 22:25:00 EDT
1978    Tag: blume-20020520-controls
1979    Description:
1980    
1981    - Added ControlUtil structure to control-lib.cm.
1982    - Use it throughout.
1983    - Used Controls facility to define MLRISC controls (as opposed to
1984      registering MLRISC control ref cells with Controls after the fact)
1985    - Fixed messed-up controls priorities.
1986    
1987    * Removed again all the stuff from config/preloads that one wouldn't
1988      be able to preload at the time the initial heap image is built.
1989      (Many libraries, e.g., CML, do not exist yet at this time.  The
1990       only libraries that can be preloaded via config/preloads are those
1991       that come bundled with the bootfiles.)
1992    
1993    ----------------------------------------------------------------------
1994    Name: Matthias Blume
1995    Date: 2002/05/20 10:59:00 EDT
1996    Tag: blume-20020520-preloads
1997    Description:
1998    
1999    Added a lot of commented-out suggestions for things to be included
2000    in config/preloads.
2001    
2002    ----------------------------------------------------------------------
2003    Name: Allen Leung
2004    Date: 2002/05/18 14:20:00 EDT
2005    Tag: leunga-20020518-mdl
2006    Description:
2007    
2008    o Made the mdl tool stuff compile and run again.
2009    o I've disabled all the stuff that depends on RTL specifications; they
2010      are all badly broken anyway.
2011    
2012    ----------------------------------------------------------------------
2013    Name: Matthias Blume
2014    Date: 2002/05/17 16:49:00 EDT
2015    Tag: blume-20020517-controls
2016    Description:
2017    
2018    0. John Reppy made several modifications to the SML/NJ library.
2019       In particular, there is a shiny new controls-lib.cm.
2020    
2021    1. Pushed new controls interface through compiler so that everything
2022       compiles again.
2023    
2024    2. Added FormatComb and FORMAT_COMB to the CML version of the
2025       SML/NJ library (so that CML compiles again).
2026    
2027    3. Modified init scripts because XXX_DEFAULT environment variables
2028       are no longer with us.  (Boot-time initialization is now done
2029       using the same environment variables that are also used for
2030       startup-time initialization of controls.)
2031    
2032    ----------------------------------------------------------------------
2033    Name: Lal George
2034    Date: 2002/05/15 09:20:10 EDT
2035    Tag: george-20020515-pseudo-op-decls
2036    Description:
2037    
2038            All pseudo-ops emitted before the first segment declaration
2039    such as TEXT, DATA, and BSS directives are assumed to be global
2040    declarations and are emitted first in the assembly file. This is
2041    useful in a number of situations where one has pseudo-ops that are not
2042    specific to any segment, and also works around the constraint that one
2043    cannot have client pseudo-ops in the TEXT segment.
2044    
2045            Because no segment is associated with these declarations it is
2046    an error to allocate any space or objects before the first segment
2047    directive and an exception will be raised. However, we cannot make
2048    this check for client pseudo-ops.
2049    
2050            These top level declarations are a field in the CFG graph_info.
2051    In theory you can continue to add to this field after the CFG has been
2052    built -- provided you know what you are doing;-)
2053    
2054    ----------------------------------------------------------------------
2055    Name: Matthias Blume
2056    Date: 2002/05/13 16:40:00 EDT
2057    Tag: blume-20020513-pp-etc
2058    Description:
2059    
2060    A few minor bugfixes:
2061    
2062      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2063        (Bogus printouts for redefined bindings still occur.  Compiler
2064        bug should no longer occur now.  We need to redo the prettyprinter
2065        from scratch.)
2066    
2067      - CM pathname printer now also adds escape sequences for ( and )
2068    
2069      - commend and docu fixes for ml-nlffi
2070    
2071    ----------------------------------------------------------------------
2072    Name: Matthias Blume
2073    Date: 2002/05/10 16:40:00 EDT
2074    Tag: blume-20020510-erg-textio
2075    Description:
2076    
2077    Applied the following bugfix provided by Emden Gansner:
2078    
2079        Output is corrupted when outputSubstr is used rather than output.
2080    
2081        The problem occurs when a substring
2082    
2083            ss = (s, dataStart, dataLen)
2084    
2085        where dataStart > 0, fills a stream buffer with avail bytes left.
2086        avail bytes of s, starting at index dataStart, are copied into the
2087        buffer, the buffer is flushed, and then the remaining dataLen-avail
2088        bytes of ss are copied into the beginning of the buffer. Instead of
2089        starting this copy at index dataStart+avail in s, the current code
2090        starts the copy at index avail.
2091    
2092        Fix:
2093        In text-io-fn.sml, change line 695 from
2094             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2095        to
2096             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2097    
2098    ----------------------------------------------------------------------
2099    Name: Matthias Blume
2100    Date: 2002/04/12 13:55:00 EDT
2101    Tag: blume-20020412-assyntax
2102    Description:
2103    
2104    1. Grabbed newer assyntax.h from the XFree86 project.
2105    2. Fiddled with how to compile X86.prim.asm without warnings.
2106    3. (Very) Minor cleanup in CM.
2107    
2108    ----------------------------------------------------------------------
2109    Name: Matthias Blume
2110    Date: 2002/04/01 (no joke!) 17:07:00 EST
2111    Tag: blume-20020401-x86div
2112    Description:
2113    
2114    Added full support for div/mod/rem/quot on the x86, using the machine
2115    instruction's two results (without clumsily recomputing the remainder)
2116    directly where appropriate.
2117    
2118    Some more extensive power-of-two support was added to the x86 instruction
2119    selector (avoiding expensive divs, mods, and muls where they can be
2120    replaced with cheaper shifts and masks).  However, this sort of thing
2121    ought to be done earlier, e.g., within the CPS optimizer so that
2122    all architectures benefit from it.
2123    
2124    The compiler compiles to a fixed point, but changes might be somewhat
2125    fragile nevertheless.  Please, report any strange things that you might
2126    see wrt. div/mod/quot/rem...
2127    
2128    ----------------------------------------------------------------------
2129    Name: Matthias Blume
2130    Date: 2002/03/29 17:22:00
2131    Tag: blume-20020329-div
2132    Description:
2133    
2134    Fixed my broken div/mod logic.  Unfortunately, this means that the
2135    inline code for div/mod now has one more comparison than before.
2136    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2137    The problem was with quotient = 0, because that alone does not tell
2138    us which way the rounding went.  One then has to look at whether
2139    remainder and divisor have the same sign...  :(
2140    
2141    Anyway, I replaced the bootfiles with fresh ones...
2142    
2143    ----------------------------------------------------------------------
2144    Name: Matthias Blume
2145    Date: 2002/03/29 14:10:00 EST
2146    Tag: blume-20020329-inlprims
2147    Description:
2148    
2149    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2150    
2151    Primops have changed. This means that the bin/boot-file formats have
2152    changed as well.
2153    
2154    To make sure that there is no confusion, I made a new version.
2155    
2156    
2157    CHANGES:
2158    
2159    * removed REMT from mltree (remainder should never overflow).
2160    
2161    * added primops to deal with divisions of all flavors to the frontend
2162    
2163    * handled these primops all the way through so they map to their respective
2164      MLRISC support
2165    
2166    * used these primops in the implementation of Int, Int32, Word, Word32
2167    
2168    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2169    
2170    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2171    
2172    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2173      floating point
2174    
2175    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2176      and Real (but Real.abs maps to a separate floating-point-only primop)
2177    
2178    
2179    TODO items:
2180    
2181    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2182      instructions because the machine instruction encoder cannot handle
2183      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2184      REM in the first place, and REM is supported by the x86 machine coder.)
2185    
2186    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2187      selection phase.  (The two can be streamlined because the hardware
2188      delivers both quotient and remainder at the same time anyway.)
2189    
2190    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2191      (Currently the behavior is inconsistent both across architectures and
2192      wrt. the draft Basis spec.)
2193    
2194    * Word8 should eventually be handled natively, too.
2195    
2196    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2197      as if there currently is no execution path that could trigger it in
2198      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2199      hold for things like multiplication and division.)
2200    
2201    ----------------------------------------------------------------------
2202    Name: Matthias Blume
2203    Date: 2002/03/27 16:27:00 EST
2204    Tag: blume-20020327-mlrisc-divisions
2205    Description:
2206    
2207    Added support for all four division operations (ML's div, mod, quot,
2208    and rem) to MLRISC.  In the course of doing so, I also rationalized
2209    the naming (no more annoying switch-around of DIV and QUOT), by
2210    parameterizing the operation by div_rounding_mode (which can be either
2211    DIV_TO_ZERO or DIV_TO_NEGINF).
2212    
2213    The generic MLTreeGen functor takes care of compiling all four
2214    operations down to only round-to-zero div.
2215    
2216    Missing pieces:
2217    
2218      * Doing something smarter than relying on MLTreeGen on architectures
2219        like, e.g., the x86 where hardware division delivers both quotient and
2220        remainder at the same time.  With this, the implementation of the
2221        round-to-neginf operations could be further streamlined.
2222    
2223      * Remove inlining support for div/mod/rem from the frontend and replace it
2224        with primops that get carried through to the backend.  Do this for all
2225        int and word types.
2226    
2227    ----------------------------------------------------------------------
2228    Name: Matthias Blume
2229    Date: 2002/03/25 17:25:00 EST
2230    Tag: blume-20020325-divmod
2231    Description:
2232    
2233    I improved (hopefully without breaking them) the implementation of Int.div,
2234    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2235    advantage of the following observations:
2236    
2237      Let  q = x quot y      r = x rem y
2238           d = x div  y      m = x mod y
2239    
2240    where "quot" is the round-to-zero version of integer division that
2241    hardware usually provides.  Then we have:
2242    
2243         r = x - q * y        where neither the * nor the - will overflow
2244         d = if q >= 0 orelse x = q * y then q else q - 1
2245                              where neither the * nor the - will overflow
2246         m = if q >= 0 orelse r = 0 then r else r + y
2247                              where the + will not overflow
2248    
2249    This results in substantial simplification of the generated code.
2250    The following table shows the number of CFG nodes and edges generated
2251    for
2252            fun f (x, y) = x OPER y
2253            (* with OPER \in div, mod, quot, rem *)
2254    
2255    
2256        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2257        --------------------------------------------------------
2258         div |         24 |         39 |         12 |         16
2259         mod |         41 |         71 |         12 |         16
2260        quot |          8 |         10 |          8 |         10
2261         rem |         10 |         14 |          8 |         10
2262    
2263    
2264    ----------------------------------------------------------------------
2265    Name: Matthias Blume
2266    Date: 2002/03/25 22:06:00 EST
2267    Tag: blume-20020325-cprotobug
2268    Description:
2269    
2270    Fixed a bug in cproto (c prototype decoder).
2271    
2272    ----------------------------------------------------------------------
2273    Name: Matthias Blume
2274    Date: 2002/03/25 16:00:00 EST
2275    Tag: blume-20020325-raw-primops
2276    Description:
2277    
2278    I did some cleanup to Allen's new primop code and
2279    replaced yesterday's bootfiles with new ones.
2280    (But they are stored in the same place.)
2281    
2282    ----------------------------------------------------------------------
2283    Name: Matthias Blume
2284    Date: 2002/03/24 22:40:00 EST
2285    Tag: blume-20020324-bootfiles
2286    Description:
2287    
2288    Made the bootfiles that Allen asked for.
2289    
2290    ----------------------------------------------------------------------
2291    Name: Allen Leung
2292    Date: 2002/03/23 15:50:00 EST
2293    Tag: leunga-20020323-flint-cps-rcc-primops
2294    Description:
2295    
2296      1. Changes to FLINT primops:
2297    
2298        (* make a call to a C-function;
2299         * The primop carries C function prototype information and specifies
2300         * which of its (ML-) arguments are floating point. C prototype
2301         * information is for use by the backend, ML information is for
2302         * use by the CPS converter. *)
2303      | RAW_CCALL of { c_proto: CTypes.c_proto,
2304                       ml_args: ccall_type list,
2305                       ml_res_opt: ccall_type option,
2306                       reentrant : bool
2307                     } option
2308       (* Allocate uninitialized storage on the heap.
2309        * The record is meant to hold short-lived C objects, i.e., they
2310        * are not ML pointers.  With the tag, the representation is
2311        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2312        *)
2313      | RAW_RECORD of {tag:bool,sz:int}
2314      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2315    
2316      2.  These CPS primops are now overloaded:
2317    
2318           rawload of {kind:numkind}
2319           rawstore of {kind:numkind}
2320    
2321          The one argument form is:
2322    
2323             rawload {kind} address
2324    
2325          The two argument form is:
2326    
2327             rawload {kind} [ml object, byte-offset]
2328    
2329      3. RAW_CCALL/RCC now takes two extra arguments:
2330    
2331         a. The first is whether the C call is reentrant, i.e., whether
2332            ML state should be saved and restored.
2333         b. The second argument is a string argument specifying the name of
2334            library and the C function.
2335    
2336         These things are currently not handled in the code generator, yet.
2337    
2338      4. In CProto,
2339    
2340         An encoding type of "bool" means "ml object" and is mapped into
2341         C prototype of PTR.  Note that "bool" is different than "string",
2342         even though "string" is also mapped into PTR, because "bool"
2343         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2344    
2345      5. Pickler/unpicker
2346    
2347         Changed to handle RAW_RECORD and newest RAW_CCALL
2348    
2349      6. MLRiscGen,
2350    
2351         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2352         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2353            in the file CodeGen/cpscompile/cps-c-calls.sml
2354    
2355      7. Added the conditional move operator
2356    
2357             condmove of branch
2358    
2359         to cps.  Generation of this is still buggy so it is currently
2360         disabled.
2361    
2362    ----------------------------------------------------------------------
2363    Name: Lal George
2364    Date: 2002/03/22 14:18:25 EST
2365    Tag: george-20020322-cps-branch-prob
2366    Description:
2367    
2368    Implemented the Ball-Larus branch prediction-heuristics, and
2369    incorporated graphical viewers for control flow graphs.
2370    
2371    Ball-Larus Heuristics:
2372    ---------------------
2373    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2374    
2375    By design it uses the Dempster-Shafer theory for combining
2376    probabilities.  For example, in the function:
2377    
2378        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2379    
2380    the ball-larus heuristics predicts that the n=0 is unlikely
2381    (OH-heuristic), and the 'then' branch is unlikely because of the
2382    RH-heuristic -- giving the 'then' branch an even lower combined
2383    probability using the Dempster-Shafer theory.
2384    
2385    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2386    probability of the 'then' branch because of the loop in the else
2387    branch.
2388    
2389    
2390    Graphical Viewing:
2391    ------------------
2392    I merely plugged in Allen's graphical viewers into the compiler. The
2393    additional code is not much. At the top level, saying:
2394    
2395            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2396    
2397    will display the graphical view of the control flow graph just before
2398    back-patching.  daVinci must be in your path for this to work. If
2399    daVinci is not available, then the default viewer can be changed
2400    using:
2401    
2402            Control.MLRISC.getString "viewer"
2403    
2404    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2405    course, these viewers must be in your path.
2406    
2407    The above will display the compilation unit at the level of clusters,
2408    many of which are small, boring, and un-interesting. Also setting:
2409    
2410            Control.MLRISC.getInt "cfg-graphical-view_size"
2411    
2412    will display clusters that are larger than the value set by the above.
2413    
2414    
2415    ----------------------------------------------------------------------
2416    Name: Matthias Blume
2417    Date: 2002/03/21 22:20:00 EST
2418    Tag: blume-20020321-kmp-bugfix
2419    Description:
2420    
2421    Changed the interface to the KMP routine in PreString and fixed
2422    a minor bug in one place where it was used.
2423    
2424    ----------------------------------------------------------------------
2425    Name: Allen Leung
2426    Date: 2002/03/21 20:30:00 EST
2427    Tag: leunga-20020321-cfg
2428    Description:
2429    
2430      Fixed a potential problem in cfg edge splitting.
2431    
2432    ----------------------------------------------------------------------
2433    Name: Allen Leung
2434    Date: 2002/03/21 17:15:00 EST
2435    Tag: leunga-20020321-x86-fp-cfg
2436    Description:
2437    
2438      1. Recoded the buggy parts of x86-fp.
2439    
2440         a. All the block reordering code has been removed.
2441            We now depend on the block placement phases to do this work.
2442    
2443         b. Critical edge splitting code has been simplified and moved into the
2444            CFG modules, as where they belong.
2445    
2446         Both of these were quite buggy and complex.  The code is now much, much
2447         simpler.
2448    
2449      2. X86 backend.
2450    
2451         a. Added instructions for 64-bit support.  Instruction selection for
2452            64-bit has not been committed, however, since that
2453            requires changes to MLTREE which haven't been approved by
2454            Lal and John.
2455    
2456         b. Added support for FUCOMI and FUCOMIP when generating code for
2457            PentiumPro and above.  We only generate these instructions in
2458            the fast-fp mode.
2459    
2460         c. Added cases for JP and JNP in X86FreqProps.
2461    
2462      3. CFG
2463    
2464         CFG now has a bunch of methods for edge splitting and merging.
2465    
2466      4. Machine description.
2467    
2468         John's simplification of MLTREE_BASIS.fcond broke a few machine
2469         description things:
2470    
2471         rtl-build.{sig,sml} and hppa.mdl fixed.
2472    
2473         NOTE: the machine description stuff in the repository is still broken.
2474               Again, I can't put my fixes in because that involves
2475               changes to MLTREE.
2476    
2477    ----------------------------------------------------------------------
2478    Name: Matthias Blume
2479    Date: 2002/03/20 15:55:00 EST
2480    Tag: blume-20020320-kmp
2481    Description:
2482    
2483    Implemented Knuth-Morris-Pratt string matching in PreString and used
2484    it for String.isSubstring, Substring.isSubstring, and
2485    Substring.position.
2486    
2487    (Might need some stress-testing.  Simple examples worked fine.)
2488    
2489    ----------------------------------------------------------------------
2490    Name: Matthias Blume
2491    Date: 2002/03/19 16:37:00 EST
2492    Tag: blume-20020319-witnesses
2493    Description:
2494    
2495    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2496    
2497    This implements a generic mechanism for changing constness qualifiers
2498    anywhere within big C types without resorting to outright "casts".
2499    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2500    modify the constness at the outermost level.)
2501    The implementation of "convert" is based on the idea of "witness"
2502    values -- values that are not used by the operation but whose types
2503    "testify" to their applicability.  On the implementation side, "convert"
2504    is simply a projection (returning its second curried argument).  With
2505    cross-module inlining, it should not result in any machine code being
2506    generated.
2507    
2508    ----------------------------------------------------------------------
2509    Name: Matthias Blume
2510    Date: 2002/03/15 16:40:00 EST
2511    Tag: blume-20020315-basis
2512    Description:
2513    
2514    Provided (preliminary?) implementations for
2515    
2516      {String,Substring}.{concatWith,isSuffix,isSubstring}
2517    
2518    and
2519    
2520      Substring.full
2521    
2522    Those are in the Basis spec but they were missing in SML/NJ.
2523    
2524    ----------------------------------------------------------------------
2525    Name: Matthias Blume
2526    Date: 2002/03/14 21:30:00 EST
2527    Tag: blume-20020314-controls
2528    Description:
2529    
2530    Controls:
2531    ---------
2532    
2533    1. Factored out the recently-added Controls : CONTROLS stuff and put
2534       it into its own library $/controls-lib.cm.  The source tree for
2535       this is under src/smlnj-lib/Controls.
2536    
2537    2. Changed the names of types and functions in this interface, so they
2538       make a bit more "sense":
2539    
2540          module -> registry
2541          'a registry -> 'a group
2542    
2543    3. The interface now deals in ref cells only.  The getter/setter interface
2544       is (mostly) gone.
2545    
2546    4. Added a function that lets one register an already-existing ref cell.
2547    
2548    5. Made the corresponding modifications to the rest of the code so that
2549       everything compiles again.
2550    
2551    6. Changed the implementation of Controls.MLRISC back to something closer
2552       to the original.  In particular, this module (and therefore MLRISC)
2553       does not depend on Controls.  There now is some link-time code in
2554       int-sys.sml that registers the MLRISC controls with the Controls
2555       module.
2556    
2557    CM:
2558    ---
2559    
2560      * One can now specify the lambda-split aggressiveness in init.cmi.
2561    
2562    ----------------------------------------------------------------------
2563    Name: Allen Leung
2564    Date: 2002/03/13 17:30:00 EST
2565    Tag: leunga-20020313-x86-fp-unary
2566    Description:
2567    
2568    Bug fix for:
2569    
2570    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2571    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2572    > - fun f(x,(y,z)) = Real.~ y;
2573    > [autoloading]
2574    > [autoloading done]
2575    >       fchsl   (%eax), 184(%esp)
2576    > Error: MLRisc bug: X86MCEmitter.emitInstr
2577    >
2578    > uncaught exception Error
2579    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2580    
2581    The problem was that the code generator did not generate any fp registers
2582    in this case, and the ra didn't know that it needed to run the X86FP phase to
2583    translate the pseudo fp instruction.   This only happened with unary fp
2584    operators in certain situations.
2585    
2586    ----------------------------------------------------------------------
2587    Name: Matthias Blume
2588    Date: 2002/03/13 14:00:00 EST
2589    Tag: blume-20020313-overload-etc
2590    Description:
2591    
2592    1. Added _overload as a synonym for overload for backward compatibility.
2593       (Control.overloadKW must be true for either version to be accepted.)
2594    
2595    2. Fixed bug in install script that caused more things to be installed
2596       than what was requested in config/targets.
2597    
2598    3. Made CM aware of the (_)overload construct so that autoloading
2599       works.
2600    
2601    ----------------------------------------------------------------------
2602    Name: Matthias Blume
2603    Date: 2002/03/12 22:03:00 EST
2604    Tag: blume-20020312-url
2605    Description:
2606    
2607    Forgot to update BOOT and srcarchiveurl.
2608    
2609    ----------------------------------------------------------------------
2610    Name: Matthias Blume
2611    Date: 2002/03/12 17:30:00 EST
2612    Tag: blume-20020312-version110392
2613    Description:
2614    
2615    Yet another version number bump (because of small changes to the
2616    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2617    
2618    Changes:
2619    
2620      The new pid generation scheme described a few weeks ago was overly
2621      complicated.  I implemented a new mechanism that is simpler and
2622      provides a bit more "stability":  Once CM has seen a compilation
2623      unit, it keeps its identity constant (as long as you do not delete
2624      those crucial CM/GUID/* files).  This means that when you change
2625      an interface, compile, then go back to the old interface, and
2626      compile again, you arrive at the original pid.
2627    
2628      There now also is a mechanism that instructs CM to use the plain
2629      environment hash as a module's pid (effectively making its GUID
2630      the empty string).  For this, "noguid" must be specified as an
2631      option to the .sml file in question within its .cm file.
2632      This is most useful for code that is being generated by tools such
2633      as ml-nlffigen (because during development programmers tend to
2634      erase the tool's entire output directory tree including CM's cached
2635      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2636      revert to the old, broken behavior of SML/NJ, but in specific cases
2637      where there is no danger of interface confusion, its use is ok
2638      (I think).
2639    
2640      ml-nlffigen by default generates "noguid" annotations.  They can be
2641      turned off by specifying -guid in its command line.
2642    
2643    ----------------------------------------------------------------------
2644    Name: Lal George
2645    Date: 2002/03/12 12 14:42:36 EST
2646    Tag: george-20020312-frequency-computation
2647    Description:
2648    
2649    Integrated jump chaining and static block frequency into the
2650    compiler. More details and numbers later.
2651    
2652    ----------------------------------------------------------------------
2653    Name: Lal George
2654    Date: 2002/03/11 11 22:38:53 EST
2655    Tag: george-20020311-jump-chain-elim
2656    Description:
2657    
2658    Tested the jump chain elimination on all architectures (except the
2659    hppa).  This is on by default right now and is profitable for the
2660    alpha and x86, however, it may not be profitable for the sparc and ppc
2661    when compiling the compiler.
2662    
2663    The gc test will typically jump to a label at the end of the cluster,
2664    where there is another jump to an external cluster containing the actual
2665    code to invoke gc. This is to allow factoring of common gc invocation
2666    sequences. That is to say, we generate:
2667    
2668            f:
2669               testgc
2670               ja   L1      % jump if above to L1
2671    
2672            L1:
2673               jmp L2
2674    
2675    
2676    After jump chain elimination the 'ja L1' instructions is converted to
2677    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2678    up being implemented in their long form (if L2 is far away) using:
2679    
2680            jbe     L3      % jump if below or equal to L3
2681            jmp     L2
2682         L3:
2683            ...
2684    
2685    
2686    For large compilation units L2  may be far away.
2687    
2688    
2689    ----------------------------------------------------------------------
2690    Name: Matthias Blume
2691    Date: 2002/03/11 13:30:00 EST
2692    Tag: blume-20020311-mltreeeval
2693    Description:
2694    
2695    A functor parameter was missing.
2696    
2697    ----------------------------------------------------------------------
2698    Name: Allen Leung
2699    Date: 2002/03/11 10:30:00 EST
2700    Tag: leunga-20020311-runtime-string0
2701    Description:
2702    
2703       The representation of the empty string now points to a
2704    legal null terminated C string instead of unit.  It is now possible
2705    to convert an ML string into C string with InlineT.CharVector.getData.
2706    This compiles into one single machine instruction.
2707    
2708    ----------------------------------------------------------------------
2709    Name: Allen Leung
2710    Date: 2002/03/10 23:55:00 EST
2711    Tag: leunga-20020310-x86-call
2712    Description:
2713    
2714       Added machine generation for CALL instruction (relative displacement mode)
2715    
2716    ----------------------------------------------------------------------
2717    Name: Matthias Blume
2718    Date: 2002/03/08 16:05:00
2719    Tag: blume-20020308-entrypoints
2720    Description:
2721    
2722    Version number bumped to 110.39.1.  NEW BOOTFILES!
2723    
2724    Entrypoints: non-zero offset into a code object where execution should begin.
2725    
2726    - Added the notion of an entrypoint to CodeObj.
2727    - Added reading/writing of entrypoint info to Binfile.
2728    - Made runtime system bootloader aware of entrypoints.
2729    - Use the address of the label of the first function given to mlriscGen
2730      as the entrypoint.  This address is currently always 0, but it will
2731      not be 0 once we turn on block placement.
2732    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2733      with entry points) from mlriscGen.
2734    
2735    ----------------------------------------------------------------------
2736    Name: Allen Leung
2737    Date: 2002/03/07 20:45:00 EST
2738    Tag: leunga-20020307-x86-cmov
2739    Description:
2740    
2741       Bug fixes for CMOVcc on x86.
2742    
2743       1. Added machine code generation for CMOVcc
2744       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2745       3. CMOVcc cannot have an immediate operand as argument.
2746    
2747    ----------------------------------------------------------------------
2748    Name: Matthias Blume
2749    Date: 2002/03/07 16:15:00 EST
2750    Tag: blume-20020307-controls
2751    Description:
2752    
2753    This is a very large but mostly boring patch which makes (almost)
2754    every tuneable compiler knob (i.e., pretty much everything under
2755    Control.* plus a few other things) configurable via both the command
2756    line and environment variables in the style CM did its configuration
2757    until now.
2758    
2759    Try starting sml with '-h' (or, if you are brave, '-H')
2760    
2761    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2762    implements the underlying generic mechanism.
2763    
2764    The interface to some of the existing such facilities has changed somewhat.
2765    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2766    (The getFoo interface is still there for backward-compatibility, but its
2767    use is deprecated.)
2768    
2769    The ml-build script passes -Cxxx=yyy command-line arguments through so
2770    that one can now twiddle the compiler settings when using this "batch"
2771    compiler.
2772    
2773    TODO items:
2774    
2775    We should go through and throw out all controls that are no longer
2776    connected to anything.  Moreover, we should go through and provide
2777    meaningful (and correct!) documentation strings for those controls
2778    that still are connected.
2779    
2780    Currently, multiple calls to Controls.new are accepted (only the first
2781    has any effect).  Eventually we should make sure that every control
2782    is being made (via Controls.new) exactly once.  Future access can then
2783    be done using Controls.acc.
2784    
2785    Finally, it would probably be a good idea to use the getter-setter
2786    interface to controls rather than ref cells.  For the time being, both
2787    styles are provided by the Controls module, but getter-setter pairs are
2788    better if thread-safety is of any concern because they can be wrapped.
2789    
2790    *****************************************
2791    
2792    One bug fix: The function blockPlacement in three of the MLRISC
2793    backpatch files used to be hard-wired to one of two possibilities at
2794    link time (according to the value of the placementFlag).  But (I
2795    think) it should rather sense the flag every time.
2796    
2797    *****************************************
2798    
2799    Other assorted changes (by other people who did not supply a HISTORY entry):
2800    
2801    1. the cross-module inliner now works much better (Monnier)
2802    2. representation of weights, frequencies, and probabilities in MLRISC
2803       changed in preparation of using those for weighted block placement
2804       (Reppy, George)
2805    
2806    ----------------------------------------------------------------------
2807    Name: Lal George
2808    Date: 2002/03/07 14:44:24 EST 2002
2809    Tag: george-20020307-weighted-block-placement
2810    
2811    Tested the weighted block placement optimization on all architectures
2812    (except the hppa) using AMPL to generate the block and edge frequencies.
2813    Changes were required in the machine properties to correctly
2814    categorize trap instructions. There is an MLRISC flag
2815    "weighted-block-placement" that can be used to enable weighted block
2816    placement, but this will be ineffective without block/edge
2817    frequencies (coming soon).
2818    
2819    
2820    ----------------------------------------------------------------------
2821    Name: Lal George
2822    Date: 2002/03/05 17:24:48 EST
2823    Tag: george-20020305-linkage-cluster
2824    
2825    In order to support the block placement optimization, a new cluster
2826    is generated as the very first cluster (called the linkage cluster).
2827    It contains a single jump to the 'real' entry point for the compilation
2828    unit. Block placement has no effect on the linkage cluster itself, but
2829    all the other clusters  have full freedom in the manner in which they
2830    reorder blocks or functions.
2831    
2832    On the x86 the typical linkage code that is generated is:
2833       ----------------------
2834            .align 2
2835       L0:
2836            addl    $L1-L0, 72(%esp)
2837            jmp     L1
2838    
2839    
2840            .align  2
2841       L1:
2842       ----------------------
2843    
2844    72(%esp) is the memory location for the stdlink register. This
2845    must contain the address of the CPS function being called. In the
2846    above example, it contains the address of  L0; before
2847    calling L1 (the real entry point for the compilation unit), it
2848    must contain the address for L1, and hence
2849    
2850            addl $L1-L0, 72(%esp)
2851    
2852    I have tested this on all architectures except the hppa.The increase
2853    in code size is of course negligible
2854    
2855    ----------------------------------------------------------------------
2856    Name: Allen Leung
2857    Date: 2002/03/03 13:20:00 EST
2858    Tag: leunga-20020303-mlrisc-tools
2859    
2860      Added #[ ... ] expressions to mlrisc tools
2861    
2862    ----------------------------------------------------------------------
2863    Name: Matthias Blume
2864    Date: 2002/02/27 12:29:00 EST
2865    Tag: blume-20020227-cdebug
2866    Description:
2867    
2868    - made types in structure C and C_Debug to be equal
2869    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2870    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2871    
2872    ----------------------------------------------------------------------
2873    Name: Matthias Blume
2874    Date: 2002/02/26 12:00:00 EST
2875    Tag: blume-20020226-ffi
2876    Description:
2877    
2878    1. Fixed a minor bug in CM's "noweb" tool:
2879       If numbering is turned off, then truly don't number (i.e., do not
2880       supply the -L option to noweb).  The previous behavior was to supply
2881       -L'' -- which caused noweb to use the "default" line numbering scheme.
2882       Thanks to Chris Richards for pointing this out (and supplying the fix).
2883    
2884    2. Once again, I reworked some aspects of the FFI:
2885    
2886       A. The incomplete/complete type business:
2887    
2888       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2889         gone!
2890       - ML types representing an incomplete type are now *equal* to
2891         ML types representing their corresponding complete types (just like
2892         in C).  This is still safe because ml-nlffigen will not generate
2893         RTTI for incomplete types, nor will it generate functions that
2894         require access to such RTTI.   But when ML code generated from both
2895         incomplete and complete versions of the C type meet, the ML types
2896         are trivially interoperable.
2897    
2898         NOTE:  These changes restore the full generality of the translation
2899         (which was previously lost when I eliminated functorization)!
2900    
2901       B. Enum types:
2902    
2903       - Structure C now has a type constructor "enum" that is similar to
2904         how the "su" constructor works.  However, "enum" is not a phantom
2905         type because each "T enum" has values (and is isomorphic to
2906         MLRep.Signed.int).
2907       - There are generic access operations for enum objects (using
2908         MLRep.Signed.int).
2909       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2910         * The structure contains the definition of type "mlrep" (the ML-side
2911         representation type of the enum).  Normally, mlrep is the same
2912         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2913         then mlrep will be defined as a datatype -- thus facilitating
2914         pattern matching on mlrep values.
2915         ("-ec" will be suppressed if there are duplicate values in an
2916          enumeration.)
2917         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2918         will be generated for each C enum constant xxx.
2919         * Conversion functions m2i and i2m convert between mlrep and
2920         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2921         * Coversion functions c and ml convert between mlrep and "tag enum".
2922         * Access functions (get/set) fetch and store mlrep values.
2923       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2924         enumerations are merged into one single enumeration represented by
2925         structure E_'.
2926    
2927    ----------------------------------------------------------------------
2928    Name: Allen Leung
2929    Date: 2002/02/25 04:45:00 EST
2930    Tag: leunga-20020225-cps-spill
2931    
2932    This is a new implementation of the CPS spill phase.
2933    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2934    In case of problems, replace it with the old file spill.sml
2935    
2936    The current compiler runs into some serious performance problems when
2937    constructing a large record.  This can happen when we try to compile a
2938    structure with many items.  Even a very simple structure like the following
2939    makes the compiler slow down.
2940    
2941        structure Foo = struct
2942           val x_1 = 0w1 : Word32.int
2943           val x_2 = 0w2 : Word32.int
2944           val x_3 = 0w3 : Word32.int
2945           ...
2946           val x_N = 0wN : Word32.int
2947        end
2948    
2949    The following table shows the compile time, from N=1000 to N=4000,
2950    with the old compiler:
2951    
2952    N
2953    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2954           MLRISC ra                               0.06u  0.00s  0.05g
2955              (spills = 0 reloads = 0)
2956           TOTAL                                   0.63u  0.07s  0.21g
2957    
2958    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2959           MLRISC ra                               5.68u  0.59s  3.93g
2960              (spills = 0 reloads = 0)
2961           TOTAL                                   14.71u  0.99s  4.81g
2962    
2963    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2964           MLRISC ra                               5.54u  0.65s  3.91g
2965              (spills = 543 reloads = 1082)
2966           TOTAL                                   65.40u  3.13s  6.00g
2967    
2968    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2969           MLRISC ra                               0.80u  0.10s  0.55g
2970              (spills = 42 reloads = 84)
2971           TOTAL                                   129.42u  5.10s  4.13g
2972    
2973    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2974           MLRISC ra                               2.69u  0.27s  1.38g
2975              (spills = 62 reloads = 124)
2976           TOTAL                                   682.48u  19.61s  13.99g
2977    
2978    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2979           MLRISC ra                               4.96u  0.27s  2.72g
2980              (spills = 85 reloads = 170)
2981           TOTAL                                   2375.26u  57.21s  48.00g
2982    
2983    As you can see the old cps spill module suffers from some serious
2984    performance problem.  But since I cannot decipher the old code fully,
2985    instead of patching the problems up, I'm reimplementing it
2986    with a different algorithm.  The new code is more modular,
2987    smaller when compiled, and substantially faster
2988    (O(n log n) time and O(n) space).  Timing of the new spill module:
2989    
2990    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2991          MLRISC ra                               0.25u  0.02s  0.15g
2992             (spills=1 reloads=3)
2993          TOTAL                                   7.74u  0.34s  1.62g
2994    
2995    Implementation details:
2996    
2997    As far as I can tell, the purpose of the CPS spill module is to make sure the
2998    number of live variables at any program point (the bandwidth)
2999    does not exceed a certain limit, which is determined by the
3000    size of the spill area.
3001    
3002    When the bandwidth is too large, we decrease the register pressure by
3003    packing live variables into spill records.  How we achieve this is
3004    completely different than what we did in the old code.
3005    
3006    First, there is something about the MLRiscGen code generator
3007    that we should be aware of:
3008    
3009    o MLRiscGen performs code motion!
3010    
3011       In particular, it will move floating point computations and
3012       address computations involving only the heap pointer to
3013       their use sites (if there is only a single use).
3014       What this means is that if we have a CPS record construction
3015       statement
3016    
3017           RECORD(k,vl,w,e)
3018    
3019       we should never count the new record address w as live if w
3020       has only one use (which is often the case).
3021    
3022       We should do something similar to floating point, but the transformation
3023       there is much more complex, so I won't deal with that.
3024    
3025    Secondly, there are now two new cps primops at our disposal:
3026    
3027     1. rawrecord of record_kind option
3028        This pure operator allocates some uninitialized storage from the heap.
3029        There are two forms:
3030    
3031         rawrecord NONE [INT n]  allocates a tagless record of length n
3032         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3033                                     and initializes the tag.
3034    
3035     2. rawupdate of cty
3036          rawupdate cty (v,i,x)
3037          Assigns to x to the ith component of record v.
3038          The storelist is not updated.
3039    
3040    We use these new primops for both spilling and increment record construction.
3041    
3042     1. Spilling.
3043    
3044        This is implemented with a linear scan algorithm (but generalized
3045        to trees).  The algorithm will create a single spill record at the
3046        beginning of the cps function and use rawupdate to spill to it,
3047        and SELECT or SELp to reload from it.  So both spills and reloads
3048        are fine-grain operations.  In contrast, in the old algorithm
3049        "spills" have to be bundled together in records.
3050    
3051        Ideally, we should sink the spill record construction to where
3052        it is needed.  We can even split the spill record into multiple ones
3053        at the places where they are needed.  But CPS is not a good
3054        representation for global code motion, so I'll keep it simple and
3055        am not attempting this.
3056    
3057     2. Incremental record construction (aka record splitting).
3058    
3059        Long records with many component values which are simulatenously live
3060        (recall that single use record addresses are not considered to
3061         be live) are constructed with rawrecord and rawupdate.
3062        We allocate space on the heap with rawrecord first, then gradually
3063        fill it in with rawupdate.  This is the technique suggested to me
3064        by Matthias.
3065    
3066        Some restrictions on when this is applicable:
3067        1. It is not a VECTOR record.  The code generator currently does not handle
3068           this case. VECTOR record uses double indirection like arrays.
3069        2. All the record component values are defined in the same "basic block"
3070           as the record constructor.  This is to prevent speculative
3071           record construction.
3072    
3073    ----------------------------------------------------------------------
3074    Name: Allen Leung
3075    Date: 2002/02/22 01:02:00 EST
3076    Tag: leunga-20020222-mlrisc-tools
3077    
3078    Minor bug fixes in the parser and rewriter
3079    
3080    ----------------------------------------------------------------------
3081    Name: Allen Leung
3082    Date: 2002/02/21 20:20:00 EST
3083    Tag: leunga-20020221-peephole
3084    
3085    Regenerated the peephole files.  Some contained typos in the specification
3086    and some didn't compile because of pretty printing bugs in the old version
3087    of 'nowhere'.
3088    
3089    ----------------------------------------------------------------------
3090    Name: Allen Leung
3091    Date: 2002/02/19 20:20:00 EST
3092    Tag: leunga-20020219-mlrisc-tools
3093    Description:
3094    
3095       Minor bug fixes to the mlrisc-tools library:
3096    
3097       1.  Fixed up parsing colon suffixed keywords
3098       2.  Added the ability to shut the error messages up
3099       3.  Reimplemented the pretty printer and fixed up/improved
3100           the pretty printing of handle and -> types.
3101       4.  Fixed up generation of literal symbols in the nowhere tool.
3102       5.  Added some SML keywords to to sml.sty
3103    
3104    ----------------------------------------------------------------------
3105    Name: Matthias Blume
3106    Date: 2002/02/19 16:20:00 EST
3107    Tag: blume-20020219-cmffi
3108    Description:
3109    
3110    A wild mix of changes, some minor, some major:
3111    
3112    * All C FFI-related libraries are now anchored under $c:
3113        $/c.cm      --> $c/c.cm
3114        $/c-int.cm  --> $c/internals/c-int.cm
3115        $/memory.cm --> $c/memory/memory.cm
3116    
3117    * "make" tool (in CM) now treats its argument pathname slightly
3118      differently:
3119        1. If the native expansion is an absolute name, then before invoking
3120           the "make" command on it, CM will apply OS.Path.mkRelative
3121           (with relativeTo = OS.FileSys.getDir()) to it.
3122        2. The argument will be passed through to subsequent phases of CM
3123           processing without "going native".  In particular, if the argument
3124           was an anchored path, then "make" will not lose track of that anchor.
3125    
3126    * Compiler backends now "know" their respective C calling conventions
3127      instead of having to be told about it by ml-nlffigen.  This relieves
3128      ml-nlffigen from one of its burdens.
3129    
3130    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3131    
3132    * Export C_DEBUG and C_Debug from $c/c.cm.
3133    
3134    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3135      subtyping relationship between incomplete pointers and their complete
3136      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3137      with the convention of instantiating 'o with ('t, 'c) obj whenever
3138      the pointer target type is complete.  In the incomplete case, 'o
3139      will be instantiated with some "'c iobj" -- a type obtained by
3140      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3141    
3142      Operations that work on both incomplete and complete pointer types are
3143      typed as taking an 'o ptr while operations that require the target to
3144      be known are typed as taking some ('t, 'c) obj ptr.
3145    
3146      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3147      where void is an eqtype without any values.  This makes it possible
3148      to work on voidptr values using functions meant to operate on light
3149      incomplete pointers.
3150    
3151    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3152      been vastly simplified.
3153    
3154    ----------------------------------------------------------------------
3155    Name: Matthias Blume
3156    Date: 2002/02/19 10:48:00 EST
3157    Tag: blume-20020219-pqfix
3158    Description:
3159    
3160    Applied Chris Okasaki's bug fix for priority queues.
3161    
3162    ----------------------------------------------------------------------
3163    Name: Matthias Blume
3164    Date: 2002/02/15 17:05:00
3165    Tag: Release_110_39
3166    Description:
3167    
3168    Last-minute retagging is becoming a tradition... :-(
3169    
3170    This is the working release 110.39.
3171    
3172    ----------------------------------------------------------------------
3173    Name: Matthias Blume
3174    Date: 2002/02/15 16:00:00 EST
3175    Tag: Release_110_39-orig
3176    Description:
3177    
3178    Working release 110.39.  New bootfiles.
3179    
3180    (Update: There was a small bug in the installer so it wouldn't work
3181    with all shells.  So I retagged. -Matthias)
3182    
3183    ----------------------------------------------------------------------
3184    Name: Matthias Blume
3185    Date: 2002/02/15 14:17:00 EST
3186    Tag: blume-20020215-showbindings
3187    Description:
3188    
3189    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3190    the latter can be useful for exploring what bindings are available at
3191    the interactive prompt.  (The first function returns only the list
3192    of symbols that are really bound, the second prints those but also the
3193    ones that CM's autoloading mechanism knows about.)
3194    
3195    ----------------------------------------------------------------------
3196    Name: Matthias Blume
3197    Date: 2002/02/15 12:08:00 EST
3198    Tag: blume-20020215-iptrs
3199    Description:
3200    
3201    Two improvements to ml-nlffigen:
3202    
3203      1. Write files only if they do not exist or if their current contents
3204         do not coincide with what's being written.  (That is, avoid messing
3205         with the time stamps unless absolutely necessary.)
3206    
3207      2. Implement a "repository" mechanism for generated files related
3208         to "incomplete pointer types".   See the README file for details.
3209    
3210    ----------------------------------------------------------------------
3211    Name: Matthias Blume
3212    Date: 2002/02/14 11:50:00 EST
3213    Tag: blume-20020214-quote
3214    Description:
3215    
3216    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3217    because of the new and improved tag generation scheme.  (Thanks to Allen
3218    Leung for pointing it out.)
3219    
3220    ----------------------------------------------------------------------
3221    Name: Lal George
3222    Date: 2002/02/14 09:55:27 EST 2002
3223    Tag: george-20020214-isabelle-bug
3224    Description:
3225    
3226    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3227    of Isabelle on the x86.
3228    
3229    From Allen:
3230    -----------
3231     I've found the problem:
3232    
3233         in ra-core.sml, I use the counter "blocked" to keep track of the
3234         true number of elements in the freeze queue.  When the counter goes
3235         to zero, I skip examining the queue.  But I've messed up the
3236         bookkeeping in combine():
3237    
3238             else ();
3239             case !ucol of
3240               PSEUDO => (if !cntv > 0 then
3241                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3242                                        ^^^^^^^^^^^^^^^^^^^^^^^
3243                      moveu := mergeMoveList(!movev, !moveu)
3244                     )
3245                  else ();
3246    
3247         combine() is called to coalesce two nodes u and v.
3248         I think I was thinking that if the move counts of u and v are both
3249         greater than zero then after they are coalesced then one node is
3250         removed from the freeze queue.  Apparently I was thinking that
3251         both u and v are of low degree, but that's clearly not necessarily true.
3252    
3253    
3254    02/12/2002:
3255        Here's the patch.  HOL now compiles.
3256    
3257        I don't know how this impact on performance (compile
3258        time or runtime).  This bug caused the RA (especially on the x86)
3259        to go thru the potential spill phase when there are still nodes on the
3260        freeze queue.
3261    
3262    
3263    
3264    
3265    ----------------------------------------------------------------------
3266    Name: Matthias Blume
3267    Date: 2002/02/13 22:40:00 EST
3268    Tag: blume-20020213-fptr-rtti
3269    Description:
3270    
3271    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3272    updates.
3273    
3274    ----------------------------------------------------------------------
3275    Name: Matthias Blume
3276    Date: 2002/02/13 16:41:00 EST
3277    Tag: blume-20020213-cmlpq
3278    Description:
3279    
3280    Added new priority queue export symbols (which have just been added to
3281    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3282    not compile and the installer would choke.)
3283    
3284    ----------------------------------------------------------------------
3285    Name: Matthias Blume
3286    Date: 2002/02/13 16:15:00 EST
3287    Tag: blume-20020213-various
3288    Description:
3289    
3290    1. More tweaks to ml-nlffigen:
3291    
3292       - better internal datastructures (resulting in slight speedup)
3293       - "-match" option requires exact match
3294       - "localized" gensym counters (untagged structs/unions nested within
3295         other structs/unions or within typedefs get a fresh counter; their
3296         tag will be prefixed by a concatenation of their parents' tags)
3297       - bug fixes (related to calculation of transitive closure of types
3298         to be included in the output)
3299    
3300    2. Minor Basis updates:
3301    
3302       - added implementations for List.collate and Option.app
3303    
3304    ----------------------------------------------------------------------
3305    Name: Matthias Blume
3306    Date: 2002/02/11 15:55:00 EST
3307    Tag: blume-20020211-gensym
3308    Description:
3309    
3310    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3311    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3312    names (ML structure names that correspond to unnamed C structs, unions,
3313    and enums), so that separate runs of ml-nlffigen do not clash.
3314    
3315    ----------------------------------------------------------------------
3316    Name: Matthias Blume
3317    Date: 2002/02/11 12:05:00 EST
3318    Tag: blume-20020211-gensml
3319    Description:
3320    
3321    A quick fix for a problem with GenSML (in the pgraph-util library):
3322    Make generation of toplevel "local" optional.  (Strictly speaking,
3323    signature definitions within "local" are not legal SML.)
3324    
3325    Other than that: updates to INSTALL and cm/TODO.
3326    
3327    ----------------------------------------------------------------------
3328    Name: Matthias Blume
3329    Date: 2002/02/08 15:00:00 EST
3330    Tag: blume-20020208-uniquepid
3331    Description:
3332    
3333    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3334    
3335    1. The installer (config/install.sh) has gotten smarter:
3336    
3337         - Configuration options are a bit easier to specify now
3338           (in config/targets).
3339         - Bug in recognizing .tar.bz2 files fixed.
3340         - Installer automatically resolves dependencies between
3341           configuration options (e.g., if you ask for eXene, you will
3342           also get cml -- regardless whether you asked for it or not).
3343         - Installer can run in "quieter mode" by setting the environment
3344           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3345           "completely silent", though.
3346         - Build HashCons library as part of smlnj-lib.
3347    
3348    2. A new scheme for assigning persistent identifiers to compilation
3349       units (and, by extension, to types etc.) has been put into place.
3350       This fixes a long-standing bug where types and even dynamic values
3351       can get internally confused, thereby compromising type safety
3352       (abstraction) and dynamic correctness.  See
3353    
3354         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3355    
3356       for an example of how things could go wrong until now.
3357    
3358       The downside of the new scheme is that pids are not quite as
3359       persistent as they used to be: CM will generate a fresh pid
3360       for every compilation unit that it thinks it sees for the first
3361       time.  That means that if you compile starting from a clean, fresh
3362       source tree at two different times, you end up with different
3363       binaries.
3364    
3365       Cutoff recompilation, however, has not been compromised because
3366       CM keeps pid information in special caches between runs.
3367    
3368    ----------------------------------------------------------------------
3369    Name: Lal George
3370    Date: 2002/02/07 15:34:13 EST 2002
3371    Tag: <none>
3372    Description:
3373    
3374    Compilers that generate assembly code may produce  global labels
3375    whose value is resolved at link time. The various peephole optimization
3376    modules did not take this in account.
3377    
3378    TODO. The Labels.addrOf function should really return an option
3379    type so that clients are forced to deal with this issue, rather
3380    than an exception being raised.
3381    
3382    ----------------------------------------------------------------------
3383    Name: Lal George
3384    Date: 2002/02/06 13:55:02 EST
3385    Tag: george-20020206-ra-breakup
3386    Description:
3387    
3388    1. A bug fix from Allen.
3389    
3390        A typo causes extra fstp %st(0)'s to be generated at compensation
3391        edges, which might cause stack underflow traps at runtime.  This
3392        occurs in fft where there are extraneous fstps right before the 'into'
3393        trap instruction (in this case they are harmless since none of the
3394        integers overflow.)
3395    
3396    2. Pulled out various utility modules that were embedded in the modules
3397       of the register allocator. I need these modules for other purposes, but
3398       they are not complete enough to put into a library (just yet).
3399    ----------------------------------------------------------------------
3400    Name: Matthias Blume
3401    Date: 2002/01/31 16:05:00 EST
3402    Tag: blume-20020131-sparc-ccalls
3403    Description:
3404    
3405    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3406       of extra stack space by mistake.  Fixed.
3407    
3408    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3409    
3410    ----------------------------------------------------------------------
3411    Name: Allen Leung
3412    Date: 2002/01/30
3413    Tag: leunga-20020130-nowhere-bug-fix
3414    Description:
3415    
3416       MLRISC bug fixes:
3417       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3418       2. MachineInt.fromString was negating its value.
3419    
3420    ----------------------------------------------------------------------
3421    Name: Matthias Blume
3422    Date: 2002/01/29
3423    Tag: blume-20020129-INSTALL
3424    Description:
3425    
3426    - Added somewhat detailed installation instructions (file INSTALL).
3427    - Fixed curl-detection bug in config/install.sh.
3428    - It is now possible to select the URL getter using the URLGETTER
3429      environment variable:
3430    
3431          not set / "unknown"      --> automatic detection (script tries wget,
3432                                       curl, and lynx)
3433          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3434                                       how to properly invoke them)
3435          other                    --> use $URLGETTER directly, it must take
3436                                       precisely two command-line arguments
3437                                       (source URL and destination file name)
3438    
3439    ----------------------------------------------------------------------
3440    Name: Matthias Blume
3441    Date: 2002/01/28
3442    Tag: blume-20020128-sparc-ccalls
3443    Description:
3444    
3445    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3446    - Make use of the allocParam argument in sparc-c-calls.
3447    
3448    ----------------------------------------------------------------------
3449    Name: Matthias Blume
3450    Date: 2002/01/28
3451    Tag: blume-20020128-allocParam
3452    Description:
3453    
3454    John Reppy:  Changes c-calls API to accept client-callback for
3455    allocating extra stack space.
3456    me: Corresponding changes to mlriscGen (using a dummy argument that
3457        does not change the current behavior).
3458    
3459    ----------------------------------------------------------------------
3460    Name: Matthias Blume
3461    Date: 2002/01/28 12:00:00
3462    Tag: Release_110_38
3463    Description:
3464    
3465    This time for real!!!
3466    
3467    ----------------------------------------------------------------------
3468    Name: Matthias Blume
3469    Date: 2002/01/28 10:56:00 EST
3470    Tag: blume-20020128-retraction
3471    Description:
3472    
3473    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3474       with blume-Release_110_38-retracted.)
3475    
3476    1. Fixed a problem with incorrect rounding modes in real64.sml.
3477       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3478    
3479    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3480       and enums fixed.  The naming of corresponding ML identifiers should
3481       now be consistent again.
3482    
3483    ----------------------------------------------------------------------
3484    Name: Allen Leung
3485    Date: 2002/01/27
3486    Tag: leunga-20020127-nowhere
3487    Description:
3488    
3489       Added a target called nowhere in the configuration scripts.
3490       Enabling this will build the MLRISC 'nowhere' tool (for translating
3491       programs with where-clauses into legal SML code) during installation.
3492    
3493    ----------------------------------------------------------------------
3494    Name: Matthias Blume
3495    Date: 2002/01/25 21:27:00 EST
3496    Tag: blume-Release_110_38-retracted
3497    Description:
3498    
3499    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3500    
3501    README will be added later.
3502    
3503    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3504    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3505    
3506    ----------------------------------------------------------------------
3507    Name: Matthias Blume
3508    Date: 2002/01/25
3509    Tag: blume-20020125-ffi
3510    Description:
3511    
3512    A large number of tweaks and improvements to ml-nlffi-lib and
3513    ml-nlffigen:
3514    
3515       - ML represenation types have been streamlined
3516       - getter and setter functions work with concrete values, not abstract
3517         ones where possible
3518       - ml-nlffigen command line more flexible (see README file there)
3519       - some bugs have been fixed (hopefully)
3520    
3521    ----------------------------------------------------------------------
3522    Name: Lal George
3523    Date: 2002/01/24
3524    Tag: george-20020124-risc-ra-interface
3525    Description:
3526    
3527       There is a dramatic simplification in the interface to the
3528       register allocator for RISC architectures as a result of making
3529       parallel copy instructions explicit.
3530    
3531    ----------------------------------------------------------------------
3532    Name: Matthias Blume
3533    Date: 2002/01/22
3534    Tag: blume-20020122-x86-ccalls
3535    Description:
3536    
3537    Bug fix for c-calls on x86 (having to do with how char- and
3538    short-arguments are being handled).
3539    
3540    ----------------------------------------------------------------------
3541    Name: Matthias Blume
3542    Date: 2002/01/21
3543    Tag: blume-20020121-ff
3544    Description:
3545    
3546    Another day of fiddling with the FFI...
3547    
3548    1. Bug fix/workaround:  CKIT does not complain about negative array
3549       dimensions, so ml-nlffigen has to guard itself against this possibility.
3550       (Otherwise a negative dimension would send it into an infinite loop.)
3551    
3552    2. Some of the abstract types (light objects, light pointers, most "base"
3553       types) in structure C are now eqtypes.
3554    
3555    3. Added constructors and test functions for NULL function pointers.
3556    
3557    ----------------------------------------------------------------------
3558    Name: Matthias Blume
3559    Date: 2002/01/18
3560    Tag: blume-20020118-ready-for-new-release
3561    Description:
3562    
3563    Made config/srcarchiveurl point to a new place.  (Will provide boot
3564    files shortly.)
3565    
3566    Maybe we christen this to be 110.38?
3567    
3568    ----------------------------------------------------------------------
3569    Name: Matthias Blume
3570    Date: 2002/01/18
3571    Tag: blume-20020118-more-ffifiddle
3572    Description:
3573    
3574    Today's FFI fiddling:
3575    
3576      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3577        C.Set.* that use concrete (MLRep.*) arguments and results instead
3578        of abstract ones.
3579    
3580      - Provided word-style bit operations etc. for "int" representation
3581        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3582        and Long.
3583    
3584    ----------------------------------------------------------------------
3585    Name: Matthias Blume
3586    Date: 2002/01/18
3587    Tag: blume-20020118-use-x86-fp
3588    Description:
3589    
3590    Now that x86-fast-fp seems to be working, I turned it back on again
3591    by default.  (Seems to work fine now, even with the FFI.)
3592    
3593    Other than that, I added some documentation about the FFI to
3594    src/ml-nlffigen/README and updated the FFI test examples in
3595    src/ml-nlffi-lib/Tests/*.
3596    
3597    ----------------------------------------------------------------------
3598    Name: Allen Leung
3599    Date: 2002/01/17
3600    Tag: leunga-20020117-x86-fast-fp-call
3601    Description:
3602    
3603       1. Fixed a problem with handling return fp values when x86's fast fp
3604          mode is turned on.
3605    
3606       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3607          of %f32.
3608    
3609       3. Added a constructor INT32lit to the ast of MLRISC tools.
3610    
3611    ----------------------------------------------------------------------
3612    Name: Matthias Blume
3613    Date: 2002/01/16
3614    Tag: blume-20020116-ffifiddle
3615    Description:
3616    
3617    More fiddling with the FFI interface:
3618    
3619     - Make constness 'c instead of rw wherever possible.  This eliminates
3620       the need for certain explicit coercions.  (However, due to ML's
3621       value polymorphism, there will still be many cases where explicit
3622       coercions are necessary.  Phantom types are not the whole answer
3623       to modeling a subtyping relationship in ML.)
3624    
3625     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3626    
3627     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3628       clumsy workaround for varargs problem.)
3629    
3630    ----------------------------------------------------------------------
3631    Name: Lal George
3632    Date: 2002/01/15
3633    Tag: <none>
3634    Description:
3635    
3636    1. Since COPY instructions are no longer native to the architecture,
3637       a generic functor can be used to implement the expandCopies function.
3638    
3639    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3640       TEXT segment.
3641    
3642    ----------------------------------------------------------------------
3643    Name: Matthias Blume
3644    Date: 2002/01/15
3645    Tag: blume-20020115-ffiupdates
3646    Description:
3647    
3648    1. Fix for bug resulting in single-precision float values being returned
3649       incorrectly from FFI calls.
3650    
3651    2. Small modifications to C FFI API:
3652    
3653        - memory-allocation routines return straight objects (no options)
3654          and raise an exception in out-of-memory situations
3655        - unsafe extensions to cast between function pointers and pointers
3656          from/to ints
3657        - added structure C_Debug as an alternative to structure C where
3658          pointer-dereferencing (|*| and |*!) always check for null-pointers
3659        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3660          but also takes a (possibly empty) list of existing library handles
3661          that the current library depends on
3662    
3663    ----------------------------------------------------------------------
3664    Name: Matthias Blume
3665    Date: 2002/01/10
3666    Tag: blume-20020110-newffigen
3667    Description:
3668    
3669    1. Updates to portable graph code.
3670    
3671    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3672       more scalable now so that even huge interfaces such as the one
3673       for GTK compile in finite time and space. :-)
3674       See src/ml-nlffigen/README for details on what's new.
3675    
3676    ----------------------------------------------------------------------
3677    Name: Lal George
3678    Date: 2001/01/09 14:31:35 EST 2002
3679    Tag: george-20011206-rm-native-copy
3680    Description:
3681    
3682            Removed the native COPY and FCOPY instructions
3683            from all the architectures and replaced it with the
3684            explicit COPY instruction from the previous commit.
3685    
3686            It is now possible to simplify many of the optimizations
3687            modules that manipulate copies. This has not been
3688            done in this change.
3689    
3690    ----------------------------------------------------------------------
3691    Name: Lal George
3692    Date: 2001/12/06 16:50:13 EST 2001
3693    Tag: george-20011206-mlrisc-instruction
3694    Description:
3695    
3696    Changed the representation of instructions from being fully abstract
3697    to being partially concrete. That is to say:
3698    
3699      from
3700            type instruction
3701    
3702      to
3703            type instr                              (* machine instruction *)
3704    
3705            datatype instruction =
3706                LIVE of {regs: C.cellset, spilled: C.cellset}
3707              | KILL of {regs: C.cellset, spilled: C.cellset}
3708              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3709              | ANNOTATION of {i: instruction, a: Annotations.annotation}
3710              | INSTR of instr
3711    
3712    This makes the handling of certain special instructions that appear on
3713    all architectures easier and uniform.
3714    
3715    LIVE and KILL say that a list of registers are live or killed at the
3716    program point where they appear. No spill code is generated when an
3717    element of the 'regs' field is spilled, but the register is moved to
3718    the 'spilled' (which is present, more for debugging than anything else).
3719    
3720    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3721    We used to generate:
3722    
3723            DEFFREG f1
3724            f1 := f2 + f3
3725            trapb
3726    
3727    but now generate:
3728    
3729            f1 := f2 + f3
3730            trapb
3731            LIVE {regs=[f1,f2,f3], spilled=[]}
3732    
3733    Furthermore, the DEFFREG (hack) required that all floating point instruction
3734    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3735    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3736    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3737    this architecture).
3738    
3739    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3740    all the architectures. This will result in further simplification of the
3741    register allocator that must be aware of them for coalescing purposes, and
3742    will also simplify certain aspects of the machine description that provides
3743    callbacks related to parallel copies.
3744    
3745    ANNOTATION should be obvious, and now INSTR represents the honest to God
3746    machine instruction set!
3747    
3748    The <arch>/instructions/<arch>Instr.sml files define certain utility
3749    functions for making porting easier -- essentially converting upper case
3750    to lower case. All machine instructions (of type instr) are in upper case,
3751    and the lower case form generates an MLRISC instruction. For example on
3752    the alpha we have:
3753    
3754      datatype instr =
3755         LDA of {r:cell, b:cell, d:operand}
3756       | ...
3757    
3758      val lda : {r:cell, b:cell, d:operand} -> instruction
3759        ...
3760    
3761    where lda is just (INSTR o LDA), etc.
3762    
3763    ----------------------------------------------------------------------
3764    Name: Matthias Blume
3765    Date: 2001/11/22 21:40:00 EST
3766    Tag: Release_110_37
3767    Description:
3768    
3769    Release 110.37.  This time for real.
3770    
3771    ----------------------------------------------------------------------
3772    Name: Matthias Blume
3773    Date: 2001/11/21 16:35:00 EST
3774    Tag: blume-20011121-foot-in-mouth
3775    Description:
3776    
3777    Removed the "Release_110_37" tag because of a serious bug.
3778    This will be re-tagged once the bug is fixed.
3779    
3780    ----------------------------------------------------------------------
3781    Name: Matthias Blume
3782    Date: 2001/11/21 16:14:00 EST
3783    Tag: blume-20011121-forgottenfile
3784    Description:
3785    
3786    Forgot to add a file.  (Just a .tex-file -- part of
3787    the CM manual source.)
3788    
3789    ----------------------------------------------------------------------
3790    Name: Matthias Blume
3791    Date: 2001/11/21 16:10:00 EST
3792    Tag: blume-20011121-invalid_110_37
3793    Description:
3794    
3795    Note: I removed the original tag "Release_110_37" from this commit
3796          because we found a serious bug in all non-x86 backends.
3797          - Matthias
3798    
3799    1. Modifications to the SML/NJ code generator and to the runtime system
3800       so that code object name strings are directly inserted into code
3801       objects at code generation time.  The only business the runtime system
3802       has with this is now to read the name strings on occasions.
3803       (The encoding of the name string has also changed somewhat.)
3804    
3805    2. CM now implements a simple "set calculus" for specifying export lists.
3806       In particular, it is now possible to refer to the export lists of
3807       other libraries/groups/sources and form unions as well as differences.
3808       See the latest CM manual for details.
3809    
3810    3. An separate notion of "proxy" libraries has again be eliminated from
3811       CM's model.  (Proxy libraries are now simply a special case of using
3812       the export list calculus.)
3813    
3814    4. Some of the existing libraries now take advantage of the new set
3815       calculus.
3816       (Notice that not all libraries have been converted because some
3817       of the existing .cm-files are supposed to be backward compatible
3818       with 110.0.x.)
3819    
3820    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3821       "General.exnMessage"!  The former relies on a certain hook to be
3822       initialized, and that often does not happen in the stand-alone case.)
3823    
3824    ----------------------------------------------------------------------
3825    Name: Lal George
3826    Date: 2001/11/21  13:56:18 EST
3827    Tag: george-2001121-pseudo-ops
3828    Description:
3829    
3830      Implemented a complete redesign of MLRISC pseudo-ops. Now there
3831      ought to never be any question of incompatabilities with
3832      pseudo-op syntax expected by host assemblers.
3833    
3834      For now, only modules supporting GAS syntax are implemented
3835      but more should follow, such as MASM, and vendor assembler
3836      syntax, e.g. IBM as, Sun as, etc.
3837    
3838    ----------------------------------------------------------------------
3839    Name: Matthias Blume
3840    Date: 2001/11/14 11:52:00 EST
3841    Tag: blume-20011114-srcname
3842    Description:
3843    
3844    1. Routed the name of the current source file to mlriscgen where it
3845       should be directly emitted into the code object.  (This last part
3846       is yet to be done.)
3847    
3848    2. Some cleanup of the pgraph code to make it match the proposal that
3849       I put out the other day.  (The proposal notwithstanding, things are
3850