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

SCM Repository

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

Diff of /sml/trunk/NOTES/HISTORY

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

revision 773, Mon Jan 8 16:18:37 2001 UTC revision 1736, Thu Dec 16 19:31:17 2004 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Date: 2004/12/16 13:35:00 CST
17    Tag: Release_110_52
18    Description:
19    
20    Last-minute changes incorporated into 110.52.  Release tag moved.
21    
22    The changes:
23       - HashString.hashString' -> HashString.hashSubstring
24       - bug fix in UnivariateStats
25    
26    ----------------------------------------------------------------------
27    Name: Matthias Blume (blume (at) tti - c (dot) org)
28    Date: 2004/12/15 23:40:00 CST
29    Tag: blume-20041215-hashSubstring
30    Description:
31    
32    - HashString.hashString' -> HashString.hashSubstring
33    - corresponding changes in atom.sml
34    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
35      added some comments
36    
37    ----------------------------------------------------------------------
38    Name: Matthias Blume (blume (at) tti - c (dot) org)
39    Date: 2004/12/15 15:30:00 CST
40    Tag: (Release_110_52)
41    Description:
42    
43    New working version (110.52).  NEW BOOTFILES!
44    
45    ----------------------------------------------------------------------
46    Name: Matthias Blume (blume (at) tti - c (dot) org)
47    Date: 2004/12/15 12:45:00 CST
48    Tag: blume-20041215-spaces
49    Description:
50    
51    More on the space problem (this time for Win32).
52    
53    ----------------------------------------------------------------------
54    Name: Matthias Blume (blume (at) tti - c (dot) org)
55    Date: 2004/12/14 17:30:00 CST
56    Tag: blume-20041214-spaces
57    Description:
58    
59    Hacked some of the scripts (in particular: the installer) to cope with
60    spaces in filenames a bit better.  But beware: the current "solution"
61    is likely still full of bugs and inherently incomplete.  (We need to
62    do away with those shell scripts for a comprehensive solution.)
63    
64    ----------------------------------------------------------------------
65    Name: Matthias Blume (blume (at) tti - c (dot) org)
66    Date: 2004/12/13 14:45:00 CST
67    Tag: blume-20041213-ml-makedepend
68    Description:
69    
70    Fixed bug in code for ml-makedepend.
71    
72    ----------------------------------------------------------------------
73    Name: Matthias Blume (blume (at) tti - c (dot) org)
74    Date: 2004/12/09 16:30:00 CST
75    Tag: blume-20041209-statistics
76    Description:
77    
78    Added two simple but potentially useful statistics modules to SML/NJ Library.
79    (See CHANGES file there.)
80    
81    ----------------------------------------------------------------------
82    Name: Matthias Blume (blume (at) tti - c (dot) org)
83    Date: 2004/12/01 16:50:00 CST
84    Tag: blume-20041201-atom
85    Description:
86    
87    smlnj-lib:
88    
89      Added function HashString.hashString' for substrings.
90      Hand-inlined CharVector.foldl into HashString (for speed).
91      Modified implementation of structure Atom to avoid extracting
92      strings from substrings unless necessary.
93      (Also see CHANGES file for smlnj-lib.)
94    
95    ----------------------------------------------------------------------
96    Name: Matthias Blume (blume (at) tti - c (dot) org)
97    Date: 2004/11/24 22:15:00 CST
98    Tag: blume-20041124-cml
99    Description:
100    
101    Made sure CML compiles when Position = Int64.
102    
103    ----------------------------------------------------------------------
104    Name: Matthias Blume (blume (at) tti - c (dot) org)
105    Date: 2004/11/24 14:30:00 CST
106    Tag: blume-20041124-position
107    Description:
108    
109    The compiler can now be compiled in a mode that makes structure Position
110    equal to Int64.  The default, however, is unchanged (Position = Int31)
111    for the time being.
112    
113    To enable 64-bit positions, use the following procedure:
114    
115    1. Start sml
116    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
117    3. Type
118          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
119    4. Run CMB.make() as usual.
120    
121    This is barely tested.  The only test so far was a little SML program
122    counting the number of characters in an 8-gigabyte file by
123    reading it character-by-character.  That test was successful.
124    
125    In support of 64-bit positions, a number of new functions have been
126    added to the runtime system.
127    
128    ----------------------------------------------------------------------
129    Name: Matthias Blume (blume (at) tti - c (dot) org)
130    Date: 2004/11/23 14:45:00 CST
131    Tag: blume-20041123-useFile
132    Description:
133    
134    Fixed a problem with unhelpful error messages related to problems with
135    .cm- or .sml files that appear as part of the sml command line.
136    
137    ----------------------------------------------------------------------
138    Name: Matthias Blume (blume (at) tti - c (dot) org)
139    Date: 2004/11/18 15:40:00 CST
140    Tag: Release_110_51
141    Description:
142    
143    New working version (110.51).  NEW BOOTFILES!
144    
145    ----------------------------------------------------------------------
146    Name: Matthias Blume (blume (at) tti - c (dot) org)
147    Date: 2004/11/18 15:35:00 CST
148    Tag: <none>
149    Description:
150    
151    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
152    Joemann).
153    
154    ----------------------------------------------------------------------
155    Name: Allen Leung (leunga (at) reservoir (dot) com)
156    Date: 2004/11/17 16:05:21 EST 2004
157    Tag: leunga-20041117-mlrisc-live-kill
158    Description:
159    
160       Added support for MLTree constructs LIVE and KILL
161    to all the architectures.
162    
163    ----------------------------------------------------------------------
164    Name: Matthias Blume (blume (at) tti - c (dot) org)
165    Date: 2004/11/13 00:20:00 CST
166    Tag: blume-20041113-versiontool
167    Description:
168    
169    - Stripped down the versiontool: It now only handles the version number.
170      The date string is generated at bootstrap time (during makeml).
171    
172    - In a previous commit, fixed a minor issue with how polyequal is being
173      translated.  In particular, the code now "looks through" abstractions.
174      This results in slightly fewer polyEqual warnings and hopefully slightly
175      more efficient code.  Important examples for where this matters are
176      the new int64 and word64 types.
177    
178    ----------------------------------------------------------------------
179    Name: Matthias Blume (blume (at) tti - c (dot) org)
180    Date: 2004/11/12 00:30:00 CST
181    Tag: blume-20041112-int64
182    Description:
183    
184    Structure Int64 fully hooked in.  (The implementation is not very
185    efficient, though.)
186    
187    ----------------------------------------------------------------------
188    Name: Matthias Blume (blume (at) tti - c (dot) org)
189    Date: 2004/11/11 17:30:00 CST
190    Tag: blume-20041111-more64
191    Description:
192    
193    All the pieces of Word64 are now there, with the exception of the
194    conversions from and to LargeWord. (Eventually these need to be identities,
195    but for the time being they don't even make sense because LargeWord is
196    32-bit wide.)
197    
198    Also started to add similar support for Int64, but major pieces of that
199    are still missing.
200    
201    ----------------------------------------------------------------------
202    Name: Matthias Blume (blume (at) tti - c (dot) org)
203    Date: 2004/11/11 00:15:00 CST
204    Tag: blume-20041111-word64
205    Description:
206    
207    Structure Word64 is now (almost) complete, word literals and patterns
208    seem to work.  There are a few odd pieces missing.  In particular,
209    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
210    at the moment.
211    
212    Making Word64 official would mean that LargeWord becomes Word64.  But
213    this requires extreme care because most word-word conversions have to
214    go through LargeWord, so making a mistake means loss of efficiency or
215    worse.  Eventually there will be a solution similar to (but actually
216    simpler than) what I did with IntInf.
217    
218    ----------------------------------------------------------------------
219    Name: Matthias Blume (blume (at) tti - c (dot) org)
220    Date: 2004/11/10 18:12:00 CST
221    Tag: blume-20041110-64bit
222    Description:
223    
224    More 64-bit hacking (but still not even half-way there yet).
225    Also, some assorted improvements to the handling of 8-bit words.
226    
227    ----------------------------------------------------------------------
228    Name: Matthias Blume (blume (at) tti - c (dot) org)
229    Date: 2004/11/09 17:50:00 CST
230    Tag: <none>
231    Description:
232    
233    Started adding some infrastructure for supporting 64-bit int- and
234    word-types.  (Still in its very early stages.)
235    
236    ----------------------------------------------------------------------
237    Name: Matthias Blume (blume (at) tti - c (dot) org)
238    Date: 2004/10/28 10:45:00 CDT
239    Tag: Release_110_50
240    Description:
241    
242    New working version (110.50).  NEW BOOTFILES!
243    
244                 =====================
245    
246    Also:
247    
248      - Changend config/srcarchiveurl from a file just
249        containing the URL string into a file containing
250        shell script code.  The code has access to the $VERSION variable.
251      - Made corresponding changes to config/install.sh and config/unpack.
252      - Default contents of config/srcarchiveurl uses $VERSION and
253        normally does not have to be edited to reflect a version change.
254    
255      (As a result, a version change can be done by just editing
256       config/version, the rest is now automatic.)
257    
258    ----------------------------------------------------------------------
259    Name: Matthias Blume (blume (at) tti - c (dot) org)
260    Date: 2004/10/27 17:50:00 CDT
261    Tag: blume-20041027-btrace-msg
262    Description:
263    
264    BackTrace.monitor now also reports the source of the exception that
265    triggered the trace.
266    
267    ----------------------------------------------------------------------
268    Name: Matthias Blume (blume (at) tti - c (dot) org)
269    Date: 2004/10/27 17:20:00 CDT
270    Tag: blume-20041027-x86-c-calls
271    Description:
272    
273    This is the HISTORY entry for two earlier commits, both concerning
274    the x86 c-calls code in MLRISC:
275    
276      - added a missing LOAD in the code that deals with struct arguments
277      - made sure the caller does not add the wrong number of bytes to the
278        stack pointer after a call of a function returning a struct
279        (the callee already pops the implicit argument which points to
280         the space reserved for the result)
281    
282    ----------------------------------------------------------------------
283    Name: Allen Leung (leunga (at) reservoir (dot) com)
284    Date: 2004/10/24 14:00:00 EST
285    Tag: leunga-20041024-x86-gas-fucomip
286    Description:
287    
288       John discovered a bug in the syntax of fucomip.
289       The opcodes FU?COMIP? have been changed to
290    
291           fu?comip? %st(i), %st
292    
293    ----------------------------------------------------------------------
294    Name: Matthias Blume (blume (at) tti - c (dot) org)
295    Date: 2004/10/20 15:06:00 CDT
296    Tag: blume-20041020-standalone-backtrace
297    Description:
298    
299    Added a mechanism for getting back-trace information from standalone
300    programs.  Here is how it works:
301    
302       1. The part of the program from which you want to get backtrace
303          information (usually the whole program) should be wrapped with
304          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
305          main program could be modified from something like
306    
307             fun main (pgm, args) = ...
308    
309         to
310    
311            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
312    
313       2. To be able to access BackTrace.monitor, you have to add
314          library $smlnj-tdp/plugins.cm to the .cm file that contains your
315          main function.
316    
317       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
318          CM might have created in the past for your project).
319    
320       4. Build the system using this command line:
321    
322             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
323                      myprog.cm MyProg.main myprog
324    
325          instead of the usual
326    
327             ml-build myprog.cm MyProg.main myprog
328    
329    I changed a library name:
330    
331       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
332    
333    New libraries:
334    
335       $smlnj-tdp/back-trace.cm
336          -- when loaded causes the back-trace plugin to be installed
337       $smlnj-tdp/coverage.cm
338          -- when loaded causes the coverage plugin to be installed
339    
340    ----------------------------------------------------------------------
341    Name: Matthias Blume (blume (at) tti - c (dot) org)
342    Date: 2004/10/18 16:45:00 CDT
343    Tag: blume-20041018-groupowner
344    Description:
345    
346    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
347    
348    Eliminated group owner specs from .cm files throughout the source tree.
349    
350    ----------------------------------------------------------------------
351    Name: Matthias Blume (blume (at) tti - c (dot) org)
352    Date: 2004/10/15 15:45:00 CDT
353    Tag: blume-20041015-coverage
354    Description:
355    
356    * Test coverage tool added!
357    * Further reorganization of tracing-, debugging-, and profiling support:
358    
359        - moved original BTImp -- now called BackTrace -- into a separate
360          library called $/trace-debug-profile.cm
361        - eliminated all mentions of BTrace from SMLofNJ.Internals
362        - only the instrumentation mechanism is now left in the compiler proper
363        - BackTrace module is a plugin which is NOT plugged in by default
364        - Coverage module is another such plugin
365    
366    To get the benefits of any of these plugin modules, the code in
367    question must be compiled with tdp instrumentation turned on.  This
368    can be done by setting SMLofNJ.Internals.TDP.mode to true.
369    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
370    
371    Plugins are selected at link time.  (Pre-compiled instrumented code
372    can be re-loaded with different plugins in effect.)  When an
373    instrumented module is linked, whatever plugins are at that time
374    enabled will come into effect for that module.
375    
376    To enable the back-trace plugin, load library $/trace-debug-profile.cm
377    and invoke BackTrace.install() (e.g., from the interactive prompt).
378    To enable the coverage plugin, load the same library and invoke
379    Coverage.install().
380    
381    Back-traces are generated automatically on uncaught exceptions and
382    when the code in question explicitly invokes BackTrace.trigger().
383    
384    Coverage (and execution frequency-) information must be queried
385    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
386    
387    ----------------------------------------------------------------------
388    Name: Matthias Blume (blume (at) tti - c (dot) org)
389    Date: 2004/10/14 17:40:00 CDT
390    Tag: blume-20041014-tdp-core
391    Description:
392    
393    Snapshot of a significant overhaul of how the trace/debug/profile support
394    is hooked into the system (specifically: Core and SMLofNJ.Internals).
395    
396    ----------------------------------------------------------------------
397    Name: Matthias Blume (blume (at) tti - c (dot) org)
398    Date: 2004/10/13 16:34:00 CDT
399    Tag: blume-20041013-tdp
400    Description:
401    
402    Some rationalization of names:
403    
404       structure BTrace -> structure TDPInstrument
405       etc.
406    
407    This is is preparation of using the original back-trace
408    instrumentation for other purposes.  "TDP" stands for
409    Trace/Debug/Profile.
410    
411    The control flag controlling whether instrumentation is on or off is now
412    registered under a different name, so instead of running sml as
413    
414     sml -Cinstrument.btrace-mode=true
415    
416    one has to say
417    
418     sml -Ctdp.instrument=true
419    
420    ----------------------------------------------------------------------
421    Name: Matthias Blume (blume (at) tti - c (dot) org)
422    Date: 2004/10/11 16:37:00 CDT
423    Tag: blume-20041011-regions
424    Description:
425    
426    Made some minor modifications to elabcore.sml to have source regions
427    be propagated more tightly -- resulting in better (i.e., smaller)
428    regions being reported in error- and debug messages.
429    
430    ----------------------------------------------------------------------
431    Name: Matthias Blume (blume (at) tti - c (dot) org)
432    Date: 2004/10/08 22:50:00 CDT
433    Tag: blume-20041008-cmkw
434    Description:
435    
436    Fixed handling of keywords in .cm files: After seeing "is" the lexer
437    treats subsequent occurrences of "group", "library", "source", "is",
438    "*", and "-" as ordinary identifiers rather than keywords.
439    
440    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
441    is supposed to accept a tool argument called "source", but this did
442    not work because of the clash with the keyword.
443    
444    ----------------------------------------------------------------------
445    Name: Matthias Blume (blume (at) tti - c (dot) org)
446    Date: 2004/10/07 16:00:00 CDT
447    Tag: blume-20041007-cleanup
448    Description:
449    
450    Assorted cleanup work:
451    
452      - got rid of intstrmap in favor of using the library's
453        hash table implementation
454      - threw out most of the pathnames stuff, as it was not used anyway
455      - simplified tokentable implementation
456      - fixed some minor spelling errors
457    
458    ----------------------------------------------------------------------
459    Name: Matthias Blume (blume (at) tti - c (dot) org)
460    Date: 2004/10/06 15:15:15 CDT
461    Tag: blume-20041006-handler
462    Description:
463    
464    Cleaned up the absyn to reflect the invariant that HANDLE always
465    carries a FNexp as part of the type definition.  This eliminates some
466    superfluous sanity checks at runtime down the road.
467    
468    Some minor cleanup of the btrace code.
469    
470    ----------------------------------------------------------------------
471    Name: Matthias Blume (blume (at) tti - c (dot) org)
472    Date: 2004/10/01 10:20:30 CDT
473    Tag: blume-20041001-slave
474    Description:
475    
476    Added hack to make slave mode work in the presence of the version
477    tool.  (Still, since the master does two passes over the code for
478    CMB.make, the release number gets bumped twice when slaves are
479    attached.  I don't know if this is worth fixing...)
480    
481    ----------------------------------------------------------------------
482    Name: Matthias Blume (blume (at) tti - c (dot) org)
483    Date: 2004/09/30 10:55:00 CDT
484    Tag: blume-20040930-version
485    Description:
486    
487    * Moved the "version" magic into its own little library under
488      src/system/smlnj/internal.  This avoids expensive reconstruction of
489      a stable src/compiler/core.cm.
490    
491    * At the same time, structure CompilerVersion is now known as
492      structure SMLNJVersion.
493    
494    * Arranged for the version tool to NOT kick in when rebuilding the system
495      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
496      Also, loading the versiontool does not work when rebuilding the system
497      because CM is not properly initialized at that time.
498    
499    ----------------------------------------------------------------------
500    Name: Matthias Blume (blume (at) tti - c (dot) org)
501    Date: 2004/09/29 14:00:00 CDT
502    Tag: blume-20040929-autoversion
503    Description:
504    
505    Implemented some CM magic to have
506        file src/compiler/TopLevel/main/version.sml
507    generated automagically.
508    The version is taken from two files: config/version and config/release.
509    The first is expected to contain a two-part version number such as 110.49.
510    The second should contain a single number, but it may be missing.
511    
512    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
513    time the version tool is loaded (which is the first time you say CMB.make),
514    then the tool will increment the value stored in config/release every
515    time CMB.make is invoked.
516    
517    The binfile format is now insensitive to anything beyond the first
518    two components of a version number, so bumping the release does not render
519    binfiles incompatible.  Auto-bumping can be used to keep track of versions
520    during development without invalidating existing binfiles.
521    
522    In any case, every CMB.make updates the date information in version.sml.
523    (This is the date that is printed in the banner.)
524    
525    ----------------------------------------------------------------------
526    Name: Matthias Blume (blume (at) tti - c (dot) org)
527    Date: 2004/09/28 10:53:00 CDT
528    Tag: blume-20040928-controls
529    Description:
530    
531    Some cleanup of the controls code.
532    
533    ----------------------------------------------------------------------
534    Name: Matthias Blume (blume (at) tti - c (dot) org)
535    Date: 2004/09/27 22:08:00 CDT
536    Tag: blume-20040927-controls
537    Description:
538    
539    Added two pieces of functionality to the Controls interface:
540    
541    1.   val save'restore: 'a control -> unit -> unit
542    
543       grabs the current value of the control in stage 1 and restores it
544       in stage 2.
545    
546    2.   val set' : 'a control * 'a -> unit -> unit
547    
548       stores the given value into the control in stage 2 (i.e., delayed)
549       but does all error checking in stage 1.
550       (This is for string controls that need to do parse their argument --
551        something that might fail.  In some cases, notably in CM, one
552        already knows the intended argument but wants to delay the actual
553        assignment until a time when error recovery would be more difficult.)
554    
555    Changed the handling of controls in tool arguments to classes "sml" and
556    "lazysml":
557       - use Controls.save'restore as a more robust way of restoring the
558         old value (in particular: without having to re-parse the string)
559       - use controls to handle the "overload" keyword in the init group
560         (I believe this change actually fixes a long-standing obscure bug.)
561    
562    ----------------------------------------------------------------------
563    Name: Matthias Blume (blume (at) tti - c (dot) org)
564    Date: 2004/09/27 17:00:00 CDT
565    Tag: blume-20040927-lazysml
566    Description:
567    
568    Added a new tool class called "lazysml" to CM's tool chest.  The only
569    difference to "sml" is that compilation is done with Control.lazysml
570    set to true.  A source of class "lazysml" is automatically recognized
571    by a file name suffix of ".lml".
572    
573    In addition to the above feature, the original class "sml" now also
574    supports a tool argument "lazy" which has the same effect.  As a
575    result, the following three lines are equivalent:
576    
577        foo.sml : lazysml
578        foo.sml : sml (lazy)
579        foo.sml (lazy)
580    
581    The setting goes into effect both during parsing and during
582    compilation.  The original setting is restored right after parsing and
583    after compilation, respectively.
584    
585    In addition to all the above, there is also a general mechanism to set
586    ANY of the "controls" that are available at the command line via
587    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
588    apply as well. (In fact, "lazy" is implemented as a special case of
589    the general mechanism.)
590    
591    The .cm file syntax uses a new keyword tool argument called "with".
592    There are several ways of indicating the desired settings:
593    
594        foo.sml (with:parser.quotations=true)
595        foo.sml (with:(name:parser.quotations value:true))
596        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
597        foo.sml (with:(name1=value1 name2=value2 ...))
598        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
599    
600        etc.
601    
602    Another possible abbreviation is to leave out the =v or value:v part
603    if the name refers to a boolean control (in which case the value is
604    taken to be true).  Thus, one could get lazy sml also by saying:
605    
606        foo.sml (with:parser.lazy-keyword=true)
607        foo.sml (with:parser.lazy-keyword)
608        foo.sml (with:(name:parser.lazy-keyword value:true))
609        foo.sml (with:(name:parser.lazy-keyword))
610    
611    ----------------------------------------------------------------------
612    Name: Matthias Blume (blume (at) tti - c (dot) org)
613    Date: 2004/09/24 16:22:00 CDT
614    Tag: blume-20040924-ppc-long-branch
615    Description:
616    
617    Turned message about "emiting long form of branch" off by default.
618    Added a control flag to turn it back on when desired.
619    
620    ----------------------------------------------------------------------
621    Name: Matthias Blume (blume (at) tti - c (dot) org)
622    Date: 2004/09/24 16:05:00 CDT
623    Tag: blume-20040924-rounding
624    Description:
625    
626    Applied patch for setting rounding modes under Mac OS X.  Thanks to
627    Melissa O'Neill for providing the code!
628    
629    ----------------------------------------------------------------------
630    Name: Matthias Blume (blume (at) tti - c (dot) org)
631    Date: 2004/09/23 17:30:00 CDT
632    Tag: blume-20040923-envvars
633    Description:
634    
635    1. Changed definition of type ControlRegistry.registry_tree to
636       include control_info (i.e., the name of the controlling environment
637       variable).
638    
639    2. Added command-line flags -e and -E to print the names of environment
640       variables that can be used to control internal settings.
641       (This uses the new API mentioned in 1.)
642    
643    ----------------------------------------------------------------------
644    Name: Matthias Blume (blume (at) tti - c (dot) org)
645    Date: 2004/09/13 16:50:00 CDT
646    Tag: Release_110_49
647    Description:
648    
649    New working version (110.49).  NEW BOOTFILES!
650    
651    ----------------------------------------------------------------------
652    Name: Matthias Blume
653    Date: 2004/09/13 16:20:00 CDT
654    Tag: blume-20040913-config-mlrisc
655    Description:
656    
657    Put target "mlrisc" back into the default list.
658    (There is no harm in having it, and some users have expressed their
659    wish to have "mlrisc" included by default.)
660    
661    ----------------------------------------------------------------------
662    Name: John Reppy
663    Date: 2004/09/13
664    Tag: jhr-20040913-signals
665    Description:
666    
667    Fixed the signal masking code to properly nest mask/unmask operations
668    on a per-signal basis.
669    
670    ----------------------------------------------------------------------
671    Name: Matthias Blume (blume (at) tti - c (dot) org)
672    Date: 2004/09/08 13:20:00 CDT
673    Tag: blume-20040908-heap-magic
674    Description:
675    
676    Bumped the heap macig number to 0x09082004 to account for the changed
677    layout of the ML frame under MacOS X.
678    
679    ----------------------------------------------------------------------
680    Name: Allen Leung (leunga (at) reservoir (dot) com)
681    Date: 2004/09/03 11:26:00 EST
682    Tag: leunga-20040903-cygwin-install
683    Description:
684    
685       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
686    The Cygwin runtime is turned on by setting the environment
687    variable SMLNJ_CYGWIN_RUNTIME to 1.
688    
689    ----------------------------------------------------------------------
690    Name: Matthias Blume (blume (at) tti - c (dot) org)
691    Date: 2004/08/31 17:15:00 CDT
692    Tag: blume-20040831-core
693    Description:
694    
695    Added some exports to src/compiler/core.cm upon request by J. Joemann.
696    
697    ----------------------------------------------------------------------
698    Name: Matthias Blume (blume (at) tti - c (dot) org)
699    Date: 2004/08/30 17:55:00 CDT
700    Tag: blume-20040830-installer
701    Description:
702    
703    Upon request by Johannes Joemann:
704    
705    - improved ML code of installer to fall back to coping when renaming
706      fails (i.e., when source and target are on different file systems);
707      the code compiles but has yet to be tested in anger
708    - removed mlrisc from list of default targets (config/targets)
709    
710    ----------------------------------------------------------------------
711    Name: Matthias Blume (blume (at) tti - c (dot) org)
712    Date: 2004/08/27 17:20:00 CDT
713    Tag: blume-20040827-ptreql
714    Description:
715    
716    Added ptreql primop to structure InlineT (upon request from Larry
717    Paulson).
718    
719    ----------------------------------------------------------------------
720    Name: Allen Leung (leunga (at) reservoir (dot) com)
721    Date: 2004/08/15 21:21:00 EST
722    Tag: leunga-110_48-udgraph
723    Description:
724    
725    Another bug fix from Carl Hauser:
726    
727    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
728    > 48c48
729    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
730    > e::L)
731    > ---
732    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
733    > Without this, any deletion of an edge in an undirected graph does severe
734    > violence to the graph.
735    
736    ----------------------------------------------------------------------
737    Name: Allen Leung (leunga (at) reservoir (dot) com)
738    Date: 2004/08/10 23:35:00 EST
739    Tag: leunga-110_48-ppc
740    Description:
741    
742       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
743    
744    ----------------------------------------------------------------------
745    Name: Matthias Blume (blume (at) tti - c (dot) org)
746    Date: 2004/08/10 12:00:00 CDT
747    Tag: Release_110_48
748    Description:
749    
750    New working version (110.48).  NEW BOOTFILES!
751    
752    ----------------------------------------------------------------------
753    Name: Allen Leung (leunga (at) reservoir (dot) com)
754    Date: 2004/08/09 12:21:00 EST
755    Tag: leunga-110_47-dijsktra
756    Description:
757    
758    Bug fix from Carl Hauser:
759    
760    single_source_shortest_paths in dijkstra.sml was observed to get wrong
761    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
762    
763    The problem is that following the expression A.update(dist,s,Num.zero)
764    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
765    
766    ----------------------------------------------------------------------
767    Name: Matthias Blume (blume (at) tti - c (dot) org)
768    Date: 2004/08/06 18:10:00 CDT
769    Tag: blume-20040806-cmdline
770    Description:
771    
772    Fiddled with handling of command-line options:
773    
774      * sml now quits after processing the command line
775        if -H, -S, -h<n>, or -s<n> appears as the last
776        command-line argument
777      * a new option -q terminates the session when encountered on
778        the command line; subsequent arguments will be ignored
779      * bug fixes: short (erroneous) arguments are no longer ignored
780        completely
781    
782    ----------------------------------------------------------------------
783    Name: Allen Leung (leunga (at) reservoir (dot) com)
784    Date: 2004/08/04 18:17:00 EST
785    Tag: leunga-110_47-ppc-ibm-asm
786    Description:
787    
788      - Added minimal IBM assembly syntax support for PowerPC.
789    
790      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
791        been altered in the runtime.  We need an automatic way to keep the file
792        in sync.
793    
794    ----------------------------------------------------------------------
795    Name: Matthias Blume (blume (at) tti - c (dot) org)
796    Date: 2004/08/04 14:00:00 CDT
797    Tag: Release_110_47
798    Description:
799    
800    New working version (110.47).  NEW BOOTFILES!
801    
802    ----------------------------------------------------------------------
803    Name: Matthias Blume (blume (at) tti - c (dot) org)
804    Date: 2004/08/03 14:25:00 CDT
805    Tag: blume-20040803-callingconv
806    Description:
807    
808    Added low-level support for choosing C calling conventions by
809    twiddling the type of rawccall.  (See
810    src/compiler/Semant/types/cproto.sml for details.)
811    
812    ----------------------------------------------------------------------
813    Name: Matthias Blume (blume (at) tti - c (dot) org)
814    Date: 2004/08/02 15:55:00 CDT
815    Tag: blume-20040802-backout
816    Description:
817    
818    Backed out of change to win32-filesys.c.  The earlier patch to
819    get_file_time caused CM to produce files with the wrong time stamp.
820    
821    ----------------------------------------------------------------------
822    Name: Matthias Blume (blume (at) tti - c (dot) org)
823    Date: 2004/08/02 14:45:00 CDT
824    Tag: blume-20040802-nlffi-win32
825    Description:
826    
827    Added NLFFI support for Win32, adapted from a patch provided by David
828    Hansel.  This is currently completely untested.  Also, the issue
829    concerning stdcall vs. ccall is still unresolved.
830    
831    ----------------------------------------------------------------------
832    Name: Matthias Blume (blume (at) tti - c (dot) org)
833    Date: 2004/07/30 17:55:00 CDT
834    Tag: blume-20040730-various
835    Description:
836    
837    Gearing up towards 110.47...
838    
839    - various minor bugfixes to ml-nlffigen
840    - a beginning of a manual for nlffi
841    
842    - eliminated 'export name=value' in config/install.sh as this does
843      not work with certain versions of /bin/sh
844      (Thanks to David King at Motorola for catching this.)
845    
846    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
847      - added a test for tm==NULL to gmtime.c and localtime.c
848      - applied patch for incorrect GetFileTime under win32
849      - toSeconds -> toMilliseconds in Win32/win32-process.sml
850    
851    ----------------------------------------------------------------------
852    Name: Matthias Blume (blume (at) tti - c (dot) org)
853    Date: 2004/07/21 18:20:00 CDT
854    Tag: blume-20040721-nlffigen
855    Description:
856    
857    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
858      for a typedef to an incomplete type, but leave out the "typ" member.
859      (This is just for consistency.)
860    - Started to produce what is supposed to become better (i.e., comprehensive)
861      documentation of what ml-nlffigen does and produces.
862    
863    ----------------------------------------------------------------------
864    Name: Matthias Blume (blume (at) tti - c (dot) org)
865    Date: 2004/07/14 16:25:00 CDT
866    Tag: blume-20040714-union
867    Description:
868    
869    Added C_UNION to c-calls/c-types.sml and updated the machinery
870    (ml-nlffigen, cproto.sml) that conveys C function interface
871    information to the code generator.
872    
873    However, the actual architecture-specific implementation of function
874    arguments and results that are C unions is still not implemented.
875    
876    ----------------------------------------------------------------------
877    Name: Allen Leung (leunga (at) reservoir (dot) com)
878    Date: 2004/07/14 14:38:00 EST
879    Tag: leunga-110_46_1-ppc-lwzu
880    Description:
881    
882        Added these instructions to the PowerPC architecture:
883        LBZU(X), LHZU(X), LWZU(X),
884        STWU(X), STFDU, STFSU
885    
886        etc...
887    
888        Note: I haven't added their instruction encoding into the description.
889    
890    ----------------------------------------------------------------------
891    Name: Allen Leung (leunga (at) reservoir (dot) com)
892    Date: 2004/07/13 15:04:00 EST
893    Tag: leunga-110_46_1-ppc-lwarx
894    Description:
895    
896       Added the two instructions LWARX and STWCX to the PowerPC
897    instruction set.
898    
899       A (untested) rewrite of loop-structure.sml.  The old version
900    is completely broken.
901    
902    ----------------------------------------------------------------------
903    Name: Matthias Blume (blume (at) tti - c (dot) org)
904    Date: 2004/07/13 13:50:00 CDT
905    Tag: blume-20040713-nlffi
906    Description:
907    
908    - use paramAlloc to report c-calls with too many arguments
909      (for PPC version where parameter area is pre-allocated)
910    - added ccall_maxargspace to machspec (to implement the above)
911    - made "make" commend in CM's "make" tool configurable
912    - added option (default: on) for passing the name of the SML/NJ's "bin"
913      directory to "make"; the call looks like this:
914    
915         make <options> SMLNJ_BINDIR=<dir> <target>
916    
917      This can be used by the Makefile to, e.g., pick the "right" version
918      of ml-nlffigen.
919    - minor code tweaks
920    
921    ----------------------------------------------------------------------
922    Name: Matthias Blume (blume (at) tti - c (dot) org)
923    Date: 2004/07/12 22:50:00 CDT
924    Tag: blume-110_46_1-macosx-nlffi
925    Description:
926    
927    NLFFI under Mac OS X now working (sort of).  This is largely untested,
928    though.
929    
930    Note:  1. You have to make a new, clean build of the runtime system.
931           2. There are new BOOTFILES, you have to use them!
932              (Doing the bootstrap process yourself would be *very* painful!
933               If you absolutely have to do it, build the system under
934               a different architecture and then cross-compile.)
935    
936    Version bumped to 110.46.1 to account for runtime data format changes.
937    
938    ----------------------------------------------------------------------
939    Name: Matthias Blume (blume (at) tti - c (dot) org)
940    Date: 2004/06/18 14:30:00 CDT
941    Tag: blume-20040618-unix
942    Description:
943    
944    Changed the implementation of structure Unix so that the same stream
945    is returned every time one of the {text,bin}{In,Out}streamOf functions
946    is invoked on the same proc.  This is not what the spec currently
947    says -- although IMO it arguably should.  (See discussion below.)
948    
949    ----------------------------------------------------------------------
950    Name: Matthias Blume (blume (at) tti - c (dot) org)
951    Date: 2004/06/17 18:15:00 CDT
952    Tag: Release_110_46
953    Description:
954    
955    New working version (110.46).  NEW BOOTFILES!
956    
957    ----------------------------------------------------------------------
958    Name: Matthias Blume (blume (at) tti - c (dot) org)
959    Date: 2004/06/17 17:20:00 CDT
960    Tag: blume-20040617-timer-unix
961    Description:
962    
963    Changed the interface of structures Timer and Unix to match the most
964    recent Basis spec.
965    
966    In the case of Unix there still seems to be an open/weird issue:
967    
968       The {text,bin}{In,Out}streamOf functions are supposed to create
969       fresh streams whenever they are called -- as opposed to have them
970       return the same stream every time.  This design is supposed to
971       prevent space leaks caused by proc values hanging on to streams.
972    
973       The reap function, on the other hand, is supposed to close the
974       streams.  This cannot be done without having a handle on the
975       stream in proc after all...
976    
977    I took the liberty to implement the following stopgap solution:
978    
979      The proc value hangs on to the most recently created stream(s).
980      Reap closes those.  If either or both of the two streams hadn't
981      been created at all yet, then reap will close the corresponding
982      file descriptors directly.
983    
984    PS: I don't understand the original space leak argument anymore.  If
985    a proc hangs on to the imperative stream, then I/O operations on those
986    will advance the state of the cached stream and avoid the space leak.
987    
988    ----------------------------------------------------------------------
989    Name: Matthias Blume (blume (at) tti - c (dot) org)
990    Date: 2004/05/28 16:45:00 CDT
991    Tag: blume-20040528-basis
992    Description:
993    
994    Added signature PACK_REAL and exported functor PrimIO.
995    
996    ----------------------------------------------------------------------
997    Name: Matthias Blume (blume (at) tti - c (dot) org)
998    Date: 2004/05/25 16:00:00 CDT
999    Tag: blume-20040525-group-owner
1000    Description:
1001    
1002    CM now ignores (but still accepts) the "owner" information in group
1003    descriptions.  The owner of a group is its next enclosing
1004    library. Each group must have a unique owner.  (There is a virtual
1005    "toplevel" library that owns groups which are not nested within a real
1006    library.)  Previously, each group had to explicitly declare its owner,
1007    and CM would check that such a declaration is correct.  The new scheme
1008    is to have CM check that for each group there is precisely one owning
1009    library.
1010    
1011    The advantage of the new scheme is that the programmer no longer needs
1012    to maintain the somewhat annoying owner information.  The downside is
1013    that CM cannot enforce the ownership rule across multiple runs of
1014    CM.make.  Fortunately, enclosing the same group in two different
1015    libraries A and B which are not part of the same program does not
1016    cause real problems.
1017    
1018    ----------------------------------------------------------------------
1019    Name: Matthias Blume (blume (at) tti - c (dot) org)
1020    Date: 2004/05/20 16:00:00 CDT
1021    Tag: blume-20040520-win32
1022    Description:
1023    
1024    Made the win32 version work again.  (Strangely, a misplaced comma had
1025    slipped into win32-process.c which prevented the runtime from being
1026    compiled correctly.)
1027    
1028    Also, included a minor addition to ml-build.bat analogous to what was
1029    done in blume-20040519-ml-build.
1030    
1031    ----------------------------------------------------------------------
1032    Name: Matthias Blume (blume (at) tti - c (dot) org)
1033    Date: 2004/05/19 22:10:00 CDT
1034    Tag: blume-20040519-ml-build
1035    Description:
1036    
1037    Arranged for ml-build to clean up after itself a little bit better.
1038    The script generates a temporary SML source file and compiles it using
1039    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1040    gets rid of those at the end, so they don't accumulate under .cm.
1041    
1042    This required a minor change to install.sh because the name of the
1043    metadata directory (default: .cm) is actually configurable at
1044    installation time.
1045    
1046    ----------------------------------------------------------------------
1047    Name: Matthias Blume (blume (at) tti - c (dot) org)
1048    Date: 2004/05/18 15:50:00 CDT
1049    Tag: blume-20040518-mkreader
1050    Description:
1051    
1052    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1053    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1054    
1055    ----------------------------------------------------------------------
1056    Name: Matthias Blume (blume (at) tti - c (dot) org)
1057    Date: 2004/05/11 14:35:00 CDT
1058    Tag: blume-20040511-win32sock
1059    Description:
1060    
1061    Added previously missing support for many socket-related functions
1062    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1063    for the voluminous patch!
1064    
1065    (I have not tested this patch under win32 yet.)
1066    
1067    Here is David's e-mail:
1068    
1069    Hi,
1070    
1071    Attached to this email you find a diff against sml/nj 110.45
1072    that will enable socket support under Windows.
1073    
1074    To apply the patch (using unix or cygwin)
1075    1) gunzip runtime.diff.gz
1076    2) "cd" into "src/runtime" in the source tree of a fresh
1077       110.45 installation.
1078    3) patch -p 1 < [your/path/to]runtime.diff
1079    
1080    The code compiles fine but has NOT yet been extensively tested.
1081    I only ran a few tests for basic socket client functionality
1082    (which worked fine).  Especially the functions that use ioctl
1083    are not tested at all and might not work (see below).
1084    
1085    I implemented this since we want to move to a newer version of sml/nj
1086    but need socket support in order to use it.  This is the first time I
1087    even had a look at the sml/nj source,  so please review my changes
1088    before making this part of the distribution!  Here are a few issues
1089    that I think might be better for someone to solve who is more
1090    familiar with the sml/nj source (and socket programming):
1091    
1092    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1093      exception since I could not figure out what the windows equivalent
1094      of these functions is
1095    
1096    - In sockets-osdep.h there are a some #include statements that are
1097      only used in a few files that include sockets-osdep.h
1098    
1099    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1100      does not process its return value since I don't know how to
1101      report an error upwards.
1102    
1103    - It would probably be good to have a call to WSACleanup() when
1104      the library is unloaded (if there is such a possibility).
1105      Otherwise I think Windows will take care of this automatically
1106      when the process finishes.
1107    
1108    - I used ioctlsocket() as a replacement for ioctl() but I have
1109      no idea if that is actually the proper replacement on Windows.
1110    
1111    - All these issues are marked in the code by "FIXME" comments.
1112    
1113    We use sml/nj extensively in our products and are quite happy
1114    with it.  I hope this contribution will help you.
1115    
1116    Keep up the good work!
1117    
1118    David
1119    
1120    ----------------------------------------------------------------------
1121    Name: Matthias Blume (blume (at) tti - c (dot) org)
1122    Date: 2004/05/11 14:20:00 CDT
1123    Tag: blume-20040511-installml
1124    Description:
1125    
1126    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1127    the patch.)
1128    
1129    ----------------------------------------------------------------------
1130    Name: Matthias Blume (blume (at) tti - c (dot) org)
1131    Date: 2004/05/11 14:05:00 CDT
1132    Tag: blume-20040511-nlffi-netbsd
1133    Description:
1134    
1135    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1136    the patch.)
1137    
1138    ----------------------------------------------------------------------
1139    Name: Matthias Blume (blume (at) tti - c (dot) org)
1140    Date: 2004/05/11 12:05:00 CDT
1141    Tag: blume-20040511-exports
1142    Description:
1143    
1144    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1145    export lists in compiler-related .cm-files.
1146    
1147    ----------------------------------------------------------------------
1148    Name: Matthias Blume (blume (at) tti - c (dot) org)
1149    Date: 2004/05/11 11:35:00 CDT
1150    Tag: blume-20040511-allsource
1151    Description:
1152    
1153    The installer now honors the "src-smlnj" target again, although its meaning
1154    has changed from "all sources required for the compiler" to "all sources
1155    the installer knows about".  In other words, if you enable "src-smlnj"
1156    in the "targets" file, then the installer will pull in sources for
1157    everything.  (Notice that this refers to source code only.  Compiled
1158    code is still only installed for modules that were requested explicitly
1159    or which are required for other modules that were requested explicitly.)
1160    
1161    ----------------------------------------------------------------------
1162    Name: Matthias Blume (blume (at) tti - c (dot) org)
1163    Date: 2004/04/23 17:40:00 CDT
1164    Tag: blume-20040423-ieee-scan
1165    Description:
1166    
1167    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1168    in the exponent calculation we get INF or ZERO (depending on the mantissa
1169    and the sign of the exponent).
1170    
1171    ----------------------------------------------------------------------
1172    Name: Matthias Blume (blume (at) tti - c (dot) org)
1173    Date: 2004/04/23 10:40:00 CDT
1174    Tag: blume-20040423-ml-build
1175    Description:
1176    
1177    The ml-build script now terminates with a non-0 status when something
1178    goes wrong.
1179    
1180    ----------------------------------------------------------------------
1181    Name: Matthias Blume (blume (at) tti - c (dot) org)
1182    Date: 2004/04/22 16:35:00 CDT
1183    Tag: blume-20040422-Option
1184    Description:
1185    
1186    Made exception Option to be the same as exception Option.Option
1187    (as it should be).
1188    
1189    ----------------------------------------------------------------------
1190    Name: Allen Leung (leunga (at) reservoir (dot) com)
1191    Date: 2004/03/19 14:40:00 EST
1192    Tag: leunga-20040319-cygwin-nlffi
1193    Description:
1194    
1195    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1196    of SML/NJ.  The problem is that
1197    
1198       lib = dlopen(NULL, ...)
1199       f   = dlsym(lib, "malloc");
1200    
1201    does not work on Windows unless we explicitly export symbols
1202    such as 'malloc' during linking.  We fixed this by explicitly
1203    exporting the required symbols with the magic gcc incantation:
1204    
1205        -Wl,--export-all cygwin.def
1206    
1207    where cygwin.def is a file containing all the symbols that
1208    we wish to export.
1209    
1210    I suspect this is a Windows problem and we'll have to
1211    do the same (somehow with windows compilers) when
1212    we build the native win32 version with the system
1213    calls LoadLibrary/GetProcAddress.
1214    
1215    ----------------------------------------------------------------------
1216    Name: Matthias Blume (blume (at) tti - c (dot) org)
1217    Date: 2004/03/04 16:35:00 CST
1218    Tag: blume-20040304-intinf-fmt
1219    Description:
1220    
1221    Fixed problem with IntInf.fmt (sign would show up on the right instead
1222    of on the left for BIN, OCT, and HEX).
1223    
1224    ----------------------------------------------------------------------
1225    Name: Matthias Blume (blume (at) tti - c (dot) org)
1226    Date: 2004/03/04 11:25:00 CST
1227    Tag: blume-20040304-symlinks
1228    Description:
1229    
1230    Fixed problem with installer script (unix only) where bin/ml-yacc and
1231    friends pointed (via symlinks) to absolute locations instead of just
1232    .run-sml.  This was reported by Vesa A Norrman.
1233    
1234    ----------------------------------------------------------------------
1235    Name: Matthias Blume (blume (at) tti - c (dot) org)
1236    Date: 2004/02/13 14:50:00 CST
1237    Tag: Release_110_45
1238    Description:
1239    
1240    New working version (110.45). New bootfiles.
1241    
1242    ----------------------------------------------------------------------
1243    Name: Matthias Blume (blume (at) tti - c (dot) org)
1244    Date: 2004/01/26 15:15:15 CST
1245    Tag: blume-20040126-toplevel
1246    Description:
1247    
1248    Improved handling of exceptions at the interactive toplevel.
1249    
1250    ----------------------------------------------------------------------
1251    Name: Matthias Blume (blume (at) tti - c (dot) org)
1252    Date: 2004/01/26 11:25:00
1253    Tag: blume-20040126-app
1254    Description:
1255    
1256    Type of top-level "app" corrected.
1257    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1258    
1259    ----------------------------------------------------------------------
1260    Name: Matthias Blume (blume (at) tti - c (dot) org)
1261    Date: 2003/11/18 17:10 CST
1262    Tag: blume-20031118-basis-fiddle
1263    Description:
1264    
1265    - changed Timer interface to what might become the spec
1266    - POSIX_FLAGS -> BIT_FLAGS according to spec
1267    - some other minor discrepancies wrt. spec eliminated
1268    
1269    ----------------------------------------------------------------------
1270    Name: Matthias Blume (blume (at) tti - c (dot) org)
1271    Date: 2003/11/06 12:00:00 CST
1272    Tag: Release_110_44
1273    Description:
1274    
1275    New working version (110.44). New bootfiles.
1276    
1277    ----------------------------------------------------------------------
1278    Name: Matthias Blume (blume (at) tti - c (dot) org)
1279    Date: 2003/11/04 11:50:00 CST
1280    Tag: blume-20031104-move-libraries
1281    Description:
1282    
1283    Eliminated the "dont_move_libraries" directive in config/targets.
1284    (The mechanism was broken and could not be fixed easily.  Moreover,
1285    there does not seem to be any reason not to move all libraries into
1286    lib during installation.  I originally implemented this directive as a
1287    backward-compatibility feature when I first introduced the new CM.
1288    Now that things have been stable for a long time and going back to the
1289    old CM is not an option, there is no reason to keep it around.)
1290    
1291    ----------------------------------------------------------------------
1292    Name: Matthias Blume (blume (at) tti - c (dot) org)
1293    Date: 2003/11/03 16:00:00 CST
1294    Tag: blume-20031103-installdir
1295    Description:
1296    
1297    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1298    Richards for pointing out the problem and providing the solution.)
1299    
1300    ----------------------------------------------------------------------
1301    Name: Matthias Blume (blume (at) tti - c (dot) org)
1302    Date: 2003/10/01 17:05:00 CDT
1303    Tag: blume-20031001-lal-mlrisc
1304    Description:
1305    
1306    MLRISC bug fix from Lal.
1307    
1308    ----------------------------------------------------------------------
1309    Name: Matthias Blume (blume (at) tti - c (dot) org)
1310    Date: 2003/09/30 16:10:00 CDT
1311    Tag: blume-20030930-primio-bat
1312    Description:
1313    
1314    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1315    2. Improved .bat files (for Win32 port) to make things work under Win95.
1316       (thanks to Aaron S. Hawley for this one)
1317    
1318    ----------------------------------------------------------------------
1319    Name: Matthias Blume (blume (at) tti - c (dot) org)
1320    Date: 2003/09/26 16:05:00 CDT
1321    Tag: blume-20030926-wrappriv
1322    Description:
1323    
1324    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1325    
1326    ----------------------------------------------------------------------
1327    Name: Matthias Blume (blume (at) tti - c (dot) org)
1328    Date: 2003/09/26 15:00:00 CDT
1329    Tag: blume-20030926-110_43_3
1330    Description:
1331    
1332    - additional cleanup
1333    - version number bump, NEW BOOTFILES
1334    
1335    ----------------------------------------------------------------------
1336    Name: Matthias Blume (blume (at) tti - c (dot) org)
1337    Date: 2003/09/26 12:00:00 CDT
1338    Tag: blume-20030926-ppautoload
1339    Description:
1340    
1341    I modified the read-eval-print loop so that the autoloader gets
1342    invoked whenever the prettyprinter tries to look up a symbol that
1343    is not currently defined in the toplevel environment but which
1344    appears in CM's autoload registry.  As a result, we see far fewer of
1345    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1346    
1347    In addition to this I tried to clean up some pieces of the Basis
1348    implementation (e.g., Socket, Word8Array) in order to prevent other
1349    instances of these ?.Foo.Bar.xxx names from being printed.
1350    
1351    The mechanism that picks names for types still needs some work, though.
1352    (Right now it seems that if there is a type A.t which is defined to
1353    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1354    "?.B.u" although the perhaps more sensible solution would be to use
1355    "A.t" in this case.  In other words, the prettyprinter should follow
1356    a chain of DEFtycs not farther than there are corresponding toplevel
1357    names in the current environment.)
1358    
1359    ----------------------------------------------------------------------
1360    Name: Matthias Blume (blume (at) tti - c (dot) org)
1361    Date: 2003/09/24 16:31:00 CDT
1362    Tag: blume-20030924-installer
1363    Description:
1364    
1365    Another installer tweak:  All the ML code for the installer is now
1366    compiled during CMB.make and put into a little library called
1367    $smlnj/installer.cm.  The installation then simply invokes
1368    
1369       sml -m $smlnj/installer.cm
1370    
1371    and everything happens automagically.
1372    
1373    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1374    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1375            and BINDIR.
1376    
1377    The new scheme guarantees that the ML code responsible for the installation
1378    is in sync with the APIs of the main system.  Also, the installer is
1379    somewhat faster because the installer script is precompiled.
1380    
1381    ----------------------------------------------------------------------
1382    Name: Matthias Blume (blume (at) tti - c (dot) org)
1383    Date: 2003/09/24 15:35:00 CDT
1384    Tag: blume-20030924-synsock
1385    Description:
1386    
1387    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1388    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1389    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1390    implementations of non-blocking operations from CML's Socket
1391    structure.
1392    
1393    ----------------------------------------------------------------------
1394    Name: Matthias Blume (blume (at) tti - c (dot) org)
1395    Date: 2003/09/24 15:10:05 CDT
1396    Tag: blume-20030924-sockets
1397    Description:
1398    
1399    1. Fixed SOCKET API and implementation to match Basis spec.
1400       This required changing the internal representation of sockets to one
1401       that remembers (for each socket file descriptor) whether it is currently
1402       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1403       call is made only if the state actually needs to change).
1404    
1405    2. OS-specific details of sockets were moved into separate files, thus
1406       making it possible to unify the bulk of the socket implementations
1407       between Unix and Win32.
1408    
1409    3. CML's socket API changed accordingly.
1410       (Note that we need to remove non-blocking functions from this API
1411       since they are redundant in the case of CML!)
1412    
1413    4. CML's socket implementation now makes use of non-blocking functions
1414       provided by Basis, thus removing all OS-dependent code from this part
1415       of CML.
1416    
1417    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1418    
1419    ----------------------------------------------------------------------
1420    Name: Matthias Blume (blume (at) tti - c (dot) org)
1421    Date: 2003/09/22 12:10:00 CDT
1422    Tag: blume-20030922-110_43_2
1423    Description:
1424    
1425    Made a new interim version and bootfiles for developer's bootstrapping
1426    convenience.
1427    
1428       110.43.2 -- NEW BOOTFILES
1429    
1430    ----------------------------------------------------------------------
1431    Name: Matthias Blume (blume (at) tti - c (dot) org)
1432    Date: 2003/09/19 15:55:00 CDT
1433    Tag: blume-20030919-cmdir
1434    Description:
1435    
1436    1. new-install.sh -> install.sh
1437    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1438    3. tweaked installer so that another name instead of .cm can be chosen
1439       at install time (by setting the CM_DIR_ARC environment variable
1440       during installation); once installation is complete, the name is
1441       fixed
1442    
1443    ----------------------------------------------------------------------
1444    Name: Matthias Blume (blume (at) tti - c (dot) org)
1445    Date: 2003/09/18 16:00:00 CDT
1446    Tag: blume-20030918-110_43_1
1447    Description:
1448    
1449    Made a new interim version and bootfiles for developer's bootstrapping
1450    convenience.
1451    
1452       110.43.1 -- NEW BOOTFILES
1453    
1454    ----------------------------------------------------------------------
1455    Name: Matthias Blume (blume (at) tti - c (dot) org)
1456    Date: 2003/09/18 15:20:00 CDT
1457    Tag: blume-20030918-misc
1458    Description:
1459    
1460    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1461       this seems to be controversial at the moment) and used those in
1462       Posix.ProcEnv.times.
1463    
1464    2. Added Time.{from,to}Nanoseconds to Time.
1465    
1466    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1467       For example, fromLargeInt never needs to look at more than 3 "big
1468       digits" to get its 53 bits of precision.
1469    
1470    ----------------------------------------------------------------------
1471    Name: Matthias Blume (blume (at) tti - c (dot) org)
1472    Date: 2003/09/17 16:30:00 CDT
1473    Tag: blume-20030917-real32-slices
1474    Description:
1475    
1476    Added an entry to the primitive environment
1477    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1478    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1479    
1480    Removed some of the "magic" constants in real64.sml and replaced them
1481    with code that generates these values from their corresponding
1482    integer counterparts.
1483    
1484    Made all(?) the slice-related changes to the Basis and made everything
1485    compile again...
1486    
1487    ----------------------------------------------------------------------
1488    Name: Matthias Blume (blume (at) tti - c (dot) org)
1489    Date: 2003/09/15 17:45:00 CDT
1490    Tag: blume-20030915-rbase
1491    Description:
1492    
1493    Fixed bug in Real.fromLargeInt.
1494    
1495    ----------------------------------------------------------------------
1496    Name: Matthias Blume (blume (at) tti - c (dot) org)
1497    Date: 2003/09/13 18:11:00 CDT
1498    Tag: blume-20030913-libinstall
1499    Description:
1500    
1501    Minor bugfix in config/libinstall (set anchor with path to
1502    standalone tool after installing it, otherwise libraries that
1503    need ml-lex or ml-yacc won't compile the first time the installer
1504    runs).
1505    
1506    ----------------------------------------------------------------------
1507    Name: Matthias Blume (blume (at) tti - c (dot) org)
1508    Date: 2003/09/12 11:45:00 CDT
1509    Tag: blume-20030912-various
1510    Description:
1511    
1512     - fixed bug in Real.toLargeInt
1513     - fixed bug in Posix.ProcEnv.times
1514     - changed inputLine functions to return an option
1515     - minor installer improvements / bugfixes
1516     - changed default @SMLalloc parameter for x86/celeron to 64k
1517    
1518    ----------------------------------------------------------------------
1519    Name: Matthias Blume (blume (at) tti - c (dot) org)
1520    Date: 2003/09/09 22:00:00 CDT
1521    Tag: Release_110_43
1522    Description:
1523    
1524        New working release 110.43.  New bootfiles.
1525    
1526    ----------------------------------------------------------------------
1527    Name: Matthias Blume (blume (at) tti - c (dot) org)
1528    Date: 2003/09/09 19:20:00 CDT
1529    Tag: blume-20030909-installer
1530    Description:
1531    
1532    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1533    Modified config/install.bat to take advantage of it.  Also modified
1534    config/install.sh (and called it config/new-install.sh) to take advantage
1535    of it on Unix systems.  (The SML code is (supposed to be) platform-
1536    independent.)
1537    
1538    The installer can now install everything under Win32
1539    as well as under *nix as long as it compiles.
1540    
1541    Other changes:
1542    
1543      - made CML compile again under Win32
1544      - made eXene compile under Win32 (by providing a fake structure UnixSock
1545        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1546      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1547        same as type int;  under Win32 it isn't
1548      - fixed some slice-related problems in the win32-specific parts of CML
1549      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1550        (under Win32, the volume name is case-insensitive, and the
1551        OS.Path code compares volume names for equality)
1552    
1553    ----------------------------------------------------------------------
1554    Name: Matthias Blume (blume (at) tti - c (dot) org)
1555    Date: 2003/09/08 11:55:00 CDT
1556    Tag: blume-20030908-fullpath
1557    Description:
1558    
1559    Made Win32 version of OS.FileSys.fullPath return current directory
1560    when given an empty string.  This is what the spec says, and incidentally,
1561    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1562    cases when presented with the name of a non-existing .cm file.)
1563    
1564    ----------------------------------------------------------------------
1565    Name: Matthias Blume (blume (at) tti - c (dot) org)
1566    Date: 2003/09/04 16:30:00 CDT
1567    Tag: blume-20030905-slices-etc
1568    Description:
1569    
1570      1. Changed interface to vectors and arrays in Basis to match
1571         (draft) Basis spec.
1572      2. Added signatures and implementations of slices according to
1573         Basis spec.
1574      3. Edited source code throughout the system to make it compile again
1575         under 1. and 2.  (In some cases code had to be added to have it
1576         match the new signatures.)
1577      4. MLRISC should be backward-compatible: the copies of the originals
1578         of files that needed to change under 3. were retained, the .cm files
1579         check the compiler version number and use old versions when
1580         appropriate.
1581      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1582         dirstream -> string option (in accordance with Basis spec).
1583      6. When generating code that counts lines, ml-lex used function
1584         CharVector.foldli, taking advantage of its old interface.
1585         This has been replaced with the corresponding code from
1586         CharVectorSlice.  (html-lex must be re-lexed!)
1587      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1588         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1589      8. Removed temporary additions (fromInternal, toInternal) from the
1590         (now obsolete) IntInf in smlnj-lib/Util.
1591      9. Cleaned up structure Byte.
1592     10. Added localOffset, scan, and fromString to Date (according to spec).
1593         Cleaned/corrected implementation of Date.
1594         (Still need to check for correctness; implement better canonicalizeDate.)
1595     11. Added "scan" to signature IEEE_REAL.
1596     12. Some improvements to IntInf [in particular: efficiency-hack for
1597         mod and rem when second operand is 2 (for parity checks).]
1598     13. Changed representation of type Time.time, using a single IntInf.int
1599         value counting microseconds.  This considerably simplified the
1600         implementation of structure Time.  We now support negative time
1601         values; scan and fromString handle signs.
1602     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1603         ArraySlice).
1604    
1605    ----------------------------------------------------------------------
1606    Name: Matthias Blume (blume (at) tti - c (dot) org)
1607    Date: 2003/08/28 17:00:00 CDT
1608    Tag: blume-20030828-intinf
1609    Description:
1610    
1611    This is a major update which comes with a version number bump
1612    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1613    and an implementation of IntInf in the Basis.
1614    
1615    There are a fairly large number of related changes and updates throughout
1616    the system:
1617    
1618    Basis:
1619            - Implemented IntInf.
1620            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1621            - Added some missing Real64 operations, most notably Real.toLargeInt.
1622            - Added FixedInt as a synonym for Int32.
1623    
1624    compiler:
1625            * Added support for a built-in intinf type.
1626              - literals
1627              - pattern matching
1628              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1629              - overloading on literals and operations
1630    
1631            This required adding a primitive type intinf, some additional
1632            primops, and implementations for several non-trivial intinf
1633            operations in Core.  (The intinf type is completely abstract
1634            to the compiler; all operations get delegated back to the Core.)
1635    
1636            * Intinf equality is handled by polyequal.  However, the compiler
1637            does not print its usual warning in this case (since polyequal
1638            is the right thing to do there).
1639    
1640            * Improved the organization of structure InlineT.
1641    
1642            * A word about conversion primops:
1643            If conversions involving intinf do not cancel out during
1644            CPS contract, then the compiler must insert calls to Core functions.
1645            Since all core access must be resolved already during the FLINT
1646            translate phase, it would be too late a the time of CPS contract
1647            to add new Core calls.  For this reason, conversion primops
1648            for intinf carry two arguments:  1. the numeric argument that
1649            they are supposed to convert, and 2. the Core function that
1650            can help with this conversion if necessary.  If CPS contract
1651            eliminates a primop, then the associated Core function becomes
1652            dead and goes away.  Intinf conversion primops that do not get
1653            eliminated by CPS contract get rewritten into calls of their
1654            core functions by a separate, new phase.
1655    
1656    interactive system:
1657            - Control.Print.intinfDepth controls max length of intinf constants
1658            being printed. (Analogous to Control.Print.stringDepth.)
1659            - Cleanup in printutil and pputil: got rid of unused stuff and
1660            duplicates; replaced some of the code with code that makes better
1661            use of library functionality.
1662    
1663    CM:
1664            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1665            are no longer silent.
1666    
1667    CKIT:
1668            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1669            in favor of LargeInt -- which is now the same as IntInf.
1670            CKIT-knowledgable people should check whether this is what's
1671            intended and otherwise change things back to using Int32 or
1672            FixedInt.
1673    
1674    Throughout the code:
1675            Started using IntInf.int literals and built-in operations
1676            (e.g., comparison with 0) where this seems appropriate.
1677    
1678    
1679    ----------------------------------------------------------------------
1680    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1681    Date: 2003/08/13 11:36:00 CDT
1682    Tag: dbm-20030813-mcz-merge1
1683    Description:
1684    
1685    Merging changes from the mcz-branch development branch into trunk.
1686    These changes involve replacement of the emulated old prettyprinter
1687    interface with direct use of the SML/NJ Lib PP library, and fixing
1688    of a couple of bugs (895, 1186) relating to error messages.  A new
1689    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1690    has been added.
1691    
1692    ----------------------------------------------------------------------
1693    Name: Matthias Blume (blume (at) tti - c (dot) org)
1694    Date: 2003/08/11 15:45:00 CDT
1695    Tag: blume-20030811-windows
1696    Description:
1697    
1698    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1699    
1700        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1701    
1702    This patch restores SML/NJ's ability to run under win32.  There are a
1703    number of changes, including fixes for several bugs that had gone
1704    unnoticed until now:
1705    
1706      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1707        trying to run the win32 version from within cygwin.)
1708    
1709      - There are a number of simple .bat scripts that substitute for their
1710        corresponding Unix shell-scripts.  (See below.)
1711    
1712      - The internals of ml-build have been modified slightly.  The main
1713        difference is that instead of calling ".link-sml" (or link-sml.bat)
1714        using OS.Process.system, the ML process delegates this task back
1715        to the script.  Otherwise problems arise in mixed environments such
1716        as Cygwin where scripts look and work like Unix scripts, but
1717        where OS.Process.system cannot run them.
1718    
1719      - In CM, the srcpath pickler used native pathname syntax -- which
1720        is incorrect in the case of cross-compilation.  The new pickle format
1721        is independent of platform-specific naming conventions.
1722    
1723      - Path configuration files (such as lib/pathconfig) can now choose
1724        between native and standard syntax.  Placing a line of the form
1725    
1726            standard!
1727    
1728        into the file causes all subsequent paths to be interpreted using
1729        CM standard pathname syntax (= Unix conventions); a line
1730    
1731            native!
1732    
1733        switches back to native style.  This was needed so that
1734        path config files can be written portably, see src/system/pathconfig.
1735    
1736      - Runtime system:
1737    
1738          - win32-filesys.c:  get_file_time and set_file_time now
1739            access modification time, not creation time.
1740    
1741          - I/O code made aware of new array representation.
1742    
1743          - Bug fixes in X86.prim.masm.
1744    
1745       - src/system/makeml made aware of win32. (For use under cygwin
1746            and other Unix-environments for windows.)
1747    
1748       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1749         which caused BinIO.inputAll to fail consistently.
1750    
1751    .bat scripts:
1752    
1753       Windows .bat scripts assume that SMLNJ_HOME is defined.
1754    
1755       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1756         applications (sml, ml-yacc, ml-lex).
1757       - ml-build.bat: analogous to ml-build.
1758       - config\install.bat:  Analogous to config/install.sh.  This requires
1759         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1760         (nmake etc. must be on the path, and vcvars32 must have been run.)
1761         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1762         and the bootfile hierarchy must have been unpacked under
1763         sml.boot.x86-win32.
1764         The script is very primitive and does a poor job at error checking.
1765         It only installs the base system, ml-lex, and ml-yacc.  No other
1766         libraries are being installed (i.e., you get only those that
1767         are part of the compiler.)
1768       - link-sml.bat:  analogous to .link-sml, but not currently used
1769    
1770    Unrelated bug fixes:
1771    
1772      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1773        types.
1774      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1775        otherwise.)
1776    
1777    ----------------------------------------------------------------------
1778    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1779    Date: 2003/06/17
1780    Tag: macqueen-20030617-bug895
1781    Description:
1782    
1783    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1784    Tag will be used for new development branch (mcz-branch) for
1785    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1786    
1787    ----------------------------------------------------------------------
1788    Name: Matthias Blume (blume (at) tti - c (dot) org)
1789    Date: 2003/05/27 16:55:00 CDT
1790    Tag: blume-20030527-polyeq
1791    Description:
1792    
1793    Tried to eliminated most cases of polymorphic equality.
1794    
1795    ----------------------------------------------------------------------
1796    Name: Matthias Blume (blume (at) tti - c (dot) org)
1797    Date: 2003/05/21 17:45:00 CDT
1798    Tag: blume-20030517-complete
1799    Description:
1800    
1801    Two changes:
1802    
1803      1. Added a flag for controlling whether non-exhaustive bindings will
1804         be treated as errors (default is false).
1805      2. Cleaned up the *entire* source tree so that CMB.make goes through
1806         without a single non-exhaustive match- or bind warning.
1807    
1808    ----------------------------------------------------------------------
1809    Name: Matthias Blume (blume (at) tti - c (dot) org)
1810    Date: 2003/05/17 10:20:00 CDT
1811    Tag: blume-20030517-absyn
1812    Description:
1813    
1814    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1815    
1816       This mainly affects the quality of error messages.  However, some
1817       of the code is now more straightforward than before. (Treatment of
1818       the above four constructs in translate.sml is much simpler than
1819       the "macro-expansion" that was going on before.  Plus, the mach-
1820       compiler no longer gets invoked just to be able to compile an
1821       if-expression.)
1822    
1823    2. The ErrorMsg.Error exception is now caught and absorbed by the
1824       interactive loop.
1825    
1826    ----------------------------------------------------------------------
1827    Name: Allen Leung
1828    Date: 2003/05/16 13:05:00 CDT
1829    Tag: leunga-20030516-cygwin-runtime
1830    Description:
1831    
1832        Ported the runtime system to cygwin, which uses the unix
1833    x86-unix bin files.  Missing/buggy features:
1834    
1835       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1836         the Cygwin library.
1837       o Ctrl-C handling may be flaky.
1838       o Windows system calls and Windows I/O are not supported.
1839    
1840       A new set of binfiles is located at:
1841    
1842         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1843    
1844       This is only needed for bootstrapping the cygwin version of smlnj.
1845       Other x86 versions can use the existing binfiles.
1846    
1847    ----------------------------------------------------------------------
1848    Name: Matthias Blume
1849    Date: 2003/04/08 15:42:00 CDT
1850    Tag: blume-20030408-listpair
1851    Description:
1852    
1853    1. Added a target 'mlrisc' to installer.
1854    
1855    2. Added missing elements to structure ListPair.
1856    
1857    ----------------------------------------------------------------------
1858    Name: Allen Leung
1859    Date: 2003/01/07 10:40:00 EST
1860    Tag: leunga-20030107-int-rem
1861    Description:
1862    
1863    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1864    The arguments to the SUBL instruction were swapped.
1865    
1866    ----------------------------------------------------------------------
1867    Name: Matthias Blume
1868    Date: 2002/12/12 16:25:00 EST
1869    Tag: blume-20021212-risc-ra
1870    Description:
1871    
1872    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1873    sent the RA into an infinite loop when floating point registers get
1874    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1875    110.37 and 110.38.)
1876    There was another set of potential problems related to the handling of
1877    MLRISC annotations (but those did not yet cause real problems, apparently).
1878    
1879    ----------------------------------------------------------------------
1880    Name: Matthias Blume
1881    Date: 2002/12/06 22:40:00 EST
1882    Tag: blume-20021206-cm-fileid
1883    Description:
1884    
1885    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1886    This fixes the problem of CM getting confused by files that suddenly
1887    change their identity (e.g., by getting unlinked and recreated by some
1888    text editor such as vi).  There might be a better/cheaper/cleaner way
1889    of doing this, but for now this will have to do.
1890    
1891    ----------------------------------------------------------------------
1892    Name: Matthias Blume
1893    Date: 2002/10/28 09:50:00 EST
1894    Tag: blume-20021028-typecheck
1895    Description:
1896    
1897    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1898    
1899    ----------------------------------------------------------------------
1900    Name: Matthias Blume
1901    Date: 2002/10/17 09:10:00 EDT
1902    Tag: Release_110_42
1903    Description:
1904    
1905    In good old tradition, there has been a slight hiccup so that we have
1906    to patch 110.42 after the fact.   The old release tag has been replaced
1907    (see below).
1908    
1909    The change solves a problem with two competing approaches the
1910    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1911    each other's way.
1912    
1913    This change only affects the runtime system code and the installer script.
1914    (No new bootfiles.)
1915    
1916    ----------------------------------------------------------------------
1917    Name: Matthias Blume
1918    Date: 2002/10/16 12:00:00 EDT
1919    Tag: Release_110_42_removed
1920    Description:
1921    
1922    New working release.  New bootfiles.
1923    
1924    ----------------------------------------------------------------------
1925    Name: Matthias Blume
1926    Date: 2002/10/10 13:10:00 EDT
1927    Tag: blume-20021010-ppc-divs
1928    Description:
1929    
1930    The mltree operator DIVS must be implemented with an overflow check on
1931    the PPC because the hardware indicates divide-by-zero using "overflow" as
1932    well.
1933    
1934    ----------------------------------------------------------------------
1935    Name: Matthias Blume
1936    Date: 2002/07/23 11:45:00
1937    Tag: blume-20020723-smlnj-home
1938    Description:
1939    
1940    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1941    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1942    CM_PATHCONFIG is also set), the path configuration file is assumed
1943    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1944    the entire tree to some other place and everything will "just work".
1945    
1946    (Companion commands such as ml-build and ml-makedepend also sense this
1947    variable.)
1948    
1949    ----------------------------------------------------------------------
1950    Name: Matthias Blume
1951    Date: 2002/07/12 21:19:00 EDT
1952    Tag: blume-20020712-liveness
1953    Description:
1954    
1955    Exported two useful "step" functions from liveness module (MLRISC).
1956    
1957    ----------------------------------------------------------------------
1958    Name: Matthias Blume
1959    Date: 2002/07/05 16:00 EDT
1960    Tag: Release_110_41
1961    Description:
1962    
1963    New working release.  New bootfiles.
1964    
1965    ----------------------------------------------------------------------
1966    Name: Matthias Blume
1967    Date: 2002/07/05 10:25:00 EDT
1968    Tag: blume-20020705-btimp
1969    Description:
1970    
1971    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1972    other clients can set up backtracing support.
1973    
1974    ----------------------------------------------------------------------
1975    Name: Matthias Blume
1976    Date: 2002/06/25 17:23:00 EDT
1977    Tag: blume-20020625-fpmax
1978    Description:
1979    
1980    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1981    case.  (The sense of the isNaN test was reversed -- which made min and
1982    max always return their first argument.)
1983    
1984    ----------------------------------------------------------------------
1985    Name: Matthias Blume
1986    Date: 2002/06/11
1987    Tag: blume-20020611-unixpath
1988    Description:
1989    
1990    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1991    
1992    ----------------------------------------------------------------------
1993    Name: Matthias Blume
1994    Date: 2002/06/10 16:35:00 EDT
1995    Tag: blume-20020610-ieeereal
1996    Description:
1997    
1998    I back-ported my implementation of IEEEReal.fromString from the
1999    idlbasis-devel branch so that we can test it.
2000    
2001    Another small change is that ppDec tries to give more information
2002    than just "<sig>" in the case of functors.  However, this code is
2003    broken in some mysterious way if the functor's body's signature
2004    has not been declared by ascription but gets inferred from the
2005    implementation.  This needs fixing...
2006    
2007    ----------------------------------------------------------------------
2008    Name: Matthias Blume
2009    Date: 2002/05/31
2010    Tag: blume-20020531-btrace-mode
2011    Description:
2012    
2013    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2014    the wayside when I switched over to using Controls everywhere.)
2015    
2016    ----------------------------------------------------------------------
2017    Name: Lal George
2018    Date: 2002/05/23 12:21:40 EDT
2019    Tag: george-20020523-visual-labels
2020    Description:
2021    
2022    Labels are now displayed in the graphical output to make
2023    the fall-through and target blocks obvious.
2024    
2025    ----------------------------------------------------------------------
2026    Name: Matthias Blume
2027    Date: 2002/05/22 11:03:00 EDT
2028    Tag: blume-20020522-shrink
2029    Description:
2030    
2031    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2032    situation that comes up when allocating huge arrays.
2033    
2034    ----------------------------------------------------------------------
2035    Name: Matthias Blume
2036    Date: 2002/05/21 16:00:00 EDT
2037    Tag: Release_110_40
2038    Description:
2039    
2040    New working release (110.40).  New bootfiles.
2041    
2042    [Also: John Reppy fixed GC bug 1131.]
2043    
2044    ----------------------------------------------------------------------
2045    Name: Matthias Blume
2046    Date: 2002/05/21 12:35:00 EDT
2047    Tag: blume-20020521-cmdoc
2048    Description:
2049    
2050    CM documentation update.
2051    
2052    ----------------------------------------------------------------------
2053    Name: Matthias Blume
2054    Date: 2002/05/21 10:55:00 EDT
2055    Tag: blume-20020521-misc
2056    Description:
2057    
2058    - John tweaked runtime to be silent on heap export (except when
2059      GC messages are on).
2060    - I added a few more things (cross-compiling versions of CMB) to
2061      config/preloads (as suggestions).
2062    
2063    ----------------------------------------------------------------------
2064    Name: Matthias Blume
2065    Date: 2002/05/20 22:25:00 EDT
2066    Tag: blume-20020520-controls
2067    Description:
2068    
2069    - Added ControlUtil structure to control-lib.cm.
2070    - Use it throughout.
2071    - Used Controls facility to define MLRISC controls (as opposed to
2072      registering MLRISC control ref cells with Controls after the fact)
2073    - Fixed messed-up controls priorities.
2074    
2075    * Removed again all the stuff from config/preloads that one wouldn't
2076      be able to preload at the time the initial heap image is built.
2077      (Many libraries, e.g., CML, do not exist yet at this time.  The
2078       only libraries that can be preloaded via config/preloads are those
2079       that come bundled with the bootfiles.)
2080    
2081    ----------------------------------------------------------------------
2082    Name: Matthias Blume
2083    Date: 2002/05/20 10:59:00 EDT
2084    Tag: blume-20020520-preloads
2085    Description:
2086    
2087    Added a lot of commented-out suggestions for things to be included
2088    in config/preloads.
2089    
2090    ----------------------------------------------------------------------
2091    Name: Allen Leung
2092    Date: 2002/05/18 14:20:00 EDT
2093    Tag: leunga-20020518-mdl
2094    Description:
2095    
2096    o Made the mdl tool stuff compile and run again.
2097    o I've disabled all the stuff that depends on RTL specifications; they
2098      are all badly broken anyway.
2099    
2100    ----------------------------------------------------------------------
2101    Name: Matthias Blume
2102    Date: 2002/05/17 16:49:00 EDT
2103    Tag: blume-20020517-controls
2104    Description:
2105    
2106    0. John Reppy made several modifications to the SML/NJ library.
2107       In particular, there is a shiny new controls-lib.cm.
2108    
2109    1. Pushed new controls interface through compiler so that everything
2110       compiles again.
2111    
2112    2. Added FormatComb and FORMAT_COMB to the CML version of the
2113       SML/NJ library (so that CML compiles again).
2114    
2115    3. Modified init scripts because XXX_DEFAULT environment variables
2116       are no longer with us.  (Boot-time initialization is now done
2117       using the same environment variables that are also used for
2118       startup-time initialization of controls.)
2119    
2120    ----------------------------------------------------------------------
2121    Name: Lal George
2122    Date: 2002/05/15 09:20:10 EDT
2123    Tag: george-20020515-pseudo-op-decls
2124    Description:
2125    
2126            All pseudo-ops emitted before the first segment declaration
2127    such as TEXT, DATA, and BSS directives are assumed to be global
2128    declarations and are emitted first in the assembly file. This is
2129    useful in a number of situations where one has pseudo-ops that are not
2130    specific to any segment, and also works around the constraint that one
2131    cannot have client pseudo-ops in the TEXT segment.
2132    
2133            Because no segment is associated with these declarations it is
2134    an error to allocate any space or objects before the first segment
2135    directive and an exception will be raised. However, we cannot make
2136    this check for client pseudo-ops.
2137    
2138            These top level declarations are a field in the CFG graph_info.
2139    In theory you can continue to add to this field after the CFG has been
2140    built -- provided you know what you are doing;-)
2141    
2142    ----------------------------------------------------------------------
2143    Name: Matthias Blume
2144    Date: 2002/05/13 16:40:00 EDT
2145    Tag: blume-20020513-pp-etc
2146    Description:
2147    
2148    A few minor bugfixes:
2149    
2150      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2151        (Bogus printouts for redefined bindings still occur.  Compiler
2152        bug should no longer occur now.  We need to redo the prettyprinter
2153        from scratch.)
2154    
2155      - CM pathname printer now also adds escape sequences for ( and )
2156    
2157      - commend and docu fixes for ml-nlffi
2158    
2159    ----------------------------------------------------------------------
2160    Name: Matthias Blume
2161    Date: 2002/05/10 16:40:00 EDT
2162    Tag: blume-20020510-erg-textio
2163    Description:
2164    
2165    Applied the following bugfix provided by Emden Gansner:
2166    
2167        Output is corrupted when outputSubstr is used rather than output.
2168    
2169        The problem occurs when a substring
2170    
2171            ss = (s, dataStart, dataLen)
2172    
2173        where dataStart > 0, fills a stream buffer with avail bytes left.
2174        avail bytes of s, starting at index dataStart, are copied into the
2175        buffer, the buffer is flushed, and then the remaining dataLen-avail
2176        bytes of ss are copied into the beginning of the buffer. Instead of
2177        starting this copy at index dataStart+avail in s, the current code
2178        starts the copy at index avail.
2179    
2180        Fix:
2181        In text-io-fn.sml, change line 695 from
2182             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2183        to
2184             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2185    
2186    ----------------------------------------------------------------------
2187    Name: Matthias Blume
2188    Date: 2002/04/12 13:55:00 EDT
2189    Tag: blume-20020412-assyntax
2190    Description:
2191    
2192    1. Grabbed newer assyntax.h from the XFree86 project.
2193    2. Fiddled with how to compile X86.prim.asm without warnings.
2194    3. (Very) Minor cleanup in CM.
2195    
2196    ----------------------------------------------------------------------
2197    Name: Matthias Blume
2198    Date: 2002/04/01 (no joke!) 17:07:00 EST
2199    Tag: blume-20020401-x86div
2200    Description:
2201    
2202    Added full support for div/mod/rem/quot on the x86, using the machine
2203    instruction's two results (without clumsily recomputing the remainder)
2204    directly where appropriate.
2205    
2206    Some more extensive power-of-two support was added to the x86 instruction
2207    selector (avoiding expensive divs, mods, and muls where they can be
2208    replaced with cheaper shifts and masks).  However, this sort of thing
2209    ought to be done earlier, e.g., within the CPS optimizer so that
2210    all architectures benefit from it.
2211    
2212    The compiler compiles to a fixed point, but changes might be somewhat
2213    fragile nevertheless.  Please, report any strange things that you might
2214    see wrt. div/mod/quot/rem...
2215    
2216    ----------------------------------------------------------------------
2217    Name: Matthias Blume
2218    Date: 2002/03/29 17:22:00
2219    Tag: blume-20020329-div
2220    Description:
2221    
2222    Fixed my broken div/mod logic.  Unfortunately, this means that the
2223    inline code for div/mod now has one more comparison than before.
2224    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2225    The problem was with quotient = 0, because that alone does not tell
2226    us which way the rounding went.  One then has to look at whether
2227    remainder and divisor have the same sign...  :(
2228    
2229    Anyway, I replaced the bootfiles with fresh ones...
2230    
2231    ----------------------------------------------------------------------
2232    Name: Matthias Blume
2233    Date: 2002/03/29 14:10:00 EST
2234    Tag: blume-20020329-inlprims
2235    Description:
2236    
2237    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2238    
2239    Primops have changed. This means that the bin/boot-file formats have
2240    changed as well.
2241    
2242    To make sure that there is no confusion, I made a new version.
2243    
2244    
2245    CHANGES:
2246    
2247    * removed REMT from mltree (remainder should never overflow).
2248    
2249    * added primops to deal with divisions of all flavors to the frontend
2250    
2251    * handled these primops all the way through so they map to their respective
2252      MLRISC support
2253    
2254    * used these primops in the implementation of Int, Int32, Word, Word32
2255    
2256    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2257    
2258    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2259    
2260    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2261      floating point
2262    
2263    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2264      and Real (but Real.abs maps to a separate floating-point-only primop)
2265    
2266    
2267    TODO items:
2268    
2269    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2270      instructions because the machine instruction encoder cannot handle
2271      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2272      REM in the first place, and REM is supported by the x86 machine coder.)
2273    
2274    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2275      selection phase.  (The two can be streamlined because the hardware
2276      delivers both quotient and remainder at the same time anyway.)
2277    
2278    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2279      (Currently the behavior is inconsistent both across architectures and
2280      wrt. the draft Basis spec.)
2281    
2282    * Word8 should eventually be handled natively, too.
2283    
2284    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2285      as if there currently is no execution path that could trigger it in
2286      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2287      hold for things like multiplication and division.)
2288    
2289    ----------------------------------------------------------------------
2290    Name: Matthias Blume
2291    Date: 2002/03/27 16:27:00 EST
2292    Tag: blume-20020327-mlrisc-divisions
2293    Description:
2294    
2295    Added support for all four division operations (ML's div, mod, quot,
2296    and rem) to MLRISC.  In the course of doing so, I also rationalized
2297    the naming (no more annoying switch-around of DIV and QUOT), by
2298    parameterizing the operation by div_rounding_mode (which can be either
2299    DIV_TO_ZERO or DIV_TO_NEGINF).
2300    
2301    The generic MLTreeGen functor takes care of compiling all four
2302    operations down to only round-to-zero div.
2303    
2304    Missing pieces:
2305    
2306      * Doing something smarter than relying on MLTreeGen on architectures
2307        like, e.g., the x86 where hardware division delivers both quotient and
2308        remainder at the same time.  With this, the implementation of the
2309        round-to-neginf operations could be further streamlined.
2310    
2311      * Remove inlining support for div/mod/rem from the frontend and replace it
2312        with primops that get carried through to the backend.  Do this for all
2313        int and word types.
2314    
2315    ----------------------------------------------------------------------
2316    Name: Matthias Blume
2317    Date: 2002/03/25 17:25:00 EST
2318    Tag: blume-20020325-divmod
2319    Description:
2320    
2321    I improved (hopefully without breaking them) the implementation of Int.div,
2322    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2323    advantage of the following observations:
2324    
2325      Let  q = x quot y      r = x rem y
2326           d = x div  y      m = x mod y
2327    
2328    where "quot" is the round-to-zero version of integer division that
2329    hardware usually provides.  Then we have:
2330    
2331         r = x - q * y        where neither the * nor the - will overflow
2332         d = if q >= 0 orelse x = q * y then q else q - 1
2333                              where neither the * nor the - will overflow
2334         m = if q >= 0 orelse r = 0 then r else r + y
2335                              where the + will not overflow
2336    
2337    This results in substantial simplification of the generated code.
2338    The following table shows the number of CFG nodes and edges generated
2339    for
2340            fun f (x, y) = x OPER y
2341            (* with OPER \in div, mod, quot, rem *)
2342    
2343    
2344        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2345        --------------------------------------------------------
2346         div |         24 |         39 |         12 |         16
2347         mod |         41 |         71 |         12 |         16
2348        quot |          8 |         10 |          8 |         10
2349         rem |         10 |         14 |          8 |         10
2350    
2351    
2352    ----------------------------------------------------------------------
2353    Name: Matthias Blume
2354    Date: 2002/03/25 22:06:00 EST
2355    Tag: blume-20020325-cprotobug
2356    Description:
2357    
2358    Fixed a bug in cproto (c prototype decoder).
2359    
2360    ----------------------------------------------------------------------
2361    Name: Matthias Blume
2362    Date: 2002/03/25 16:00:00 EST
2363    Tag: blume-20020325-raw-primops
2364    Description:
2365    
2366    I did some cleanup to Allen's new primop code and
2367    replaced yesterday's bootfiles with new ones.
2368    (But they are stored in the same place.)
2369    
2370    ----------------------------------------------------------------------
2371    Name: Matthias Blume
2372    Date: 2002/03/24 22:40:00 EST
2373    Tag: blume-20020324-bootfiles
2374    Description:
2375    
2376    Made the bootfiles that Allen asked for.
2377    
2378    ----------------------------------------------------------------------
2379    Name: Allen Leung
2380    Date: 2002/03/23 15:50:00 EST
2381    Tag: leunga-20020323-flint-cps-rcc-primops
2382    Description:
2383    
2384      1. Changes to FLINT primops:
2385    
2386        (* make a call to a C-function;
2387         * The primop carries C function prototype information and specifies
2388         * which of its (ML-) arguments are floating point. C prototype
2389         * information is for use by the backend, ML information is for
2390         * use by the CPS converter. *)
2391      | RAW_CCALL of { c_proto: CTypes.c_proto,
2392                       ml_args: ccall_type list,
2393                       ml_res_opt: ccall_type option,
2394                       reentrant : bool
2395                     } option
2396       (* Allocate uninitialized storage on the heap.
2397        * The record is meant to hold short-lived C objects, i.e., they
2398        * are not ML pointers.  With the tag, the representation is
2399        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2400        *)
2401      | RAW_RECORD of {tag:bool,sz:int}
2402      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2403    
2404      2.  These CPS primops are now overloaded:
2405    
2406           rawload of {kind:numkind}
2407           rawstore of {kind:numkind}
2408    
2409          The one argument form is:
2410    
2411             rawload {kind} address
2412    
2413          The two argument form is:
2414    
2415             rawload {kind} [ml object, byte-offset]
2416    
2417      3. RAW_CCALL/RCC now takes two extra arguments:
2418    
2419         a. The first is whether the C call is reentrant, i.e., whether
2420            ML state should be saved and restored.
2421         b. The second argument is a string argument specifying the name of
2422            library and the C function.
2423    
2424         These things are currently not handled in the code generator, yet.
2425    
2426      4. In CProto,
2427    
2428         An encoding type of "bool" means "ml object" and is mapped into
2429         C prototype of PTR.  Note that "bool" is different than "string",
2430         even though "string" is also mapped into PTR, because "bool"
2431         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2432    
2433      5. Pickler/unpicker
2434    
2435         Changed to handle RAW_RECORD and newest RAW_CCALL
2436    
2437      6. MLRiscGen,
2438    
2439         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2440         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2441            in the file CodeGen/cpscompile/cps-c-calls.sml
2442    
2443      7. Added the conditional move operator
2444    
2445             condmove of branch
2446    
2447         to cps.  Generation of this is still buggy so it is currently
2448         disabled.
2449    
2450    ----------------------------------------------------------------------
2451    Name: Lal George
2452    Date: 2002/03/22 14:18:25 EST
2453    Tag: george-20020322-cps-branch-prob
2454    Description:
2455    
2456    Implemented the Ball-Larus branch prediction-heuristics, and
2457    incorporated graphical viewers for control flow graphs.
2458    
2459    Ball-Larus Heuristics:
2460    ---------------------
2461    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2462    
2463    By design it uses the Dempster-Shafer theory for combining
2464    probabilities.  For example, in the function:
2465    
2466        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2467    
2468    the ball-larus heuristics predicts that the n=0 is unlikely
2469    (OH-heuristic), and the 'then' branch is unlikely because of the
2470    RH-heuristic -- giving the 'then' branch an even lower combined
2471    probability using the Dempster-Shafer theory.
2472    
2473    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2474    probability of the 'then' branch because of the loop in the else
2475    branch.
2476    
2477    
2478    Graphical Viewing:
2479    ------------------
2480    I merely plugged in Allen's graphical viewers into the compiler. The
2481    additional code is not much. At the top level, saying:
2482    
2483            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2484    
2485    will display the graphical view of the control flow graph just before
2486    back-patching.  daVinci must be in your path for this to work. If
2487    daVinci is not available, then the default viewer can be changed
2488    using:
2489    
2490            Control.MLRISC.getString "viewer"
2491    
2492    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2493    course, these viewers must be in your path.
2494    
2495    The above will display the compilation unit at the level of clusters,
2496    many of which are small, boring, and un-interesting. Also setting:
2497    
2498            Control.MLRISC.getInt "cfg-graphical-view_size"
2499    
2500    will display clusters that are larger than the value set by the above.
2501    
2502    
2503    ----------------------------------------------------------------------
2504    Name: Matthias Blume
2505    Date: 2002/03/21 22:20:00 EST
2506    Tag: blume-20020321-kmp-bugfix
2507    Description:
2508    
2509    Changed the interface to the KMP routine in PreString and fixed
2510    a minor bug in one place where it was used.
2511    
2512    ----------------------------------------------------------------------
2513    Name: Allen Leung
2514    Date: 2002/03/21 20:30:00 EST
2515    Tag: leunga-20020321-cfg
2516    Description:
2517    
2518      Fixed a potential problem in cfg edge splitting.
2519    
2520    ----------------------------------------------------------------------
2521    Name: Allen Leung
2522    Date: 2002/03/21 17:15:00 EST
2523    Tag: leunga-20020321-x86-fp-cfg
2524    Description:
2525    
2526      1. Recoded the buggy parts of x86-fp.
2527    
2528         a. All the block reordering code has been removed.
2529            We now depend on the block placement phases to do this work.
2530    
2531         b. Critical edge splitting code has been simplified and moved into the
2532            CFG modules, as where they belong.
2533    
2534         Both of these were quite buggy and complex.  The code is now much, much
2535         simpler.
2536    
2537      2. X86 backend.
2538    
2539         a. Added instructions for 64-bit support.  Instruction selection for
2540            64-bit has not been committed, however, since that
2541            requires changes to MLTREE which haven't been approved by
2542            Lal and John.
2543    
2544         b. Added support for FUCOMI and FUCOMIP when generating code for
2545            PentiumPro and above.  We only generate these instructions in
2546            the fast-fp mode.
2547    
2548         c. Added cases for JP and JNP in X86FreqProps.
2549    
2550      3. CFG
2551    
2552         CFG now has a bunch of methods for edge splitting and merging.
2553    
2554      4. Machine description.
2555    
2556         John's simplification of MLTREE_BASIS.fcond broke a few machine
2557         description things:
2558    
2559         rtl-build.{sig,sml} and hppa.mdl fixed.
2560    
2561         NOTE: the machine description stuff in the repository is still broken.
2562               Again, I can't put my fixes in because that involves
2563               changes to MLTREE.
2564    
2565    ----------------------------------------------------------------------
2566    Name: Matthias Blume
2567    Date: 2002/03/20 15:55:00 EST
2568    Tag: blume-20020320-kmp
2569    Description:
2570    
2571    Implemented Knuth-Morris-Pratt string matching in PreString and used
2572    it for String.isSubstring, Substring.isSubstring, and
2573    Substring.position.
2574    
2575    (Might need some stress-testing.  Simple examples worked fine.)
2576    
2577    ----------------------------------------------------------------------
2578    Name: Matthias Blume
2579    Date: 2002/03/19 16:37:00 EST
2580    Tag: blume-20020319-witnesses
2581    Description:
2582    
2583    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2584    
2585    This implements a generic mechanism for changing constness qualifiers
2586    anywhere within big C types without resorting to outright "casts".
2587    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2588    modify the constness at the outermost level.)
2589    The implementation of "convert" is based on the idea of "witness"
2590    values -- values that are not used by the operation but whose types
2591    "testify" to their applicability.  On the implementation side, "convert"
2592    is simply a projection (returning its second curried argument).  With
2593    cross-module inlining, it should not result in any machine code being
2594    generated.
2595    
2596    ----------------------------------------------------------------------
2597    Name: Matthias Blume
2598    Date: 2002/03/15 16:40:00 EST
2599    Tag: blume-20020315-basis
2600    Description:
2601    
2602    Provided (preliminary?) implementations for
2603    
2604      {String,Substring}.{concatWith,isSuffix,isSubstring}
2605    
2606    and
2607    
2608      Substring.full
2609    
2610    Those are in the Basis spec but they were missing in SML/NJ.
2611    
2612    ----------------------------------------------------------------------
2613    Name: Matthias Blume
2614    Date: 2002/03/14 21:30:00 EST
2615    Tag: blume-20020314-controls
2616    Description:
2617    
2618    Controls:
2619    ---------
2620    
2621    1. Factored out the recently-added Controls : CONTROLS stuff and put
2622       it into its own library $/controls-lib.cm.  The source tree for
2623       this is under src/smlnj-lib/Controls.
2624    
2625    2. Changed the names of types and functions in this interface, so they
2626       make a bit more "sense":
2627    
2628          module -> registry
2629          'a registry -> 'a group
2630    
2631    3. The interface now deals in ref cells only.  The getter/setter interface
2632       is (mostly) gone.
2633    
2634    4. Added a function that lets one register an already-existing ref cell.
2635    
2636    5. Made the corresponding modifications to the rest of the code so that
2637       everything compiles again.
2638    
2639    6. Changed the implementation of Controls.MLRISC back to something closer
2640       to the original.  In particular, this module (and therefore MLRISC)
2641       does not depend on Controls.  There now is some link-time code in
2642       int-sys.sml that registers the MLRISC controls with the Controls
2643       module.
2644    
2645    CM:
2646    ---
2647    
2648      * One can now specify the lambda-split aggressiveness in init.cmi.
2649    
2650    ----------------------------------------------------------------------
2651    Name: Allen Leung
2652    Date: 2002/03/13 17:30:00 EST
2653    Tag: leunga-20020313-x86-fp-unary
2654    Description:
2655    
2656    Bug fix for:
2657    
2658    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2659    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2660    > - fun f(x,(y,z)) = Real.~ y;
2661    > [autoloading]
2662    > [autoloading done]
2663    >       fchsl   (%eax), 184(%esp)
2664    > Error: MLRisc bug: X86MCEmitter.emitInstr
2665    >
2666    > uncaught exception Error
2667    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2668    
2669    The problem was that the code generator did not generate any fp registers
2670    in this case, and the ra didn't know that it needed to run the X86FP phase to
2671    translate the pseudo fp instruction.   This only happened with unary fp
2672    operators in certain situations.
2673    
2674    ----------------------------------------------------------------------
2675    Name: Matthias Blume
2676    Date: 2002/03/13 14:00:00 EST
2677    Tag: blume-20020313-overload-etc
2678    Description:
2679    
2680    1. Added _overload as a synonym for overload for backward compatibility.
2681       (Control.overloadKW must be true for either version to be accepted.)
2682    
2683    2. Fixed bug in install script that caused more things to be installed
2684       than what was requested in config/targets.
2685    
2686    3. Made CM aware of the (_)overload construct so that autoloading
2687       works.
2688    
2689    ----------------------------------------------------------------------
2690    Name: Matthias Blume
2691    Date: 2002/03/12 22:03:00 EST
2692    Tag: blume-20020312-url
2693    Description:
2694    
2695    Forgot to update BOOT and srcarchiveurl.
2696    
2697    ----------------------------------------------------------------------
2698    Name: Matthias Blume
2699    Date: 2002/03/12 17:30:00 EST
2700    Tag: blume-20020312-version110392
2701    Description:
2702    
2703    Yet another version number bump (because of small changes to the
2704    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2705    
2706    Changes:
2707    
2708      The new pid generation scheme described a few weeks ago was overly
2709      complicated.  I implemented a new mechanism that is simpler and
2710      provides a bit more "stability":  Once CM has seen a compilation
2711      unit, it keeps its identity constant (as long as you do not delete
2712      those crucial CM/GUID/* files).  This means that when you change
2713      an interface, compile, then go back to the old interface, and
2714      compile again, you arrive at the original pid.
2715    
2716      There now also is a mechanism that instructs CM to use the plain
2717      environment hash as a module's pid (effectively making its GUID
2718      the empty string).  For this, "noguid" must be specified as an
2719      option to the .sml file in question within its .cm file.
2720      This is most useful for code that is being generated by tools such
2721      as ml-nlffigen (because during development programmers tend to
2722      erase the tool's entire output directory tree including CM's cached
2723      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2724      revert to the old, broken behavior of SML/NJ, but in specific cases
2725      where there is no danger of interface confusion, its use is ok
2726      (I think).
2727    
2728      ml-nlffigen by default generates "noguid" annotations.  They can be
2729      turned off by specifying -guid in its command line.
2730    
2731    ----------------------------------------------------------------------
2732    Name: Lal George
2733    Date: 2002/03/12 12 14:42:36 EST
2734    Tag: george-20020312-frequency-computation
2735    Description:
2736    
2737    Integrated jump chaining and static block frequency into the
2738    compiler. More details and numbers later.
2739    
2740    ----------------------------------------------------------------------
2741    Name: Lal George
2742    Date: 2002/03/11 11 22:38:53 EST
2743    Tag: george-20020311-jump-chain-elim
2744    Description:
2745    
2746    Tested the jump chain elimination on all architectures (except the
2747    hppa).  This is on by default right now and is profitable for the
2748    alpha and x86, however, it may not be profitable for the sparc and ppc
2749    when compiling the compiler.
2750    
2751    The gc test will typically jump to a label at the end of the cluster,
2752    where there is another jump to an external cluster containing the actual
2753    code to invoke gc. This is to allow factoring of common gc invocation
2754    sequences. That is to say, we generate:
2755    
2756            f:
2757               testgc
2758               ja   L1      % jump if above to L1
2759    
2760            L1:
2761               jmp L2
2762    
2763    
2764    After jump chain elimination the 'ja L1' instructions is converted to
2765    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2766    up being implemented in their long form (if L2 is far away) using:
2767    
2768            jbe     L3      % jump if below or equal to L3
2769            jmp     L2
2770         L3:
2771            ...
2772    
2773    
2774    For large compilation units L2  may be far away.
2775    
2776    
2777    ----------------------------------------------------------------------
2778    Name: Matthias Blume
2779    Date: 2002/03/11 13:30:00 EST
2780    Tag: blume-20020311-mltreeeval
2781    Description:
2782    
2783    A functor parameter was missing.
2784    
2785    ----------------------------------------------------------------------
2786    Name: Allen Leung
2787    Date: 2002/03/11 10:30:00 EST
2788    Tag: leunga-20020311-runtime-string0
2789    Description:
2790    
2791       The representation of the empty string now points to a
2792    legal null terminated C string instead of unit.  It is now possible
2793    to convert an ML string into C string with InlineT.CharVector.getData.
2794    This compiles into one single machine instruction.
2795    
2796    ----------------------------------------------------------------------
2797    Name: Allen Leung
2798    Date: 2002/03/10 23:55:00 EST
2799    Tag: leunga-20020310-x86-call
2800    Description:
2801    
2802       Added machine generation for CALL instruction (relative displacement mode)
2803    
2804    ----------------------------------------------------------------------
2805    Name: Matthias Blume
2806    Date: 2002/03/08 16:05:00
2807    Tag: blume-20020308-entrypoints
2808    Description:
2809    
2810    Version number bumped to 110.39.1.  NEW BOOTFILES!
2811    
2812    Entrypoints: non-zero offset into a code object where execution should begin.
2813    
2814    - Added the notion of an entrypoint to CodeObj.
2815    - Added reading/writing of entrypoint info to Binfile.
2816    - Made runtime system bootloader aware of entrypoints.
2817    - Use the address of the label of the first function given to mlriscGen
2818      as the entrypoint.  This address is currently always 0, but it will
2819      not be 0 once we turn on block placement.
2820    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2821      with entry points) from mlriscGen.
2822    
2823    ----------------------------------------------------------------------
2824    Name: Allen Leung
2825    Date: 2002/03/07 20:45:00 EST
2826    Tag: leunga-20020307-x86-cmov
2827    Description:
2828    
2829       Bug fixes for CMOVcc on x86.
2830    
2831       1. Added machine code generation for CMOVcc
2832       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2833       3. CMOVcc cannot have an immediate operand as argument.
2834    
2835    ----------------------------------------------------------------------
2836    Name: Matthias Blume
2837    Date: 2002/03/07 16:15:00 EST
2838    Tag: blume-20020307-controls
2839    Description:
2840    
2841    This is a very large but mostly boring patch which makes (almost)
2842    every tuneable compiler knob (i.e., pretty much everything under
2843    Control.* plus a few other things) configurable via both the command
2844    line and environment variables in the style CM did its configuration
2845    until now.
2846    
2847    Try starting sml with '-h' (or, if you are brave, '-H')
2848    
2849    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2850    implements the underlying generic mechanism.
2851    
2852    The interface to some of the existing such facilities has changed somewhat.
2853    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2854    (The getFoo interface is still there for backward-compatibility, but its
2855    use is deprecated.)
2856    
2857    The ml-build script passes -Cxxx=yyy command-line arguments through so
2858    that one can now twiddle the compiler settings when using this "batch"
2859    compiler.
2860    
2861    TODO items:
2862    
2863    We should go through and throw out all controls that are no longer
2864    connected to anything.  Moreover, we should go through and provide
2865    meaningful (and correct!) documentation strings for those controls
2866    that still are connected.
2867    
2868    Currently, multiple calls to Controls.new are accepted (only the first
2869    has any effect).  Eventually we should make sure that every control
2870    is being made (via Controls.new) exactly once.  Future access can then
2871    be done using Controls.acc.
2872    
2873    Finally, it would probably be a good idea to use the getter-setter
2874    interface to controls rather than ref cells.  For the time being, both
2875    styles are provided by the Controls module, but getter-setter pairs are
2876    better if thread-safety is of any concern because they can be wrapped.
2877    
2878    *****************************************
2879    
2880    One bug fix: The function blockPlacement in three of the MLRISC
2881    backpatch files used to be hard-wired to one of two possibilities at
2882    link time (according to the value of the placementFlag).  But (I
2883    think) it should rather sense the flag every time.
2884    
2885    *****************************************
2886    
2887    Other assorted changes (by other people who did not supply a HISTORY entry):
2888    
2889    1. the cross-module inliner now works much better (Monnier)
2890    2. representation of weights, frequencies, and probabilities in MLRISC
2891       changed in preparation of using those for weighted block placement
2892       (Reppy, George)
2893    
2894    ----------------------------------------------------------------------
2895    Name: Lal George
2896    Date: 2002/03/07 14:44:24 EST 2002
2897    Tag: george-20020307-weighted-block-placement
2898    
2899    Tested the weighted block placement optimization on all architectures
2900    (except the hppa) using AMPL to generate the block and edge frequencies.
2901    Changes were required in the machine properties to correctly
2902    categorize trap instructions. There is an MLRISC flag
2903    "weighted-block-placement" that can be used to enable weighted block
2904    placement, but this will be ineffective without block/edge
2905    frequencies (coming soon).
2906    
2907    
2908    ----------------------------------------------------------------------
2909    Name: Lal George
2910    Date: 2002/03/05 17:24:48 EST
2911    Tag: george-20020305-linkage-cluster
2912    
2913    In order to support the block placement optimization, a new cluster
2914    is generated as the very first cluster (called the linkage cluster).
2915    It contains a single jump to the 'real' entry point for the compilation
2916    unit. Block placement has no effect on the linkage cluster itself, but
2917    all the other clusters  have full freedom in the manner in which they
2918    reorder blocks or functions.
2919    
2920    On the x86 the typical linkage code that is generated is:
2921       ----------------------
2922            .align 2
2923       L0:
2924            addl    $L1-L0, 72(%esp)
2925            jmp     L1
2926    
2927    
2928            .align  2
2929       L1:
2930       ----------------------
2931    
2932    72(%esp) is the memory location for the stdlink register. This
2933    must contain the address of the CPS function being called. In the
2934    above example, it contains the address of  L0; before
2935    calling L1 (the real entry point for the compilation unit), it
2936    must contain the address for L1, and hence
2937    
2938            addl $L1-L0, 72(%esp)
2939    
2940    I have tested this on all architectures except the hppa.The increase
2941    in code size is of course negligible
2942    
2943    ----------------------------------------------------------------------
2944    Name: Allen Leung
2945    Date: 2002/03/03 13:20:00 EST
2946    Tag: leunga-20020303-mlrisc-tools
2947    
2948      Added #[ ... ] expressions to mlrisc tools
2949    
2950    ----------------------------------------------------------------------
2951    Name: Matthias Blume
2952    Date: 2002/02/27 12:29:00 EST
2953    Tag: blume-20020227-cdebug
2954    Description:
2955    
2956    - made types in structure C and C_Debug to be equal
2957    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2958    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2959    
2960    ----------------------------------------------------------------------
2961    Name: Matthias Blume
2962    Date: 2002/02/26 12:00:00 EST
2963    Tag: blume-20020226-ffi
2964    Description:
2965    
2966    1. Fixed a minor bug in CM's "noweb" tool:
2967       If numbering is turned off, then truly don't number (i.e., do not
2968       supply the -L option to noweb).  The previous behavior was to supply
2969       -L'' -- which caused noweb to use the "default" line numbering scheme.
2970       Thanks to Chris Richards for pointing this out (and supplying the fix).
2971    
2972    2. Once again, I reworked some aspects of the FFI:
2973    
2974       A. The incomplete/complete type business:
2975    
2976       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2977         gone!
2978       - ML types representing an incomplete type are now *equal* to
2979         ML types representing their corresponding complete types (just like
2980         in C).  This is still safe because ml-nlffigen will not generate
2981         RTTI for incomplete types, nor will it generate functions that
2982         require access to such RTTI.   But when ML code generated from both
2983         incomplete and complete versions of the C type meet, the ML types
2984         are trivially interoperable.
2985    
2986         NOTE:  These changes restore the full generality of the translation
2987         (which was previously lost when I eliminated functorization)!
2988    
2989       B. Enum types:
2990    
2991       - Structure C now has a type constructor "enum" that is similar to
2992         how the "su" constructor works.  However, "enum" is not a phantom
2993         type because each "T enum" has values (and is isomorphic to
2994         MLRep.Signed.int).
2995       - There are generic access operations for enum objects (using
2996         MLRep.Signed.int).
2997       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2998         * The structure contains the definition of type "mlrep" (the ML-side
2999         representation type of the enum).  Normally, mlrep is the same
3000         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3001         then mlrep will be defined as a datatype -- thus facilitating
3002         pattern matching on mlrep values.
3003         ("-ec" will be suppressed if there are duplicate values in an
3004          enumeration.)
3005         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3006         will be generated for each C enum constant xxx.
3007         * Conversion functions m2i and i2m convert between mlrep and
3008         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
3009         * Coversion functions c and ml convert between mlrep and "tag enum".
3010         * Access functions (get/set) fetch and store mlrep values.
3011       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3012         enumerations are merged into one single enumeration represented by
3013         structure E_'.
3014    
3015    ----------------------------------------------------------------------
3016    Name: Allen Leung
3017    Date: 2002/02/25 04:45:00 EST
3018    Tag: leunga-20020225-cps-spill
3019    
3020    This is a new implementation of the CPS spill phase.
3021    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3022    In case of problems, replace it with the old file spill.sml
3023    
3024    The current compiler runs into some serious performance problems when
3025    constructing a large record.  This can happen when we try to compile a
3026    structure with many items.  Even a very simple structure like the following
3027    makes the compiler slow down.
3028    
3029        structure Foo = struct
3030           val x_1 = 0w1 : Word32.int
3031           val x_2 = 0w2 : Word32.int
3032           val x_3 = 0w3 : Word32.int
3033           ...
3034           val x_N = 0wN : Word32.int
3035        end
3036    
3037    The following table shows the compile time, from N=1000 to N=4000,
3038    with the old compiler:
3039    
3040    N
3041    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3042           MLRISC ra                               0.06u  0.00s  0.05g
3043              (spills = 0 reloads = 0)
3044           TOTAL                                   0.63u  0.07s  0.21g
3045    
3046    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3047           MLRISC ra                               5.68u  0.59s  3.93g
3048              (spills = 0 reloads = 0)
3049           TOTAL                                   14.71u  0.99s  4.81g
3050    
3051    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3052           MLRISC ra                               5.54u  0.65s  3.91g
3053              (spills = 543 reloads = 1082)
3054           TOTAL                                   65.40u  3.13s  6.00g
3055    
3056    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3057           MLRISC ra                               0.80u  0.10s  0.55g
3058              (spills = 42 reloads = 84)
3059           TOTAL                                   129.42u  5.10s  4.13g
3060    
3061    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3062           MLRISC ra                               2.69u  0.27s  1.38g
3063              (spills = 62 reloads = 124)
3064           TOTAL                                   682.48u  19.61s  13.99g
3065    
3066    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3067           MLRISC ra                               4.96u  0.27s  2.72g
3068              (spills = 85 reloads = 170)
3069           TOTAL                                   2375.26u  57.21s  48.00g
3070    
3071    As you can see the old cps spill module suffers from some serious
3072    performance problem.  But since I cannot decipher the old code fully,
3073    instead of patching the problems up, I'm reimplementing it
3074    with a different algorithm.  The new code is more modular,
3075    smaller when compiled, and substantially faster
3076    (O(n log n) time and O(n) space).  Timing of the new spill module:
3077    
3078    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3079          MLRISC ra                               0.25u  0.02s  0.15g
3080             (spills=1 reloads=3)
3081          TOTAL                                   7.74u  0.34s  1.62g
3082    
3083    Implementation details:
3084    
3085    As far as I can tell, the purpose of the CPS spill module is to make sure the
3086    number of live variables at any program point (the bandwidth)
3087    does not exceed a certain limit, which is determined by the
3088    size of the spill area.
3089    
3090    When the bandwidth is too large, we decrease the register pressure by
3091    packing live variables into spill records.  How we achieve this is
3092    completely different than what we did in the old code.
3093    
3094    First, there is something about the MLRiscGen code generator
3095    that we should be aware of:
3096    
3097    o MLRiscGen performs code motion!
3098    
3099       In particular, it will move floating point computations and
3100       address computations involving only the heap pointer to
3101       their use sites (if there is only a single use).
3102       What this means is that if we have a CPS record construction
3103       statement
3104    
3105           RECORD(k,vl,w,e)
3106    
3107       we should never count the new record address w as live if w
3108       has only one use (which is often the case).
3109    
3110       We should do something similar to floating point, but the transformation
3111       there is much more complex, so I won't deal with that.
3112    
3113    Secondly, there are now two new cps primops at our disposal:
3114    
3115     1. rawrecord of record_kind option
3116        This pure operator allocates some uninitialized storage from the heap.
3117        There are two forms:
3118    
3119         rawrecord NONE [INT n]  allocates a tagless record of length n
3120         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3121                                     and initializes the tag.
3122    
3123     2. rawupdate of cty
3124          rawupdate cty (v,i,x)
3125          Assigns to x to the ith component of record v.
3126          The storelist is not updated.
3127    
3128    We use these new primops for both spilling and increment record construction.
3129    
3130     1. Spilling.
3131    
3132        This is implemented with a linear scan algorithm (but generalized
3133        to trees).  The algorithm will create a single spill record at the
3134        beginning of the cps function and use rawupdate to spill to it,
3135        and SELECT or SELp to reload from it.  So both spills and reloads
3136        are fine-grain operations.  In contrast, in the old algorithm
3137        "spills" have to be bundled together in records.
3138    
3139        Ideally, we should sink the spill record construction to where
3140        it is needed.  We can even split the spill record into multiple ones
3141        at the places where they are needed.  But CPS is not a good
3142        representation for global code motion, so I'll keep it simple and
3143        am not attempting this.
3144    
3145     2. Incremental record construction (aka record splitting).
3146    
3147        Long records with many component values which are simulatenously live
3148        (recall that single use record addresses are not considered to
3149         be live) are constructed with rawrecord and rawupdate.
3150        We allocate space on the heap with rawrecord first, then gradually
3151        fill it in with rawupdate.  This is the technique suggested to me
3152        by Matthias.
3153    
3154        Some restrictions on when this is applicable:
3155        1. It is not a VECTOR record.  The code generator currently does not handle
3156           this case. VECTOR record uses double indirection like arrays.
3157        2. All the record component values are defined in the same "basic block"
3158           as the record constructor.  This is to prevent speculative
3159           record construction.
3160    
3161    ----------------------------------------------------------------------
3162    Name: Allen Leung
3163    Date: 2002/02/22 01:02:00 EST
3164    Tag: leunga-20020222-mlrisc-tools
3165    
3166    Minor bug fixes in the parser and rewriter
3167    
3168    ----------------------------------------------------------------------
3169    Name: Allen Leung
3170    Date: 2002/02/21 20:20:00 EST
3171    Tag: leunga-20020221-peephole
3172    
3173    Regenerated the peephole files.  Some contained typos in the specification
3174    and some didn't compile because of pretty printing bugs in the old version
3175    of 'nowhere'.
3176    
3177    ----------------------------------------------------------------------
3178    Name: Allen Leung
3179    Date: 2002/02/19 20:20:00 EST
3180    Tag: leunga-20020219-mlrisc-tools
3181    Description:
3182    
3183       Minor bug fixes to the mlrisc-tools library:
3184    
3185       1.  Fixed up parsing colon suffixed keywords
3186       2.  Added the ability to shut the error messages up
3187       3.  Reimplemented the pretty printer and fixed up/improved
3188           the pretty printing of handle and -> types.
3189       4.  Fixed up generation of literal symbols in the nowhere tool.
3190       5.  Added some SML keywords to to sml.sty
3191    
3192    ----------------------------------------------------------------------
3193    Name: Matthias Blume
3194    Date: 2002/02/19 16:20:00 EST
3195    Tag: blume-20020219-cmffi
3196    Description:
3197    
3198    A wild mix of changes, some minor, some major:
3199    
3200    * All C FFI-related libraries are now anchored under $c:
3201        $/c.cm      --> $c/c.cm
3202        $/c-int.cm  --> $c/internals/c-int.cm
3203        $/memory.cm --> $c/memory/memory.cm
3204    
3205    * "make" tool (in CM) now treats its argument pathname slightly
3206      differently:
3207        1. If the native expansion is an absolute name, then before invoking
3208           the "make" command on it, CM will apply OS.Path.mkRelative
3209           (with relativeTo = OS.FileSys.getDir()) to it.
3210        2. The argument will be passed through to subsequent phases of CM
3211           processing without "going native".  In particular, if the argument
3212           was an anchored path, then "make" will not lose track of that anchor.
3213    
3214    * Compiler backends now "know" their respective C calling conventions
3215      instead of having to be told about it by ml-nlffigen.  This relieves
3216      ml-nlffigen from one of its burdens.
3217    
3218    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3219    
3220    * Export C_DEBUG and C_Debug from $c/c.cm.
3221    
3222    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3223      subtyping relationship between incomplete pointers and their complete
3224      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3225      with the convention of instantiating 'o with ('t, 'c) obj whenever
3226      the pointer target type is complete.  In the incomplete case, 'o
3227      will be instantiated with some "'c iobj" -- a type obtained by
3228      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3229    
3230      Operations that work on both incomplete and complete pointer types are
3231      typed as taking an 'o ptr while operations that require the target to
3232      be known are typed as taking some ('t, 'c) obj ptr.
3233    
3234      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3235      where void is an eqtype without any values.  This makes it possible
3236      to work on voidptr values using functions meant to operate on light
3237      incomplete pointers.
3238    
3239    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3240      been vastly simplified.
3241    
3242    ----------------------------------------------------------------------
3243    Name: Matthias Blume
3244    Date: 2002/02/19 10:48:00 EST
3245    Tag: blume-20020219-pqfix
3246    Description:
3247    
3248    Applied Chris Okasaki's bug fix for priority queues.
3249    
3250    ----------------------------------------------------------------------
3251    Name: Matthias Blume
3252    Date: 2002/02/15 17:05:00
3253    Tag: Release_110_39
3254    Description:
3255    
3256    Last-minute retagging is becoming a tradition... :-(
3257    
3258    This is the working release 110.39.
3259    
3260    ----------------------------------------------------------------------
3261    Name: Matthias Blume
3262    Date: 2002/02/15 16:00:00 EST
3263    Tag: Release_110_39-orig
3264    Description:
3265    
3266    Working release 110.39.  New bootfiles.
3267    
3268    (Update: There was a small bug in the installer so it wouldn't work
3269    with all shells.  So I retagged. -Matthias)
3270    
3271    ----------------------------------------------------------------------
3272    Name: Matthias Blume
3273    Date: 2002/02/15 14:17:00 EST
3274    Tag: blume-20020215-showbindings
3275    Description:
3276    
3277    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3278    the latter can be useful for exploring what bindings are available at
3279    the interactive prompt.  (The first function returns only the list
3280    of symbols that are really bound, the second prints those but also the
3281    ones that CM's autoloading mechanism knows about.)
3282    
3283    ----------------------------------------------------------------------
3284    Name: Matthias Blume
3285    Date: 2002/02/15 12:08:00 EST
3286    Tag: blume-20020215-iptrs
3287    Description:
3288    
3289    Two improvements to ml-nlffigen:
3290    
3291      1. Write files only if they do not exist or if their current contents
3292         do not coincide with what's being written.  (That is, avoid messing
3293         with the time stamps unless absolutely necessary.)
3294    
3295      2. Implement a "repository" mechanism for generated files related
3296         to "incomplete pointer types".   See the README file for details.
3297    
3298    ----------------------------------------------------------------------
3299    Name: Matthias Blume
3300    Date: 2002/02/14 11:50:00 EST
3301    Tag: blume-20020214-quote
3302    Description:
3303    
3304    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3305    because of the new and improved tag generation scheme.  (Thanks to Allen
3306    Leung for pointing it out.)
3307    
3308    ----------------------------------------------------------------------
3309    Name: Lal George
3310    Date: 2002/02/14 09:55:27 EST 2002
3311    Tag: george-20020214-isabelle-bug
3312    Description:
3313    
3314    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3315    of Isabelle on the x86.
3316    
3317    From Allen:
3318    -----------
3319     I've found the problem:
3320    
3321         in ra-core.sml, I use the counter "blocked" to keep track of the
3322         true number of elements in the freeze queue.  When the counter goes
3323         to zero, I skip examining the queue.  But I've messed up the
3324         bookkeeping in combine():
3325    
3326             else ();
3327             case !ucol of
3328               PSEUDO => (if !cntv > 0 then
3329                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3330                                        ^^^^^^^^^^^^^^^^^^^^^^^
3331                      moveu := mergeMoveList(!movev, !moveu)
3332                     )
3333                  else ();
3334    
3335         combine() is called to coalesce two nodes u and v.
3336         I think I was thinking that if the move counts of u and v are both
3337         greater than zero then after they are coalesced then one node is
3338         removed from the freeze queue.  Apparently I was thinking that
3339         both u and v are of low degree, but that's clearly not necessarily true.
3340    
3341    
3342    02/12/2002:
3343        Here's the patch.  HOL now compiles.
3344    
3345        I don't know how this impact on performance (compile
3346        time or runtime).  This bug caused the RA (especially on the x86)
3347        to go thru the potential spill phase when there are still nodes on the
3348        freeze queue.
3349    
3350    
3351    
3352    
3353    ----------------------------------------------------------------------
3354    Name: Matthias Blume
3355    Date: 2002/02/13 22:40:00 EST
3356    Tag: blume-20020213-fptr-rtti
3357    Description:
3358    
3359    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3360    updates.
3361    
3362    ----------------------------------------------------------------------
3363    Name: Matthias Blume
3364    Date: 2002/02/13 16:41:00 EST
3365    Tag: blume-20020213-cmlpq
3366    Description:
3367    
3368    Added new priority queue export symbols (which have just been added to
3369    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3370    not compile and the installer would choke.)
3371    
3372    ----------------------------------------------------------------------
3373    Name: Matthias Blume
3374    Date: 2002/02/13 16:15:00 EST
3375    Tag: blume-20020213-various
3376    Description:
3377    
3378    1. More tweaks to ml-nlffigen:
3379    
3380       - better internal datastructures (resulting in slight speedup)
3381       - "-match" option requires exact match
3382       - "localized" gensym counters (untagged structs/unions nested within
3383         other structs/unions or within typedefs get a fresh counter; their
3384         tag will be prefixed by a concatenation of their parents' tags)
3385       - bug fixes (related to calculation of transitive closure of types
3386         to be included in the output)
3387    
3388    2. Minor Basis updates:
3389    
3390       - added implementations for List.collate and Option.app
3391    
3392    ----------------------------------------------------------------------
3393    Name: Matthias Blume
3394    Date: 2002/02/11 15:55:00 EST
3395    Tag: blume-20020211-gensym
3396    Description:
3397    
3398    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3399    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3400    names (ML structure names that correspond to unnamed C structs, unions,
3401    and enums), so that separate runs of ml-nlffigen do not clash.
3402    
3403    ----------------------------------------------------------------------
3404    Name: Matthias Blume
3405    Date: 2002/02/11 12:05:00 EST
3406    Tag: blume-20020211-gensml
3407    Description:
3408    
3409    A quick fix for a problem with GenSML (in the pgraph-util library):
3410    Make generation of toplevel "local" optional.  (Strictly speaking,
3411    signature definitions within "local" are not legal SML.)
3412    
3413    Other than that: updates to INSTALL and cm/TODO.
3414    
3415    ----------------------------------------------------------------------
3416    Name: Matthias Blume
3417    Date: 2002/02/08 15:00:00 EST
3418    Tag: blume-20020208-uniquepid
3419    Description:
3420    
3421    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3422    
3423    1. The installer (config/install.sh) has gotten smarter:
3424    
3425         - Configuration options are a bit easier to specify now
3426           (in config/targets).
3427         - Bug in recognizing .tar.bz2 files fixed.
3428         - Installer automatically resolves dependencies between
3429           configuration options (e.g., if you ask for eXene, you will
3430           also get cml -- regardless whether you asked for it or not).
3431         - Installer can run in "quieter mode" by setting the environment
3432           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3433           "completely silent", though.
3434         - Build HashCons library as part of smlnj-lib.
3435    
3436    2. A new scheme for assigning persistent identifiers to compilation
3437       units (and, by extension, to types etc.) has been put into place.
3438       This fixes a long-standing bug where types and even dynamic values
3439       can get internally confused, thereby compromising type safety
3440       (abstraction) and dynamic correctness.  See
3441    
3442         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3443    
3444       for an example of how things could go wrong until now.
3445    
3446       The downside of the new scheme is that pids are not quite as
3447       persistent as they used to be: CM will generate a fresh pid
3448       for every compilation unit that it thinks it sees for the first
3449       time.  That means that if you compile starting from a clean, fresh
3450       source tree at two different times, you end up with different
3451       binaries.
3452    
3453       Cutoff recompilation, however, has not been compromised because
3454       CM keeps pid information in special caches between runs.
3455    
3456    ----------------------------------------------------------------------
3457    Name: Lal George
3458    Date: 2002/02/07 15:34:13 EST 2002
3459    Tag: <none>
3460    Description:
3461    
3462    Compilers that generate assembly code may produce  global labels
3463    whose value is resolved at link time. The various peephole optimization
3464    modules did not take this in account.
3465    
3466    TODO. The Labels.addrOf function should really return an option
3467    type so that clients are forced to deal with this issue, rather
3468    than an exception being raised.
3469    
3470    ----------------------------------------------------------------------
3471    Name: Lal George
3472    Date: 2002/02/06 13:55:02 EST
3473    Tag: george-20020206-ra-breakup
3474    Description:
3475    
3476    1. A bug fix from Allen.
3477    
3478        A typo causes extra fstp %st(0)'s to be generated at compensation
3479        edges, which might cause stack underflow traps at runtime.  This
3480        occurs in fft where there are extraneous fstps right before the 'into'
3481        trap instruction (in this case they are harmless since none of the
3482        integers overflow.)
3483    
3484    2. Pulled out various utility modules that were embedded in the modules
3485       of the register allocator. I need these modules for other purposes, but
3486       they are not complete enough to put into a library (just yet).
3487    ----------------------------------------------------------------------
3488    Name: Matthias Blume
3489    Date: 2002/01/31 16:05:00 EST
3490    Tag: blume-20020131-sparc-ccalls
3491    Description:
3492    
3493    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3494       of extra stack space by mistake.  Fixed.
3495    
3496    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3497    
3498    ----------------------------------------------------------------------
3499    Name: Allen Leung
3500    Date: 2002/01/30
3501    Tag: leunga-20020130-nowhere-bug-fix
3502    Description:
3503    
3504       MLRISC bug fixes:
3505       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3506       2. MachineInt.fromString was negating its value.
3507    
3508    ----------------------------------------------------------------------
3509    Name: Matthias Blume
3510    Date: 2002/01/29
3511    Tag: blume-20020129-INSTALL
3512    Description:
3513    
3514    - Added somewhat detailed installation instructions (file INSTALL).
3515    - Fixed curl-detection bug in config/install.sh.
3516    - It is now possible to select the URL getter using the URLGETTER
3517      environment variable:
3518    
3519          not set / "unknown"      --> automatic detection (script tries wget,
3520                                       curl, and lynx)
3521          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3522                                       how to properly invoke them)
3523          other                    --> use $URLGETTER directly, it must take
3524                                       precisely two command-line arguments
3525                                       (source URL and destination file name)
3526    
3527    ----------------------------------------------------------------------
3528    Name: Matthias Blume
3529    Date: 2002/01/28
3530    Tag: blume-20020128-sparc-ccalls
3531    Description:
3532    
3533    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3534    - Make use of the allocParam argument in sparc-c-calls.
3535    
3536    ----------------------------------------------------------------------
3537    Name: Matthias Blume
3538    Date: 2002/01/28
3539    Tag: blume-20020128-allocParam
3540    Description:
3541    
3542    John Reppy:  Changes c-calls API to accept client-callback for
3543    allocating extra stack space.
3544    me: Corresponding changes to mlriscGen (using a dummy argument that
3545        does not change the current behavior).
3546    
3547    ----------------------------------------------------------------------
3548    Name: Matthias Blume
3549    Date: 2002/01/28 12:00:00
3550    Tag: Release_110_38
3551    Description:
3552    
3553    This time for real!!!
3554    
3555    ----------------------------------------------------------------------
3556    Name: Matthias Blume
3557    Date: 2002/01/28 10:56:00 EST
3558    Tag: blume-20020128-retraction
3559    Description:
3560    
3561    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3562       with blume-Release_110_38-retracted.)
3563    
3564    1. Fixed a problem with incorrect rounding modes in real64.sml.
3565       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3566    
3567    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3568       and enums fixed.  The naming of corresponding ML identifiers should
3569       now be consistent again.
3570    
3571    ----------------------------------------------------------------------
3572    Name: Allen Leung
3573    Date: 2002/01/27
3574    Tag: leunga-20020127-nowhere
3575    Description:
3576    
3577       Added a target called nowhere in the configuration scripts.
3578       Enabling this will build the MLRISC 'nowhere' tool (for translating
3579       programs with where-clauses into legal SML code) during installation.
3580    
3581    ----------------------------------------------------------------------
3582    Name: Matthias Blume
3583    Date: 2002/01/25 21:27:00 EST
3584    Tag: blume-Release_110_38-retracted
3585    Description:
3586    
3587    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3588    
3589    README will be added later.
3590    
3591    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3592    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3593    
3594    ----------------------------------------------------------------------
3595    Name: Matthias Blume
3596    Date: 2002/01/25
3597    Tag: blume-20020125-ffi
3598    Description:
3599    
3600    A large number of tweaks and improvements to ml-nlffi-lib and
3601    ml-nlffigen:
3602    
3603       - ML represenation types have been streamlined
3604       - getter and setter functions work with concrete values, not abstract
3605         ones where possible
3606       - ml-nlffigen command line more flexible (see README file there)
3607       - some bugs have been fixed (hopefully)
3608    
3609    ----------------------------------------------------------------------
3610    Name: Lal George
3611    Date: 2002/01/24
3612    Tag: george-20020124-risc-ra-interface
3613    Description:
3614    
3615       There is a dramatic simplification in the interface to the
3616       register allocator for RISC architectures as a result of making
3617       parallel copy instructions explicit.
3618    
3619    ----------------------------------------------------------------------
3620    Name: Matthias Blume
3621    Date: 2002/01/22