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

SCM Repository

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

Diff of /sml/trunk/HISTORY

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

revision 683, Wed Jul 5 00:42:19 2000 UTC revision 1737, Thu Dec 16 19:38:22 2004 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/12/16 13:35:00 CST
18    Tag: Release_110_52
19    Description:
20    
21    Last-minute changes incorporated into 110.52.  Release tag moved.
22    
23    The changes:
24       - HashString.hashString' -> HashString.hashSubstring
25       - bug fix in UnivariateStats
26    
27    ----------------------------------------------------------------------
28    Name: Matthias Blume (blume (at) tti - c (dot) org)
29    Date: 2004/12/15 23:40:00 CST
30    Tag: blume-20041215-hashSubstring
31    Description:
32    
33    - HashString.hashString' -> HashString.hashSubstring
34    - corresponding changes in atom.sml
35    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
36      added some comments
37    
38    ----------------------------------------------------------------------
39    Name: Matthias Blume (blume (at) tti - c (dot) org)
40    Date: 2004/12/15 15:30:00 CST
41    Tag: (Release_110_52)
42    Description:
43    
44    New working version (110.52).  NEW BOOTFILES!
45    
46    ----------------------------------------------------------------------
47    Name: Matthias Blume (blume (at) tti - c (dot) org)
48    Date: 2004/12/15 12:45:00 CST
49    Tag: blume-20041215-spaces
50    Description:
51    
52    More on the space problem (this time for Win32).
53    
54    ----------------------------------------------------------------------
55    Name: Matthias Blume (blume (at) tti - c (dot) org)
56    Date: 2004/12/14 17:30:00 CST
57    Tag: blume-20041214-spaces
58    Description:
59    
60    Hacked some of the scripts (in particular: the installer) to cope with
61    spaces in filenames a bit better.  But beware: the current "solution"
62    is likely still full of bugs and inherently incomplete.  (We need to
63    do away with those shell scripts for a comprehensive solution.)
64    
65    ----------------------------------------------------------------------
66    Name: Matthias Blume (blume (at) tti - c (dot) org)
67    Date: 2004/12/13 14:45:00 CST
68    Tag: blume-20041213-ml-makedepend
69    Description:
70    
71    Fixed bug in code for ml-makedepend.
72    
73    ----------------------------------------------------------------------
74    Name: Matthias Blume (blume (at) tti - c (dot) org)
75    Date: 2004/12/09 16:30:00 CST
76    Tag: blume-20041209-statistics
77    Description:
78    
79    Added two simple but potentially useful statistics modules to SML/NJ Library.
80    (See CHANGES file there.)
81    
82    ----------------------------------------------------------------------
83    Name: Matthias Blume (blume (at) tti - c (dot) org)
84    Date: 2004/12/01 16:50:00 CST
85    Tag: blume-20041201-atom
86    Description:
87    
88    smlnj-lib:
89    
90      Added function HashString.hashString' for substrings.
91      Hand-inlined CharVector.foldl into HashString (for speed).
92      Modified implementation of structure Atom to avoid extracting
93      strings from substrings unless necessary.
94      (Also see CHANGES file for smlnj-lib.)
95    
96    ----------------------------------------------------------------------
97    Name: Matthias Blume (blume (at) tti - c (dot) org)
98    Date: 2004/11/24 22:15:00 CST
99    Tag: blume-20041124-cml
100    Description:
101    
102    Made sure CML compiles when Position = Int64.
103    
104    ----------------------------------------------------------------------
105    Name: Matthias Blume (blume (at) tti - c (dot) org)
106    Date: 2004/11/24 14:30:00 CST
107    Tag: blume-20041124-position
108    Description:
109    
110    The compiler can now be compiled in a mode that makes structure Position
111    equal to Int64.  The default, however, is unchanged (Position = Int31)
112    for the time being.
113    
114    To enable 64-bit positions, use the following procedure:
115    
116    1. Start sml
117    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
118    3. Type
119          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
120    4. Run CMB.make() as usual.
121    
122    This is barely tested.  The only test so far was a little SML program
123    counting the number of characters in an 8-gigabyte file by
124    reading it character-by-character.  That test was successful.
125    
126    In support of 64-bit positions, a number of new functions have been
127    added to the runtime system.
128    
129    ----------------------------------------------------------------------
130    Name: Matthias Blume (blume (at) tti - c (dot) org)
131    Date: 2004/11/23 14:45:00 CST
132    Tag: blume-20041123-useFile
133    Description:
134    
135    Fixed a problem with unhelpful error messages related to problems with
136    .cm- or .sml files that appear as part of the sml command line.
137    
138    ----------------------------------------------------------------------
139    Name: Matthias Blume (blume (at) tti - c (dot) org)
140    Date: 2004/11/18 15:40:00 CST
141    Tag: Release_110_51
142    Description:
143    
144    New working version (110.51).  NEW BOOTFILES!
145    
146    ----------------------------------------------------------------------
147    Name: Matthias Blume (blume (at) tti - c (dot) org)
148    Date: 2004/11/18 15:35:00 CST
149    Tag: <none>
150    Description:
151    
152    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
153    Joemann).
154    
155    ----------------------------------------------------------------------
156    Name: Allen Leung (leunga (at) reservoir (dot) com)
157    Date: 2004/11/17 16:05:21 EST 2004
158    Tag: leunga-20041117-mlrisc-live-kill
159    Description:
160    
161       Added support for MLTree constructs LIVE and KILL
162    to all the architectures.
163    
164    ----------------------------------------------------------------------
165    Name: Matthias Blume (blume (at) tti - c (dot) org)
166    Date: 2004/11/13 00:20:00 CST
167    Tag: blume-20041113-versiontool
168    Description:
169    
170    - Stripped down the versiontool: It now only handles the version number.
171      The date string is generated at bootstrap time (during makeml).
172    
173    - In a previous commit, fixed a minor issue with how polyequal is being
174      translated.  In particular, the code now "looks through" abstractions.
175      This results in slightly fewer polyEqual warnings and hopefully slightly
176      more efficient code.  Important examples for where this matters are
177      the new int64 and word64 types.
178    
179    ----------------------------------------------------------------------
180    Name: Matthias Blume (blume (at) tti - c (dot) org)
181    Date: 2004/11/12 00:30:00 CST
182    Tag: blume-20041112-int64
183    Description:
184    
185    Structure Int64 fully hooked in.  (The implementation is not very
186    efficient, though.)
187    
188    ----------------------------------------------------------------------
189    Name: Matthias Blume (blume (at) tti - c (dot) org)
190    Date: 2004/11/11 17:30:00 CST
191    Tag: blume-20041111-more64
192    Description:
193    
194    All the pieces of Word64 are now there, with the exception of the
195    conversions from and to LargeWord. (Eventually these need to be identities,
196    but for the time being they don't even make sense because LargeWord is
197    32-bit wide.)
198    
199    Also started to add similar support for Int64, but major pieces of that
200    are still missing.
201    
202    ----------------------------------------------------------------------
203    Name: Matthias Blume (blume (at) tti - c (dot) org)
204    Date: 2004/11/11 00:15:00 CST
205    Tag: blume-20041111-word64
206    Description:
207    
208    Structure Word64 is now (almost) complete, word literals and patterns
209    seem to work.  There are a few odd pieces missing.  In particular,
210    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
211    at the moment.
212    
213    Making Word64 official would mean that LargeWord becomes Word64.  But
214    this requires extreme care because most word-word conversions have to
215    go through LargeWord, so making a mistake means loss of efficiency or
216    worse.  Eventually there will be a solution similar to (but actually
217    simpler than) what I did with IntInf.
218    
219    ----------------------------------------------------------------------
220    Name: Matthias Blume (blume (at) tti - c (dot) org)
221    Date: 2004/11/10 18:12:00 CST
222    Tag: blume-20041110-64bit
223    Description:
224    
225    More 64-bit hacking (but still not even half-way there yet).
226    Also, some assorted improvements to the handling of 8-bit words.
227    
228    ----------------------------------------------------------------------
229    Name: Matthias Blume (blume (at) tti - c (dot) org)
230    Date: 2004/11/09 17:50:00 CST
231    Tag: <none>
232    Description:
233    
234    Started adding some infrastructure for supporting 64-bit int- and
235    word-types.  (Still in its very early stages.)
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume (blume (at) tti - c (dot) org)
239    Date: 2004/10/28 10:45:00 CDT
240    Tag: Release_110_50
241    Description:
242    
243    New working version (110.50).  NEW BOOTFILES!
244    
245                 =====================
246    
247    Also:
248    
249      - Changend config/srcarchiveurl from a file just
250        containing the URL string into a file containing
251        shell script code.  The code has access to the $VERSION variable.
252      - Made corresponding changes to config/install.sh and config/unpack.
253      - Default contents of config/srcarchiveurl uses $VERSION and
254        normally does not have to be edited to reflect a version change.
255    
256      (As a result, a version change can be done by just editing
257       config/version, the rest is now automatic.)
258    
259    ----------------------------------------------------------------------
260    Name: Matthias Blume (blume (at) tti - c (dot) org)
261    Date: 2004/10/27 17:50:00 CDT
262    Tag: blume-20041027-btrace-msg
263    Description:
264    
265    BackTrace.monitor now also reports the source of the exception that
266    triggered the trace.
267    
268    ----------------------------------------------------------------------
269    Name: Matthias Blume (blume (at) tti - c (dot) org)
270    Date: 2004/10/27 17:20:00 CDT
271    Tag: blume-20041027-x86-c-calls
272    Description:
273    
274    This is the HISTORY entry for two earlier commits, both concerning
275    the x86 c-calls code in MLRISC:
276    
277      - added a missing LOAD in the code that deals with struct arguments
278      - made sure the caller does not add the wrong number of bytes to the
279        stack pointer after a call of a function returning a struct
280        (the callee already pops the implicit argument which points to
281         the space reserved for the result)
282    
283    ----------------------------------------------------------------------
284    Name: Allen Leung (leunga (at) reservoir (dot) com)
285    Date: 2004/10/24 14:00:00 EST
286    Tag: leunga-20041024-x86-gas-fucomip
287    Description:
288    
289       John discovered a bug in the syntax of fucomip.
290       The opcodes FU?COMIP? have been changed to
291    
292           fu?comip? %st(i), %st
293    
294    ----------------------------------------------------------------------
295    Name: Matthias Blume (blume (at) tti - c (dot) org)
296    Date: 2004/10/20 15:06:00 CDT
297    Tag: blume-20041020-standalone-backtrace
298    Description:
299    
300    Added a mechanism for getting back-trace information from standalone
301    programs.  Here is how it works:
302    
303       1. The part of the program from which you want to get backtrace
304          information (usually the whole program) should be wrapped with
305          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
306          main program could be modified from something like
307    
308             fun main (pgm, args) = ...
309    
310         to
311    
312            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
313    
314       2. To be able to access BackTrace.monitor, you have to add
315          library $smlnj-tdp/plugins.cm to the .cm file that contains your
316          main function.
317    
318       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
319          CM might have created in the past for your project).
320    
321       4. Build the system using this command line:
322    
323             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
324                      myprog.cm MyProg.main myprog
325    
326          instead of the usual
327    
328             ml-build myprog.cm MyProg.main myprog
329    
330    I changed a library name:
331    
332       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
333    
334    New libraries:
335    
336       $smlnj-tdp/back-trace.cm
337          -- when loaded causes the back-trace plugin to be installed
338       $smlnj-tdp/coverage.cm
339          -- when loaded causes the coverage plugin to be installed
340    
341    ----------------------------------------------------------------------
342    Name: Matthias Blume (blume (at) tti - c (dot) org)
343    Date: 2004/10/18 16:45:00 CDT
344    Tag: blume-20041018-groupowner
345    Description:
346    
347    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
348    
349    Eliminated group owner specs from .cm files throughout the source tree.
350    
351    ----------------------------------------------------------------------
352    Name: Matthias Blume (blume (at) tti - c (dot) org)
353    Date: 2004/10/15 15:45:00 CDT
354    Tag: blume-20041015-coverage
355    Description:
356    
357    * Test coverage tool added!
358    * Further reorganization of tracing-, debugging-, and profiling support:
359    
360        - moved original BTImp -- now called BackTrace -- into a separate
361          library called $/trace-debug-profile.cm
362        - eliminated all mentions of BTrace from SMLofNJ.Internals
363        - only the instrumentation mechanism is now left in the compiler proper
364        - BackTrace module is a plugin which is NOT plugged in by default
365        - Coverage module is another such plugin
366    
367    To get the benefits of any of these plugin modules, the code in
368    question must be compiled with tdp instrumentation turned on.  This
369    can be done by setting SMLofNJ.Internals.TDP.mode to true.
370    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
371    
372    Plugins are selected at link time.  (Pre-compiled instrumented code
373    can be re-loaded with different plugins in effect.)  When an
374    instrumented module is linked, whatever plugins are at that time
375    enabled will come into effect for that module.
376    
377    To enable the back-trace plugin, load library $/trace-debug-profile.cm
378    and invoke BackTrace.install() (e.g., from the interactive prompt).
379    To enable the coverage plugin, load the same library and invoke
380    Coverage.install().
381    
382    Back-traces are generated automatically on uncaught exceptions and
383    when the code in question explicitly invokes BackTrace.trigger().
384    
385    Coverage (and execution frequency-) information must be queried
386    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
387    
388    ----------------------------------------------------------------------
389    Name: Matthias Blume (blume (at) tti - c (dot) org)
390    Date: 2004/10/14 17:40:00 CDT
391    Tag: blume-20041014-tdp-core
392    Description:
393    
394    Snapshot of a significant overhaul of how the trace/debug/profile support
395    is hooked into the system (specifically: Core and SMLofNJ.Internals).
396    
397    ----------------------------------------------------------------------
398    Name: Matthias Blume (blume (at) tti - c (dot) org)
399    Date: 2004/10/13 16:34:00 CDT
400    Tag: blume-20041013-tdp
401    Description:
402    
403    Some rationalization of names:
404    
405       structure BTrace -> structure TDPInstrument
406       etc.
407    
408    This is is preparation of using the original back-trace
409    instrumentation for other purposes.  "TDP" stands for
410    Trace/Debug/Profile.
411    
412    The control flag controlling whether instrumentation is on or off is now
413    registered under a different name, so instead of running sml as
414    
415     sml -Cinstrument.btrace-mode=true
416    
417    one has to say
418    
419     sml -Ctdp.instrument=true
420    
421    ----------------------------------------------------------------------
422    Name: Matthias Blume (blume (at) tti - c (dot) org)
423    Date: 2004/10/11 16:37:00 CDT
424    Tag: blume-20041011-regions
425    Description:
426    
427    Made some minor modifications to elabcore.sml to have source regions
428    be propagated more tightly -- resulting in better (i.e., smaller)
429    regions being reported in error- and debug messages.
430    
431    ----------------------------------------------------------------------
432    Name: Matthias Blume (blume (at) tti - c (dot) org)
433    Date: 2004/10/08 22:50:00 CDT
434    Tag: blume-20041008-cmkw
435    Description:
436    
437    Fixed handling of keywords in .cm files: After seeing "is" the lexer
438    treats subsequent occurrences of "group", "library", "source", "is",
439    "*", and "-" as ordinary identifiers rather than keywords.
440    
441    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
442    is supposed to accept a tool argument called "source", but this did
443    not work because of the clash with the keyword.
444    
445    ----------------------------------------------------------------------
446    Name: Matthias Blume (blume (at) tti - c (dot) org)
447    Date: 2004/10/07 16:00:00 CDT
448    Tag: blume-20041007-cleanup
449    Description:
450    
451    Assorted cleanup work:
452    
453      - got rid of intstrmap in favor of using the library's
454        hash table implementation
455      - threw out most of the pathnames stuff, as it was not used anyway
456      - simplified tokentable implementation
457      - fixed some minor spelling errors
458    
459    ----------------------------------------------------------------------
460    Name: Matthias Blume (blume (at) tti - c (dot) org)
461    Date: 2004/10/06 15:15:15 CDT
462    Tag: blume-20041006-handler
463    Description:
464    
465    Cleaned up the absyn to reflect the invariant that HANDLE always
466    carries a FNexp as part of the type definition.  This eliminates some
467    superfluous sanity checks at runtime down the road.
468    
469    Some minor cleanup of the btrace code.
470    
471    ----------------------------------------------------------------------
472    Name: Matthias Blume (blume (at) tti - c (dot) org)
473    Date: 2004/10/01 10:20:30 CDT
474    Tag: blume-20041001-slave
475    Description:
476    
477    Added hack to make slave mode work in the presence of the version
478    tool.  (Still, since the master does two passes over the code for
479    CMB.make, the release number gets bumped twice when slaves are
480    attached.  I don't know if this is worth fixing...)
481    
482    ----------------------------------------------------------------------
483    Name: Matthias Blume (blume (at) tti - c (dot) org)
484    Date: 2004/09/30 10:55:00 CDT
485    Tag: blume-20040930-version
486    Description:
487    
488    * Moved the "version" magic into its own little library under
489      src/system/smlnj/internal.  This avoids expensive reconstruction of
490      a stable src/compiler/core.cm.
491    
492    * At the same time, structure CompilerVersion is now known as
493      structure SMLNJVersion.
494    
495    * Arranged for the version tool to NOT kick in when rebuilding the system
496      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
497      Also, loading the versiontool does not work when rebuilding the system
498      because CM is not properly initialized at that time.
499    
500    ----------------------------------------------------------------------
501    Name: Matthias Blume (blume (at) tti - c (dot) org)
502    Date: 2004/09/29 14:00:00 CDT
503    Tag: blume-20040929-autoversion
504    Description:
505    
506    Implemented some CM magic to have
507        file src/compiler/TopLevel/main/version.sml
508    generated automagically.
509    The version is taken from two files: config/version and config/release.
510    The first is expected to contain a two-part version number such as 110.49.
511    The second should contain a single number, but it may be missing.
512    
513    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
514    time the version tool is loaded (which is the first time you say CMB.make),
515    then the tool will increment the value stored in config/release every
516    time CMB.make is invoked.
517    
518    The binfile format is now insensitive to anything beyond the first
519    two components of a version number, so bumping the release does not render
520    binfiles incompatible.  Auto-bumping can be used to keep track of versions
521    during development without invalidating existing binfiles.
522    
523    In any case, every CMB.make updates the date information in version.sml.
524    (This is the date that is printed in the banner.)
525    
526    ----------------------------------------------------------------------
527    Name: Matthias Blume (blume (at) tti - c (dot) org)
528    Date: 2004/09/28 10:53:00 CDT
529    Tag: blume-20040928-controls
530    Description:
531    
532    Some cleanup of the controls code.
533    
534    ----------------------------------------------------------------------
535    Name: Matthias Blume (blume (at) tti - c (dot) org)
536    Date: 2004/09/27 22:08:00 CDT
537    Tag: blume-20040927-controls
538    Description:
539    
540    Added two pieces of functionality to the Controls interface:
541    
542    1.   val save'restore: 'a control -> unit -> unit
543    
544       grabs the current value of the control in stage 1 and restores it
545       in stage 2.
546    
547    2.   val set' : 'a control * 'a -> unit -> unit
548    
549       stores the given value into the control in stage 2 (i.e., delayed)
550       but does all error checking in stage 1.
551       (This is for string controls that need to do parse their argument --
552        something that might fail.  In some cases, notably in CM, one
553        already knows the intended argument but wants to delay the actual
554        assignment until a time when error recovery would be more difficult.)
555    
556    Changed the handling of controls in tool arguments to classes "sml" and
557    "lazysml":
558       - use Controls.save'restore as a more robust way of restoring the
559         old value (in particular: without having to re-parse the string)
560       - use controls to handle the "overload" keyword in the init group
561         (I believe this change actually fixes a long-standing obscure bug.)
562    
563    ----------------------------------------------------------------------
564    Name: Matthias Blume (blume (at) tti - c (dot) org)
565    Date: 2004/09/27 17:00:00 CDT
566    Tag: blume-20040927-lazysml
567    Description:
568    
569    Added a new tool class called "lazysml" to CM's tool chest.  The only
570    difference to "sml" is that compilation is done with Control.lazysml
571    set to true.  A source of class "lazysml" is automatically recognized
572    by a file name suffix of ".lml".
573    
574    In addition to the above feature, the original class "sml" now also
575    supports a tool argument "lazy" which has the same effect.  As a
576    result, the following three lines are equivalent:
577    
578        foo.sml : lazysml
579        foo.sml : sml (lazy)
580        foo.sml (lazy)
581    
582    The setting goes into effect both during parsing and during
583    compilation.  The original setting is restored right after parsing and
584    after compilation, respectively.
585    
586    In addition to all the above, there is also a general mechanism to set
587    ANY of the "controls" that are available at the command line via
588    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
589    apply as well. (In fact, "lazy" is implemented as a special case of
590    the general mechanism.)
591    
592    The .cm file syntax uses a new keyword tool argument called "with".
593    There are several ways of indicating the desired settings:
594    
595        foo.sml (with:parser.quotations=true)
596        foo.sml (with:(name:parser.quotations value:true))
597        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
598        foo.sml (with:(name1=value1 name2=value2 ...))
599        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
600    
601        etc.
602    
603    Another possible abbreviation is to leave out the =v or value:v part
604    if the name refers to a boolean control (in which case the value is
605    taken to be true).  Thus, one could get lazy sml also by saying:
606    
607        foo.sml (with:parser.lazy-keyword=true)
608        foo.sml (with:parser.lazy-keyword)
609        foo.sml (with:(name:parser.lazy-keyword value:true))
610        foo.sml (with:(name:parser.lazy-keyword))
611    
612    ----------------------------------------------------------------------
613    Name: Matthias Blume (blume (at) tti - c (dot) org)
614    Date: 2004/09/24 16:22:00 CDT
615    Tag: blume-20040924-ppc-long-branch
616    Description:
617    
618    Turned message about "emiting long form of branch" off by default.
619    Added a control flag to turn it back on when desired.
620    
621    ----------------------------------------------------------------------
622    Name: Matthias Blume (blume (at) tti - c (dot) org)
623    Date: 2004/09/24 16:05:00 CDT
624    Tag: blume-20040924-rounding
625    Description:
626    
627    Applied patch for setting rounding modes under Mac OS X.  Thanks to
628    Melissa O'Neill for providing the code!
629    
630    ----------------------------------------------------------------------
631    Name: Matthias Blume (blume (at) tti - c (dot) org)
632    Date: 2004/09/23 17:30:00 CDT
633    Tag: blume-20040923-envvars
634    Description:
635    
636    1. Changed definition of type ControlRegistry.registry_tree to
637       include control_info (i.e., the name of the controlling environment
638       variable).
639    
640    2. Added command-line flags -e and -E to print the names of environment
641       variables that can be used to control internal settings.
642       (This uses the new API mentioned in 1.)
643    
644    ----------------------------------------------------------------------
645    Name: Matthias Blume (blume (at) tti - c (dot) org)
646    Date: 2004/09/13 16:50:00 CDT
647    Tag: Release_110_49
648    Description:
649    
650    New working version (110.49).  NEW BOOTFILES!
651    
652    ----------------------------------------------------------------------
653    Name: Matthias Blume
654    Date: 2004/09/13 16:20:00 CDT
655    Tag: blume-20040913-config-mlrisc
656    Description:
657    
658    Put target "mlrisc" back into the default list.
659    (There is no harm in having it, and some users have expressed their
660    wish to have "mlrisc" included by default.)
661    
662    ----------------------------------------------------------------------
663    Name: John Reppy
664    Date: 2004/09/13
665    Tag: jhr-20040913-signals
666    Description:
667    
668    Fixed the signal masking code to properly nest mask/unmask operations
669    on a per-signal basis.
670    
671    ----------------------------------------------------------------------
672    Name: Matthias Blume (blume (at) tti - c (dot) org)
673    Date: 2004/09/08 13:20:00 CDT
674    Tag: blume-20040908-heap-magic
675    Description:
676    
677    Bumped the heap macig number to 0x09082004 to account for the changed
678    layout of the ML frame under MacOS X.
679    
680    ----------------------------------------------------------------------
681    Name: Allen Leung (leunga (at) reservoir (dot) com)
682    Date: 2004/09/03 11:26:00 EST
683    Tag: leunga-20040903-cygwin-install
684    Description:
685    
686       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
687    The Cygwin runtime is turned on by setting the environment
688    variable SMLNJ_CYGWIN_RUNTIME to 1.
689    
690    ----------------------------------------------------------------------
691    Name: Matthias Blume (blume (at) tti - c (dot) org)
692    Date: 2004/08/31 17:15:00 CDT
693    Tag: blume-20040831-core
694    Description:
695    
696    Added some exports to src/compiler/core.cm upon request by J. Joemann.
697    
698    ----------------------------------------------------------------------
699    Name: Matthias Blume (blume (at) tti - c (dot) org)
700    Date: 2004/08/30 17:55:00 CDT
701    Tag: blume-20040830-installer
702    Description:
703    
704    Upon request by Johannes Joemann:
705    
706    - improved ML code of installer to fall back to coping when renaming
707      fails (i.e., when source and target are on different file systems);
708      the code compiles but has yet to be tested in anger
709    - removed mlrisc from list of default targets (config/targets)
710    
711    ----------------------------------------------------------------------
712    Name: Matthias Blume (blume (at) tti - c (dot) org)
713    Date: 2004/08/27 17:20:00 CDT
714    Tag: blume-20040827-ptreql
715    Description:
716    
717    Added ptreql primop to structure InlineT (upon request from Larry
718    Paulson).
719    
720    ----------------------------------------------------------------------
721    Name: Allen Leung (leunga (at) reservoir (dot) com)
722    Date: 2004/08/15 21:21:00 EST
723    Tag: leunga-110_48-udgraph
724    Description:
725    
726    Another bug fix from Carl Hauser:
727    
728    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
729    > 48c48
730    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
731    > e::L)
732    > ---
733    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
734    > Without this, any deletion of an edge in an undirected graph does severe
735    > violence to the graph.
736    
737    ----------------------------------------------------------------------
738    Name: Allen Leung (leunga (at) reservoir (dot) com)
739    Date: 2004/08/10 23:35:00 EST
740    Tag: leunga-110_48-ppc
741    Description:
742    
743       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
744    
745    ----------------------------------------------------------------------
746    Name: Matthias Blume (blume (at) tti - c (dot) org)
747    Date: 2004/08/10 12:00:00 CDT
748    Tag: Release_110_48
749    Description:
750    
751    New working version (110.48).  NEW BOOTFILES!
752    
753    ----------------------------------------------------------------------
754    Name: Allen Leung (leunga (at) reservoir (dot) com)
755    Date: 2004/08/09 12:21:00 EST
756    Tag: leunga-110_47-dijsktra
757    Description:
758    
759    Bug fix from Carl Hauser:
760    
761    single_source_shortest_paths in dijkstra.sml was observed to get wrong
762    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
763    
764    The problem is that following the expression A.update(dist,s,Num.zero)
765    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
766    
767    ----------------------------------------------------------------------
768    Name: Matthias Blume (blume (at) tti - c (dot) org)
769    Date: 2004/08/06 18:10:00 CDT
770    Tag: blume-20040806-cmdline
771    Description:
772    
773    Fiddled with handling of command-line options:
774    
775      * sml now quits after processing the command line
776        if -H, -S, -h<n>, or -s<n> appears as the last
777        command-line argument
778      * a new option -q terminates the session when encountered on
779        the command line; subsequent arguments will be ignored
780      * bug fixes: short (erroneous) arguments are no longer ignored
781        completely
782    
783    ----------------------------------------------------------------------
784    Name: Allen Leung (leunga (at) reservoir (dot) com)
785    Date: 2004/08/04 18:17:00 EST
786    Tag: leunga-110_47-ppc-ibm-asm
787    Description:
788    
789      - Added minimal IBM assembly syntax support for PowerPC.
790    
791      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
792        been altered in the runtime.  We need an automatic way to keep the file
793        in sync.
794    
795    ----------------------------------------------------------------------
796    Name: Matthias Blume (blume (at) tti - c (dot) org)
797    Date: 2004/08/04 14:00:00 CDT
798    Tag: Release_110_47
799    Description:
800    
801    New working version (110.47).  NEW BOOTFILES!
802    
803    ----------------------------------------------------------------------
804    Name: Matthias Blume (blume (at) tti - c (dot) org)
805    Date: 2004/08/03 14:25:00 CDT
806    Tag: blume-20040803-callingconv
807    Description:
808    
809    Added low-level support for choosing C calling conventions by
810    twiddling the type of rawccall.  (See
811    src/compiler/Semant/types/cproto.sml for details.)
812    
813    ----------------------------------------------------------------------
814    Name: Matthias Blume (blume (at) tti - c (dot) org)
815    Date: 2004/08/02 15:55:00 CDT
816    Tag: blume-20040802-backout
817    Description:
818    
819    Backed out of change to win32-filesys.c.  The earlier patch to
820    get_file_time caused CM to produce files with the wrong time stamp.
821    
822    ----------------------------------------------------------------------
823    Name: Matthias Blume (blume (at) tti - c (dot) org)
824    Date: 2004/08/02 14:45:00 CDT
825    Tag: blume-20040802-nlffi-win32
826    Description:
827    
828    Added NLFFI support for Win32, adapted from a patch provided by David
829    Hansel.  This is currently completely untested.  Also, the issue
830    concerning stdcall vs. ccall is still unresolved.
831    
832    ----------------------------------------------------------------------
833    Name: Matthias Blume (blume (at) tti - c (dot) org)
834    Date: 2004/07/30 17:55:00 CDT
835    Tag: blume-20040730-various
836    Description:
837    
838    Gearing up towards 110.47...
839    
840    - various minor bugfixes to ml-nlffigen
841    - a beginning of a manual for nlffi
842    
843    - eliminated 'export name=value' in config/install.sh as this does
844      not work with certain versions of /bin/sh
845      (Thanks to David King at Motorola for catching this.)
846    
847    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
848      - added a test for tm==NULL to gmtime.c and localtime.c
849      - applied patch for incorrect GetFileTime under win32
850      - toSeconds -> toMilliseconds in Win32/win32-process.sml
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume (blume (at) tti - c (dot) org)
854    Date: 2004/07/21 18:20:00 CDT
855    Tag: blume-20040721-nlffigen
856    Description:
857    
858    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
859      for a typedef to an incomplete type, but leave out the "typ" member.
860      (This is just for consistency.)
861    - Started to produce what is supposed to become better (i.e., comprehensive)
862      documentation of what ml-nlffigen does and produces.
863    
864    ----------------------------------------------------------------------
865    Name: Matthias Blume (blume (at) tti - c (dot) org)
866    Date: 2004/07/14 16:25:00 CDT
867    Tag: blume-20040714-union
868    Description:
869    
870    Added C_UNION to c-calls/c-types.sml and updated the machinery
871    (ml-nlffigen, cproto.sml) that conveys C function interface
872    information to the code generator.
873    
874    However, the actual architecture-specific implementation of function
875    arguments and results that are C unions is still not implemented.
876    
877    ----------------------------------------------------------------------
878    Name: Allen Leung (leunga (at) reservoir (dot) com)
879    Date: 2004/07/14 14:38:00 EST
880    Tag: leunga-110_46_1-ppc-lwzu
881    Description:
882    
883        Added these instructions to the PowerPC architecture:
884        LBZU(X), LHZU(X), LWZU(X),
885        STWU(X), STFDU, STFSU
886    
887        etc...
888    
889        Note: I haven't added their instruction encoding into the description.
890    
891    ----------------------------------------------------------------------
892    Name: Allen Leung (leunga (at) reservoir (dot) com)
893    Date: 2004/07/13 15:04:00 EST
894    Tag: leunga-110_46_1-ppc-lwarx
895    Description:
896    
897       Added the two instructions LWARX and STWCX to the PowerPC
898    instruction set.
899    
900       A (untested) rewrite of loop-structure.sml.  The old version
901    is completely broken.
902    
903    ----------------------------------------------------------------------
904    Name: Matthias Blume (blume (at) tti - c (dot) org)
905    Date: 2004/07/13 13:50:00 CDT
906    Tag: blume-20040713-nlffi
907    Description:
908    
909    - use paramAlloc to report c-calls with too many arguments
910      (for PPC version where parameter area is pre-allocated)
911    - added ccall_maxargspace to machspec (to implement the above)
912    - made "make" commend in CM's "make" tool configurable
913    - added option (default: on) for passing the name of the SML/NJ's "bin"
914      directory to "make"; the call looks like this:
915    
916         make <options> SMLNJ_BINDIR=<dir> <target>
917    
918      This can be used by the Makefile to, e.g., pick the "right" version
919      of ml-nlffigen.
920    - minor code tweaks
921    
922    ----------------------------------------------------------------------
923    Name: Matthias Blume (blume (at) tti - c (dot) org)
924    Date: 2004/07/12 22:50:00 CDT
925    Tag: blume-110_46_1-macosx-nlffi
926    Description:
927    
928    NLFFI under Mac OS X now working (sort of).  This is largely untested,
929    though.
930    
931    Note:  1. You have to make a new, clean build of the runtime system.
932           2. There are new BOOTFILES, you have to use them!
933              (Doing the bootstrap process yourself would be *very* painful!
934               If you absolutely have to do it, build the system under
935               a different architecture and then cross-compile.)
936    
937    Version bumped to 110.46.1 to account for runtime data format changes.
938    
939    ----------------------------------------------------------------------
940    Name: Matthias Blume (blume (at) tti - c (dot) org)
941    Date: 2004/06/18 14:30:00 CDT
942    Tag: blume-20040618-unix
943    Description:
944    
945    Changed the implementation of structure Unix so that the same stream
946    is returned every time one of the {text,bin}{In,Out}streamOf functions
947    is invoked on the same proc.  This is not what the spec currently
948    says -- although IMO it arguably should.  (See discussion below.)
949    
950    ----------------------------------------------------------------------
951    Name: Matthias Blume (blume (at) tti - c (dot) org)
952    Date: 2004/06/17 18:15:00 CDT
953    Tag: Release_110_46
954    Description:
955    
956    New working version (110.46).  NEW BOOTFILES!
957    
958    ----------------------------------------------------------------------
959    Name: Matthias Blume (blume (at) tti - c (dot) org)
960    Date: 2004/06/17 17:20:00 CDT
961    Tag: blume-20040617-timer-unix
962    Description:
963    
964    Changed the interface of structures Timer and Unix to match the most
965    recent Basis spec.
966    
967    In the case of Unix there still seems to be an open/weird issue:
968    
969       The {text,bin}{In,Out}streamOf functions are supposed to create
970       fresh streams whenever they are called -- as opposed to have them
971       return the same stream every time.  This design is supposed to
972       prevent space leaks caused by proc values hanging on to streams.
973    
974       The reap function, on the other hand, is supposed to close the
975       streams.  This cannot be done without having a handle on the
976       stream in proc after all...
977    
978    I took the liberty to implement the following stopgap solution:
979    
980      The proc value hangs on to the most recently created stream(s).
981      Reap closes those.  If either or both of the two streams hadn't
982      been created at all yet, then reap will close the corresponding
983      file descriptors directly.
984    
985    PS: I don't understand the original space leak argument anymore.  If
986    a proc hangs on to the imperative stream, then I/O operations on those
987    will advance the state of the cached stream and avoid the space leak.
988    
989    ----------------------------------------------------------------------
990    Name: Matthias Blume (blume (at) tti - c (dot) org)
991    Date: 2004/05/28 16:45:00 CDT
992    Tag: blume-20040528-basis
993    Description:
994    
995    Added signature PACK_REAL and exported functor PrimIO.
996    
997    ----------------------------------------------------------------------
998    Name: Matthias Blume (blume (at) tti - c (dot) org)
999    Date: 2004/05/25 16:00:00 CDT
1000    Tag: blume-20040525-group-owner
1001    Description:
1002    
1003    CM now ignores (but still accepts) the "owner" information in group
1004    descriptions.  The owner of a group is its next enclosing
1005    library. Each group must have a unique owner.  (There is a virtual
1006    "toplevel" library that owns groups which are not nested within a real
1007    library.)  Previously, each group had to explicitly declare its owner,
1008    and CM would check that such a declaration is correct.  The new scheme
1009    is to have CM check that for each group there is precisely one owning
1010    library.
1011    
1012    The advantage of the new scheme is that the programmer no longer needs
1013    to maintain the somewhat annoying owner information.  The downside is
1014    that CM cannot enforce the ownership rule across multiple runs of
1015    CM.make.  Fortunately, enclosing the same group in two different
1016    libraries A and B which are not part of the same program does not
1017    cause real problems.
1018    
1019    ----------------------------------------------------------------------
1020    Name: Matthias Blume (blume (at) tti - c (dot) org)
1021    Date: 2004/05/20 16:00:00 CDT
1022    Tag: blume-20040520-win32
1023    Description:
1024    
1025    Made the win32 version work again.  (Strangely, a misplaced comma had
1026    slipped into win32-process.c which prevented the runtime from being
1027    compiled correctly.)
1028    
1029    Also, included a minor addition to ml-build.bat analogous to what was
1030    done in blume-20040519-ml-build.
1031    
1032    ----------------------------------------------------------------------
1033    Name: Matthias Blume (blume (at) tti - c (dot) org)
1034    Date: 2004/05/19 22:10:00 CDT
1035    Tag: blume-20040519-ml-build
1036    Description:
1037    
1038    Arranged for ml-build to clean up after itself a little bit better.
1039    The script generates a temporary SML source file and compiles it using
1040    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1041    gets rid of those at the end, so they don't accumulate under .cm.
1042    
1043    This required a minor change to install.sh because the name of the
1044    metadata directory (default: .cm) is actually configurable at
1045    installation time.
1046    
1047    ----------------------------------------------------------------------
1048    Name: Matthias Blume (blume (at) tti - c (dot) org)
1049    Date: 2004/05/18 15:50:00 CDT
1050    Tag: blume-20040518-mkreader
1051    Description:
1052    
1053    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1054    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1055    
1056    ----------------------------------------------------------------------
1057    Name: Matthias Blume (blume (at) tti - c (dot) org)
1058    Date: 2004/05/11 14:35:00 CDT
1059    Tag: blume-20040511-win32sock
1060    Description:
1061    
1062    Added previously missing support for many socket-related functions
1063    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1064    for the voluminous patch!
1065    
1066    (I have not tested this patch under win32 yet.)
1067    
1068    Here is David's e-mail:
1069    
1070    Hi,
1071    
1072    Attached to this email you find a diff against sml/nj 110.45
1073    that will enable socket support under Windows.
1074    
1075    To apply the patch (using unix or cygwin)
1076    1) gunzip runtime.diff.gz
1077    2) "cd" into "src/runtime" in the source tree of a fresh
1078       110.45 installation.
1079    3) patch -p 1 < [your/path/to]runtime.diff
1080    
1081    The code compiles fine but has NOT yet been extensively tested.
1082    I only ran a few tests for basic socket client functionality
1083    (which worked fine).  Especially the functions that use ioctl
1084    are not tested at all and might not work (see below).
1085    
1086    I implemented this since we want to move to a newer version of sml/nj
1087    but need socket support in order to use it.  This is the first time I
1088    even had a look at the sml/nj source,  so please review my changes
1089    before making this part of the distribution!  Here are a few issues
1090    that I think might be better for someone to solve who is more
1091    familiar with the sml/nj source (and socket programming):
1092    
1093    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1094      exception since I could not figure out what the windows equivalent
1095      of these functions is
1096    
1097    - In sockets-osdep.h there are a some #include statements that are
1098      only used in a few files that include sockets-osdep.h
1099    
1100    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1101      does not process its return value since I don't know how to
1102      report an error upwards.
1103    
1104    - It would probably be good to have a call to WSACleanup() when
1105      the library is unloaded (if there is such a possibility).
1106      Otherwise I think Windows will take care of this automatically
1107      when the process finishes.
1108    
1109    - I used ioctlsocket() as a replacement for ioctl() but I have
1110      no idea if that is actually the proper replacement on Windows.
1111    
1112    - All these issues are marked in the code by "FIXME" comments.
1113    
1114    We use sml/nj extensively in our products and are quite happy
1115    with it.  I hope this contribution will help you.
1116    
1117    Keep up the good work!
1118    
1119    David
1120    
1121    ----------------------------------------------------------------------
1122    Name: Matthias Blume (blume (at) tti - c (dot) org)
1123    Date: 2004/05/11 14:20:00 CDT
1124    Tag: blume-20040511-installml
1125    Description:
1126    
1127    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1128    the patch.)
1129    
1130    ----------------------------------------------------------------------
1131    Name: Matthias Blume (blume (at) tti - c (dot) org)
1132    Date: 2004/05/11 14:05:00 CDT
1133    Tag: blume-20040511-nlffi-netbsd
1134    Description:
1135    
1136    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1137    the patch.)
1138    
1139    ----------------------------------------------------------------------
1140    Name: Matthias Blume (blume (at) tti - c (dot) org)
1141    Date: 2004/05/11 12:05:00 CDT
1142    Tag: blume-20040511-exports
1143    Description:
1144    
1145    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1146    export lists in compiler-related .cm-files.
1147    
1148    ----------------------------------------------------------------------
1149    Name: Matthias Blume (blume (at) tti - c (dot) org)
1150    Date: 2004/05/11 11:35:00 CDT
1151    Tag: blume-20040511-allsource
1152    Description:
1153    
1154    The installer now honors the "src-smlnj" target again, although its meaning
1155    has changed from "all sources required for the compiler" to "all sources
1156    the installer knows about".  In other words, if you enable "src-smlnj"
1157    in the "targets" file, then the installer will pull in sources for
1158    everything.  (Notice that this refers to source code only.  Compiled
1159    code is still only installed for modules that were requested explicitly
1160    or which are required for other modules that were requested explicitly.)
1161    
1162    ----------------------------------------------------------------------
1163    Name: Matthias Blume (blume (at) tti - c (dot) org)
1164    Date: 2004/04/23 17:40:00 CDT
1165    Tag: blume-20040423-ieee-scan
1166    Description:
1167    
1168    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1169    in the exponent calculation we get INF or ZERO (depending on the mantissa
1170    and the sign of the exponent).
1171    
1172    ----------------------------------------------------------------------
1173    Name: Matthias Blume (blume (at) tti - c (dot) org)
1174    Date: 2004/04/23 10:40:00 CDT
1175    Tag: blume-20040423-ml-build
1176    Description:
1177    
1178    The ml-build script now terminates with a non-0 status when something
1179    goes wrong.
1180    
1181    ----------------------------------------------------------------------
1182    Name: Matthias Blume (blume (at) tti - c (dot) org)
1183    Date: 2004/04/22 16:35:00 CDT
1184    Tag: blume-20040422-Option
1185    Description:
1186    
1187    Made exception Option to be the same as exception Option.Option
1188    (as it should be).
1189    
1190    ----------------------------------------------------------------------
1191    Name: Allen Leung (leunga (at) reservoir (dot) com)
1192    Date: 2004/03/19 14:40:00 EST
1193    Tag: leunga-20040319-cygwin-nlffi
1194    Description:
1195    
1196    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1197    of SML/NJ.  The problem is that
1198    
1199       lib = dlopen(NULL, ...)
1200       f   = dlsym(lib, "malloc");
1201    
1202    does not work on Windows unless we explicitly export symbols
1203    such as 'malloc' during linking.  We fixed this by explicitly
1204    exporting the required symbols with the magic gcc incantation:
1205    
1206        -Wl,--export-all cygwin.def
1207    
1208    where cygwin.def is a file containing all the symbols that
1209    we wish to export.
1210    
1211    I suspect this is a Windows problem and we'll have to
1212    do the same (somehow with windows compilers) when
1213    we build the native win32 version with the system
1214    calls LoadLibrary/GetProcAddress.
1215    
1216    ----------------------------------------------------------------------
1217    Name: Matthias Blume (blume (at) tti - c (dot) org)
1218    Date: 2004/03/04 16:35:00 CST
1219    Tag: blume-20040304-intinf-fmt
1220    Description:
1221    
1222    Fixed problem with IntInf.fmt (sign would show up on the right instead
1223    of on the left for BIN, OCT, and HEX).
1224    
1225    ----------------------------------------------------------------------
1226    Name: Matthias Blume (blume (at) tti - c (dot) org)
1227    Date: 2004/03/04 11:25:00 CST
1228    Tag: blume-20040304-symlinks
1229    Description:
1230    
1231    Fixed problem with installer script (unix only) where bin/ml-yacc and
1232    friends pointed (via symlinks) to absolute locations instead of just
1233    .run-sml.  This was reported by Vesa A Norrman.
1234    
1235    ----------------------------------------------------------------------
1236    Name: Matthias Blume (blume (at) tti - c (dot) org)
1237    Date: 2004/02/13 14:50:00 CST
1238    Tag: Release_110_45
1239    Description:
1240    
1241    New working version (110.45). New bootfiles.
1242    
1243    ----------------------------------------------------------------------
1244    Name: Matthias Blume (blume (at) tti - c (dot) org)
1245    Date: 2004/01/26 15:15:15 CST
1246    Tag: blume-20040126-toplevel
1247    Description:
1248    
1249    Improved handling of exceptions at the interactive toplevel.
1250    
1251    ----------------------------------------------------------------------
1252    Name: Matthias Blume (blume (at) tti - c (dot) org)
1253    Date: 2004/01/26 11:25:00
1254    Tag: blume-20040126-app
1255    Description:
1256    
1257    Type of top-level "app" corrected.
1258    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1259    
1260    ----------------------------------------------------------------------
1261    Name: Matthias Blume (blume (at) tti - c (dot) org)
1262    Date: 2003/11/18 17:10 CST
1263    Tag: blume-20031118-basis-fiddle
1264    Description:
1265    
1266    - changed Timer interface to what might become the spec
1267    - POSIX_FLAGS -> BIT_FLAGS according to spec
1268    - some other minor discrepancies wrt. spec eliminated
1269    
1270    ----------------------------------------------------------------------
1271    Name: Matthias Blume (blume (at) tti - c (dot) org)
1272    Date: 2003/11/06 12:00:00 CST
1273    Tag: Release_110_44
1274    Description:
1275    
1276    New working version (110.44). New bootfiles.
1277    
1278    ----------------------------------------------------------------------
1279    Name: Matthias Blume (blume (at) tti - c (dot) org)
1280    Date: 2003/11/04 11:50:00 CST
1281    Tag: blume-20031104-move-libraries
1282    Description:
1283    
1284    Eliminated the "dont_move_libraries" directive in config/targets.
1285    (The mechanism was broken and could not be fixed easily.  Moreover,
1286    there does not seem to be any reason not to move all libraries into
1287    lib during installation.  I originally implemented this directive as a
1288    backward-compatibility feature when I first introduced the new CM.
1289    Now that things have been stable for a long time and going back to the
1290    old CM is not an option, there is no reason to keep it around.)
1291    
1292    ----------------------------------------------------------------------
1293    Name: Matthias Blume (blume (at) tti - c (dot) org)
1294    Date: 2003/11/03 16:00:00 CST
1295    Tag: blume-20031103-installdir
1296    Description:
1297    
1298    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1299    Richards for pointing out the problem and providing the solution.)
1300    
1301    ----------------------------------------------------------------------
1302    Name: Matthias Blume (blume (at) tti - c (dot) org)
1303    Date: 2003/10/01 17:05:00 CDT
1304    Tag: blume-20031001-lal-mlrisc
1305    Description:
1306    
1307    MLRISC bug fix from Lal.
1308    
1309    ----------------------------------------------------------------------
1310    Name: Matthias Blume (blume (at) tti - c (dot) org)
1311    Date: 2003/09/30 16:10:00 CDT
1312    Tag: blume-20030930-primio-bat
1313    Description:
1314    
1315    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1316    2. Improved .bat files (for Win32 port) to make things work under Win95.
1317       (thanks to Aaron S. Hawley for this one)
1318    
1319    ----------------------------------------------------------------------
1320    Name: Matthias Blume (blume (at) tti - c (dot) org)
1321    Date: 2003/09/26 16:05:00 CDT
1322    Tag: blume-20030926-wrappriv
1323    Description:
1324    
1325    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1326    
1327    ----------------------------------------------------------------------
1328    Name: Matthias Blume (blume (at) tti - c (dot) org)
1329    Date: 2003/09/26 15:00:00 CDT
1330    Tag: blume-20030926-110_43_3
1331    Description:
1332    
1333    - additional cleanup
1334    - version number bump, NEW BOOTFILES
1335    
1336    ----------------------------------------------------------------------
1337    Name: Matthias Blume (blume (at) tti - c (dot) org)
1338    Date: 2003/09/26 12:00:00 CDT
1339    Tag: blume-20030926-ppautoload
1340    Description:
1341    
1342    I modified the read-eval-print loop so that the autoloader gets
1343    invoked whenever the prettyprinter tries to look up a symbol that
1344    is not currently defined in the toplevel environment but which
1345    appears in CM's autoload registry.  As a result, we see far fewer of
1346    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1347    
1348    In addition to this I tried to clean up some pieces of the Basis
1349    implementation (e.g., Socket, Word8Array) in order to prevent other
1350    instances of these ?.Foo.Bar.xxx names from being printed.
1351    
1352    The mechanism that picks names for types still needs some work, though.
1353    (Right now it seems that if there is a type A.t which is defined to
1354    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1355    "?.B.u" although the perhaps more sensible solution would be to use
1356    "A.t" in this case.  In other words, the prettyprinter should follow
1357    a chain of DEFtycs not farther than there are corresponding toplevel
1358    names in the current environment.)
1359    
1360    ----------------------------------------------------------------------
1361    Name: Matthias Blume (blume (at) tti - c (dot) org)
1362    Date: 2003/09/24 16:31:00 CDT
1363    Tag: blume-20030924-installer
1364    Description:
1365    
1366    Another installer tweak:  All the ML code for the installer is now
1367    compiled during CMB.make and put into a little library called
1368    $smlnj/installer.cm.  The installation then simply invokes
1369    
1370       sml -m $smlnj/installer.cm
1371    
1372    and everything happens automagically.
1373    
1374    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1375    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1376            and BINDIR.
1377    
1378    The new scheme guarantees that the ML code responsible for the installation
1379    is in sync with the APIs of the main system.  Also, the installer is
1380    somewhat faster because the installer script is precompiled.
1381    
1382    ----------------------------------------------------------------------
1383    Name: Matthias Blume (blume (at) tti - c (dot) org)
1384    Date: 2003/09/24 15:35:00 CDT
1385    Tag: blume-20030924-synsock
1386    Description:
1387    
1388    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1389    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1390    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1391    implementations of non-blocking operations from CML's Socket
1392    structure.
1393    
1394    ----------------------------------------------------------------------
1395    Name: Matthias Blume (blume (at) tti - c (dot) org)
1396    Date: 2003/09/24 15:10:05 CDT
1397    Tag: blume-20030924-sockets
1398    Description:
1399    
1400    1. Fixed SOCKET API and implementation to match Basis spec.
1401       This required changing the internal representation of sockets to one
1402       that remembers (for each socket file descriptor) whether it is currently
1403       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1404       call is made only if the state actually needs to change).
1405    
1406    2. OS-specific details of sockets were moved into separate files, thus
1407       making it possible to unify the bulk of the socket implementations
1408       between Unix and Win32.
1409    
1410    3. CML's socket API changed accordingly.
1411       (Note that we need to remove non-blocking functions from this API
1412       since they are redundant in the case of CML!)
1413    
1414    4. CML's socket implementation now makes use of non-blocking functions
1415       provided by Basis, thus removing all OS-dependent code from this part
1416       of CML.
1417    
1418    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1419    
1420    ----------------------------------------------------------------------
1421    Name: Matthias Blume (blume (at) tti - c (dot) org)
1422    Date: 2003/09/22 12:10:00 CDT
1423    Tag: blume-20030922-110_43_2
1424    Description:
1425    
1426    Made a new interim version and bootfiles for developer's bootstrapping
1427    convenience.
1428    
1429       110.43.2 -- NEW BOOTFILES
1430    
1431    ----------------------------------------------------------------------
1432    Name: Matthias Blume (blume (at) tti - c (dot) org)
1433    Date: 2003/09/19 15:55:00 CDT
1434    Tag: blume-20030919-cmdir
1435    Description:
1436    
1437    1. new-install.sh -> install.sh
1438    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1439    3. tweaked installer so that another name instead of .cm can be chosen
1440       at install time (by setting the CM_DIR_ARC environment variable
1441       during installation); once installation is complete, the name is
1442       fixed
1443    
1444    ----------------------------------------------------------------------
1445    Name: Matthias Blume (blume (at) tti - c (dot) org)
1446    Date: 2003/09/18 16:00:00 CDT
1447    Tag: blume-20030918-110_43_1
1448    Description:
1449    
1450    Made a new interim version and bootfiles for developer's bootstrapping
1451    convenience.
1452    
1453       110.43.1 -- NEW BOOTFILES
1454    
1455    ----------------------------------------------------------------------
1456    Name: Matthias Blume (blume (at) tti - c (dot) org)
1457    Date: 2003/09/18 15:20:00 CDT
1458    Tag: blume-20030918-misc
1459    Description:
1460    
1461    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1462       this seems to be controversial at the moment) and used those in
1463       Posix.ProcEnv.times.
1464    
1465    2. Added Time.{from,to}Nanoseconds to Time.
1466    
1467    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1468       For example, fromLargeInt never needs to look at more than 3 "big
1469       digits" to get its 53 bits of precision.
1470    
1471    ----------------------------------------------------------------------
1472    Name: Matthias Blume (blume (at) tti - c (dot) org)
1473    Date: 2003/09/17 16:30:00 CDT
1474    Tag: blume-20030917-real32-slices
1475    Description:
1476    
1477    Added an entry to the primitive environment
1478    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1479    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1480    
1481    Removed some of the "magic" constants in real64.sml and replaced them
1482    with code that generates these values from their corresponding
1483    integer counterparts.
1484    
1485    Made all(?) the slice-related changes to the Basis and made everything
1486    compile again...
1487    
1488    ----------------------------------------------------------------------
1489    Name: Matthias Blume (blume (at) tti - c (dot) org)
1490    Date: 2003/09/15 17:45:00 CDT
1491    Tag: blume-20030915-rbase
1492    Description:
1493    
1494    Fixed bug in Real.fromLargeInt.
1495    
1496    ----------------------------------------------------------------------
1497    Name: Matthias Blume (blume (at) tti - c (dot) org)
1498    Date: 2003/09/13 18:11:00 CDT
1499    Tag: blume-20030913-libinstall
1500    Description:
1501    
1502    Minor bugfix in config/libinstall (set anchor with path to
1503    standalone tool after installing it, otherwise libraries that
1504    need ml-lex or ml-yacc won't compile the first time the installer
1505    runs).
1506    
1507    ----------------------------------------------------------------------
1508    Name: Matthias Blume (blume (at) tti - c (dot) org)
1509    Date: 2003/09/12 11:45:00 CDT
1510    Tag: blume-20030912-various
1511    Description:
1512    
1513     - fixed bug in Real.toLargeInt
1514     - fixed bug in Posix.ProcEnv.times
1515     - changed inputLine functions to return an option
1516     - minor installer improvements / bugfixes
1517     - changed default @SMLalloc parameter for x86/celeron to 64k
1518    
1519    ----------------------------------------------------------------------
1520    Name: Matthias Blume (blume (at) tti - c (dot) org)
1521    Date: 2003/09/09 22:00:00 CDT
1522    Tag: Release_110_43
1523    Description:
1524    
1525        New working release 110.43.  New bootfiles.
1526    
1527    ----------------------------------------------------------------------
1528    Name: Matthias Blume (blume (at) tti - c (dot) org)
1529    Date: 2003/09/09 19:20:00 CDT
1530    Tag: blume-20030909-installer
1531    Description:
1532    
1533    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1534    Modified config/install.bat to take advantage of it.  Also modified
1535    config/install.sh (and called it config/new-install.sh) to take advantage
1536    of it on Unix systems.  (The SML code is (supposed to be) platform-
1537    independent.)
1538    
1539    The installer can now install everything under Win32
1540    as well as under *nix as long as it compiles.
1541    
1542    Other changes:
1543    
1544      - made CML compile again under Win32
1545      - made eXene compile under Win32 (by providing a fake structure UnixSock
1546        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1547      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1548        same as type int;  under Win32 it isn't
1549      - fixed some slice-related problems in the win32-specific parts of CML
1550      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1551        (under Win32, the volume name is case-insensitive, and the
1552        OS.Path code compares volume names for equality)
1553    
1554    ----------------------------------------------------------------------
1555    Name: Matthias Blume (blume (at) tti - c (dot) org)
1556    Date: 2003/09/08 11:55:00 CDT
1557    Tag: blume-20030908-fullpath
1558    Description:
1559    
1560    Made Win32 version of OS.FileSys.fullPath return current directory
1561    when given an empty string.  This is what the spec says, and incidentally,
1562    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1563    cases when presented with the name of a non-existing .cm file.)
1564    
1565    ----------------------------------------------------------------------
1566    Name: Matthias Blume (blume (at) tti - c (dot) org)
1567    Date: 2003/09/04 16:30:00 CDT
1568    Tag: blume-20030905-slices-etc
1569    Description:
1570    
1571      1. Changed interface to vectors and arrays in Basis to match
1572         (draft) Basis spec.
1573      2. Added signatures and implementations of slices according to
1574         Basis spec.
1575      3. Edited source code throughout the system to make it compile again
1576         under 1. and 2.  (In some cases code had to be added to have it
1577         match the new signatures.)
1578      4. MLRISC should be backward-compatible: the copies of the originals
1579         of files that needed to change under 3. were retained, the .cm files
1580         check the compiler version number and use old versions when
1581         appropriate.
1582      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1583         dirstream -> string option (in accordance with Basis spec).
1584      6. When generating code that counts lines, ml-lex used function
1585         CharVector.foldli, taking advantage of its old interface.
1586         This has been replaced with the corresponding code from
1587         CharVectorSlice.  (html-lex must be re-lexed!)
1588      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1589         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1590      8. Removed temporary additions (fromInternal, toInternal) from the
1591         (now obsolete) IntInf in smlnj-lib/Util.
1592      9. Cleaned up structure Byte.
1593     10. Added localOffset, scan, and fromString to Date (according to spec).
1594         Cleaned/corrected implementation of Date.
1595         (Still need to check for correctness; implement better canonicalizeDate.)
1596     11. Added "scan" to signature IEEE_REAL.
1597     12. Some improvements to IntInf [in particular: efficiency-hack for
1598         mod and rem when second operand is 2 (for parity checks).]
1599     13. Changed representation of type Time.time, using a single IntInf.int
1600         value counting microseconds.  This considerably simplified the
1601         implementation of structure Time.  We now support negative time
1602         values; scan and fromString handle signs.
1603     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1604         ArraySlice).
1605    
1606    ----------------------------------------------------------------------
1607    Name: Matthias Blume (blume (at) tti - c (dot) org)
1608    Date: 2003/08/28 17:00:00 CDT
1609    Tag: blume-20030828-intinf
1610    Description:
1611    
1612    This is a major update which comes with a version number bump
1613    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1614    and an implementation of IntInf in the Basis.
1615    
1616    There are a fairly large number of related changes and updates throughout
1617    the system:
1618    
1619    Basis:
1620            - Implemented IntInf.
1621            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1622            - Added some missing Real64 operations, most notably Real.toLargeInt.
1623            - Added FixedInt as a synonym for Int32.
1624    
1625    compiler:
1626            * Added support for a built-in intinf type.
1627              - literals
1628              - pattern matching
1629              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1630              - overloading on literals and operations
1631    
1632            This required adding a primitive type intinf, some additional
1633            primops, and implementations for several non-trivial intinf
1634            operations in Core.  (The intinf type is completely abstract
1635            to the compiler; all operations get delegated back to the Core.)
1636    
1637            * Intinf equality is handled by polyequal.  However, the compiler
1638            does not print its usual warning in this case (since polyequal
1639            is the right thing to do there).
1640    
1641            * Improved the organization of structure InlineT.
1642    
1643            * A word about conversion primops:
1644            If conversions involving intinf do not cancel out during
1645            CPS contract, then the compiler must insert calls to Core functions.
1646            Since all core access must be resolved already during the FLINT
1647            translate phase, it would be too late a the time of CPS contract
1648            to add new Core calls.  For this reason, conversion primops
1649            for intinf carry two arguments:  1. the numeric argument that
1650            they are supposed to convert, and 2. the Core function that
1651            can help with this conversion if necessary.  If CPS contract
1652            eliminates a primop, then the associated Core function becomes
1653            dead and goes away.  Intinf conversion primops that do not get
1654            eliminated by CPS contract get rewritten into calls of their
1655            core functions by a separate, new phase.
1656    
1657    interactive system:
1658            - Control.Print.intinfDepth controls max length of intinf constants
1659            being printed. (Analogous to Control.Print.stringDepth.)
1660            - Cleanup in printutil and pputil: got rid of unused stuff and
1661            duplicates; replaced some of the code with code that makes better
1662            use of library functionality.
1663    
1664    CM:
1665            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1666            are no longer silent.
1667    
1668    CKIT:
1669            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1670            in favor of LargeInt -- which is now the same as IntInf.
1671            CKIT-knowledgable people should check whether this is what's
1672            intended and otherwise change things back to using Int32 or
1673            FixedInt.
1674    
1675    Throughout the code:
1676            Started using IntInf.int literals and built-in operations
1677            (e.g., comparison with 0) where this seems appropriate.
1678    
1679    
1680    ----------------------------------------------------------------------
1681    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1682    Date: 2003/08/13 11:36:00 CDT
1683    Tag: dbm-20030813-mcz-merge1
1684    Description:
1685    
1686    Merging changes from the mcz-branch development branch into trunk.
1687    These changes involve replacement of the emulated old prettyprinter
1688    interface with direct use of the SML/NJ Lib PP library, and fixing
1689    of a couple of bugs (895, 1186) relating to error messages.  A new
1690    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1691    has been added.
1692    
1693    ----------------------------------------------------------------------
1694    Name: Matthias Blume (blume (at) tti - c (dot) org)
1695    Date: 2003/08/11 15:45:00 CDT
1696    Tag: blume-20030811-windows
1697    Description:
1698    
1699    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1700    
1701        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1702    
1703    This patch restores SML/NJ's ability to run under win32.  There are a
1704    number of changes, including fixes for several bugs that had gone
1705    unnoticed until now:
1706    
1707      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1708        trying to run the win32 version from within cygwin.)
1709    
1710      - There are a number of simple .bat scripts that substitute for their
1711        corresponding Unix shell-scripts.  (See below.)
1712    
1713      - The internals of ml-build have been modified slightly.  The main
1714        difference is that instead of calling ".link-sml" (or link-sml.bat)
1715        using OS.Process.system, the ML process delegates this task back
1716        to the script.  Otherwise problems arise in mixed environments such
1717        as Cygwin where scripts look and work like Unix scripts, but
1718        where OS.Process.system cannot run them.
1719    
1720      - In CM, the srcpath pickler used native pathname syntax -- which
1721        is incorrect in the case of cross-compilation.  The new pickle format
1722        is independent of platform-specific naming conventions.
1723    
1724      - Path configuration files (such as lib/pathconfig) can now choose
1725        between native and standard syntax.  Placing a line of the form
1726    
1727            standard!
1728    
1729        into the file causes all subsequent paths to be interpreted using
1730        CM standard pathname syntax (= Unix conventions); a line
1731    
1732            native!
1733    
1734        switches back to native style.  This was needed so that
1735        path config files can be written portably, see src/system/pathconfig.
1736    
1737      - Runtime system:
1738    
1739          - win32-filesys.c:  get_file_time and set_file_time now
1740            access modification time, not creation time.
1741    
1742          - I/O code made aware of new array representation.
1743    
1744          - Bug fixes in X86.prim.masm.
1745    
1746       - src/system/makeml made aware of win32. (For use under cygwin
1747            and other Unix-environments for windows.)
1748    
1749       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1750         which caused BinIO.inputAll to fail consistently.
1751    
1752    .bat scripts:
1753    
1754       Windows .bat scripts assume that SMLNJ_HOME is defined.
1755    
1756       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1757         applications (sml, ml-yacc, ml-lex).
1758       - ml-build.bat: analogous to ml-build.
1759       - config\install.bat:  Analogous to config/install.sh.  This requires
1760         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1761         (nmake etc. must be on the path, and vcvars32 must have been run.)
1762         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1763         and the bootfile hierarchy must have been unpacked under
1764         sml.boot.x86-win32.
1765         The script is very primitive and does a poor job at error checking.
1766         It only installs the base system, ml-lex, and ml-yacc.  No other
1767         libraries are being installed (i.e., you get only those that
1768         are part of the compiler.)
1769       - link-sml.bat:  analogous to .link-sml, but not currently used
1770    
1771    Unrelated bug fixes:
1772    
1773      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1774        types.
1775      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1776        otherwise.)
1777    
1778    ----------------------------------------------------------------------
1779    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1780    Date: 2003/06/17
1781    Tag: macqueen-20030617-bug895
1782    Description:
1783    
1784    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1785    Tag will be used for new development branch (mcz-branch) for
1786    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1787    
1788    ----------------------------------------------------------------------
1789    Name: Matthias Blume (blume (at) tti - c (dot) org)
1790    Date: 2003/05/27 16:55:00 CDT
1791    Tag: blume-20030527-polyeq
1792    Description:
1793    
1794    Tried to eliminated most cases of polymorphic equality.
1795    
1796    ----------------------------------------------------------------------
1797    Name: Matthias Blume (blume (at) tti - c (dot) org)
1798    Date: 2003/05/21 17:45:00 CDT
1799    Tag: blume-20030517-complete
1800    Description:
1801    
1802    Two changes:
1803    
1804      1. Added a flag for controlling whether non-exhaustive bindings will
1805         be treated as errors (default is false).
1806      2. Cleaned up the *entire* source tree so that CMB.make goes through
1807         without a single non-exhaustive match- or bind warning.
1808    
1809    ----------------------------------------------------------------------
1810    Name: Matthias Blume (blume (at) tti - c (dot) org)
1811    Date: 2003/05/17 10:20:00 CDT
1812    Tag: blume-20030517-absyn
1813    Description:
1814    
1815    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1816    
1817       This mainly affects the quality of error messages.  However, some
1818       of the code is now more straightforward than before. (Treatment of
1819       the above four constructs in translate.sml is much simpler than
1820       the "macro-expansion" that was going on before.  Plus, the mach-
1821       compiler no longer gets invoked just to be able to compile an
1822       if-expression.)
1823    
1824    2. The ErrorMsg.Error exception is now caught and absorbed by the
1825       interactive loop.
1826    
1827    ----------------------------------------------------------------------
1828    Name: Allen Leung
1829    Date: 2003/05/16 13:05:00 CDT
1830    Tag: leunga-20030516-cygwin-runtime
1831    Description:
1832    
1833        Ported the runtime system to cygwin, which uses the unix
1834    x86-unix bin files.  Missing/buggy features:
1835    
1836       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1837         the Cygwin library.
1838       o Ctrl-C handling may be flaky.
1839       o Windows system calls and Windows I/O are not supported.
1840    
1841       A new set of binfiles is located at:
1842    
1843         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1844    
1845       This is only needed for bootstrapping the cygwin version of smlnj.
1846       Other x86 versions can use the existing binfiles.
1847    
1848    ----------------------------------------------------------------------
1849    Name: Matthias Blume
1850    Date: 2003/04/08 15:42:00 CDT
1851    Tag: blume-20030408-listpair
1852    Description:
1853    
1854    1. Added a target 'mlrisc' to installer.
1855    
1856    2. Added missing elements to structure ListPair.
1857    
1858    ----------------------------------------------------------------------
1859    Name: Allen Leung
1860    Date: 2003/01/07 10:40:00 EST
1861    Tag: leunga-20030107-int-rem
1862    Description:
1863    
1864    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1865    The arguments to the SUBL instruction were swapped.
1866    
1867    ----------------------------------------------------------------------
1868    Name: Matthias Blume
1869    Date: 2002/12/12 16:25:00 EST
1870    Tag: blume-20021212-risc-ra
1871    Description:
1872    
1873    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1874    sent the RA into an infinite loop when floating point registers get
1875    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1876    110.37 and 110.38.)
1877    There was another set of potential problems related to the handling of
1878    MLRISC annotations (but those did not yet cause real problems, apparently).
1879    
1880    ----------------------------------------------------------------------
1881    Name: Matthias Blume
1882    Date: 2002/12/06 22:40:00 EST
1883    Tag: blume-20021206-cm-fileid
1884    Description:
1885    
1886    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1887    This fixes the problem of CM getting confused by files that suddenly
1888    change their identity (e.g., by getting unlinked and recreated by some
1889    text editor such as vi).  There might be a better/cheaper/cleaner way
1890    of doing this, but for now this will have to do.
1891    
1892    ----------------------------------------------------------------------
1893    Name: Matthias Blume
1894    Date: 2002/10/28 09:50:00 EST
1895    Tag: blume-20021028-typecheck
1896    Description:
1897    
1898    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1899    
1900    ----------------------------------------------------------------------
1901    Name: Matthias Blume
1902    Date: 2002/10/17 09:10:00 EDT
1903    Tag: Release_110_42
1904    Description:
1905    
1906    In good old tradition, there has been a slight hiccup so that we have
1907    to patch 110.42 after the fact.   The old release tag has been replaced
1908    (see below).
1909    
1910    The change solves a problem with two competing approaches the
1911    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1912    each other's way.
1913    
1914    This change only affects the runtime system code and the installer script.
1915    (No new bootfiles.)
1916    
1917    ----------------------------------------------------------------------
1918    Name: Matthias Blume
1919    Date: 2002/10/16 12:00:00 EDT
1920    Tag: Release_110_42_removed
1921    Description:
1922    
1923    New working release.  New bootfiles.
1924    
1925    ----------------------------------------------------------------------
1926    Name: Matthias Blume
1927    Date: 2002/10/10 13:10:00 EDT
1928    Tag: blume-20021010-ppc-divs
1929    Description:
1930    
1931    The mltree operator DIVS must be implemented with an overflow check on
1932    the PPC because the hardware indicates divide-by-zero using "overflow" as
1933    well.
1934    
1935    ----------------------------------------------------------------------
1936    Name: Matthias Blume
1937    Date: 2002/07/23 11:45:00
1938    Tag: blume-20020723-smlnj-home
1939    Description:
1940    
1941    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1942    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1943    CM_PATHCONFIG is also set), the path configuration file is assumed
1944    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1945    the entire tree to some other place and everything will "just work".
1946    
1947    (Companion commands such as ml-build and ml-makedepend also sense this
1948    variable.)
1949    
1950    ----------------------------------------------------------------------
1951    Name: Matthias Blume
1952    Date: 2002/07/12 21:19:00 EDT
1953    Tag: blume-20020712-liveness
1954    Description:
1955    
1956    Exported two useful "step" functions from liveness module (MLRISC).
1957    
1958    ----------------------------------------------------------------------
1959    Name: Matthias Blume
1960    Date: 2002/07/05 16:00 EDT
1961    Tag: Release_110_41
1962    Description:
1963    
1964    New working release.  New bootfiles.
1965    
1966    ----------------------------------------------------------------------
1967    Name: Matthias Blume
1968    Date: 2002/07/05 10:25:00 EDT
1969    Tag: blume-20020705-btimp
1970    Description:
1971    
1972    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1973    other clients can set up backtracing support.
1974    
1975    ----------------------------------------------------------------------
1976    Name: Matthias Blume
1977    Date: 2002/06/25 17:23:00 EDT
1978    Tag: blume-20020625-fpmax
1979    Description:
1980    
1981    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1982    case.  (The sense of the isNaN test was reversed -- which made min and
1983    max always return their first argument.)
1984    
1985    ----------------------------------------------------------------------
1986    Name: Matthias Blume
1987    Date: 2002/06/11
1988    Tag: blume-20020611-unixpath
1989    Description:
1990    
1991    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1992    
1993    ----------------------------------------------------------------------
1994    Name: Matthias Blume
1995    Date: 2002/06/10 16:35:00 EDT
1996    Tag: blume-20020610-ieeereal
1997    Description:
1998    
1999    I back-ported my implementation of IEEEReal.fromString from the
2000    idlbasis-devel branch so that we can test it.
2001    
2002    Another small change is that ppDec tries to give more information
2003    than just "<sig>" in the case of functors.  However, this code is
2004    broken in some mysterious way if the functor's body's signature
2005    has not been declared by ascription but gets inferred from the
2006    implementation.  This needs fixing...
2007    
2008    ----------------------------------------------------------------------
2009    Name: Matthias Blume
2010    Date: 2002/05/31
2011    Tag: blume-20020531-btrace-mode
2012    Description:
2013    
2014    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2015    the wayside when I switched over to using Controls everywhere.)
2016    
2017    ----------------------------------------------------------------------
2018    Name: Lal George
2019    Date: 2002/05/23 12:21:40 EDT
2020    Tag: george-20020523-visual-labels
2021    Description:
2022    
2023    Labels are now displayed in the graphical output to make
2024    the fall-through and target blocks obvious.
2025    
2026    ----------------------------------------------------------------------
2027    Name: Matthias Blume
2028    Date: 2002/05/22 11:03:00 EDT
2029    Tag: blume-20020522-shrink
2030    Description:
2031    
2032    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2033    situation that comes up when allocating huge arrays.
2034    
2035    ----------------------------------------------------------------------
2036    Name: Matthias Blume
2037    Date: 2002/05/21 16:00:00 EDT
2038    Tag: Release_110_40
2039    Description:
2040    
2041    New working release (110.40).  New bootfiles.
2042    
2043    [Also: John Reppy fixed GC bug 1131.]
2044    
2045    ----------------------------------------------------------------------
2046    Name: Matthias Blume
2047    Date: 2002/05/21 12:35:00 EDT
2048    Tag: blume-20020521-cmdoc
2049    Description:
2050    
2051    CM documentation update.
2052    
2053    ----------------------------------------------------------------------
2054    Name: Matthias Blume
2055    Date: 2002/05/21 10:55:00 EDT
2056    Tag: blume-20020521-misc
2057    Description:
2058    
2059    - John tweaked runtime to be silent on heap export (except when
2060      GC messages are on).
2061    - I added a few more things (cross-compiling versions of CMB) to
2062      config/preloads (as suggestions).
2063    
2064    ----------------------------------------------------------------------
2065    Name: Matthias Blume
2066    Date: 2002/05/20 22:25:00 EDT
2067    Tag: blume-20020520-controls
2068    Description:
2069    
2070    - Added ControlUtil structure to control-lib.cm.
2071    - Use it throughout.
2072    - Used Controls facility to define MLRISC controls (as opposed to
2073      registering MLRISC control ref cells with Controls after the fact)
2074    - Fixed messed-up controls priorities.
2075    
2076    * Removed again all the stuff from config/preloads that one wouldn't
2077      be able to preload at the time the initial heap image is built.
2078      (Many libraries, e.g., CML, do not exist yet at this time.  The
2079       only libraries that can be preloaded via config/preloads are those
2080       that come bundled with the bootfiles.)
2081    
2082    ----------------------------------------------------------------------
2083    Name: Matthias Blume
2084    Date: 2002/05/20 10:59:00 EDT
2085    Tag: blume-20020520-preloads
2086    Description:
2087    
2088    Added a lot of commented-out suggestions for things to be included
2089    in config/preloads.
2090    
2091    ----------------------------------------------------------------------
2092    Name: Allen Leung
2093    Date: 2002/05/18 14:20:00 EDT
2094    Tag: leunga-20020518-mdl
2095    Description:
2096    
2097    o Made the mdl tool stuff compile and run again.
2098    o I've disabled all the stuff that depends on RTL specifications; they
2099      are all badly broken anyway.
2100    
2101    ----------------------------------------------------------------------
2102    Name: Matthias Blume
2103    Date: 2002/05/17 16:49:00 EDT
2104    Tag: blume-20020517-controls
2105    Description:
2106    
2107    0. John Reppy made several modifications to the SML/NJ library.
2108       In particular, there is a shiny new controls-lib.cm.
2109    
2110    1. Pushed new controls interface through compiler so that everything
2111       compiles again.
2112    
2113    2. Added FormatComb and FORMAT_COMB to the CML version of the
2114       SML/NJ library (so that CML compiles again).
2115    
2116    3. Modified init scripts because XXX_DEFAULT environment variables
2117       are no longer with us.  (Boot-time initialization is now done
2118       using the same environment variables that are also used for
2119       startup-time initialization of controls.)
2120    
2121    ----------------------------------------------------------------------
2122    Name: Lal George
2123    Date: 2002/05/15 09:20:10 EDT
2124    Tag: george-20020515-pseudo-op-decls
2125    Description:
2126    
2127            All pseudo-ops emitted before the first segment declaration
2128    such as TEXT, DATA, and BSS directives are assumed to be global
2129    declarations and are emitted first in the assembly file. This is
2130    useful in a number of situations where one has pseudo-ops that are not
2131    specific to any segment, and also works around the constraint that one
2132    cannot have client pseudo-ops in the TEXT segment.
2133    
2134            Because no segment is associated with these declarations it is
2135    an error to allocate any space or objects before the first segment
2136    directive and an exception will be raised. However, we cannot make
2137    this check for client pseudo-ops.
2138    
2139            These top level declarations are a field in the CFG graph_info.
2140    In theory you can continue to add to this field after the CFG has been
2141    built -- provided you know what you are doing;-)
2142    
2143    ----------------------------------------------------------------------
2144    Name: Matthias Blume
2145    Date: 2002/05/13 16:40:00 EDT
2146    Tag: blume-20020513-pp-etc
2147    Description:
2148    
2149    A few minor bugfixes:
2150    
2151      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2152        (Bogus printouts for redefined bindings still occur.  Compiler
2153        bug should no longer occur now.  We need to redo the prettyprinter
2154        from scratch.)
2155    
2156      - CM pathname printer now also adds escape sequences for ( and )
2157    
2158      - commend and docu fixes for ml-nlffi
2159    
2160    ----------------------------------------------------------------------
2161    Name: Matthias Blume
2162    Date: 2002/05/10 16:40:00 EDT
2163    Tag: blume-20020510-erg-textio
2164    Description:
2165    
2166    Applied the following bugfix provided by Emden Gansner:
2167    
2168        Output is corrupted when outputSubstr is used rather than output.
2169    
2170        The problem occurs when a substring
2171    
2172            ss = (s, dataStart, dataLen)
2173    
2174        where dataStart > 0, fills a stream buffer with avail bytes left.
2175        avail bytes of s, starting at index dataStart, are copied into the
2176        buffer, the buffer is flushed, and then the remaining dataLen-avail
2177        bytes of ss are copied into the beginning of the buffer. Instead of
2178        starting this copy at index dataStart+avail in s, the current code
2179        starts the copy at index avail.
2180    
2181        Fix:
2182        In text-io-fn.sml, change line 695 from
2183             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2184        to
2185             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2186    
2187    ----------------------------------------------------------------------
2188    Name: Matthias Blume
2189    Date: 2002/04/12 13:55:00 EDT
2190    Tag: blume-20020412-assyntax
2191    Description:
2192    
2193    1. Grabbed newer assyntax.h from the XFree86 project.
2194    2. Fiddled with how to compile X86.prim.asm without warnings.
2195    3. (Very) Minor cleanup in CM.
2196    
2197    ----------------------------------------------------------------------
2198    Name: Matthias Blume
2199    Date: 2002/04/01 (no joke!) 17:07:00 EST
2200    Tag: blume-20020401-x86div
2201    Description:
2202    
2203    Added full support for div/mod/rem/quot on the x86, using the machine
2204    instruction's two results (without clumsily recomputing the remainder)
2205    directly where appropriate.
2206    
2207    Some more extensive power-of-two support was added to the x86 instruction
2208    selector (avoiding expensive divs, mods, and muls where they can be
2209    replaced with cheaper shifts and masks).  However, this sort of thing
2210    ought to be done earlier, e.g., within the CPS optimizer so that
2211    all architectures benefit from it.
2212    
2213    The compiler compiles to a fixed point, but changes might be somewhat
2214    fragile nevertheless.  Please, report any strange things that you might
2215    see wrt. div/mod/quot/rem...
2216    
2217    ----------------------------------------------------------------------
2218    Name: Matthias Blume
2219    Date: 2002/03/29 17:22:00
2220    Tag: blume-20020329-div
2221    Description:
2222    
2223    Fixed my broken div/mod logic.  Unfortunately, this means that the
2224    inline code for div/mod now has one more comparison than before.
2225    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2226    The problem was with quotient = 0, because that alone does not tell
2227    us which way the rounding went.  One then has to look at whether
2228    remainder and divisor have the same sign...  :(
2229    
2230    Anyway, I replaced the bootfiles with fresh ones...
2231    
2232    ----------------------------------------------------------------------
2233    Name: Matthias Blume
2234    Date: 2002/03/29 14:10:00 EST
2235    Tag: blume-20020329-inlprims
2236    Description:
2237    
2238    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2239    
2240    Primops have changed. This means that the bin/boot-file formats have
2241    changed as well.
2242    
2243    To make sure that there is no confusion, I made a new version.
2244    
2245    
2246    CHANGES:
2247    
2248    * removed REMT from mltree (remainder should never overflow).
2249    
2250    * added primops to deal with divisions of all flavors to the frontend
2251    
2252    * handled these primops all the way through so they map to their respective
2253      MLRISC support
2254    
2255    * used these primops in the implementation of Int, Int32, Word, Word32
2256    
2257    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2258    
2259    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2260    
2261    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2262      floating point
2263    
2264    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2265      and Real (but Real.abs maps to a separate floating-point-only primop)
2266    
2267    
2268    TODO items:
2269    
2270    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2271      instructions because the machine instruction encoder cannot handle
2272      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2273      REM in the first place, and REM is supported by the x86 machine coder.)
2274    
2275    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2276      selection phase.  (The two can be streamlined because the hardware
2277      delivers both quotient and remainder at the same time anyway.)
2278    
2279    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2280      (Currently the behavior is inconsistent both across architectures and
2281      wrt. the draft Basis spec.)
2282    
2283    * Word8 should eventually be handled natively, too.
2284    
2285    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2286      as if there currently is no execution path that could trigger it in
2287      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2288      hold for things like multiplication and division.)
2289    
2290    ----------------------------------------------------------------------
2291    Name: Matthias Blume
2292    Date: 2002/03/27 16:27:00 EST
2293    Tag: blume-20020327-mlrisc-divisions
2294    Description:
2295    
2296    Added support for all four division operations (ML's div, mod, quot,
2297    and rem) to MLRISC.  In the course of doing so, I also rationalized
2298    the naming (no more annoying switch-around of DIV and QUOT), by
2299    parameterizing the operation by div_rounding_mode (which can be either
2300    DIV_TO_ZERO or DIV_TO_NEGINF).
2301    
2302    The generic MLTreeGen functor takes care of compiling all four
2303    operations down to only round-to-zero div.
2304    
2305    Missing pieces:
2306    
2307      * Doing something smarter than relying on MLTreeGen on architectures
2308        like, e.g., the x86 where hardware division delivers both quotient and
2309        remainder at the same time.  With this, the implementation of the
2310        round-to-neginf operations could be further streamlined.
2311    
2312      * Remove inlining support for div/mod/rem from the frontend and replace it
2313        with primops that get carried through to the backend.  Do this for all
2314        int and word types.
2315    
2316    ----------------------------------------------------------------------
2317    Name: Matthias Blume
2318    Date: 2002/03/25 17:25:00 EST
2319    Tag: blume-20020325-divmod
2320    Description:
2321    
2322    I improved (hopefully without breaking them) the implementation of Int.div,
2323    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2324    advantage of the following observations:
2325    
2326      Let  q = x quot y      r = x rem y
2327           d = x div  y      m = x mod y
2328    
2329    where "quot" is the round-to-zero version of integer division that
2330    hardware usually provides.  Then we have:
2331    
2332         r = x - q * y        where neither the * nor the - will overflow
2333         d = if q >= 0 orelse x = q * y then q else q - 1
2334                              where neither the * nor the - will overflow
2335         m = if q >= 0 orelse r = 0 then r else r + y
2336                              where the + will not overflow
2337    
2338    This results in substantial simplification of the generated code.
2339    The following table shows the number of CFG nodes and edges generated
2340    for
2341            fun f (x, y) = x OPER y
2342            (* with OPER \in div, mod, quot, rem *)
2343    
2344    
2345        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2346        --------------------------------------------------------
2347         div |         24 |         39 |         12 |         16
2348         mod |         41 |         71 |         12 |         16
2349        quot |          8 |         10 |          8 |         10
2350         rem |         10 |         14 |          8 |         10
2351    
2352    
2353    ----------------------------------------------------------------------
2354    Name: Matthias Blume
2355    Date: 2002/03/25 22:06:00 EST
2356    Tag: blume-20020325-cprotobug
2357    Description:
2358    
2359    Fixed a bug in cproto (c prototype decoder).
2360    
2361    ----------------------------------------------------------------------
2362    Name: Matthias Blume
2363    Date: 2002/03/25 16:00:00 EST
2364    Tag: blume-20020325-raw-primops
2365    Description:
2366    
2367    I did some cleanup to Allen's new primop code and
2368    replaced yesterday's bootfiles with new ones.
2369    (But they are stored in the same place.)
2370    
2371    ----------------------------------------------------------------------
2372    Name: Matthias Blume
2373    Date: 2002/03/24 22:40:00 EST
2374    Tag: blume-20020324-bootfiles
2375    Description:
2376    
2377    Made the bootfiles that Allen asked for.
2378    
2379    ----------------------------------------------------------------------
2380    Name: Allen Leung
2381    Date: 2002/03/23 15:50:00 EST
2382    Tag: leunga-20020323-flint-cps-rcc-primops
2383    Description:
2384    
2385      1. Changes to FLINT primops:
2386    
2387        (* make a call to a C-function;
2388         * The primop carries C function prototype information and specifies
2389         * which of its (ML-) arguments are floating point. C prototype
2390         * information is for use by the backend, ML information is for
2391         * use by the CPS converter. *)
2392      | RAW_CCALL of { c_proto: CTypes.c_proto,
2393                       ml_args: ccall_type list,
2394                       ml_res_opt: ccall_type option,
2395                       reentrant : bool
2396                     } option
2397       (* Allocate uninitialized storage on the heap.
2398        * The record is meant to hold short-lived C objects, i.e., they
2399        * are not ML pointers.  With the tag, the representation is
2400        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2401        *)
2402      | RAW_RECORD of {tag:bool,sz:int}
2403      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2404    
2405      2.  These CPS primops are now overloaded:
2406    
2407           rawload of {kind:numkind}
2408           rawstore of {kind:numkind}
2409    
2410          The one argument form is:
2411    
2412             rawload {kind} address
2413    
2414          The two argument form is:
2415    
2416             rawload {kind} [ml object, byte-offset]
2417    
2418      3. RAW_CCALL/RCC now takes two extra arguments:
2419    
2420         a. The first is whether the C call is reentrant, i.e., whether
2421            ML state should be saved and restored.
2422         b. The second argument is a string argument specifying the name of
2423            library and the C function.
2424    
2425         These things are currently not handled in the code generator, yet.
2426    
2427      4. In CProto,
2428    
2429         An encoding type of "bool" means "ml object" and is mapped into
2430         C prototype of PTR.  Note that "bool" is different than "string",
2431         even though "string" is also mapped into PTR, because "bool"
2432         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2433    
2434      5. Pickler/unpicker
2435    
2436         Changed to handle RAW_RECORD and newest RAW_CCALL
2437    
2438      6. MLRiscGen,
2439    
2440         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2441         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2442            in the file CodeGen/cpscompile/cps-c-calls.sml
2443    
2444      7. Added the conditional move operator
2445    
2446             condmove of branch
2447    
2448         to cps.  Generation of this is still buggy so it is currently
2449         disabled.
2450    
2451    ----------------------------------------------------------------------
2452    Name: Lal George
2453    Date: 2002/03/22 14:18:25 EST
2454    Tag: george-20020322-cps-branch-prob
2455    Description:
2456    
2457    Implemented the Ball-Larus branch prediction-heuristics, and
2458    incorporated graphical viewers for control flow graphs.
2459    
2460    Ball-Larus Heuristics:
2461    ---------------------
2462    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2463    
2464    By design it uses the Dempster-Shafer theory for combining
2465    probabilities.  For example, in the function:
2466    
2467        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2468    
2469    the ball-larus heuristics predicts that the n=0 is unlikely
2470    (OH-heuristic), and the 'then' branch is unlikely because of the
2471    RH-heuristic -- giving the 'then' branch an even lower combined
2472    probability using the Dempster-Shafer theory.
2473    
2474    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2475    probability of the 'then' branch because of the loop in the else
2476    branch.
2477    
2478    
2479    Graphical Viewing:
2480    ------------------
2481    I merely plugged in Allen's graphical viewers into the compiler. The
2482    additional code is not much. At the top level, saying:
2483    
2484            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2485    
2486    will display the graphical view of the control flow graph just before
2487    back-patching.  daVinci must be in your path for this to work. If
2488    daVinci is not available, then the default viewer can be changed
2489    using:
2490    
2491            Control.MLRISC.getString "viewer"
2492    
2493    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2494    course, these viewers must be in your path.
2495    
2496    The above will display the compilation unit at the level of clusters,
2497    many of which are small, boring, and un-interesting. Also setting:
2498    
2499            Control.MLRISC.getInt "cfg-graphical-view_size"
2500    
2501    will display clusters that are larger than the value set by the above.
2502    
2503    
2504    ----------------------------------------------------------------------
2505    Name: Matthias Blume
2506    Date: 2002/03/21 22:20:00 EST
2507    Tag: blume-20020321-kmp-bugfix
2508    Description:
2509    
2510    Changed the interface to the KMP routine in PreString and fixed
2511    a minor bug in one place where it was used.
2512    
2513    ----------------------------------------------------------------------
2514    Name: Allen Leung
2515    Date: 2002/03/21 20:30:00 EST
2516    Tag: leunga-20020321-cfg
2517    Description:
2518    
2519      Fixed a potential problem in cfg edge splitting.
2520    
2521    ----------------------------------------------------------------------
2522    Name: Allen Leung
2523    Date: 2002/03/21 17:15:00 EST
2524    Tag: leunga-20020321-x86-fp-cfg
2525    Description:
2526    
2527      1. Recoded the buggy parts of x86-fp.
2528    
2529         a. All the block reordering code has been removed.
2530            We now depend on the block placement phases to do this work.
2531    
2532         b. Critical edge splitting code has been simplified and moved into the
2533            CFG modules, as where they belong.
2534    
2535         Both of these were quite buggy and complex.  The code is now much, much
2536         simpler.
2537    
2538      2. X86 backend.
2539    
2540         a. Added instructions for 64-bit support.  Instruction selection for
2541            64-bit has not been committed, however, since that
2542            requires changes to MLTREE which haven't been approved by
2543            Lal and John.
2544    
2545         b. Added support for FUCOMI and FUCOMIP when generating code for
2546            PentiumPro and above.  We only generate these instructions in
2547            the fast-fp mode.
2548    
2549         c. Added cases for JP and JNP in X86FreqProps.
2550    
2551      3. CFG
2552    
2553         CFG now has a bunch of methods for edge splitting and merging.
2554    
2555      4. Machine description.
2556    
2557         John's simplification of MLTREE_BASIS.fcond broke a few machine
2558         description things:
2559    
2560         rtl-build.{sig,sml} and hppa.mdl fixed.
2561    
2562         NOTE: the machine description stuff in the repository is still broken.
2563               Again, I can't put my fixes in because that involves
2564               changes to MLTREE.
2565    
2566    ----------------------------------------------------------------------
2567    Name: Matthias Blume
2568    Date: 2002/03/20 15:55:00 EST
2569    Tag: blume-20020320-kmp
2570    Description:
2571    
2572    Implemented Knuth-Morris-Pratt string matching in PreString and used
2573    it for String.isSubstring, Substring.isSubstring, and
2574    Substring.position.
2575    
2576    (Might need some stress-testing.  Simple examples worked fine.)
2577    
2578    ----------------------------------------------------------------------
2579    Name: Matthias Blume
2580    Date: 2002/03/19 16:37:00 EST
2581    Tag: blume-20020319-witnesses
2582    Description:
2583    
2584    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2585    
2586    This implements a generic mechanism for changing constness qualifiers
2587    anywhere within big C types without resorting to outright "casts".
2588    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2589    modify the constness at the outermost level.)
2590    The implementation of "convert" is based on the idea of "witness"
2591    values -- values that are not used by the operation but whose types
2592    "testify" to their applicability.  On the implementation side, "convert"
2593    is simply a projection (returning its second curried argument).  With
2594    cross-module inlining, it should not result in any machine code being
2595    generated.
2596    
2597    ----------------------------------------------------------------------
2598    Name: Matthias Blume
2599    Date: 2002/03/15 16:40:00 EST
2600    Tag: blume-20020315-basis
2601    Description:
2602    
2603    Provided (preliminary?) implementations for
2604    
2605      {String,Substring}.{concatWith,isSuffix,isSubstring}
2606    
2607    and
2608    
2609      Substring.full
2610    
2611    Those are in the Basis spec but they were missing in SML/NJ.
2612    
2613    ----------------------------------------------------------------------
2614    Name: Matthias Blume
2615    Date: 2002/03/14 21:30:00 EST
2616    Tag: blume-20020314-controls
2617    Description:
2618    
2619    Controls:
2620    ---------
2621    
2622    1. Factored out the recently-added Controls : CONTROLS stuff and put
2623       it into its own library $/controls-lib.cm.  The source tree for
2624       this is under src/smlnj-lib/Controls.
2625    
2626    2. Changed the names of types and functions in this interface, so they
2627       make a bit more "sense":
2628    
2629          module -> registry
2630          'a registry -> 'a group
2631    
2632    3. The interface now deals in ref cells only.  The getter/setter interface
2633       is (mostly) gone.
2634    
2635    4. Added a function that lets one register an already-existing ref cell.
2636    
2637    5. Made the corresponding modifications to the rest of the code so that
2638       everything compiles again.
2639    
2640    6. Changed the implementation of Controls.MLRISC back to something closer
2641       to the original.  In particular, this module (and therefore MLRISC)
2642       does not depend on Controls.  There now is some link-time code in
2643       int-sys.sml that registers the MLRISC controls with the Controls
2644       module.
2645    
2646    CM:
2647    ---
2648    
2649      * One can now specify the lambda-split aggressiveness in init.cmi.
2650    
2651    ----------------------------------------------------------------------
2652    Name: Allen Leung
2653    Date: 2002/03/13 17:30:00 EST
2654    Tag: leunga-20020313-x86-fp-unary
2655    Description:
2656    
2657    Bug fix for:
2658    
2659    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2660    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2661    > - fun f(x,(y,z)) = Real.~ y;
2662    > [autoloading]
2663    > [autoloading done]
2664    >       fchsl   (%eax), 184(%esp)
2665    > Error: MLRisc bug: X86MCEmitter.emitInstr
2666    >
2667    > uncaught exception Error
2668    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2669    
2670    The problem was that the code generator did not generate any fp registers
2671    in this case, and the ra didn't know that it needed to run the X86FP phase to
2672    translate the pseudo fp instruction.   This only happened with unary fp
2673    operators in certain situations.
2674    
2675    ----------------------------------------------------------------------
2676    Name: Matthias Blume
2677    Date: 2002/03/13 14:00:00 EST
2678    Tag: blume-20020313-overload-etc
2679    Description:
2680    
2681    1. Added _overload as a synonym for overload for backward compatibility.
2682       (Control.overloadKW must be true for either version to be accepted.)
2683    
2684    2. Fixed bug in install script that caused more things to be installed
2685       than what was requested in config/targets.
2686    
2687    3. Made CM aware of the (_)overload construct so that autoloading
2688       works.
2689    
2690    ----------------------------------------------------------------------
2691    Name: Matthias Blume
2692    Date: 2002/03/12 22:03:00 EST
2693    Tag: blume-20020312-url
2694    Description:
2695    
2696    Forgot to update BOOT and srcarchiveurl.
2697    
2698    ----------------------------------------------------------------------
2699    Name: Matthias Blume
2700    Date: 2002/03/12 17:30:00 EST
2701    Tag: blume-20020312-version110392
2702    Description:
2703    
2704    Yet another version number bump (because of small changes to the
2705    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2706    
2707    Changes:
2708    
2709      The new pid generation scheme described a few weeks ago was overly
2710      complicated.  I implemented a new mechanism that is simpler and
2711      provides a bit more "stability":  Once CM has seen a compilation
2712      unit, it keeps its identity constant (as long as you do not delete
2713      those crucial CM/GUID/* files).  This means that when you change
2714      an interface, compile, then go back to the old interface, and
2715      compile again, you arrive at the original pid.
2716    
2717      There now also is a mechanism that instructs CM to use the plain
2718      environment hash as a module's pid (effectively making its GUID
2719      the empty string).  For this, "noguid" must be specified as an
2720      option to the .sml file in question within its .cm file.
2721      This is most useful for code that is being generated by tools such
2722      as ml-nlffigen (because during development programmers tend to
2723      erase the tool's entire output directory tree including CM's cached
2724      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2725      revert to the old, broken behavior of SML/NJ, but in specific cases
2726      where there is no danger of interface confusion, its use is ok
2727      (I think).
2728    
2729      ml-nlffigen by default generates "noguid" annotations.  They can be
2730      turned off by specifying -guid in its command line.
2731    
2732    ----------------------------------------------------------------------
2733    Name: Lal George
2734    Date: 2002/03/12 12 14:42:36 EST
2735    Tag: george-20020312-frequency-computation
2736    Description:
2737    
2738    Integrated jump chaining and static block frequency into the
2739    compiler. More details and numbers later.
2740    
2741    ----------------------------------------------------------------------
2742    Name: Lal George
2743    Date: 2002/03/11 11 22:38:53 EST
2744    Tag: george-20020311-jump-chain-elim
2745    Description:
2746    
2747    Tested the jump chain elimination on all architectures (except the
2748    hppa).  This is on by default right now and is profitable for the
2749    alpha and x86, however, it may not be profitable for the sparc and ppc
2750    when compiling the compiler.
2751    
2752    The gc test will typically jump to a label at the end of the cluster,
2753    where there is another jump to an external cluster containing the actual
2754    code to invoke gc. This is to allow factoring of common gc invocation
2755    sequences. That is to say, we generate:
2756    
2757            f:
2758               testgc
2759               ja   L1      % jump if above to L1
2760    
2761            L1:
2762               jmp L2
2763    
2764    
2765    After jump chain elimination the 'ja L1' instructions is converted to
2766    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2767    up being implemented in their long form (if L2 is far away) using:
2768    
2769            jbe     L3      % jump if below or equal to L3
2770            jmp     L2
2771         L3:
2772            ...
2773    
2774    
2775    For large compilation units L2  may be far away.
2776    
2777    
2778    ----------------------------------------------------------------------
2779    Name: Matthias Blume
2780    Date: 2002/03/11 13:30:00 EST
2781    Tag: blume-20020311-mltreeeval
2782    Description:
2783    
2784    A functor parameter was missing.
2785    
2786    ----------------------------------------------------------------------
2787    Name: Allen Leung
2788    Date: 2002/03/11 10:30:00 EST
2789    Tag: leunga-20020311-runtime-string0
2790    Description:
2791    
2792       The representation of the empty string now points to a
2793    legal null terminated C string instead of unit.  It is now possible
2794    to convert an ML string into C string with InlineT.CharVector.getData.
2795    This compiles into one single machine instruction.
2796    
2797    ----------------------------------------------------------------------
2798    Name: Allen Leung
2799    Date: 2002/03/10 23:55:00 EST
2800    Tag: leunga-20020310-x86-call
2801    Description:
2802    
2803       Added machine generation for CALL instruction (relative displacement mode)
2804    
2805    ----------------------------------------------------------------------
2806    Name: Matthias Blume
2807    Date: 2002/03/08 16:05:00
2808    Tag: blume-20020308-entrypoints
2809    Description:
2810    
2811    Version number bumped to 110.39.1.  NEW BOOTFILES!
2812    
2813    Entrypoints: non-zero offset into a code object where execution should begin.
2814    
2815    - Added the notion of an entrypoint to CodeObj.
2816    - Added reading/writing of entrypoint info to Binfile.
2817    - Made runtime system bootloader aware of entrypoints.
2818    - Use the address of the label of the first function given to mlriscGen
2819      as the entrypoint.  This address is currently always 0, but it will
2820      not be 0 once we turn on block placement.
2821    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2822      with entry points) from mlriscGen.
2823    
2824    ----------------------------------------------------------------------
2825    Name: Allen Leung
2826    Date: 2002/03/07 20:45:00 EST
2827    Tag: leunga-20020307-x86-cmov
2828    Description:
2829    
2830       Bug fixes for CMOVcc on x86.
2831    
2832       1. Added machine code generation for CMOVcc
2833       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2834       3. CMOVcc cannot have an immediate operand as argument.
2835    
2836    ----------------------------------------------------------------------
2837    Name: Matthias Blume
2838    Date: 2002/03/07 16:15:00 EST
2839    Tag: blume-20020307-controls
2840    Description:
2841    
2842    This is a very large but mostly boring patch which makes (almost)
2843    every tuneable compiler knob (i.e., pretty much everything under
2844    Control.* plus a few other things) configurable via both the command
2845    line and environment variables in the style CM did its configuration
2846    until now.
2847    
2848    Try starting sml with '-h' (or, if you are brave, '-H')
2849    
2850    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2851    implements the underlying generic mechanism.
2852    
2853    The interface to some of the existing such facilities has changed somewhat.
2854    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2855    (The getFoo interface is still there for backward-compatibility, but its
2856    use is deprecated.)
2857    
2858    The ml-build script passes -Cxxx=yyy command-line arguments through so
2859    that one can now twiddle the compiler settings when using this "batch"
2860    compiler.
2861    
2862    TODO items:
2863    
2864    We should go through and throw out all controls that are no longer
2865    connected to anything.  Moreover, we should go through and provide
2866    meaningful (and correct!) documentation strings for those controls
2867    that still are connected.
2868    
2869    Currently, multiple calls to Controls.new are accepted (only the first
2870    has any effect).  Eventually we should make sure that every control
2871    is being made (via Controls.new) exactly once.  Future access can then
2872    be done using Controls.acc.
2873    
2874    Finally, it would probably be a good idea to use the getter-setter
2875    interface to controls rather than ref cells.  For the time being, both
2876    styles are provided by the Controls module, but getter-setter pairs are
2877    better if thread-safety is of any concern because they can be wrapped.
2878    
2879    *****************************************
2880    
2881    One bug fix: The function blockPlacement in three of the MLRISC
2882    backpatch files used to be hard-wired to one of two possibilities at
2883    link time (according to the value of the placementFlag).  But (I
2884    think) it should rather sense the flag every time.
2885    
2886    *****************************************
2887    
2888    Other assorted changes (by other people who did not supply a HISTORY entry):
2889    
2890    1. the cross-module inliner now works much better (Monnier)
2891    2. representation of weights, frequencies, and probabilities in MLRISC
2892       changed in preparation of using those for weighted block placement
2893       (Reppy, George)
2894    
2895    ----------------------------------------------------------------------
2896    Name: Lal George
2897    Date: 2002/03/07 14:44:24 EST 2002
2898    Tag: george-20020307-weighted-block-placement
2899    
2900    Tested the weighted block placement optimization on all architectures
2901    (except the hppa) using AMPL to generate the block and edge frequencies.
2902    Changes were required in the machine properties to correctly
2903    categorize trap instructions. There is an MLRISC flag
2904    "weighted-block-placement" that can be used to enable weighted block
2905    placement, but this will be ineffective without block/edge
2906    frequencies (coming soon).
2907    
2908    
2909    ----------------------------------------------------------------------
2910    Name: Lal George
2911    Date: 2002/03/05 17:24:48 EST
2912    Tag: george-20020305-linkage-cluster
2913    
2914    In order to support the block placement optimization, a new cluster
2915    is generated as the very first cluster (called the linkage cluster).
2916    It contains a single jump to the 'real' entry point for the compilation
2917    unit. Block placement has no effect on the linkage cluster itself, but
2918    all the other clusters  have full freedom in the manner in which they
2919    reorder blocks or functions.
2920    
2921    On the x86 the typical linkage code that is generated is:
2922       ----------------------
2923            .align 2
2924       L0:
2925            addl    $L1-L0, 72(%esp)
2926            jmp     L1
2927    
2928    
2929            .align  2
2930       L1:
2931       ----------------------
2932    
2933    72(%esp) is the memory location for the stdlink register. This
2934    must contain the address of the CPS function being called. In the
2935    above example, it contains the address of  L0; before
2936    calling L1 (the real entry point for the compilation unit), it
2937    must contain the address for L1, and hence
2938    
2939            addl $L1-L0, 72(%esp)
2940    
2941    I have tested this on all architectures except the hppa.The increase
2942    in code size is of course negligible
2943    
2944    ----------------------------------------------------------------------
2945    Name: Allen Leung
2946    Date: 2002/03/03 13:20:00 EST
2947    Tag: leunga-20020303-mlrisc-tools
2948    
2949      Added #[ ... ] expressions to mlrisc tools
2950    
2951    ----------------------------------------------------------------------
2952    Name: Matthias Blume
2953    Date: 2002/02/27 12:29:00 EST
2954    Tag: blume-20020227-cdebug
2955    Description:
2956    
2957    - made types in structure C and C_Debug to be equal
2958    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2959    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2960    
2961    ----------------------------------------------------------------------
2962    Name: Matthias Blume
2963    Date: 2002/02/26 12:00:00 EST
2964    Tag: blume-20020226-ffi
2965    Description:
2966    
2967    1. Fixed a minor bug in CM's "noweb" tool:
2968       If numbering is turned off, then truly don't number (i.e., do not
2969       supply the -L option to noweb).  The previous behavior was to supply
2970       -L'' -- which caused noweb to use the "default" line numbering scheme.
2971       Thanks to Chris Richards for pointing this out (and supplying the fix).
2972    
2973    2. Once again, I reworked some aspects of the FFI:
2974    
2975       A. The incomplete/complete type business:
2976    
2977       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2978         gone!
2979       - ML types representing an incomplete type are now *equal* to
2980         ML types representing their corresponding complete types (just like
2981         in C).  This is still safe because ml-nlffigen will not generate
2982         RTTI for incomplete types, nor will it generate functions that
2983         require access to such RTTI.   But when ML code generated from both
2984         incomplete and complete versions of the C type meet, the ML types
2985         are trivially interoperable.
2986    
2987         NOTE:  These changes restore the full generality of the translation
2988         (which was previously lost when I eliminated functorization)!
2989    
2990       B. Enum types:
2991    
2992       - Structure C now has a type constructor "enum" that is similar to
2993         how the "su" constructor works.  However, "enum" is not a phantom
2994         type because each "T enum" has values (and is isomorphic to
2995         MLRep.Signed.int).
2996       - There are generic access operations for enum objects (using
2997         MLRep.Signed.int).
2998       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2999         * The structure contains the definition of type "mlrep" (the ML-side
3000         representation type of the enum).  Normally, mlrep is the same
3001         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3002         then mlrep will be defined as a datatype -- thus facilitating
3003         pattern matching on mlrep values.
3004         ("-ec" will be suppressed if there are duplicate values in an
3005          enumeration.)
3006         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3007         will be generated for each C enum constant xxx.
3008         * Conversion functions m2i and i2m convert between mlrep and
3009         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
3010         * Coversion functions c and ml convert between mlrep and "tag enum".
3011         * Access functions (get/set) fetch and store mlrep values.
3012       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3013         enumerations are merged into one single enumeration represented by
3014         structure E_'.
3015    
3016    ----------------------------------------------------------------------
3017    Name: Allen Leung
3018    Date: 2002/02/25 04:45:00 EST
3019    Tag: leunga-20020225-cps-spill
3020    
3021    This is a new implementation of the CPS spill phase.
3022    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3023    In case of problems, replace it with the old file spill.sml
3024    
3025    The current compiler runs into some serious performance problems when
3026    constructing a large record.  This can happen when we try to compile a
3027    structure with many items.  Even a very simple structure like the following
3028    makes the compiler slow down.
3029    
3030        structure Foo = struct
3031           val x_1 = 0w1 : Word32.int
3032           val x_2 = 0w2 : Word32.int
3033           val x_3 = 0w3 : Word32.int
3034           ...
3035           val x_N = 0wN : Word32.int
3036        end
3037    
3038    The following table shows the compile time, from N=1000 to N=4000,
3039    with the old compiler:
3040    
3041    N
3042    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3043           MLRISC ra                               0.06u  0.00s  0.05g
3044              (spills = 0 reloads = 0)
3045           TOTAL                                   0.63u  0.07s  0.21g
3046    
3047    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3048           MLRISC ra                               5.68u  0.59s  3.93g
3049              (spills = 0 reloads = 0)
3050           TOTAL                                   14.71u  0.99s  4.81g
3051    
3052    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3053           MLRISC ra                               5.54u  0.65s  3.91g
3054              (spills = 543 reloads = 1082)
3055           TOTAL                                   65.40u  3.13s  6.00g
3056    
3057    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3058           MLRISC ra                               0.80u  0.10s  0.55g
3059              (spills = 42 reloads = 84)
3060           TOTAL                                   129.42u  5.10s  4.13g
3061    
3062    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3063           MLRISC ra                               2.69u  0.27s  1.38g
3064              (spills = 62 reloads = 124)
3065           TOTAL                                   682.48u  19.61s  13.99g
3066    
3067    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3068           MLRISC ra                               4.96u  0.27s  2.72g
3069              (spills = 85 reloads = 170)
3070           TOTAL                                   2375.26u  57.21s  48.00g
3071    
3072    As you can see the old cps spill module suffers from some serious
3073    performance problem.  But since I cannot decipher the old code fully,
3074    instead of patching the problems up, I'm reimplementing it
3075    with a different algorithm.  The new code is more modular,
3076    smaller when compiled, and substantially faster
3077    (O(n log n) time and O(n) space).  Timing of the new spill module:
3078    
3079    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3080          MLRISC ra                               0.25u  0.02s  0.15g
3081             (spills=1 reloads=3)
3082          TOTAL                                   7.74u  0.34s  1.62g
3083    
3084    Implementation details:
3085    
3086    As far as I can tell, the purpose of the CPS spill module is to make sure the
3087    number of live variables at any program point (the bandwidth)
3088    does not exceed a certain limit, which is determined by the
3089    size of the spill area.
3090    
3091    When the bandwidth is too large, we decrease the register pressure by
3092    packing live variables into spill records.  How we achieve this is
3093    completely different than what we did in the old code.
3094    
3095    First, there is something about the MLRiscGen code generator
3096    that we should be aware of:
3097    
3098    o MLRiscGen performs code motion!
3099    
3100       In particular, it will move floating point computations and
3101       address computations involving only the heap pointer to
3102       their use sites (if there is only a single use).
3103       What this means is that if we have a CPS record construction
3104       statement
3105    
3106           RECORD(k,vl,w,e)
3107    
3108       we should never count the new record address w as live if w
3109       has only one use (which is often the case).
3110    
3111       We should do something similar to floating point, but the transformation
3112       there is much more complex, so I won't deal with that.
3113    
3114    Secondly, there are now two new cps primops at our disposal:
3115    
3116     1. rawrecord of record_kind option
3117        This pure operator allocates some uninitialized storage from the heap.
3118        There are two forms:
3119    
3120         rawrecord NONE [INT n]  allocates a tagless record of length n
3121         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3122                                     and initializes the tag.
3123    
3124     2. rawupdate of cty
3125          rawupdate cty (v,i,x)
3126          Assigns to x to the ith component of record v.
3127          The storelist is not updated.
3128    
3129    We use these new primops for both spilling and increment record construction.
3130    
3131     1. Spilling.
3132    
3133        This is implemented with a linear scan algorithm (but generalized
3134        to trees).  The algorithm will create a single spill record at the
3135        beginning of the cps function and use rawupdate to spill to it,
3136        and SELECT or SELp to reload from it.  So both spills and reloads
3137        are fine-grain operations.  In contrast, in the old algorithm
3138        "spills" have to be bundled together in records.
3139    
3140        Ideally, we should sink the spill record construction to where
3141        it is needed.  We can even split the spill record into multiple ones
3142        at the places where they are needed.  But CPS is not a good
3143        representation for global code motion, so I'll keep it simple and
3144        am not attempting this.
3145    
3146     2. Incremental record construction (aka record splitting).
3147    
3148        Long records with many component values which are simulatenously live
3149        (recall that single use record addresses are not considered to
3150         be live) are constructed with rawrecord and rawupdate.
3151        We allocate space on the heap with rawrecord first, then gradually
3152        fill it in with rawupdate.  This is the technique suggested to me
3153        by Matthias.
3154    
3155        Some restrictions on when this is applicable:
3156        1. It is not a VECTOR record.  The code generator currently does not handle
3157           this case. VECTOR record uses double indirection like arrays.
3158        2. All the record component values are defined in the same "basic block"
3159           as the record constructor.  This is to prevent speculative
3160           record construction.
3161    
3162    ----------------------------------------------------------------------
3163    Name: Allen Leung
3164    Date: 2002/02/22 01:02:00 EST
3165    Tag: leunga-20020222-mlrisc-tools
3166    
3167    Minor bug fixes in the parser and rewriter
3168    
3169    ----------------------------------------------------------------------
3170    Name: Allen Leung
3171    Date: 2002/02/21 20:20:00 EST
3172    Tag: leunga-20020221-peephole
3173    
3174    Regenerated the peephole files.  Some contained typos in the specification
3175    and some didn't compile because of pretty printing bugs in the old version
3176    of 'nowhere'.
3177    
3178    ----------------------------------------------------------------------
3179    Name: Allen Leung
3180    Date: 2002/02/19 20:20:00 EST
3181    Tag: leunga-20020219-mlrisc-tools
3182    Description:
3183    
3184       Minor bug fixes to the mlrisc-tools library:
3185    
3186       1.  Fixed up parsing colon suffixed keywords
3187       2.  Added the ability to shut the error messages up
3188       3.  Reimplemented the pretty printer and fixed up/improved
3189           the pretty printing of handle and -> types.
3190       4.  Fixed up generation of literal symbols in the nowhere tool.
3191       5.  Added some SML keywords to to sml.sty
3192    
3193    ----------------------------------------------------------------------
3194    Name: Matthias Blume
3195    Date: 2002/02/19 16:20:00 EST
3196    Tag: blume-20020219-cmffi
3197    Description:
3198    
3199    A wild mix of changes, some minor, some major:
3200    
3201    * All C FFI-related libraries are now anchored under $c:
3202        $/c.cm      --> $c/c.cm
3203        $/c-int.cm  --> $c/internals/c-int.cm
3204        $/memory.cm --> $c/memory/memory.cm
3205    
3206    * "make" tool (in CM) now treats its argument pathname slightly
3207      differently:
3208        1. If the native expansion is an absolute name, then before invoking
3209           the "make" command on it, CM will apply OS.Path.mkRelative
3210           (with relativeTo = OS.FileSys.getDir()) to it.
3211        2. The argument will be passed through to subsequent phases of CM
3212           processing without "going native".  In particular, if the argument
3213           was an anchored path, then "make" will not lose track of that anchor.
3214    
3215    * Compiler backends now "know" their respective C calling conventions
3216      instead of having to be told about it by ml-nlffigen.  This relieves
3217      ml-nlffigen from one of its burdens.
3218    
3219    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3220    
3221    * Export C_DEBUG and C_Debug from $c/c.cm.
3222    
3223    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3224      subtyping relationship between incomplete pointers and their complete
3225      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3226      with the convention of instantiating 'o with ('t, 'c) obj whenever
3227      the pointer target type is complete.  In the incomplete case, 'o
3228      will be instantiated with some "'c iobj" -- a type obtained by
3229      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3230    
3231      Operations that work on both incomplete and complete pointer types are
3232      typed as taking an 'o ptr while operations that require the target to
3233      be known are typed as taking some ('t, 'c) obj ptr.
3234    
3235      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3236      where void is an eqtype without any values.  This makes it possible
3237      to work on voidptr values using functions meant to operate on light
3238      incomplete pointers.
3239    
3240    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3241      been vastly simplified.
3242    
3243    ----------------------------------------------------------------------
3244    Name: Matthias Blume
3245    Date: 2002/02/19 10:48:00 EST
3246    Tag: blume-20020219-pqfix
3247    Description:
3248    
3249    Applied Chris Okasaki's bug fix for priority queues.
3250    
3251    ----------------------------------------------------------------------
3252    Name: Matthias Blume
3253    Date: 2002/02/15 17:05:00
3254    Tag: Release_110_39
3255    Description:
3256    
3257    Last-minute retagging is becoming a tradition... :-(
3258    
3259    This is the working release 110.39.
3260    
3261    ----------------------------------------------------------------------
3262    Name: Matthias Blume
3263    Date: 2002/02/15 16:00:00 EST
3264    Tag: Release_110_39-orig
3265    Description:
3266    
3267    Working release 110.39.  New bootfiles.
3268    
3269    (Update: There was a small bug in the installer so it wouldn't work
3270    with all shells.  So I retagged. -Matthias)
3271    
3272    ----------------------------------------------------------------------
3273    Name: Matthias Blume
3274    Date: 2002/02/15 14:17:00 EST
3275    Tag: blume-20020215-showbindings
3276    Description:
3277    
3278    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3279    the latter can be useful for exploring what bindings are available at
3280    the interactive prompt.  (The first function returns only the list
3281    of symbols that are really bound, the second prints those but also the
3282    ones that CM's autoloading mechanism knows about.)
3283    
3284    ----------------------------------------------------------------------
3285    Name: Matthias Blume
3286    Date: 2002/02/15 12:08:00 EST
3287    Tag: blume-20020215-iptrs
3288    Description:
3289    
3290    Two improvements to ml-nlffigen:
3291    
3292      1. Write files only if they do not exist or if their current contents
3293         do not coincide with what's being written.  (That is, avoid messing
3294         with the time stamps unless absolutely necessary.)
3295    
3296      2. Implement a "repository" mechanism for generated files related
3297         to "incomplete pointer types".   See the README file for details.
3298    
3299    ----------------------------------------------------------------------
3300    Name: Matthias Blume
3301    Date: 2002/02/14 11:50:00 EST
3302    Tag: blume-20020214-quote
3303    Description:
3304    
3305    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3306    because of the new and improved tag generation scheme.  (Thanks to Allen
3307    Leung for pointing it out.)
3308    
3309    ----------------------------------------------------------------------
3310    Name: Lal George
3311    Date: 2002/02/14 09:55:27 EST 2002
3312    Tag: george-20020214-isabelle-bug
3313    Description:
3314    
3315    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3316    of Isabelle on the x86.
3317    
3318    From Allen:
3319    -----------
3320     I've found the problem:
3321    
3322         in ra-core.sml, I use the counter "blocked" to keep track of the
3323         true number of elements in the freeze queue.  When the counter goes
3324         to zero, I skip examining the queue.  But I've messed up the
3325         bookkeeping in combine():
3326    
3327             else ();
3328             case !ucol of
3329               PSEUDO => (if !cntv > 0 then
3330                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3331                                        ^^^^^^^^^^^^^^^^^^^^^^^
3332                      moveu := mergeMoveList(!movev, !moveu)
3333                     )
3334                  else ();
3335    
3336         combine() is called to coalesce two nodes u and v.
3337         I think I was thinking that if the move counts of u and v are both
3338         greater than zero then after they are coalesced then one node is
3339         removed from the freeze queue.  Apparently I was thinking that
3340         both u and v are of low degree, but that's clearly not necessarily true.
3341    
3342    
3343    02/12/2002:
3344        Here's the patch.  HOL now compiles.
3345    
3346        I don't know how this impact on performance (compile
3347        time or runtime).  This bug caused the RA (especially on the x86)
3348        to go thru the potential spill phase when there are still nodes on the
3349        freeze queue.
3350    
3351    
3352    
3353    
3354    ----------------------------------------------------------------------
3355    Name: Matthias Blume
3356    Date: 2002/02/13 22:40:00 EST
3357    Tag: blume-20020213-fptr-rtti
3358    Description:
3359    
3360    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3361    updates.
3362    
3363    ----------------------------------------------------------------------
3364    Name: Matthias Blume
3365    Date: 2002/02/13 16:41:00 EST
3366    Tag: blume-20020213-cmlpq
3367    Description:
3368    
3369    Added new priority queue export symbols (which have just been added to
3370    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3371    not compile and the installer would choke.)
3372    
3373    ----------------------------------------------------------------------
3374    Name: Matthias Blume
3375    Date: 2002/02/13 16:15:00 EST
3376    Tag: blume-20020213-various
3377    Description:
3378    
3379    1. More tweaks to ml-nlffigen:
3380    
3381       - better internal datastructures (resulting in slight speedup)
3382       - "-match" option requires exact match
3383       - "localized" gensym counters (untagged structs/unions nested within
3384         other structs/unions or within typedefs get a fresh counter; their
3385         tag will be prefixed by a concatenation of their parents' tags)
3386       - bug fixes (related to calculation of transitive closure of types
3387         to be included in the output)
3388    
3389    2. Minor Basis updates:
3390    
3391       - added implementations for List.collate and Option.app
3392    
3393    ----------------------------------------------------------------------
3394    Name: Matthias Blume
3395    Date: 2002/02/11 15:55:00 EST
3396    Tag: blume-20020211-gensym
3397    Description:
3398    
3399    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3400    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3401    names (ML structure names that correspond to unnamed C structs, unions,
3402    and enums), so that separate runs of ml-nlffigen do not clash.
3403    
3404    ----------------------------------------------------------------------
3405    Name: Matthias Blume
3406    Date: 2002/02/11 12:05:00 EST
3407    Tag: blume-20020211-gensml
3408    Description:
3409    
3410    A quick fix for a problem with GenSML (in the pgraph-util library):
3411    Make generation of toplevel "local" optional.  (Strictly speaking,
3412    signature definitions within "local" are not legal SML.)
3413    
3414    Other than that: updates to INSTALL and cm/TODO.
3415    
3416    ----------------------------------------------------------------------
3417    Name: Matthias Blume
3418    Date: 2002/02/08 15:00:00 EST
3419    Tag: blume-20020208-uniquepid
3420    Description:
3421    
3422    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3423    
3424    1. The installer (config/install.sh) has gotten smarter:
3425    
3426         - Configuration options are a bit easier to specify now
3427           (in config/targets).
3428         - Bug in recognizing .tar.bz2 files fixed.
3429         - Installer automatically resolves dependencies between
3430           configuration options (e.g., if you ask for eXene, you will
3431           also get cml -- regardless whether you asked for it or not).
3432         - Installer can run in "quieter mode" by setting the environment
3433           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3434           "completely silent", though.
3435         - Build HashCons library as part of smlnj-lib.
3436    
3437    2. A new scheme for assigning persistent identifiers to compilation
3438       units (and, by extension, to types etc.) has been put into place.
3439       This fixes a long-standing bug where types and even dynamic values
3440       can get internally confused, thereby compromising type safety
3441       (abstraction) and dynamic correctness.  See
3442    
3443         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3444    
3445       for an example of how things could go wrong until now.
3446    
3447       The downside of the new scheme is that pids are not quite as
3448       persistent as they used to be: CM will generate a fresh pid
3449       for every compilation unit that it thinks it sees for the first
3450       time.  That means that if you compile starting from a clean, fresh
3451       source tree at two different times, you end up with different
3452       binaries.
3453    
3454       Cutoff recompilation, however, has not been compromised because
3455       CM keeps pid information in special caches between runs.
3456    
3457    ----------------------------------------------------------------------
3458    Name: Lal George
3459    Date: 2002/02/07 15:34:13 EST 2002
3460    Tag: <none>
3461    Description:
3462    
3463    Compilers that generate assembly code may produce  global labels
3464    whose value is resolved at link time. The various peephole optimization
3465    modules did not take this in account.
3466    
3467    TODO. The Labels.addrOf function should really return an option
3468    type so that clients are forced to deal with this issue, rather
3469    than an exception being raised.
3470    
3471    ----------------------------------------------------------------------
3472    Name: Lal George
3473    Date: 2002/02/06 13:55:02 EST
3474    Tag: george-20020206-ra-breakup
3475    Description:
3476    
3477    1. A bug fix from Allen.
3478    
3479        A typo causes extra fstp %st(0)'s to be generated at compensation
3480        edges, which might cause stack underflow traps at runtime.  This
3481        occurs in fft where there are extraneous fstps right before the 'into'
3482        trap instruction (in this case they are harmless since none of the
3483        integers overflow.)
3484    
3485    2. Pulled out various utility modules that were embedded in the modules
3486       of the register allocator. I need these modules for other purposes, but
3487       they are not complete enough to put into a library (just yet).
3488    ----------------------------------------------------------------------
3489    Name: Matthias Blume
3490    Date: 2002/01/31 16:05:00 EST
3491    Tag: blume-20020131-sparc-ccalls
3492    Description:
3493    
3494    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3495       of extra stack space by mistake.  Fixed.
3496    
3497    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3498    
3499    ----------------------------------------------------------------------
3500    Name: Allen Leung
3501    Date: 2002/01/30
3502    Tag: leunga-20020130-nowhere-bug-fix
3503    Description:
3504    
3505       MLRISC bug fixes:
3506       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3507       2. MachineInt.fromString was negating its value.
3508    
3509    ----------------------------------------------------------------------
3510    Name: Matthias Blume
3511    Date: 2002/01/29
3512    Tag: blume-20020129-INSTALL
3513    Description:
3514    
3515    - Added somewhat detailed installation instructions (file INSTALL).
3516    - Fixed curl-detection bug in config/install.sh.
3517    - It is now possible to select the URL getter using the URLGETTER
3518      environment variable:
3519    
3520          not set / "unknown"      --> automatic detection (script tries wget,
3521                                       curl, and lynx)
3522          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3523                                       how to properly invoke them)
3524          other                    --> use $URLGETTER directly, it must take
3525                                       precisely two command-line arguments
3526                                       (source URL and destination file name)
3527    
3528    ----------------------------------------------------------------------
3529    Name: Matthias Blume
3530    Date: 2002/01/28
3531    Tag: blume-20020128-sparc-ccalls
3532    Description:
3533    
3534    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3535    - Make use of the allocParam argument in sparc-c-calls.
3536    
3537    ----------------------------------------------------------------------
3538    Name: Matthias Blume
3539    Date: 2002/01/28
3540    Tag: blume-20020128-allocParam
3541    Description:
3542    
3543    John Reppy:  Changes c-calls API to accept client-callback for
3544    allocating extra stack space.
3545    me: Corresponding changes to mlriscGen (using a dummy argument that
3546        does not change the current behavior).
3547    
3548    ----------------------------------------------------------------------
3549    Name: Matthias Blume
3550    Date: 2002/01/28 12:00:00
3551    Tag: Release_110_38
3552    Description:
3553    
3554    This time for real!!!
3555    
3556    ----------------------------------------------------------------------
3557    Name: Matthias Blume
3558    Date: 2002/01/28 10:56:00 EST
3559    Tag: blume-20020128-retraction
3560    Description:
3561    
3562    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3563       with blume-Release_110_38-retracted.)
3564    
3565    1. Fixed a problem with incorrect rounding modes in real64.sml.
3566       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3567    
3568    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3569       and enums fixed.  The naming of corresponding ML identifiers should
3570       now be consistent again.
3571    
3572    ----------------------------------------------------------------------
3573    Name: Allen Leung
3574    Date: 2002/01/27
3575    Tag: leunga-20020127-nowhere
3576    Description:
3577    
3578       Added a target called nowhere in the configuration scripts.
3579       Enabling this will build the MLRISC 'nowhere' tool (for translating
3580       programs with where-clauses into legal SML code) during installation.
3581    
3582    ----------------------------------------------------------------------
3583    Name: Matthias Blume
3584    Date: 2002/01/25 21:27:00 EST
3585    Tag: blume-Release_110_38-retracted
3586    Description:
3587    
3588    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3589    
3590    README will be added later.
3591    
3592    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3593    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3594    
3595    ----------------------------------------------------------------------
3596    Name: Matthias Blume
3597    Date: 2002/01/25
3598    Tag: blume-20020125-ffi
3599    Description:
3600    
3601    A large number of tweaks and improvements to ml-nlffi-lib and
3602    ml-nlffigen:
3603    
3604       - ML represenation types have been streamlined
3605       - getter and setter functions work with concrete values, not abstract
3606         ones where possible
3607       - ml-nlffigen command line more flexible (see README file there)
3608       - some bugs have been fixed (hopefully)
3609    
3610    ----------------------------------------------------------------------
3611    Name: Lal George
3612    Date: 2002/01/24
3613    Tag: george-20020124-risc-ra-interface
3614    Description:
3615    
3616       There is a dramatic simplification in the interface to the
3617       register allocator for RISC architectures as a result of making
3618       parallel copy instructions explicit.
3619    
3620    ----------------------------------------------------------------------
3621    Name: Matthias Blume
3622    Date: 2002