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 972, Sun Oct 28 00:37:02 2001 UTC revision 1672, Thu Oct 28 15:44:02 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: 2004/10/28 10:45:00 CDT
18    Tag: Release_110_50
19    Description:
20    
21    New working version (110.50).  NEW BOOTFILES!
22    
23                 =====================
24    
25    Also:
26    
27      - Changend config/srcarchiveurl from a file just
28        containing the URL string into a file containing
29        shell script code.  The code has access to the $VERSION variable.
30      - Made corresponding changes to config/install.sh and config/unpack.
31      - Default contents of config/srcarchiveurl uses $VERSION and
32        normally does not have to be edited to reflect a version change.
33    
34      (As a result, a version change can be done by just editing
35       config/version, the rest is now automatic.)
36    
37    ----------------------------------------------------------------------
38    Name: Matthias Blume (blume (at) tti - c (dot) org)
39    Date: 2004/10/27 17:50:00 CDT
40    Tag: blume-20041027-btrace-msg
41    Description:
42    
43    BackTrace.monitor now also reports the source of the exception that
44    triggered the trace.
45    
46    ----------------------------------------------------------------------
47    Name: Matthias Blume (blume (at) tti - c (dot) org)
48    Date: 2004/10/27 17:20:00 CDT
49    Tag: blume-20041027-x86-c-calls
50    Description:
51    
52    This is the HISTORY entry for two earlier commits, both concerning
53    the x86 c-calls code in MLRISC:
54    
55      - added a missing LOAD in the code that deals with struct arguments
56      - made sure the caller does not add the wrong number of bytes to the
57        stack pointer after a call of a function returning a struct
58        (the callee already pops the implicit argument which points to
59         the space reserved for the result)
60    
61    ----------------------------------------------------------------------
62    Name: Allen Leung (leunga (at) reservoir (dot) com)
63    Date: 2004/10/24 14:00:00 EST
64    Tag: leunga-20041024-x86-gas-fucomip
65    Description:
66    
67       John discovered a bug in the syntax of fucomip.
68       The opcodes FU?COMIP? have been changed to
69    
70           fu?comip? %st(i), %st
71    
72    ----------------------------------------------------------------------
73    Name: Matthias Blume (blume (at) tti - c (dot) org)
74    Date: 2004/10/20 15:06:00 CDT
75    Tag: blume-20041020-standalone-backtrace
76    Description:
77    
78    Added a mechanism for getting back-trace information from standalone
79    programs.  Here is how it works:
80    
81       1. The part of the program from which you want to get backtrace
82          information (usually the whole program) should be wrapped with
83          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
84          main program could be modified from something like
85    
86             fun main (pgm, args) = ...
87    
88         to
89    
90            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
91    
92       2. To be able to access BackTrace.monitor, you have to add
93          library $smlnj-tdp/plugins.cm to the .cm file that contains your
94          main function.
95    
96       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
97          CM might have created in the past for your project).
98    
99       4. Build the system using this command line:
100    
101             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
102                      myprog.cm MyProg.main myprog
103    
104          instead of the usual
105    
106             ml-build myprog.cm MyProg.main myprog
107    
108    I changed a library name:
109    
110       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
111    
112    New libraries:
113    
114       $smlnj-tdp/back-trace.cm
115          -- when loaded causes the back-trace plugin to be installed
116       $smlnj-tdp/coverage.cm
117          -- when loaded causes the coverage plugin to be installed
118    
119    ----------------------------------------------------------------------
120    Name: Matthias Blume (blume (at) tti - c (dot) org)
121    Date: 2004/10/18 16:45:00 CDT
122    Tag: blume-20041018-groupowner
123    Description:
124    
125    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
126    
127    Eliminated group owner specs from .cm files throughout the source tree.
128    
129    ----------------------------------------------------------------------
130    Name: Matthias Blume (blume (at) tti - c (dot) org)
131    Date: 2004/10/15 15:45:00 CDT
132    Tag: blume-20041015-coverage
133    Description:
134    
135    * Test coverage tool added!
136    * Further reorganization of tracing-, debugging-, and profiling support:
137    
138        - moved original BTImp -- now called BackTrace -- into a separate
139          library called $/trace-debug-profile.cm
140        - eliminated all mentions of BTrace from SMLofNJ.Internals
141        - only the instrumentation mechanism is now left in the compiler proper
142        - BackTrace module is a plugin which is NOT plugged in by default
143        - Coverage module is another such plugin
144    
145    To get the benefits of any of these plugin modules, the code in
146    question must be compiled with tdp instrumentation turned on.  This
147    can be done by setting SMLofNJ.Internals.TDP.mode to true.
148    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
149    
150    Plugins are selected at link time.  (Pre-compiled instrumented code
151    can be re-loaded with different plugins in effect.)  When an
152    instrumented module is linked, whatever plugins are at that time
153    enabled will come into effect for that module.
154    
155    To enable the back-trace plugin, load library $/trace-debug-profile.cm
156    and invoke BackTrace.install() (e.g., from the interactive prompt).
157    To enable the coverage plugin, load the same library and invoke
158    Coverage.install().
159    
160    Back-traces are generated automatically on uncaught exceptions and
161    when the code in question explicitly invokes BackTrace.trigger().
162    
163    Coverage (and execution frequency-) information must be queried
164    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
165    
166    ----------------------------------------------------------------------
167    Name: Matthias Blume (blume (at) tti - c (dot) org)
168    Date: 2004/10/14 17:40:00 CDT
169    Tag: blume-20041014-tdp-core
170    Description:
171    
172    Snapshot of a significant overhaul of how the trace/debug/profile support
173    is hooked into the system (specifically: Core and SMLofNJ.Internals).
174    
175    ----------------------------------------------------------------------
176    Name: Matthias Blume (blume (at) tti - c (dot) org)
177    Date: 2004/10/13 16:34:00 CDT
178    Tag: blume-20041013-tdp
179    Description:
180    
181    Some rationalization of names:
182    
183       structure BTrace -> structure TDPInstrument
184       etc.
185    
186    This is is preparation of using the original back-trace
187    instrumentation for other purposes.  "TDP" stands for
188    Trace/Debug/Profile.
189    
190    The control flag controlling whether instrumentation is on or off is now
191    registered under a different name, so instead of running sml as
192    
193     sml -Cinstrument.btrace-mode=true
194    
195    one has to say
196    
197     sml -Ctdp.instrument=true
198    
199    ----------------------------------------------------------------------
200    Name: Matthias Blume (blume (at) tti - c (dot) org)
201    Date: 2004/10/11 16:37:00 CDT
202    Tag: blume-20041011-regions
203    Description:
204    
205    Made some minor modifications to elabcore.sml to have source regions
206    be propagated more tightly -- resulting in better (i.e., smaller)
207    regions being reported in error- and debug messages.
208    
209    ----------------------------------------------------------------------
210    Name: Matthias Blume (blume (at) tti - c (dot) org)
211    Date: 2004/10/08 22:50:00 CDT
212    Tag: blume-20041008-cmkw
213    Description:
214    
215    Fixed handling of keywords in .cm files: After seeing "is" the lexer
216    treats subsequent occurrences of "group", "library", "source", "is",
217    "*", and "-" as ordinary identifiers rather than keywords.
218    
219    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
220    is supposed to accept a tool argument called "source", but this did
221    not work because of the clash with the keyword.
222    
223    ----------------------------------------------------------------------
224    Name: Matthias Blume (blume (at) tti - c (dot) org)
225    Date: 2004/10/07 16:00:00 CDT
226    Tag: blume-20041007-cleanup
227    Description:
228    
229    Assorted cleanup work:
230    
231      - got rid of intstrmap in favor of using the library's
232        hash table implementation
233      - threw out most of the pathnames stuff, as it was not used anyway
234      - simplified tokentable implementation
235      - fixed some minor spelling errors
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume (blume (at) tti - c (dot) org)
239    Date: 2004/10/06 15:15:15 CDT
240    Tag: blume-20041006-handler
241    Description:
242    
243    Cleaned up the absyn to reflect the invariant that HANDLE always
244    carries a FNexp as part of the type definition.  This eliminates some
245    superfluous sanity checks at runtime down the road.
246    
247    Some minor cleanup of the btrace code.
248    
249    ----------------------------------------------------------------------
250    Name: Matthias Blume (blume (at) tti - c (dot) org)
251    Date: 2004/10/01 10:20:30 CDT
252    Tag: blume-20041001-slave
253    Description:
254    
255    Added hack to make slave mode work in the presence of the version
256    tool.  (Still, since the master does two passes over the code for
257    CMB.make, the release number gets bumped twice when slaves are
258    attached.  I don't know if this is worth fixing...)
259    
260    ----------------------------------------------------------------------
261    Name: Matthias Blume (blume (at) tti - c (dot) org)
262    Date: 2004/09/30 10:55:00 CDT
263    Tag: blume-20040930-version
264    Description:
265    
266    * Moved the "version" magic into its own little library under
267      src/system/smlnj/internal.  This avoids expensive reconstruction of
268      a stable src/compiler/core.cm.
269    
270    * At the same time, structure CompilerVersion is now known as
271      structure SMLNJVersion.
272    
273    * Arranged for the version tool to NOT kick in when rebuilding the system
274      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
275      Also, loading the versiontool does not work when rebuilding the system
276      because CM is not properly initialized at that time.
277    
278    ----------------------------------------------------------------------
279    Name: Matthias Blume (blume (at) tti - c (dot) org)
280    Date: 2004/09/29 14:00:00 CDT
281    Tag: blume-20040929-autoversion
282    Description:
283    
284    Implemented some CM magic to have
285        file src/compiler/TopLevel/main/version.sml
286    generated automagically.
287    The version is taken from two files: config/version and config/release.
288    The first is expected to contain a two-part version number such as 110.49.
289    The second should contain a single number, but it may be missing.
290    
291    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
292    time the version tool is loaded (which is the first time you say CMB.make),
293    then the tool will increment the value stored in config/release every
294    time CMB.make is invoked.
295    
296    The binfile format is now insensitive to anything beyond the first
297    two components of a version number, so bumping the release does not render
298    binfiles incompatible.  Auto-bumping can be used to keep track of versions
299    during development without invalidating existing binfiles.
300    
301    In any case, every CMB.make updates the date information in version.sml.
302    (This is the date that is printed in the banner.)
303    
304    ----------------------------------------------------------------------
305    Name: Matthias Blume (blume (at) tti - c (dot) org)
306    Date: 2004/09/28 10:53:00 CDT
307    Tag: blume-20040928-controls
308    Description:
309    
310    Some cleanup of the controls code.
311    
312    ----------------------------------------------------------------------
313    Name: Matthias Blume (blume (at) tti - c (dot) org)
314    Date: 2004/09/27 22:08:00 CDT
315    Tag: blume-20040927-controls
316    Description:
317    
318    Added two pieces of functionality to the Controls interface:
319    
320    1.   val save'restore: 'a control -> unit -> unit
321    
322       grabs the current value of the control in stage 1 and restores it
323       in stage 2.
324    
325    2.   val set' : 'a control * 'a -> unit -> unit
326    
327       stores the given value into the control in stage 2 (i.e., delayed)
328       but does all error checking in stage 1.
329       (This is for string controls that need to do parse their argument --
330        something that might fail.  In some cases, notably in CM, one
331        already knows the intended argument but wants to delay the actual
332        assignment until a time when error recovery would be more difficult.)
333    
334    Changed the handling of controls in tool arguments to classes "sml" and
335    "lazysml":
336       - use Controls.save'restore as a more robust way of restoring the
337         old value (in particular: without having to re-parse the string)
338       - use controls to handle the "overload" keyword in the init group
339         (I believe this change actually fixes a long-standing obscure bug.)
340    
341    ----------------------------------------------------------------------
342    Name: Matthias Blume (blume (at) tti - c (dot) org)
343    Date: 2004/09/27 17:00:00 CDT
344    Tag: blume-20040927-lazysml
345    Description:
346    
347    Added a new tool class called "lazysml" to CM's tool chest.  The only
348    difference to "sml" is that compilation is done with Control.lazysml
349    set to true.  A source of class "lazysml" is automatically recognized
350    by a file name suffix of ".lml".
351    
352    In addition to the above feature, the original class "sml" now also
353    supports a tool argument "lazy" which has the same effect.  As a
354    result, the following three lines are equivalent:
355    
356        foo.sml : lazysml
357        foo.sml : sml (lazy)
358        foo.sml (lazy)
359    
360    The setting goes into effect both during parsing and during
361    compilation.  The original setting is restored right after parsing and
362    after compilation, respectively.
363    
364    In addition to all the above, there is also a general mechanism to set
365    ANY of the "controls" that are available at the command line via
366    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
367    apply as well. (In fact, "lazy" is implemented as a special case of
368    the general mechanism.)
369    
370    The .cm file syntax uses a new keyword tool argument called "with".
371    There are several ways of indicating the desired settings:
372    
373        foo.sml (with:parser.quotations=true)
374        foo.sml (with:(name:parser.quotations value:true))
375        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
376        foo.sml (with:(name1=value1 name2=value2 ...))
377        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
378    
379        etc.
380    
381    Another possible abbreviation is to leave out the =v or value:v part
382    if the name refers to a boolean control (in which case the value is
383    taken to be true).  Thus, one could get lazy sml also by saying:
384    
385        foo.sml (with:parser.lazy-keyword=true)
386        foo.sml (with:parser.lazy-keyword)
387        foo.sml (with:(name:parser.lazy-keyword value:true))
388        foo.sml (with:(name:parser.lazy-keyword))
389    
390    ----------------------------------------------------------------------
391    Name: Matthias Blume (blume (at) tti - c (dot) org)
392    Date: 2004/09/24 16:22:00 CDT
393    Tag: blume-20040924-ppc-long-branch
394    Description:
395    
396    Turned message about "emiting long form of branch" off by default.
397    Added a control flag to turn it back on when desired.
398    
399    ----------------------------------------------------------------------
400    Name: Matthias Blume (blume (at) tti - c (dot) org)
401    Date: 2004/09/24 16:05:00 CDT
402    Tag: blume-20040924-rounding
403    Description:
404    
405    Applied patch for setting rounding modes under Mac OS X.  Thanks to
406    Melissa O'Neill for providing the code!
407    
408    ----------------------------------------------------------------------
409    Name: Matthias Blume (blume (at) tti - c (dot) org)
410    Date: 2004/09/23 17:30:00 CDT
411    Tag: blume-20040923-envvars
412    Description:
413    
414    1. Changed definition of type ControlRegistry.registry_tree to
415       include control_info (i.e., the name of the controlling environment
416       variable).
417    
418    2. Added command-line flags -e and -E to print the names of environment
419       variables that can be used to control internal settings.
420       (This uses the new API mentioned in 1.)
421    
422    ----------------------------------------------------------------------
423    Name: Matthias Blume (blume (at) tti - c (dot) org)
424    Date: 2004/09/13 16:50:00 CDT
425    Tag: Release_110_49
426    Description:
427    
428    New working version (110.49).  NEW BOOTFILES!
429    
430    ----------------------------------------------------------------------
431    Name: Matthias Blume
432    Date: 2004/09/13 16:20:00 CDT
433    Tag: blume-20040913-config-mlrisc
434    Description:
435    
436    Put target "mlrisc" back into the default list.
437    (There is no harm in having it, and some users have expressed their
438    wish to have "mlrisc" included by default.)
439    
440    ----------------------------------------------------------------------
441    Name: John Reppy
442    Date: 2004/09/13
443    Tag: jhr-20040913-signals
444    Description:
445    
446    Fixed the signal masking code to properly nest mask/unmask operations
447    on a per-signal basis.
448    
449    ----------------------------------------------------------------------
450    Name: Matthias Blume (blume (at) tti - c (dot) org)
451    Date: 2004/09/08 13:20:00 CDT
452    Tag: blume-20040908-heap-magic
453    Description:
454    
455    Bumped the heap macig number to 0x09082004 to account for the changed
456    layout of the ML frame under MacOS X.
457    
458    ----------------------------------------------------------------------
459    Name: Allen Leung (leunga (at) reservoir (dot) com)
460    Date: 2004/09/03 11:26:00 EST
461    Tag: leunga-20040903-cygwin-install
462    Description:
463    
464       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
465    The Cygwin runtime is turned on by setting the environment
466    variable SMLNJ_CYGWIN_RUNTIME to 1.
467    
468    ----------------------------------------------------------------------
469    Name: Matthias Blume (blume (at) tti - c (dot) org)
470    Date: 2004/08/31 17:15:00 CDT
471    Tag: blume-20040831-core
472    Description:
473    
474    Added some exports to src/compiler/core.cm upon request by J. Joemann.
475    
476    ----------------------------------------------------------------------
477    Name: Matthias Blume (blume (at) tti - c (dot) org)
478    Date: 2004/08/30 17:55:00 CDT
479    Tag: blume-20040830-installer
480    Description:
481    
482    Upon request by Johannes Joemann:
483    
484    - improved ML code of installer to fall back to coping when renaming
485      fails (i.e., when source and target are on different file systems);
486      the code compiles but has yet to be tested in anger
487    - removed mlrisc from list of default targets (config/targets)
488    
489    ----------------------------------------------------------------------
490    Name: Matthias Blume (blume (at) tti - c (dot) org)
491    Date: 2004/08/27 17:20:00 CDT
492    Tag: blume-20040827-ptreql
493    Description:
494    
495    Added ptreql primop to structure InlineT (upon request from Larry
496    Paulson).
497    
498    ----------------------------------------------------------------------
499    Name: Allen Leung (leunga (at) reservoir (dot) com)
500    Date: 2004/08/15 21:21:00 EST
501    Tag: leunga-110_48-udgraph
502    Description:
503    
504    Another bug fix from Carl Hauser:
505    
506    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
507    > 48c48
508    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
509    > e::L)
510    > ---
511    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
512    > Without this, any deletion of an edge in an undirected graph does severe
513    > violence to the graph.
514    
515    ----------------------------------------------------------------------
516    Name: Allen Leung (leunga (at) reservoir (dot) com)
517    Date: 2004/08/10 23:35:00 EST
518    Tag: leunga-110_48-ppc
519    Description:
520    
521       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
522    
523    ----------------------------------------------------------------------
524    Name: Matthias Blume (blume (at) tti - c (dot) org)
525    Date: 2004/08/10 12:00:00 CDT
526    Tag: Release_110_48
527    Description:
528    
529    New working version (110.48).  NEW BOOTFILES!
530    
531    ----------------------------------------------------------------------
532    Name: Allen Leung (leunga (at) reservoir (dot) com)
533    Date: 2004/08/09 12:21:00 EST
534    Tag: leunga-110_47-dijsktra
535    Description:
536    
537    Bug fix from Carl Hauser:
538    
539    single_source_shortest_paths in dijkstra.sml was observed to get wrong
540    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
541    
542    The problem is that following the expression A.update(dist,s,Num.zero)
543    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
544    
545    ----------------------------------------------------------------------
546    Name: Matthias Blume (blume (at) tti - c (dot) org)
547    Date: 2004/08/06 18:10:00 CDT
548    Tag: blume-20040806-cmdline
549    Description:
550    
551    Fiddled with handling of command-line options:
552    
553      * sml now quits after processing the command line
554        if -H, -S, -h<n>, or -s<n> appears as the last
555        command-line argument
556      * a new option -q terminates the session when encountered on
557        the command line; subsequent arguments will be ignored
558      * bug fixes: short (erroneous) arguments are no longer ignored
559        completely
560    
561    ----------------------------------------------------------------------
562    Name: Allen Leung (leunga (at) reservoir (dot) com)
563    Date: 2004/08/04 18:17:00 EST
564    Tag: leunga-110_47-ppc-ibm-asm
565    Description:
566    
567      - Added minimal IBM assembly syntax support for PowerPC.
568    
569      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
570        been altered in the runtime.  We need an automatic way to keep the file
571        in sync.
572    
573    ----------------------------------------------------------------------
574    Name: Matthias Blume (blume (at) tti - c (dot) org)
575    Date: 2004/08/04 14:00:00 CDT
576    Tag: Release_110_47
577    Description:
578    
579    New working version (110.47).  NEW BOOTFILES!
580    
581    ----------------------------------------------------------------------
582    Name: Matthias Blume (blume (at) tti - c (dot) org)
583    Date: 2004/08/03 14:25:00 CDT
584    Tag: blume-20040803-callingconv
585    Description:
586    
587    Added low-level support for choosing C calling conventions by
588    twiddling the type of rawccall.  (See
589    src/compiler/Semant/types/cproto.sml for details.)
590    
591    ----------------------------------------------------------------------
592    Name: Matthias Blume (blume (at) tti - c (dot) org)
593    Date: 2004/08/02 15:55:00 CDT
594    Tag: blume-20040802-backout
595    Description:
596    
597    Backed out of change to win32-filesys.c.  The earlier patch to
598    get_file_time caused CM to produce files with the wrong time stamp.
599    
600    ----------------------------------------------------------------------
601    Name: Matthias Blume (blume (at) tti - c (dot) org)
602    Date: 2004/08/02 14:45:00 CDT
603    Tag: blume-20040802-nlffi-win32
604    Description:
605    
606    Added NLFFI support for Win32, adapted from a patch provided by David
607    Hansel.  This is currently completely untested.  Also, the issue
608    concerning stdcall vs. ccall is still unresolved.
609    
610    ----------------------------------------------------------------------
611    Name: Matthias Blume (blume (at) tti - c (dot) org)
612    Date: 2004/07/30 17:55:00 CDT
613    Tag: blume-20040730-various
614    Description:
615    
616    Gearing up towards 110.47...
617    
618    - various minor bugfixes to ml-nlffigen
619    - a beginning of a manual for nlffi
620    
621    - eliminated 'export name=value' in config/install.sh as this does
622      not work with certain versions of /bin/sh
623      (Thanks to David King at Motorola for catching this.)
624    
625    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
626      - added a test for tm==NULL to gmtime.c and localtime.c
627      - applied patch for incorrect GetFileTime under win32
628      - toSeconds -> toMilliseconds in Win32/win32-process.sml
629    
630    ----------------------------------------------------------------------
631    Name: Matthias Blume (blume (at) tti - c (dot) org)
632    Date: 2004/07/21 18:20:00 CDT
633    Tag: blume-20040721-nlffigen
634    Description:
635    
636    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
637      for a typedef to an incomplete type, but leave out the "typ" member.
638      (This is just for consistency.)
639    - Started to produce what is supposed to become better (i.e., comprehensive)
640      documentation of what ml-nlffigen does and produces.
641    
642    ----------------------------------------------------------------------
643    Name: Matthias Blume (blume (at) tti - c (dot) org)
644    Date: 2004/07/14 16:25:00 CDT
645    Tag: blume-20040714-union
646    Description:
647    
648    Added C_UNION to c-calls/c-types.sml and updated the machinery
649    (ml-nlffigen, cproto.sml) that conveys C function interface
650    information to the code generator.
651    
652    However, the actual architecture-specific implementation of function
653    arguments and results that are C unions is still not implemented.
654    
655    ----------------------------------------------------------------------
656    Name: Allen Leung (leunga (at) reservoir (dot) com)
657    Date: 2004/07/14 14:38:00 EST
658    Tag: leunga-110_46_1-ppc-lwzu
659    Description:
660    
661        Added these instructions to the PowerPC architecture:
662        LBZU(X), LHZU(X), LWZU(X),
663        STWU(X), STFDU, STFSU
664    
665        etc...
666    
667        Note: I haven't added their instruction encoding into the description.
668    
669    ----------------------------------------------------------------------
670    Name: Allen Leung (leunga (at) reservoir (dot) com)
671    Date: 2004/07/13 15:04:00 EST
672    Tag: leunga-110_46_1-ppc-lwarx
673    Description:
674    
675       Added the two instructions LWARX and STWCX to the PowerPC
676    instruction set.
677    
678       A (untested) rewrite of loop-structure.sml.  The old version
679    is completely broken.
680    
681    ----------------------------------------------------------------------
682    Name: Matthias Blume (blume (at) tti - c (dot) org)
683    Date: 2004/07/13 13:50:00 CDT
684    Tag: blume-20040713-nlffi
685    Description:
686    
687    - use paramAlloc to report c-calls with too many arguments
688      (for PPC version where parameter area is pre-allocated)
689    - added ccall_maxargspace to machspec (to implement the above)
690    - made "make" commend in CM's "make" tool configurable
691    - added option (default: on) for passing the name of the SML/NJ's "bin"
692      directory to "make"; the call looks like this:
693    
694         make <options> SMLNJ_BINDIR=<dir> <target>
695    
696      This can be used by the Makefile to, e.g., pick the "right" version
697      of ml-nlffigen.
698    - minor code tweaks
699    
700    ----------------------------------------------------------------------
701    Name: Matthias Blume (blume (at) tti - c (dot) org)
702    Date: 2004/07/12 22:50:00 CDT
703    Tag: blume-110_46_1-macosx-nlffi
704    Description:
705    
706    NLFFI under Mac OS X now working (sort of).  This is largely untested,
707    though.
708    
709    Note:  1. You have to make a new, clean build of the runtime system.
710           2. There are new BOOTFILES, you have to use them!
711              (Doing the bootstrap process yourself would be *very* painful!
712               If you absolutely have to do it, build the system under
713               a different architecture and then cross-compile.)
714    
715    Version bumped to 110.46.1 to account for runtime data format changes.
716    
717    ----------------------------------------------------------------------
718    Name: Matthias Blume (blume (at) tti - c (dot) org)
719    Date: 2004/06/18 14:30:00 CDT
720    Tag: blume-20040618-unix
721    Description:
722    
723    Changed the implementation of structure Unix so that the same stream
724    is returned every time one of the {text,bin}{In,Out}streamOf functions
725    is invoked on the same proc.  This is not what the spec currently
726    says -- although IMO it arguably should.  (See discussion below.)
727    
728    ----------------------------------------------------------------------
729    Name: Matthias Blume (blume (at) tti - c (dot) org)
730    Date: 2004/06/17 18:15:00 CDT
731    Tag: Release_110_46
732    Description:
733    
734    New working version (110.46).  NEW BOOTFILES!
735    
736    ----------------------------------------------------------------------
737    Name: Matthias Blume (blume (at) tti - c (dot) org)
738    Date: 2004/06/17 17:20:00 CDT
739    Tag: blume-20040617-timer-unix
740    Description:
741    
742    Changed the interface of structures Timer and Unix to match the most
743    recent Basis spec.
744    
745    In the case of Unix there still seems to be an open/weird issue:
746    
747       The {text,bin}{In,Out}streamOf functions are supposed to create
748       fresh streams whenever they are called -- as opposed to have them
749       return the same stream every time.  This design is supposed to
750       prevent space leaks caused by proc values hanging on to streams.
751    
752       The reap function, on the other hand, is supposed to close the
753       streams.  This cannot be done without having a handle on the
754       stream in proc after all...
755    
756    I took the liberty to implement the following stopgap solution:
757    
758      The proc value hangs on to the most recently created stream(s).
759      Reap closes those.  If either or both of the two streams hadn't
760      been created at all yet, then reap will close the corresponding
761      file descriptors directly.
762    
763    PS: I don't understand the original space leak argument anymore.  If
764    a proc hangs on to the imperative stream, then I/O operations on those
765    will advance the state of the cached stream and avoid the space leak.
766    
767    ----------------------------------------------------------------------
768    Name: Matthias Blume (blume (at) tti - c (dot) org)
769    Date: 2004/05/28 16:45:00 CDT
770    Tag: blume-20040528-basis
771    Description:
772    
773    Added signature PACK_REAL and exported functor PrimIO.
774    
775    ----------------------------------------------------------------------
776    Name: Matthias Blume (blume (at) tti - c (dot) org)
777    Date: 2004/05/25 16:00:00 CDT
778    Tag: blume-20040525-group-owner
779    Description:
780    
781    CM now ignores (but still accepts) the "owner" information in group
782    descriptions.  The owner of a group is its next enclosing
783    library. Each group must have a unique owner.  (There is a virtual
784    "toplevel" library that owns groups which are not nested within a real
785    library.)  Previously, each group had to explicitly declare its owner,
786    and CM would check that such a declaration is correct.  The new scheme
787    is to have CM check that for each group there is precisely one owning
788    library.
789    
790    The advantage of the new scheme is that the programmer no longer needs
791    to maintain the somewhat annoying owner information.  The downside is
792    that CM cannot enforce the ownership rule across multiple runs of
793    CM.make.  Fortunately, enclosing the same group in two different
794    libraries A and B which are not part of the same program does not
795    cause real problems.
796    
797    ----------------------------------------------------------------------
798    Name: Matthias Blume (blume (at) tti - c (dot) org)
799    Date: 2004/05/20 16:00:00 CDT
800    Tag: blume-20040520-win32
801    Description:
802    
803    Made the win32 version work again.  (Strangely, a misplaced comma had
804    slipped into win32-process.c which prevented the runtime from being
805    compiled correctly.)
806    
807    Also, included a minor addition to ml-build.bat analogous to what was
808    done in blume-20040519-ml-build.
809    
810    ----------------------------------------------------------------------
811    Name: Matthias Blume (blume (at) tti - c (dot) org)
812    Date: 2004/05/19 22:10:00 CDT
813    Tag: blume-20040519-ml-build
814    Description:
815    
816    Arranged for ml-build to clean up after itself a little bit better.
817    The script generates a temporary SML source file and compiles it using
818    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
819    gets rid of those at the end, so they don't accumulate under .cm.
820    
821    This required a minor change to install.sh because the name of the
822    metadata directory (default: .cm) is actually configurable at
823    installation time.
824    
825    ----------------------------------------------------------------------
826    Name: Matthias Blume (blume (at) tti - c (dot) org)
827    Date: 2004/05/18 15:50:00 CDT
828    Tag: blume-20040518-mkreader
829    Description:
830    
831    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
832    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
833    
834    ----------------------------------------------------------------------
835    Name: Matthias Blume (blume (at) tti - c (dot) org)
836    Date: 2004/05/11 14:35:00 CDT
837    Tag: blume-20040511-win32sock
838    Description:
839    
840    Added previously missing support for many socket-related functions
841    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
842    for the voluminous patch!
843    
844    (I have not tested this patch under win32 yet.)
845    
846    Here is David's e-mail:
847    
848    Hi,
849    
850    Attached to this email you find a diff against sml/nj 110.45
851    that will enable socket support under Windows.
852    
853    To apply the patch (using unix or cygwin)
854    1) gunzip runtime.diff.gz
855    2) "cd" into "src/runtime" in the source tree of a fresh
856       110.45 installation.
857    3) patch -p 1 < [your/path/to]runtime.diff
858    
859    The code compiles fine but has NOT yet been extensively tested.
860    I only ran a few tests for basic socket client functionality
861    (which worked fine).  Especially the functions that use ioctl
862    are not tested at all and might not work (see below).
863    
864    I implemented this since we want to move to a newer version of sml/nj
865    but need socket support in order to use it.  This is the first time I
866    even had a look at the sml/nj source,  so please review my changes
867    before making this part of the distribution!  Here are a few issues
868    that I think might be better for someone to solve who is more
869    familiar with the sml/nj source (and socket programming):
870    
871    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
872      exception since I could not figure out what the windows equivalent
873      of these functions is
874    
875    - In sockets-osdep.h there are a some #include statements that are
876      only used in a few files that include sockets-osdep.h
877    
878    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
879      does not process its return value since I don't know how to
880      report an error upwards.
881    
882    - It would probably be good to have a call to WSACleanup() when
883      the library is unloaded (if there is such a possibility).
884      Otherwise I think Windows will take care of this automatically
885      when the process finishes.
886    
887    - I used ioctlsocket() as a replacement for ioctl() but I have
888      no idea if that is actually the proper replacement on Windows.
889    
890    - All these issues are marked in the code by "FIXME" comments.
891    
892    We use sml/nj extensively in our products and are quite happy
893    with it.  I hope this contribution will help you.
894    
895    Keep up the good work!
896    
897    David
898    
899    ----------------------------------------------------------------------
900    Name: Matthias Blume (blume (at) tti - c (dot) org)
901    Date: 2004/05/11 14:20:00 CDT
902    Tag: blume-20040511-installml
903    Description:
904    
905    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
906    the patch.)
907    
908    ----------------------------------------------------------------------
909    Name: Matthias Blume (blume (at) tti - c (dot) org)
910    Date: 2004/05/11 14:05:00 CDT
911    Tag: blume-20040511-nlffi-netbsd
912    Description:
913    
914    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
915    the patch.)
916    
917    ----------------------------------------------------------------------
918    Name: Matthias Blume (blume (at) tti - c (dot) org)
919    Date: 2004/05/11 12:05:00 CDT
920    Tag: blume-20040511-exports
921    Description:
922    
923    As per request by Adam Chlipala <adam@hcoop.net>, extended various
924    export lists in compiler-related .cm-files.
925    
926    ----------------------------------------------------------------------
927    Name: Matthias Blume (blume (at) tti - c (dot) org)
928    Date: 2004/05/11 11:35:00 CDT
929    Tag: blume-20040511-allsource
930    Description:
931    
932    The installer now honors the "src-smlnj" target again, although its meaning
933    has changed from "all sources required for the compiler" to "all sources
934    the installer knows about".  In other words, if you enable "src-smlnj"
935    in the "targets" file, then the installer will pull in sources for
936    everything.  (Notice that this refers to source code only.  Compiled
937    code is still only installed for modules that were requested explicitly
938    or which are required for other modules that were requested explicitly.)
939    
940    ----------------------------------------------------------------------
941    Name: Matthias Blume (blume (at) tti - c (dot) org)
942    Date: 2004/04/23 17:40:00 CDT
943    Tag: blume-20040423-ieee-scan
944    Description:
945    
946    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
947    in the exponent calculation we get INF or ZERO (depending on the mantissa
948    and the sign of the exponent).
949    
950    ----------------------------------------------------------------------
951    Name: Matthias Blume (blume (at) tti - c (dot) org)
952    Date: 2004/04/23 10:40:00 CDT
953    Tag: blume-20040423-ml-build
954    Description:
955    
956    The ml-build script now terminates with a non-0 status when something
957    goes wrong.
958    
959    ----------------------------------------------------------------------
960    Name: Matthias Blume (blume (at) tti - c (dot) org)
961    Date: 2004/04/22 16:35:00 CDT
962    Tag: blume-20040422-Option
963    Description:
964    
965    Made exception Option to be the same as exception Option.Option
966    (as it should be).
967    
968    ----------------------------------------------------------------------
969    Name: Allen Leung (leunga (at) reservoir (dot) com)
970    Date: 2004/03/19 14:40:00 EST
971    Tag: leunga-20040319-cygwin-nlffi
972    Description:
973    
974    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
975    of SML/NJ.  The problem is that
976    
977       lib = dlopen(NULL, ...)
978       f   = dlsym(lib, "malloc");
979    
980    does not work on Windows unless we explicitly export symbols
981    such as 'malloc' during linking.  We fixed this by explicitly
982    exporting the required symbols with the magic gcc incantation:
983    
984        -Wl,--export-all cygwin.def
985    
986    where cygwin.def is a file containing all the symbols that
987    we wish to export.
988    
989    I suspect this is a Windows problem and we'll have to
990    do the same (somehow with windows compilers) when
991    we build the native win32 version with the system
992    calls LoadLibrary/GetProcAddress.
993    
994    ----------------------------------------------------------------------
995    Name: Matthias Blume (blume (at) tti - c (dot) org)
996    Date: 2004/03/04 16:35:00 CST
997    Tag: blume-20040304-intinf-fmt
998    Description:
999    
1000    Fixed problem with IntInf.fmt (sign would show up on the right instead
1001    of on the left for BIN, OCT, and HEX).
1002    
1003    ----------------------------------------------------------------------
1004    Name: Matthias Blume (blume (at) tti - c (dot) org)
1005    Date: 2004/03/04 11:25:00 CST
1006    Tag: blume-20040304-symlinks
1007    Description:
1008    
1009    Fixed problem with installer script (unix only) where bin/ml-yacc and
1010    friends pointed (via symlinks) to absolute locations instead of just
1011    .run-sml.  This was reported by Vesa A Norrman.
1012    
1013    ----------------------------------------------------------------------
1014    Name: Matthias Blume (blume (at) tti - c (dot) org)
1015    Date: 2004/02/13 14:50:00 CST
1016    Tag: Release_110_45
1017    Description:
1018    
1019    New working version (110.45). New bootfiles.
1020    
1021    ----------------------------------------------------------------------
1022    Name: Matthias Blume (blume (at) tti - c (dot) org)
1023    Date: 2004/01/26 15:15:15 CST
1024    Tag: blume-20040126-toplevel
1025    Description:
1026    
1027    Improved handling of exceptions at the interactive toplevel.
1028    
1029    ----------------------------------------------------------------------
1030    Name: Matthias Blume (blume (at) tti - c (dot) org)
1031    Date: 2004/01/26 11:25:00
1032    Tag: blume-20040126-app
1033    Description:
1034    
1035    Type of top-level "app" corrected.
1036    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1037    
1038    ----------------------------------------------------------------------
1039    Name: Matthias Blume (blume (at) tti - c (dot) org)
1040    Date: 2003/11/18 17:10 CST
1041    Tag: blume-20031118-basis-fiddle
1042    Description:
1043    
1044    - changed Timer interface to what might become the spec
1045    - POSIX_FLAGS -> BIT_FLAGS according to spec
1046    - some other minor discrepancies wrt. spec eliminated
1047    
1048    ----------------------------------------------------------------------
1049    Name: Matthias Blume (blume (at) tti - c (dot) org)
1050    Date: 2003/11/06 12:00:00 CST
1051    Tag: Release_110_44
1052    Description:
1053    
1054    New working version (110.44). New bootfiles.
1055    
1056    ----------------------------------------------------------------------
1057    Name: Matthias Blume (blume (at) tti - c (dot) org)
1058    Date: 2003/11/04 11:50:00 CST
1059    Tag: blume-20031104-move-libraries
1060    Description:
1061    
1062    Eliminated the "dont_move_libraries" directive in config/targets.
1063    (The mechanism was broken and could not be fixed easily.  Moreover,
1064    there does not seem to be any reason not to move all libraries into
1065    lib during installation.  I originally implemented this directive as a
1066    backward-compatibility feature when I first introduced the new CM.
1067    Now that things have been stable for a long time and going back to the
1068    old CM is not an option, there is no reason to keep it around.)
1069    
1070    ----------------------------------------------------------------------
1071    Name: Matthias Blume (blume (at) tti - c (dot) org)
1072    Date: 2003/11/03 16:00:00 CST
1073    Tag: blume-20031103-installdir
1074    Description:
1075    
1076    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1077    Richards for pointing out the problem and providing the solution.)
1078    
1079    ----------------------------------------------------------------------
1080    Name: Matthias Blume (blume (at) tti - c (dot) org)
1081    Date: 2003/10/01 17:05:00 CDT
1082    Tag: blume-20031001-lal-mlrisc
1083    Description:
1084    
1085    MLRISC bug fix from Lal.
1086    
1087    ----------------------------------------------------------------------
1088    Name: Matthias Blume (blume (at) tti - c (dot) org)
1089    Date: 2003/09/30 16:10:00 CDT
1090    Tag: blume-20030930-primio-bat
1091    Description:
1092    
1093    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1094    2. Improved .bat files (for Win32 port) to make things work under Win95.
1095       (thanks to Aaron S. Hawley for this one)
1096    
1097    ----------------------------------------------------------------------
1098    Name: Matthias Blume (blume (at) tti - c (dot) org)
1099    Date: 2003/09/26 16:05:00 CDT
1100    Tag: blume-20030926-wrappriv
1101    Description:
1102    
1103    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1104    
1105    ----------------------------------------------------------------------
1106    Name: Matthias Blume (blume (at) tti - c (dot) org)
1107    Date: 2003/09/26 15:00:00 CDT
1108    Tag: blume-20030926-110_43_3
1109    Description:
1110    
1111    - additional cleanup
1112    - version number bump, NEW BOOTFILES
1113    
1114    ----------------------------------------------------------------------
1115    Name: Matthias Blume (blume (at) tti - c (dot) org)
1116    Date: 2003/09/26 12:00:00 CDT
1117    Tag: blume-20030926-ppautoload
1118    Description:
1119    
1120    I modified the read-eval-print loop so that the autoloader gets
1121    invoked whenever the prettyprinter tries to look up a symbol that
1122    is not currently defined in the toplevel environment but which
1123    appears in CM's autoload registry.  As a result, we see far fewer of
1124    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1125    
1126    In addition to this I tried to clean up some pieces of the Basis
1127    implementation (e.g., Socket, Word8Array) in order to prevent other
1128    instances of these ?.Foo.Bar.xxx names from being printed.
1129    
1130    The mechanism that picks names for types still needs some work, though.
1131    (Right now it seems that if there is a type A.t which is defined to
1132    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1133    "?.B.u" although the perhaps more sensible solution would be to use
1134    "A.t" in this case.  In other words, the prettyprinter should follow
1135    a chain of DEFtycs not farther than there are corresponding toplevel
1136    names in the current environment.)
1137    
1138    ----------------------------------------------------------------------
1139    Name: Matthias Blume (blume (at) tti - c (dot) org)
1140    Date: 2003/09/24 16:31:00 CDT
1141    Tag: blume-20030924-installer
1142    Description:
1143    
1144    Another installer tweak:  All the ML code for the installer is now
1145    compiled during CMB.make and put into a little library called
1146    $smlnj/installer.cm.  The installation then simply invokes
1147    
1148       sml -m $smlnj/installer.cm
1149    
1150    and everything happens automagically.
1151    
1152    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1153    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1154            and BINDIR.
1155    
1156    The new scheme guarantees that the ML code responsible for the installation
1157    is in sync with the APIs of the main system.  Also, the installer is
1158    somewhat faster because the installer script is precompiled.
1159    
1160    ----------------------------------------------------------------------
1161    Name: Matthias Blume (blume (at) tti - c (dot) org)
1162    Date: 2003/09/24 15:35:00 CDT
1163    Tag: blume-20030924-synsock
1164    Description:
1165    
1166    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1167    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1168    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1169    implementations of non-blocking operations from CML's Socket
1170    structure.
1171    
1172    ----------------------------------------------------------------------
1173    Name: Matthias Blume (blume (at) tti - c (dot) org)
1174    Date: 2003/09/24 15:10:05 CDT
1175    Tag: blume-20030924-sockets
1176    Description:
1177    
1178    1. Fixed SOCKET API and implementation to match Basis spec.
1179       This required changing the internal representation of sockets to one
1180       that remembers (for each socket file descriptor) whether it is currently
1181       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1182       call is made only if the state actually needs to change).
1183    
1184    2. OS-specific details of sockets were moved into separate files, thus
1185       making it possible to unify the bulk of the socket implementations
1186       between Unix and Win32.
1187    
1188    3. CML's socket API changed accordingly.
1189       (Note that we need to remove non-blocking functions from this API
1190       since they are redundant in the case of CML!)
1191    
1192    4. CML's socket implementation now makes use of non-blocking functions
1193       provided by Basis, thus removing all OS-dependent code from this part
1194       of CML.
1195    
1196    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1197    
1198    ----------------------------------------------------------------------
1199    Name: Matthias Blume (blume (at) tti - c (dot) org)
1200    Date: 2003/09/22 12:10:00 CDT
1201    Tag: blume-20030922-110_43_2
1202    Description:
1203    
1204    Made a new interim version and bootfiles for developer's bootstrapping
1205    convenience.
1206    
1207       110.43.2 -- NEW BOOTFILES
1208    
1209    ----------------------------------------------------------------------
1210    Name: Matthias Blume (blume (at) tti - c (dot) org)
1211    Date: 2003/09/19 15:55:00 CDT
1212    Tag: blume-20030919-cmdir
1213    Description:
1214    
1215    1. new-install.sh -> install.sh
1216    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1217    3. tweaked installer so that another name instead of .cm can be chosen
1218       at install time (by setting the CM_DIR_ARC environment variable
1219       during installation); once installation is complete, the name is
1220       fixed
1221    
1222    ----------------------------------------------------------------------
1223    Name: Matthias Blume (blume (at) tti - c (dot) org)
1224    Date: 2003/09/18 16:00:00 CDT
1225    Tag: blume-20030918-110_43_1
1226    Description:
1227    
1228    Made a new interim version and bootfiles for developer's bootstrapping
1229    convenience.
1230    
1231       110.43.1 -- NEW BOOTFILES
1232    
1233    ----------------------------------------------------------------------
1234    Name: Matthias Blume (blume (at) tti - c (dot) org)
1235    Date: 2003/09/18 15:20:00 CDT
1236    Tag: blume-20030918-misc
1237    Description:
1238    
1239    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1240       this seems to be controversial at the moment) and used those in
1241       Posix.ProcEnv.times.
1242    
1243    2. Added Time.{from,to}Nanoseconds to Time.
1244    
1245    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1246       For example, fromLargeInt never needs to look at more than 3 "big
1247       digits" to get its 53 bits of precision.
1248    
1249    ----------------------------------------------------------------------
1250    Name: Matthias Blume (blume (at) tti - c (dot) org)
1251    Date: 2003/09/17 16:30:00 CDT
1252    Tag: blume-20030917-real32-slices
1253    Description:
1254    
1255    Added an entry to the primitive environment
1256    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1257    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1258    
1259    Removed some of the "magic" constants in real64.sml and replaced them
1260    with code that generates these values from their corresponding
1261    integer counterparts.
1262    
1263    Made all(?) the slice-related changes to the Basis and made everything
1264    compile again...
1265    
1266    ----------------------------------------------------------------------
1267    Name: Matthias Blume (blume (at) tti - c (dot) org)
1268    Date: 2003/09/15 17:45:00 CDT
1269    Tag: blume-20030915-rbase
1270    Description:
1271    
1272    Fixed bug in Real.fromLargeInt.
1273    
1274    ----------------------------------------------------------------------
1275    Name: Matthias Blume (blume (at) tti - c (dot) org)
1276    Date: 2003/09/13 18:11:00 CDT
1277    Tag: blume-20030913-libinstall
1278    Description:
1279    
1280    Minor bugfix in config/libinstall (set anchor with path to
1281    standalone tool after installing it, otherwise libraries that
1282    need ml-lex or ml-yacc won't compile the first time the installer
1283    runs).
1284    
1285    ----------------------------------------------------------------------
1286    Name: Matthias Blume (blume (at) tti - c (dot) org)
1287    Date: 2003/09/12 11:45:00 CDT
1288    Tag: blume-20030912-various
1289    Description:
1290    
1291     - fixed bug in Real.toLargeInt
1292     - fixed bug in Posix.ProcEnv.times
1293     - changed inputLine functions to return an option
1294     - minor installer improvements / bugfixes
1295     - changed default @SMLalloc parameter for x86/celeron to 64k
1296    
1297    ----------------------------------------------------------------------
1298    Name: Matthias Blume (blume (at) tti - c (dot) org)
1299    Date: 2003/09/09 22:00:00 CDT
1300    Tag: Release_110_43
1301    Description:
1302    
1303        New working release 110.43.  New bootfiles.
1304    
1305    ----------------------------------------------------------------------
1306    Name: Matthias Blume (blume (at) tti - c (dot) org)
1307    Date: 2003/09/09 19:20:00 CDT
1308    Tag: blume-20030909-installer
1309    Description:
1310    
1311    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1312    Modified config/install.bat to take advantage of it.  Also modified
1313    config/install.sh (and called it config/new-install.sh) to take advantage
1314    of it on Unix systems.  (The SML code is (supposed to be) platform-
1315    independent.)
1316    
1317    The installer can now install everything under Win32
1318    as well as under *nix as long as it compiles.
1319    
1320    Other changes:
1321    
1322      - made CML compile again under Win32
1323      - made eXene compile under Win32 (by providing a fake structure UnixSock
1324        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1325      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1326        same as type int;  under Win32 it isn't
1327      - fixed some slice-related problems in the win32-specific parts of CML
1328      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1329        (under Win32, the volume name is case-insensitive, and the
1330        OS.Path code compares volume names for equality)
1331    
1332    ----------------------------------------------------------------------
1333    Name: Matthias Blume (blume (at) tti - c (dot) org)
1334    Date: 2003/09/08 11:55:00 CDT
1335    Tag: blume-20030908-fullpath
1336    Description:
1337    
1338    Made Win32 version of OS.FileSys.fullPath return current directory
1339    when given an empty string.  This is what the spec says, and incidentally,
1340    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1341    cases when presented with the name of a non-existing .cm file.)
1342    
1343    ----------------------------------------------------------------------
1344    Name: Matthias Blume (blume (at) tti - c (dot) org)
1345    Date: 2003/09/04 16:30:00 CDT
1346    Tag: blume-20030905-slices-etc
1347    Description:
1348    
1349      1. Changed interface to vectors and arrays in Basis to match
1350         (draft) Basis spec.
1351      2. Added signatures and implementations of slices according to
1352         Basis spec.
1353      3. Edited source code throughout the system to make it compile again
1354         under 1. and 2.  (In some cases code had to be added to have it
1355         match the new signatures.)
1356      4. MLRISC should be backward-compatible: the copies of the originals
1357         of files that needed to change under 3. were retained, the .cm files
1358         check the compiler version number and use old versions when
1359         appropriate.
1360      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1361         dirstream -> string option (in accordance with Basis spec).
1362      6. When generating code that counts lines, ml-lex used function
1363         CharVector.foldli, taking advantage of its old interface.
1364         This has been replaced with the corresponding code from
1365         CharVectorSlice.  (html-lex must be re-lexed!)
1366      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1367         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1368      8. Removed temporary additions (fromInternal, toInternal) from the
1369         (now obsolete) IntInf in smlnj-lib/Util.
1370      9. Cleaned up structure Byte.
1371     10. Added localOffset, scan, and fromString to Date (according to spec).
1372         Cleaned/corrected implementation of Date.
1373         (Still need to check for correctness; implement better canonicalizeDate.)
1374     11. Added "scan" to signature IEEE_REAL.
1375     12. Some improvements to IntInf [in particular: efficiency-hack for
1376         mod and rem when second operand is 2 (for parity checks).]
1377     13. Changed representation of type Time.time, using a single IntInf.int
1378         value counting microseconds.  This considerably simplified the
1379         implementation of structure Time.  We now support negative time
1380         values; scan and fromString handle signs.
1381     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1382         ArraySlice).
1383    
1384    ----------------------------------------------------------------------
1385    Name: Matthias Blume (blume (at) tti - c (dot) org)
1386    Date: 2003/08/28 17:00:00 CDT
1387    Tag: blume-20030828-intinf
1388    Description:
1389    
1390    This is a major update which comes with a version number bump
1391    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1392    and an implementation of IntInf in the Basis.
1393    
1394    There are a fairly large number of related changes and updates throughout
1395    the system:
1396    
1397    Basis:
1398            - Implemented IntInf.
1399            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1400            - Added some missing Real64 operations, most notably Real.toLargeInt.
1401            - Added FixedInt as a synonym for Int32.
1402    
1403    compiler:
1404            * Added support for a built-in intinf type.
1405              - literals
1406              - pattern matching
1407              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1408              - overloading on literals and operations
1409    
1410            This required adding a primitive type intinf, some additional
1411            primops, and implementations for several non-trivial intinf
1412            operations in Core.  (The intinf type is completely abstract
1413            to the compiler; all operations get delegated back to the Core.)
1414    
1415            * Intinf equality is handled by polyequal.  However, the compiler
1416            does not print its usual warning in this case (since polyequal
1417            is the right thing to do there).
1418    
1419            * Improved the organization of structure InlineT.
1420    
1421            * A word about conversion primops:
1422            If conversions involving intinf do not cancel out during
1423            CPS contract, then the compiler must insert calls to Core functions.
1424            Since all core access must be resolved already during the FLINT
1425            translate phase, it would be too late a the time of CPS contract
1426            to add new Core calls.  For this reason, conversion primops
1427            for intinf carry two arguments:  1. the numeric argument that
1428            they are supposed to convert, and 2. the Core function that
1429            can help with this conversion if necessary.  If CPS contract
1430            eliminates a primop, then the associated Core function becomes
1431            dead and goes away.  Intinf conversion primops that do not get
1432            eliminated by CPS contract get rewritten into calls of their
1433            core functions by a separate, new phase.
1434    
1435    interactive system:
1436            - Control.Print.intinfDepth controls max length of intinf constants
1437            being printed. (Analogous to Control.Print.stringDepth.)
1438            - Cleanup in printutil and pputil: got rid of unused stuff and
1439            duplicates; replaced some of the code with code that makes better
1440            use of library functionality.
1441    
1442    CM:
1443            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1444            are no longer silent.
1445    
1446    CKIT:
1447            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1448            in favor of LargeInt -- which is now the same as IntInf.
1449            CKIT-knowledgable people should check whether this is what's
1450            intended and otherwise change things back to using Int32 or
1451            FixedInt.
1452    
1453    Throughout the code:
1454            Started using IntInf.int literals and built-in operations
1455            (e.g., comparison with 0) where this seems appropriate.
1456    
1457    
1458    ----------------------------------------------------------------------
1459    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1460    Date: 2003/08/13 11:36:00 CDT
1461    Tag: dbm-20030813-mcz-merge1
1462    Description:
1463    
1464    Merging changes from the mcz-branch development branch into trunk.
1465    These changes involve replacement of the emulated old prettyprinter
1466    interface with direct use of the SML/NJ Lib PP library, and fixing
1467    of a couple of bugs (895, 1186) relating to error messages.  A new
1468    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1469    has been added.
1470    
1471    ----------------------------------------------------------------------
1472    Name: Matthias Blume (blume (at) tti - c (dot) org)
1473    Date: 2003/08/11 15:45:00 CDT
1474    Tag: blume-20030811-windows
1475    Description:
1476    
1477    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1478    
1479        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1480    
1481    This patch restores SML/NJ's ability to run under win32.  There are a
1482    number of changes, including fixes for several bugs that had gone
1483    unnoticed until now:
1484    
1485      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1486        trying to run the win32 version from within cygwin.)
1487    
1488      - There are a number of simple .bat scripts that substitute for their
1489        corresponding Unix shell-scripts.  (See below.)
1490    
1491      - The internals of ml-build have been modified slightly.  The main
1492        difference is that instead of calling ".link-sml" (or link-sml.bat)
1493        using OS.Process.system, the ML process delegates this task back
1494        to the script.  Otherwise problems arise in mixed environments such
1495        as Cygwin where scripts look and work like Unix scripts, but
1496        where OS.Process.system cannot run them.
1497    
1498      - In CM, the srcpath pickler used native pathname syntax -- which
1499        is incorrect in the case of cross-compilation.  The new pickle format
1500        is independent of platform-specific naming conventions.
1501    
1502      - Path configuration files (such as lib/pathconfig) can now choose
1503        between native and standard syntax.  Placing a line of the form
1504    
1505            standard!
1506    
1507        into the file causes all subsequent paths to be interpreted using
1508        CM standard pathname syntax (= Unix conventions); a line
1509    
1510            native!
1511    
1512        switches back to native style.  This was needed so that
1513        path config files can be written portably, see src/system/pathconfig.
1514    
1515      - Runtime system:
1516    
1517          - win32-filesys.c:  get_file_time and set_file_time now
1518            access modification time, not creation time.
1519    
1520          - I/O code made aware of new array representation.
1521    
1522          - Bug fixes in X86.prim.masm.
1523    
1524       - src/system/makeml made aware of win32. (For use under cygwin
1525            and other Unix-environments for windows.)
1526    
1527       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1528         which caused BinIO.inputAll to fail consistently.
1529    
1530    .bat scripts:
1531    
1532       Windows .bat scripts assume that SMLNJ_HOME is defined.
1533    
1534       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1535         applications (sml, ml-yacc, ml-lex).
1536       - ml-build.bat: analogous to ml-build.
1537       - config\install.bat:  Analogous to config/install.sh.  This requires
1538         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1539         (nmake etc. must be on the path, and vcvars32 must have been run.)
1540         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1541         and the bootfile hierarchy must have been unpacked under
1542         sml.boot.x86-win32.
1543         The script is very primitive and does a poor job at error checking.
1544         It only installs the base system, ml-lex, and ml-yacc.  No other
1545         libraries are being installed (i.e., you get only those that
1546         are part of the compiler.)
1547       - link-sml.bat:  analogous to .link-sml, but not currently used
1548    
1549    Unrelated bug fixes:
1550    
1551      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1552        types.
1553      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1554        otherwise.)
1555    
1556    ----------------------------------------------------------------------
1557    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1558    Date: 2003/06/17
1559    Tag: macqueen-20030617-bug895
1560    Description:
1561    
1562    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1563    Tag will be used for new development branch (mcz-branch) for
1564    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1565    
1566    ----------------------------------------------------------------------
1567    Name: Matthias Blume (blume (at) tti - c (dot) org)
1568    Date: 2003/05/27 16:55:00 CDT
1569    Tag: blume-20030527-polyeq
1570    Description:
1571    
1572    Tried to eliminated most cases of polymorphic equality.
1573    
1574    ----------------------------------------------------------------------
1575    Name: Matthias Blume (blume (at) tti - c (dot) org)
1576    Date: 2003/05/21 17:45:00 CDT
1577    Tag: blume-20030517-complete
1578    Description:
1579    
1580    Two changes:
1581    
1582      1. Added a flag for controlling whether non-exhaustive bindings will
1583         be treated as errors (default is false).
1584      2. Cleaned up the *entire* source tree so that CMB.make goes through
1585         without a single non-exhaustive match- or bind warning.
1586    
1587    ----------------------------------------------------------------------
1588    Name: Matthias Blume (blume (at) tti - c (dot) org)
1589    Date: 2003/05/17 10:20:00 CDT
1590    Tag: blume-20030517-absyn
1591    Description:
1592    
1593    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1594    
1595       This mainly affects the quality of error messages.  However, some
1596       of the code is now more straightforward than before. (Treatment of
1597       the above four constructs in translate.sml is much simpler than
1598       the "macro-expansion" that was going on before.  Plus, the mach-
1599       compiler no longer gets invoked just to be able to compile an
1600       if-expression.)
1601    
1602    2. The ErrorMsg.Error exception is now caught and absorbed by the
1603       interactive loop.
1604    
1605    ----------------------------------------------------------------------
1606    Name: Allen Leung
1607    Date: 2003/05/16 13:05:00 CDT
1608    Tag: leunga-20030516-cygwin-runtime
1609    Description:
1610    
1611        Ported the runtime system to cygwin, which uses the unix
1612    x86-unix bin files.  Missing/buggy features:
1613    
1614       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1615         the Cygwin library.
1616       o Ctrl-C handling may be flaky.
1617       o Windows system calls and Windows I/O are not supported.
1618    
1619       A new set of binfiles is located at:
1620    
1621         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1622    
1623       This is only needed for bootstrapping the cygwin version of smlnj.
1624       Other x86 versions can use the existing binfiles.
1625    
1626    ----------------------------------------------------------------------
1627    Name: Matthias Blume
1628    Date: 2003/04/08 15:42:00 CDT
1629    Tag: blume-20030408-listpair
1630    Description:
1631    
1632    1. Added a target 'mlrisc' to installer.
1633    
1634    2. Added missing elements to structure ListPair.
1635    
1636    ----------------------------------------------------------------------
1637    Name: Allen Leung
1638    Date: 2003/01/07 10:40:00 EST
1639    Tag: leunga-20030107-int-rem
1640    Description:
1641    
1642    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1643    The arguments to the SUBL instruction were swapped.
1644    
1645    ----------------------------------------------------------------------
1646    Name: Matthias Blume
1647    Date: 2002/12/12 16:25:00 EST
1648    Tag: blume-20021212-risc-ra
1649    Description:
1650    
1651    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1652    sent the RA into an infinite loop when floating point registers get
1653    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1654    110.37 and 110.38.)
1655    There was another set of potential problems related to the handling of
1656    MLRISC annotations (but those did not yet cause real problems, apparently).
1657    
1658    ----------------------------------------------------------------------
1659    Name: Matthias Blume
1660    Date: 2002/12/06 22:40:00 EST
1661    Tag: blume-20021206-cm-fileid
1662    Description:
1663    
1664    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1665    This fixes the problem of CM getting confused by files that suddenly
1666    change their identity (e.g., by getting unlinked and recreated by some
1667    text editor such as vi).  There might be a better/cheaper/cleaner way
1668    of doing this, but for now this will have to do.
1669    
1670    ----------------------------------------------------------------------
1671    Name: Matthias Blume
1672    Date: 2002/10/28 09:50:00 EST
1673    Tag: blume-20021028-typecheck
1674    Description:
1675    
1676    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1677    
1678    ----------------------------------------------------------------------
1679    Name: Matthias Blume
1680    Date: 2002/10/17 09:10:00 EDT
1681    Tag: Release_110_42
1682    Description:
1683    
1684    In good old tradition, there has been a slight hiccup so that we have
1685    to patch 110.42 after the fact.   The old release tag has been replaced
1686    (see below).
1687    
1688    The change solves a problem with two competing approaches the
1689    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1690    each other's way.
1691    
1692    This change only affects the runtime system code and the installer script.
1693    (No new bootfiles.)
1694    
1695    ----------------------------------------------------------------------
1696    Name: Matthias Blume
1697    Date: 2002/10/16 12:00:00 EDT
1698    Tag: Release_110_42_removed
1699    Description:
1700    
1701    New working release.  New bootfiles.
1702    
1703    ----------------------------------------------------------------------
1704    Name: Matthias Blume
1705    Date: 2002/10/10 13:10:00 EDT
1706    Tag: blume-20021010-ppc-divs
1707    Description:
1708    
1709    The mltree operator DIVS must be implemented with an overflow check on
1710    the PPC because the hardware indicates divide-by-zero using "overflow" as
1711    well.
1712    
1713    ----------------------------------------------------------------------
1714    Name: Matthias Blume
1715    Date: 2002/07/23 11:45:00
1716    Tag: blume-20020723-smlnj-home
1717    Description:
1718    
1719    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1720    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1721    CM_PATHCONFIG is also set), the path configuration file is assumed
1722    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1723    the entire tree to some other place and everything will "just work".
1724    
1725    (Companion commands such as ml-build and ml-makedepend also sense this
1726    variable.)
1727    
1728    ----------------------------------------------------------------------
1729    Name: Matthias Blume
1730    Date: 2002/07/12 21:19:00 EDT
1731    Tag: blume-20020712-liveness
1732    Description:
1733    
1734    Exported two useful "step" functions from liveness module (MLRISC).
1735    
1736    ----------------------------------------------------------------------
1737    Name: Matthias Blume
1738    Date: 2002/07/05 16:00 EDT
1739    Tag: Release_110_41
1740    Description:
1741    
1742    New working release.  New bootfiles.
1743    
1744    ----------------------------------------------------------------------
1745    Name: Matthias Blume
1746    Date: 2002/07/05 10:25:00 EDT
1747    Tag: blume-20020705-btimp
1748    Description:
1749    
1750    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1751    other clients can set up backtracing support.
1752    
1753    ----------------------------------------------------------------------
1754    Name: Matthias Blume
1755    Date: 2002/06/25 17:23:00 EDT
1756    Tag: blume-20020625-fpmax
1757    Description:
1758    
1759    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1760    case.  (The sense of the isNaN test was reversed -- which made min and
1761    max always return their first argument.)
1762    
1763    ----------------------------------------------------------------------
1764    Name: Matthias Blume
1765    Date: 2002/06/11
1766    Tag: blume-20020611-unixpath
1767    Description:
1768    
1769    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1770    
1771    ----------------------------------------------------------------------
1772    Name: Matthias Blume
1773    Date: 2002/06/10 16:35:00 EDT
1774    Tag: blume-20020610-ieeereal
1775    Description:
1776    
1777    I back-ported my implementation of IEEEReal.fromString from the
1778    idlbasis-devel branch so that we can test it.
1779    
1780    Another small change is that ppDec tries to give more information
1781    than just "<sig>" in the case of functors.  However, this code is
1782    broken in some mysterious way if the functor's body's signature
1783    has not been declared by ascription but gets inferred from the
1784    implementation.  This needs fixing...
1785    
1786    ----------------------------------------------------------------------
1787    Name: Matthias Blume
1788    Date: 2002/05/31
1789    Tag: blume-20020531-btrace-mode
1790    Description:
1791    
1792    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1793    the wayside when I switched over to using Controls everywhere.)
1794    
1795    ----------------------------------------------------------------------
1796    Name: Lal George
1797    Date: 2002/05/23 12:21:40 EDT
1798    Tag: george-20020523-visual-labels
1799    Description:
1800    
1801    Labels are now displayed in the graphical output to make
1802    the fall-through and target blocks obvious.
1803    
1804    ----------------------------------------------------------------------
1805    Name: Matthias Blume
1806    Date: 2002/05/22 11:03:00 EDT
1807    Tag: blume-20020522-shrink
1808    Description:
1809    
1810    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1811    situation that comes up when allocating huge arrays.
1812    
1813    ----------------------------------------------------------------------
1814    Name: Matthias Blume
1815    Date: 2002/05/21 16:00:00 EDT
1816    Tag: Release_110_40
1817    Description:
1818    
1819    New working release (110.40).  New bootfiles.
1820    
1821    [Also: John Reppy fixed GC bug 1131.]
1822    
1823    ----------------------------------------------------------------------
1824    Name: Matthias Blume
1825    Date: 2002/05/21 12:35:00 EDT
1826    Tag: blume-20020521-cmdoc
1827    Description:
1828    
1829    CM documentation update.
1830    
1831    ----------------------------------------------------------------------
1832    Name: Matthias Blume
1833    Date: 2002/05/21 10:55:00 EDT
1834    Tag: blume-20020521-misc
1835    Description:
1836    
1837    - John tweaked runtime to be silent on heap export (except when
1838      GC messages are on).
1839    - I added a few more things (cross-compiling versions of CMB) to
1840      config/preloads (as suggestions).
1841    
1842    ----------------------------------------------------------------------
1843    Name: Matthias Blume
1844    Date: 2002/05/20 22:25:00 EDT
1845    Tag: blume-20020520-controls
1846    Description:
1847    
1848    - Added ControlUtil structure to control-lib.cm.
1849    - Use it throughout.
1850    - Used Controls facility to define MLRISC controls (as opposed to
1851      registering MLRISC control ref cells with Controls after the fact)
1852    - Fixed messed-up controls priorities.
1853    
1854    * Removed again all the stuff from config/preloads that one wouldn't
1855      be able to preload at the time the initial heap image is built.
1856      (Many libraries, e.g., CML, do not exist yet at this time.  The
1857       only libraries that can be preloaded via config/preloads are those
1858       that come bundled with the bootfiles.)
1859    
1860    ----------------------------------------------------------------------
1861    Name: Matthias Blume
1862    Date: 2002/05/20 10:59:00 EDT
1863    Tag: blume-20020520-preloads
1864    Description:
1865    
1866    Added a lot of commented-out suggestions for things to be included
1867    in config/preloads.
1868    
1869    ----------------------------------------------------------------------
1870    Name: Allen Leung
1871    Date: 2002/05/18 14:20:00 EDT
1872    Tag: leunga-20020518-mdl
1873    Description:
1874    
1875    o Made the mdl tool stuff compile and run again.
1876    o I've disabled all the stuff that depends on RTL specifications; they
1877      are all badly broken anyway.
1878    
1879    ----------------------------------------------------------------------
1880    Name: Matthias Blume
1881    Date: 2002/05/17 16:49:00 EDT
1882    Tag: blume-20020517-controls
1883    Description:
1884    
1885    0. John Reppy made several modifications to the SML/NJ library.
1886       In particular, there is a shiny new controls-lib.cm.
1887    
1888    1. Pushed new controls interface through compiler so that everything
1889       compiles again.
1890    
1891    2. Added FormatComb and FORMAT_COMB to the CML version of the
1892       SML/NJ library (so that CML compiles again).
1893    
1894    3. Modified init scripts because XXX_DEFAULT environment variables
1895       are no longer with us.  (Boot-time initialization is now done
1896       using the same environment variables that are also used for
1897       startup-time initialization of controls.)
1898    
1899    ----------------------------------------------------------------------
1900    Name: Lal George
1901    Date: 2002/05/15 09:20:10 EDT
1902    Tag: george-20020515-pseudo-op-decls
1903    Description:
1904    
1905            All pseudo-ops emitted before the first segment declaration
1906    such as TEXT, DATA, and BSS directives are assumed to be global
1907    declarations and are emitted first in the assembly file. This is
1908    useful in a number of situations where one has pseudo-ops that are not
1909    specific to any segment, and also works around the constraint that one
1910    cannot have client pseudo-ops in the TEXT segment.
1911    
1912            Because no segment is associated with these declarations it is
1913    an error to allocate any space or objects before the first segment
1914    directive and an exception will be raised. However, we cannot make
1915    this check for client pseudo-ops.
1916    
1917            These top level declarations are a field in the CFG graph_info.
1918    In theory you can continue to add to this field after the CFG has been
1919    built -- provided you know what you are doing;-)
1920    
1921    ----------------------------------------------------------------------
1922    Name: Matthias Blume
1923    Date: 2002/05/13 16:40:00 EDT
1924    Tag: blume-20020513-pp-etc
1925    Description:
1926    
1927    A few minor bugfixes:
1928    
1929      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1930        (Bogus printouts for redefined bindings still occur.  Compiler
1931        bug should no longer occur now.  We need to redo the prettyprinter
1932        from scratch.)
1933    
1934      - CM pathname printer now also adds escape sequences for ( and )
1935    
1936      - commend and docu fixes for ml-nlffi
1937    
1938    ----------------------------------------------------------------------
1939    Name: Matthias Blume
1940    Date: 2002/05/10 16:40:00 EDT
1941    Tag: blume-20020510-erg-textio
1942    Description:
1943    
1944    Applied the following bugfix provided by Emden Gansner:
1945    
1946        Output is corrupted when outputSubstr is used rather than output.
1947    
1948        The problem occurs when a substring
1949    
1950            ss = (s, dataStart, dataLen)
1951    
1952        where dataStart > 0, fills a stream buffer with avail bytes left.
1953        avail bytes of s, starting at index dataStart, are copied into the
1954        buffer, the buffer is flushed, and then the remaining dataLen-avail
1955        bytes of ss are copied into the beginning of the buffer. Instead of
1956        starting this copy at index dataStart+avail in s, the current code
1957        starts the copy at index avail.
1958    
1959        Fix:
1960        In text-io-fn.sml, change line 695 from
1961             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1962        to
1963             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1964    
1965    ----------------------------------------------------------------------
1966    Name: Matthias Blume
1967    Date: 2002/04/12 13:55:00 EDT
1968    Tag: blume-20020412-assyntax
1969    Description:
1970    
1971    1. Grabbed newer assyntax.h from the XFree86 project.
1972    2. Fiddled with how to compile X86.prim.asm without warnings.
1973    3. (Very) Minor cleanup in CM.
1974    
1975    ----------------------------------------------------------------------
1976    Name: Matthias Blume
1977    Date: 2002/04/01 (no joke!) 17:07:00 EST
1978    Tag: blume-20020401-x86div
1979    Description:
1980    
1981    Added full support for div/mod/rem/quot on the x86, using the machine
1982    instruction's two results (without clumsily recomputing the remainder)
1983    directly where appropriate.
1984    
1985    Some more extensive power-of-two support was added to the x86 instruction
1986    selector (avoiding expensive divs, mods, and muls where they can be
1987    replaced with cheaper shifts and masks).  However, this sort of thing
1988    ought to be done earlier, e.g., within the CPS optimizer so that
1989    all architectures benefit from it.
1990    
1991    The compiler compiles to a fixed point, but changes might be somewhat
1992    fragile nevertheless.  Please, report any strange things that you might
1993    see wrt. div/mod/quot/rem...
1994    
1995    ----------------------------------------------------------------------
1996    Name: Matthias Blume
1997    Date: 2002/03/29 17:22:00
1998    Tag: blume-20020329-div
1999    Description:
2000    
2001    Fixed my broken div/mod logic.  Unfortunately, this means that the
2002    inline code for div/mod now has one more comparison than before.
2003    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2004    The problem was with quotient = 0, because that alone does not tell
2005    us which way the rounding went.  One then has to look at whether
2006    remainder and divisor have the same sign...  :(
2007    
2008    Anyway, I replaced the bootfiles with fresh ones...
2009    
2010    ----------------------------------------------------------------------
2011    Name: Matthias Blume
2012    Date: 2002/03/29 14:10:00 EST
2013    Tag: blume-20020329-inlprims
2014    Description:
2015    
2016    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2017    
2018    Primops have changed. This means that the bin/boot-file formats have
2019    changed as well.
2020    
2021    To make sure that there is no confusion, I made a new version.
2022    
2023    
2024    CHANGES:
2025    
2026    * removed REMT from mltree (remainder should never overflow).
2027    
2028    * added primops to deal with divisions of all flavors to the frontend
2029    
2030    * handled these primops all the way through so they map to their respective
2031      MLRISC support
2032    
2033    * used these primops in the implementation of Int, Int32, Word, Word32
2034    
2035    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2036    
2037    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2038    
2039    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2040      floating point
2041    
2042    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2043      and Real (but Real.abs maps to a separate floating-point-only primop)
2044    
2045    
2046    TODO items:
2047    
2048    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2049      instructions because the machine instruction encoder cannot handle
2050      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2051      REM in the first place, and REM is supported by the x86 machine coder.)
2052    
2053    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2054      selection phase.  (The two can be streamlined because the hardware
2055      delivers both quotient and remainder at the same time anyway.)
2056    
2057    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2058      (Currently the behavior is inconsistent both across architectures and
2059      wrt. the draft Basis spec.)
2060    
2061    * Word8 should eventually be handled natively, too.
2062    
2063    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2064      as if there currently is no execution path that could trigger it in
2065      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2066      hold for things like multiplication and division.)
2067    
2068    ----------------------------------------------------------------------
2069    Name: Matthias Blume
2070    Date: 2002/03/27 16:27:00 EST
2071    Tag: blume-20020327-mlrisc-divisions
2072    Description:
2073    
2074    Added support for all four division operations (ML's div, mod, quot,
2075    and rem) to MLRISC.  In the course of doing so, I also rationalized
2076    the naming (no more annoying switch-around of DIV and QUOT), by
2077    parameterizing the operation by div_rounding_mode (which can be either
2078    DIV_TO_ZERO or DIV_TO_NEGINF).
2079    
2080    The generic MLTreeGen functor takes care of compiling all four
2081    operations down to only round-to-zero div.
2082    
2083    Missing pieces:
2084    
2085      * Doing something smarter than relying on MLTreeGen on architectures
2086        like, e.g., the x86 where hardware division delivers both quotient and
2087        remainder at the same time.  With this, the implementation of the
2088        round-to-neginf operations could be further streamlined.
2089    
2090      * Remove inlining support for div/mod/rem from the frontend and replace it
2091        with primops that get carried through to the backend.  Do this for all
2092        int and word types.
2093    
2094    ----------------------------------------------------------------------
2095    Name: Matthias Blume
2096    Date: 2002/03/25 17:25:00 EST
2097    Tag: blume-20020325-divmod
2098    Description:
2099    
2100    I improved (hopefully without breaking them) the implementation of Int.div,
2101    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2102    advantage of the following observations:
2103    
2104      Let  q = x quot y      r = x rem y
2105           d = x div  y      m = x mod y
2106    
2107    where "quot" is the round-to-zero version of integer division that
2108    hardware usually provides.  Then we have:
2109    
2110         r = x - q * y        where neither the * nor the - will overflow
2111         d = if q >= 0 orelse x = q * y then q else q - 1
2112                              where neither the * nor the - will overflow
2113         m = if q >= 0 orelse r = 0 then r else r + y
2114                              where the + will not overflow
2115    
2116    This results in substantial simplification of the generated code.
2117    The following table shows the number of CFG nodes and edges generated
2118    for
2119            fun f (x, y) = x OPER y
2120            (* with OPER \in div, mod, quot, rem *)
2121    
2122    
2123        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2124        --------------------------------------------------------
2125         div |         24 |         39 |         12 |         16
2126         mod |         41 |         71 |         12 |         16
2127        quot |          8 |         10 |          8 |         10
2128         rem |         10 |         14 |          8 |         10
2129    
2130    
2131    ----------------------------------------------------------------------
2132    Name: Matthias Blume
2133    Date: 2002/03/25 22:06:00 EST
2134    Tag: blume-20020325-cprotobug
2135    Description:
2136    
2137    Fixed a bug in cproto (c prototype decoder).
2138    
2139    ----------------------------------------------------------------------
2140    Name: Matthias Blume
2141    Date: 2002/03/25 16:00:00 EST
2142    Tag: blume-20020325-raw-primops
2143    Description:
2144    
2145    I did some cleanup to Allen's new primop code and
2146    replaced yesterday's bootfiles with new ones.
2147    (But they are stored in the same place.)
2148    
2149    ----------------------------------------------------------------------
2150    Name: Matthias Blume
2151    Date: 2002/03/24 22:40:00 EST
2152    Tag: blume-20020324-bootfiles
2153    Description:
2154    
2155    Made the bootfiles that Allen asked for.
2156    
2157    ----------------------------------------------------------------------
2158    Name: Allen Leung
2159    Date: 2002/03/23 15:50:00 EST
2160    Tag: leunga-20020323-flint-cps-rcc-primops
2161    Description:
2162    
2163      1. Changes to FLINT primops:
2164    
2165        (* make a call to a C-function;
2166         * The primop carries C function prototype information and specifies
2167         * which of its (ML-) arguments are floating point. C prototype
2168         * information is for use by the backend, ML information is for
2169         * use by the CPS converter. *)
2170      | RAW_CCALL of { c_proto: CTypes.c_proto,
2171                       ml_args: ccall_type list,
2172                       ml_res_opt: ccall_type option,
2173                       reentrant : bool
2174                     } option
2175       (* Allocate uninitialized storage on the heap.
2176        * The record is meant to hold short-lived C objects, i.e., they
2177        * are not ML pointers.  With the tag, the representation is
2178        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2179        *)
2180      | RAW_RECORD of {tag:bool,sz:int}
2181      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2182    
2183      2.  These CPS primops are now overloaded:
2184    
2185           rawload of {kind:numkind}
2186           rawstore of {kind:numkind}
2187    
2188          The one argument form is:
2189    
2190             rawload {kind} address
2191    
2192          The two argument form is:
2193    
2194             rawload {kind} [ml object, byte-offset]
2195    
2196      3. RAW_CCALL/RCC now takes two extra arguments:
2197    
2198         a. The first is whether the C call is reentrant, i.e., whether
2199            ML state should be saved and restored.
2200         b. The second argument is a string argument specifying the name of
2201            library and the C function.
2202    
2203         These things are currently not handled in the code generator, yet.
2204    
2205      4. In CProto,
2206    
2207         An encoding type of "bool" means "ml object" and is mapped into
2208         C prototype of PTR.  Note that "bool" is different than "string",
2209         even though "string" is also mapped into PTR, because "bool"
2210         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2211    
2212      5. Pickler/unpicker
2213    
2214         Changed to handle RAW_RECORD and newest RAW_CCALL
2215    
2216      6. MLRiscGen,
2217    
2218         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2219         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2220            in the file CodeGen/cpscompile/cps-c-calls.sml
2221    
2222      7. Added the conditional move operator
2223    
2224             condmove of branch
2225    
2226         to cps.  Generation of this is still buggy so it is currently
2227         disabled.
2228    
2229    ----------------------------------------------------------------------
2230    Name: Lal George
2231    Date: 2002/03/22 14:18:25 EST
2232    Tag: george-20020322-cps-branch-prob
2233    Description:
2234    
2235    Implemented the Ball-Larus branch prediction-heuristics, and
2236    incorporated graphical viewers for control flow graphs.
2237    
2238    Ball-Larus Heuristics:
2239    ---------------------
2240    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2241    
2242    By design it uses the Dempster-Shafer theory for combining
2243    probabilities.  For example, in the function:
2244    
2245        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2246    
2247    the ball-larus heuristics predicts that the n=0 is unlikely
2248    (OH-heuristic), and the 'then' branch is unlikely because of the
2249    RH-heuristic -- giving the 'then' branch an even lower combined
2250    probability using the Dempster-Shafer theory.
2251    
2252    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2253    probability of the 'then' branch because of the loop in the else
2254    branch.
2255    
2256    
2257    Graphical Viewing:
2258    ------------------
2259    I merely plugged in Allen's graphical viewers into the compiler. The
2260    additional code is not much. At the top level, saying:
2261    
2262            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2263    
2264    will display the graphical view of the control flow graph just before
2265    back-patching.  daVinci must be in your path for this to work. If
2266    daVinci is not available, then the default viewer can be changed
2267    using:
2268    
2269            Control.MLRISC.getString "viewer"
2270    
2271    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2272    course, these viewers must be in your path.
2273    
2274    The above will display the compilation unit at the level of clusters,
2275    many of which are small, boring, and un-interesting. Also setting:
2276    
2277            Control.MLRISC.getInt "cfg-graphical-view_size"
2278    
2279    will display clusters that are larger than the value set by the above.
2280    
2281    
2282    ----------------------------------------------------------------------
2283    Name: Matthias Blume
2284    Date: 2002/03/21 22:20:00 EST
2285    Tag: blume-20020321-kmp-bugfix
2286    Description:
2287    
2288    Changed the interface to the KMP routine in PreString and fixed
2289    a minor bug in one place where it was used.
2290    
2291    ----------------------------------------------------------------------
2292    Name: Allen Leung
2293    Date: 2002/03/21 20:30:00 EST
2294    Tag: leunga-20020321-cfg
2295    Description:
2296    
2297      Fixed a potential problem in cfg edge splitting.
2298    
2299    ----------------------------------------------------------------------
2300    Name: Allen Leung
2301    Date: 2002/03/21 17:15:00 EST
2302    Tag: leunga-20020321-x86-fp-cfg
2303    Description:
2304    
2305      1. Recoded the buggy parts of x86-fp.
2306    
2307         a. All the block reordering code has been removed.
2308            We now depend on the block placement phases to do this work.
2309    
2310         b. Critical edge splitting code has been simplified and moved into the
2311            CFG modules, as where they belong.
2312    
2313         Both of these were quite buggy and complex.  The code is now much, much
2314         simpler.
2315    
2316      2. X86 backend.
2317    
2318         a. Added instructions for 64-bit support.  Instruction selection for
2319            64-bit has not been committed, however, since that
2320            requires changes to MLTREE which haven't been approved by
2321            Lal and John.
2322    
2323         b. Added support for FUCOMI and FUCOMIP when generating code for
2324            PentiumPro and above.  We only generate these instructions in
2325            the fast-fp mode.
2326    
2327         c. Added cases for JP and JNP in X86FreqProps.
2328    
2329      3. CFG
2330    
2331         CFG now has a bunch of methods for edge splitting and merging.
2332    
2333      4. Machine description.
2334    
2335         John's simplification of MLTREE_BASIS.fcond broke a few machine
2336         description things:
2337    
2338         rtl-build.{sig,sml} and hppa.mdl fixed.
2339    
2340         NOTE: the machine description stuff in the repository is still broken.
2341               Again, I can't put my fixes in because that involves
2342               changes to MLTREE.
2343    
2344    ----------------------------------------------------------------------
2345    Name: Matthias Blume
2346    Date: 2002/03/20 15:55:00 EST
2347    Tag: blume-20020320-kmp
2348    Description:
2349    
2350    Implemented Knuth-Morris-Pratt string matching in PreString and used
2351    it for String.isSubstring, Substring.isSubstring, and
2352    Substring.position.
2353    
2354    (Might need some stress-testing.  Simple examples worked fine.)
2355    
2356    ----------------------------------------------------------------------
2357    Name: Matthias Blume
2358    Date: 2002/03/19 16:37:00 EST
2359    Tag: blume-20020319-witnesses
2360    Description:
2361    
2362    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2363    
2364    This implements a generic mechanism for changing constness qualifiers
2365    anywhere within big C types without resorting to outright "casts".
2366    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2367    modify the constness at the outermost level.)
2368    The implementation of "convert" is based on the idea of "witness"
2369    values -- values that are not used by the operation but whose types
2370    "testify" to their applicability.  On the implementation side, "convert"
2371    is simply a projection (returning its second curried argument).  With
2372    cross-module inlining, it should not result in any machine code being
2373    generated.
2374    
2375    ----------------------------------------------------------------------
2376    Name: Matthias Blume
2377    Date: 2002/03/15 16:40:00 EST
2378    Tag: blume-20020315-basis
2379    Description:
2380    
2381    Provided (preliminary?) implementations for
2382    
2383      {String,Substring}.{concatWith,isSuffix,isSubstring}
2384    
2385    and
2386    
2387      Substring.full
2388    
2389    Those are in the Basis spec but they were missing in SML/NJ.
2390    
2391    ----------------------------------------------------------------------
2392    Name: Matthias Blume
2393    Date: 2002/03/14 21:30:00 EST
2394    Tag: blume-20020314-controls
2395    Description:
2396    
2397    Controls:
2398    ---------
2399    
2400    1. Factored out the recently-added Controls : CONTROLS stuff and put
2401       it into its own library $/controls-lib.cm.  The source tree for
2402       this is under src/smlnj-lib/Controls.
2403    
2404    2. Changed the names of types and functions in this interface, so they
2405       make a bit more "sense":
2406    
2407          module -> registry
2408          'a registry -> 'a group
2409    
2410    3. The interface now deals in ref cells only.  The getter/setter interface
2411       is (mostly) gone.
2412    
2413    4. Added a function that lets one register an already-existing ref cell.
2414    
2415    5. Made the corresponding modifications to the rest of the code so that
2416       everything compiles again.
2417    
2418    6. Changed the implementation of Controls.MLRISC back to something closer
2419       to the original.  In particular, this module (and therefore MLRISC)
2420       does not depend on Controls.  There now is some link-time code in
2421       int-sys.sml that registers the MLRISC controls with the Controls
2422       module.
2423    
2424    CM:
2425    ---
2426    
2427      * One can now specify the lambda-split aggressiveness in init.cmi.
2428    
2429    ----------------------------------------------------------------------
2430    Name: Allen Leung
2431    Date: 2002/03/13 17:30:00 EST
2432    Tag: leunga-20020313-x86-fp-unary
2433    Description:
2434    
2435    Bug fix for:
2436    
2437    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2438    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2439    > - fun f(x,(y,z)) = Real.~ y;
2440    > [autoloading]
2441    > [autoloading done]
2442    >       fchsl   (%eax), 184(%esp)
2443    > Error: MLRisc bug: X86MCEmitter.emitInstr
2444    >
2445    > uncaught exception Error
2446    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2447    
2448    The problem was that the code generator did not generate any fp registers
2449    in this case, and the ra didn't know that it needed to run the X86FP phase to
2450    translate the pseudo fp instruction.   This only happened with unary fp
2451    operators in certain situations.
2452    
2453    ----------------------------------------------------------------------
2454    Name: Matthias Blume
2455    Date: 2002/03/13 14:00:00 EST
2456    Tag: blume-20020313-overload-etc
2457    Description:
2458    
2459    1. Added _overload as a synonym for overload for backward compatibility.
2460       (Control.overloadKW must be true for either version to be accepted.)
2461    
2462    2. Fixed bug in install script that caused more things to be installed
2463       than what was requested in config/targets.
2464    
2465    3. Made CM aware of the (_)overload construct so that autoloading
2466       works.
2467    
2468    ----------------------------------------------------------------------
2469    Name: Matthias Blume
2470    Date: 2002/03/12 22:03:00 EST
2471    Tag: blume-20020312-url
2472    Description:
2473    
2474    Forgot to update BOOT and srcarchiveurl.
2475    
2476    ----------------------------------------------------------------------
2477    Name: Matthias Blume
2478    Date: 2002/03/12 17:30:00 EST
2479    Tag: blume-20020312-version110392
2480    Description:
2481    
2482    Yet another version number bump (because of small changes to the
2483    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2484    
2485    Changes:
2486    
2487      The new pid generation scheme described a few weeks ago was overly
2488      complicated.  I implemented a new mechanism that is simpler and
2489      provides a bit more "stability":  Once CM has seen a compilation
2490      unit, it keeps its identity constant (as long as you do not delete
2491      those crucial CM/GUID/* files).  This means that when you change
2492      an interface, compile, then go back to the old interface, and
2493      compile again, you arrive at the original pid.
2494    
2495      There now also is a mechanism that instructs CM to use the plain
2496      environment hash as a module's pid (effectively making its GUID
2497      the empty string).  For this, "noguid" must be specified as an
2498      option to the .sml file in question within its .cm file.
2499      This is most useful for code that is being generated by tools such
2500      as ml-nlffigen (because during development programmers tend to
2501      erase the tool's entire output directory tree including CM's cached
2502      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2503      revert to the old, broken behavior of SML/NJ, but in specific cases
2504      where there is no danger of interface confusion, its use is ok
2505      (I think).
2506    
2507      ml-nlffigen by default generates "noguid" annotations.  They can be
2508      turned off by specifying -guid in its command line.
2509    
2510    ----------------------------------------------------------------------
2511    Name: Lal George
2512    Date: 2002/03/12 12 14:42:36 EST
2513    Tag: george-20020312-frequency-computation
2514    Description:
2515    
2516    Integrated jump chaining and static block frequency into the
2517    compiler. More details and numbers later.
2518    
2519    ----------------------------------------------------------------------
2520    Name: Lal George
2521    Date: 2002/03/11 11 22:38:53 EST
2522    Tag: george-20020311-jump-chain-elim
2523    Description:
2524    
2525    Tested the jump chain elimination on all architectures (except the
2526    hppa).  This is on by default right now and is profitable for the
2527    alpha and x86, however, it may not be profitable for the sparc and ppc
2528    when compiling the compiler.
2529    
2530    The gc test will typically jump to a label at the end of the cluster,
2531    where there is another jump to an external cluster containing the actual
2532    code to invoke gc. This is to allow factoring of common gc invocation
2533    sequences. That is to say, we generate:
2534    
2535            f:
2536               testgc
2537               ja   L1      % jump if above to L1
2538    
2539            L1:
2540               jmp L2
2541    
2542    
2543    After jump chain elimination the 'ja L1' instructions is converted to
2544    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2545    up being implemented in their long form (if L2 is far away) using:
2546    
2547            jbe     L3      % jump if below or equal to L3
2548            jmp     L2
2549         L3:
2550            ...
2551    
2552    
2553    For large compilation units L2  may be far away.
2554    
2555    
2556    ----------------------------------------------------------------------
2557    Name: Matthias Blume
2558    Date: 2002/03/11 13:30:00 EST
2559    Tag: blume-20020311-mltreeeval
2560    Description:
2561    
2562    A functor parameter was missing.
2563    
2564    ----------------------------------------------------------------------
2565    Name: Allen Leung
2566    Date: 2002/03/11 10:30:00 EST
2567    Tag: leunga-20020311-runtime-string0
2568    Description:
2569    
2570       The representation of the empty string now points to a
2571    legal null terminated C string instead of unit.  It is now possible
2572    to convert an ML string into C string with InlineT.CharVector.getData.
2573    This compiles into one single machine instruction.
2574    
2575    ----------------------------------------------------------------------
2576    Name: Allen Leung
2577    Date: 2002/03/10 23:55:00 EST
2578    Tag: leunga-20020310-x86-call
2579    Description:
2580    
2581       Added machine generation for CALL instruction (relative displacement mode)
2582    
2583    ----------------------------------------------------------------------
2584    Name: Matthias Blume
2585    Date: 2002/03/08 16:05:00
2586    Tag: blume-20020308-entrypoints
2587    Description:
2588    
2589    Version number bumped to 110.39.1.  NEW BOOTFILES!
2590    
2591    Entrypoints: non-zero offset into a code object where execution should begin.
2592    
2593    - Added the notion of an entrypoint to CodeObj.
2594    - Added reading/writing of entrypoint info to Binfile.
2595    - Made runtime system bootloader aware of entrypoints.
2596    - Use the address of the label of the first function given to mlriscGen
2597      as the entrypoint.  This address is currently always 0, but it will
2598      not be 0 once we turn on block placement.
2599    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2600      with entry points) from mlriscGen.
2601    
2602    ----------------------------------------------------------------------
2603    Name: Allen Leung
2604    Date: 2002/03/07 20:45:00 EST
2605    Tag: leunga-20020307-x86-cmov
2606    Description:
2607    
2608       Bug fixes for CMOVcc on x86.
2609    
2610       1. Added machine code generation for CMOVcc
2611       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2612       3. CMOVcc cannot have an immediate operand as argument.
2613    
2614    ----------------------------------------------------------------------
2615    Name: Matthias Blume
2616    Date: 2002/03/07 16:15:00 EST
2617    Tag: blume-20020307-controls
2618    Description:
2619    
2620    This is a very large but mostly boring patch which makes (almost)
2621    every tuneable compiler knob (i.e., pretty much everything under
2622    Control.* plus a few other things) configurable via both the command
2623    line and environment variables in the style CM did its configuration
2624    until now.
2625    
2626    Try starting sml with '-h' (or, if you are brave, '-H')
2627    
2628    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2629    implements the underlying generic mechanism.
2630    
2631    The interface to some of the existing such facilities has changed somewhat.
2632    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2633    (The getFoo interface is still there for backward-compatibility, but its
2634    use is deprecated.)
2635    
2636    The ml-build script passes -Cxxx=yyy command-line arguments through so
2637    that one can now twiddle the compiler settings when using this "batch"
2638    compiler.
2639    
2640    TODO items:
2641    
2642    We should go through and throw out all controls that are no longer
2643    connected to anything.  Moreover, we should go through and provide
2644    meaningful (and correct!) documentation strings for those controls
2645    that still are connected.
2646    
2647    Currently, multiple calls to Controls.new are accepted (only the first
2648    has any effect).  Eventually we should make sure that every control
2649    is being made (via Controls.new) exactly once.  Future access can then
2650    be done using Controls.acc.
2651    
2652    Finally, it would probably be a good idea to use the getter-setter
2653    interface to controls rather than ref cells.  For the time being, both
2654    styles are provided by the Controls module, but getter-setter pairs are
2655    better if thread-safety is of any concern because they can be wrapped.
2656    
2657    *****************************************
2658    
2659    One bug fix: The function blockPlacement in three of the MLRISC
2660    backpatch files used to be hard-wired to one of two possibilities at
2661    link time (according to the value of the placementFlag).  But (I
2662    think) it should rather sense the flag every time.
2663    
2664    *****************************************
2665    
2666    Other assorted changes (by other people who did not supply a HISTORY entry):
2667    
2668    1. the cross-module inliner now works much better (Monnier)
2669    2. representation of weights, frequencies, and probabilities in MLRISC
2670       changed in preparation of using those for weighted block placement
2671       (Reppy, George)
2672    
2673    ----------------------------------------------------------------------
2674    Name: Lal George
2675    Date: 2002/03/07 14:44:24 EST 2002
2676    Tag: george-20020307-weighted-block-placement
2677    
2678    Tested the weighted block placement optimization on all architectures
2679    (except the hppa) using AMPL to generate the block and edge frequencies.
2680    Changes were required in the machine properties to correctly
2681    categorize trap instructions. There is an MLRISC flag
2682    "weighted-block-placement" that can be used to enable weighted block
2683    placement, but this will be ineffective without block/edge
2684    frequencies (coming soon).
2685    
2686    
2687    ----------------------------------------------------------------------
2688    Name: Lal George
2689    Date: 2002/03/05 17:24:48 EST
2690    Tag: george-20020305-linkage-cluster
2691    
2692    In order to support the block placement optimization, a new cluster
2693    is generated as the very first cluster (called the linkage cluster).
2694    It contains a single jump to the 'real' entry point for the compilation
2695    unit. Block placement has no effect on the linkage cluster itself, but
2696    all the other clusters  have full freedom in the manner in which they
2697    reorder blocks or functions.
2698    
2699    On the x86 the typical linkage code that is generated is:
2700       ----------------------
2701            .align 2
2702       L0:
2703            addl    $L1-L0, 72(%esp)
2704            jmp     L1
2705    
2706    
2707            .align  2
2708       L1:
2709       ----------------------
2710    
2711    72(%esp) is the memory location for the stdlink register. This
2712    must contain the address of the CPS function being called. In the
2713    above example, it contains the address of  L0; before
2714    calling L1 (the real entry point for the compilation unit), it
2715    must contain the address for L1, and hence
2716    
2717            addl $L1-L0, 72(%esp)
2718    
2719    I have tested this on all architectures except the hppa.The increase
2720    in code size is of course negligible
2721    
2722    ----------------------------------------------------------------------
2723    Name: Allen Leung
2724    Date: 2002/03/03 13:20:00 EST
2725    Tag: leunga-20020303-mlrisc-tools
2726    
2727      Added #[ ... ] expressions to mlrisc tools
2728    
2729    ----------------------------------------------------------------------
2730    Name: Matthias Blume
2731    Date: 2002/02/27 12:29:00 EST
2732    Tag: blume-20020227-cdebug
2733    Description:
2734    
2735    - made types in structure C and C_Debug to be equal
2736    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2737    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2738    
2739    ----------------------------------------------------------------------
2740    Name: Matthias Blume
2741    Date: 2002/02/26 12:00:00 EST
2742    Tag: blume-20020226-ffi
2743    Description:
2744    
2745    1. Fixed a minor bug in CM's "noweb" tool:
2746       If numbering is turned off, then truly don't number (i.e., do not
2747       supply the -L option to noweb).  The previous behavior was to supply
2748       -L'' -- which caused noweb to use the "default" line numbering scheme.
2749       Thanks to Chris Richards for pointing this out (and supplying the fix).
2750    
2751    2. Once again, I reworked some aspects of the FFI:
2752    
2753       A. The incomplete/complete type business:
2754    
2755       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2756         gone!
2757       - ML types representing an incomplete type are now *equal* to
2758         ML types representing their corresponding complete types (just like
2759         in C).  This is still safe because ml-nlffigen will not generate
2760         RTTI for incomplete types, nor will it generate functions that
2761         require access to such RTTI.   But when ML code generated from both
2762         incomplete and complete versions of the C type meet, the ML types
2763         are trivially interoperable.
2764    
2765         NOTE:  These changes restore the full generality of the translation
2766         (which was previously lost when I eliminated functorization)!
2767    
2768       B. Enum types:
2769    
2770       - Structure C now has a type constructor "enum" that is similar to
2771         how the "su" constructor works.  However, "enum" is not a phantom
2772         type because each "T enum" has values (and is isomorphic to
2773         MLRep.Signed.int).
2774       - There are generic access operations for enum objects (using
2775         MLRep.Signed.int).
2776       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2777         * The structure contains the definition of type "mlrep" (the ML-side
2778         representation type of the enum).  Normally, mlrep is the same
2779         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2780         then mlrep will be defined as a datatype -- thus facilitating
2781         pattern matching on mlrep values.
2782         ("-ec" will be suppressed if there are duplicate values in an
2783          enumeration.)
2784         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2785         will be generated for each C enum constant xxx.
2786         * Conversion functions m2i and i2m convert between mlrep and
2787         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2788         * Coversion functions c and ml convert between mlrep and "tag enum".
2789         * Access functions (get/set) fetch and store mlrep values.
2790       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2791         enumerations are merged into one single enumeration represented by
2792         structure E_'.
2793    
2794    ----------------------------------------------------------------------
2795    Name: Allen Leung
2796    Date: 2002/02/25 04:45:00 EST
2797    Tag: leunga-20020225-cps-spill
2798    
2799    This is a new implementation of the CPS spill phase.
2800    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2801    In case of problems, replace it with the old file spill.sml
2802    
2803    The current compiler runs into some serious performance problems when
2804    constructing a large record.  This can happen when we try to compile a
2805    structure with many items.  Even a very simple structure like the following
2806    makes the compiler slow down.
2807    
2808        structure Foo = struct
2809           val x_1 = 0w1 : Word32.int
2810           val x_2 = 0w2 : Word32.int
2811           val x_3 = 0w3 : Word32.int
2812           ...
2813           val x_N = 0wN : Word32.int
2814        end
2815    
2816    The following table shows the compile time, from N=1000 to N=4000,
2817    with the old compiler:
2818    
2819    N
2820    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2821           MLRISC ra                               0.06u  0.00s  0.05g
2822              (spills = 0 reloads = 0)
2823           TOTAL                                   0.63u  0.07s  0.21g
2824    
2825    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2826           MLRISC ra                               5.68u  0.59s  3.93g
2827              (spills = 0 reloads = 0)
2828           TOTAL                                   14.71u  0.99s  4.81g
2829    
2830    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2831           MLRISC ra                               5.54u  0.65s  3.91g
2832              (spills = 543 reloads = 1082)
2833           TOTAL                                   65.40u  3.13s  6.00g
2834    
2835    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2836           MLRISC ra                               0.80u  0.10s  0.55g
2837              (spills = 42 reloads = 84)
2838           TOTAL                                   129.42u  5.10s  4.13g
2839    
2840    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2841           MLRISC ra                               2.69u  0.27s  1.38g
2842              (spills = 62 reloads = 124)
2843           TOTAL                                   682.48u  19.61s  13.99g
2844    
2845    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2846           MLRISC ra                               4.96u  0.27s  2.72g
2847              (spills = 85 reloads = 170)
2848           TOTAL                                   2375.26u  57.21s  48.00g
2849    
2850    As you can see the old cps spill module suffers from some serious
2851    performance problem.  But since I cannot decipher the old code fully,
2852    instead of patching the problems up, I'm reimplementing it
2853    with a different algorithm.  The new code is more modular,
2854    smaller when compiled, and substantially faster
2855    (O(n log n) time and O(n) space).  Timing of the new spill module:
2856    
2857    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2858          MLRISC ra                               0.25u  0.02s  0.15g
2859             (spills=1 reloads=3)
2860          TOTAL                                   7.74u  0.34s  1.62g
2861    
2862    Implementation details:
2863    
2864    As far as I can tell, the purpose of the CPS spill module is to make sure the
2865    number of live variables at any program point (the bandwidth)
2866    does not exceed a certain limit, which is determined by the
2867    size of the spill area.
2868    
2869    When the bandwidth is too large, we decrease the register pressure by
2870    packing live variables into spill records.  How we achieve this is
2871    completely different than what we did in the old code.
2872    
2873    First, there is something about the MLRiscGen code generator
2874    that we should be aware of:
2875    
2876    o MLRiscGen performs code motion!
2877    
2878       In particular, it will move floating point computations and
2879       address computations involving only the heap pointer to
2880       their use sites (if there is only a single use).
2881       What this means is that if we have a CPS record construction
2882       statement
2883    
2884           RECORD(k,vl,w,e)
2885    
2886       we should never count the new record address w as live if w
2887       has only one use (which is often the case).
2888    
2889       We should do something similar to floating point, but the transformation
2890       there is much more complex, so I won't deal with that.
2891    
2892    Secondly, there are now two new cps primops at our disposal:
2893    
2894     1. rawrecord of record_kind option
2895        This pure operator allocates some uninitialized storage from the heap.
2896        There are two forms:
2897    
2898         rawrecord NONE [INT n]  allocates a tagless record of length n
2899         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2900                                     and initializes the tag.
2901    
2902     2. rawupdate of cty
2903          rawupdate cty (v,i,x)
2904          Assigns to x to the ith component of record v.
2905          The storelist is not updated.
2906    
2907    We use these new primops for both spilling and increment record construction.
2908    
2909     1. Spilling.
2910    
2911        This is implemented with a linear scan algorithm (but generalized
2912        to trees).  The algorithm will create a single spill record at the
2913        beginning of the cps function and use rawupdate to spill to it,
2914        and SELECT or SELp to reload from it.  So both spills and reloads
2915        are fine-grain operations.  In contrast, in the old algorithm
2916        "spills" have to be bundled together in records.
2917    
2918        Ideally, we should sink the spill record construction to where
2919        it is needed.  We can even split the spill record into multiple ones
2920        at the places where they are needed.  But CPS is not a good
2921        representation for global code motion, so I'll keep it simple and
2922        am not attempting this.
2923    
2924     2. Incremental record construction (aka record splitting).
2925    
2926        Long records with many component values which are simulatenously live
2927        (recall that single use record addresses are not considered to
2928         be live) are constructed with rawrecord and rawupdate.
2929        We allocate space on the heap with rawrecord first, then gradually
2930        fill it in with rawupdate.  This is the technique suggested to me
2931        by Matthias.
2932    
2933        Some restrictions on when this is applicable:
2934        1. It is not a VECTOR record.  The code generator currently does not handle
2935           this case. VECTOR record uses double indirection like arrays.
2936        2. All the record component values are defined in the same "basic block"
2937           as the record constructor.  This is to prevent speculative
2938           record construction.
2939    
2940    ----------------------------------------------------------------------
2941    Name: Allen Leung
2942    Date: 2002/02/22 01:02:00 EST
2943    Tag: leunga-20020222-mlrisc-tools
2944    
2945    Minor bug fixes in the parser and rewriter
2946    
2947    ----------------------------------------------------------------------
2948    Name: Allen Leung
2949    Date: 2002/02/21 20:20:00 EST
2950    Tag: leunga-20020221-peephole
2951    
2952    Regenerated the peephole files.  Some contained typos in the specification
2953    and some didn't compile because of pretty printing bugs in the old version
2954    of 'nowhere'.
2955    
2956    ----------------------------------------------------------------------
2957    Name: Allen Leung
2958    Date: 2002/02/19 20:20:00 EST
2959    Tag: leunga-20020219-mlrisc-tools
2960    Description:
2961    
2962       Minor bug fixes to the mlrisc-tools library:
2963    
2964       1.  Fixed up parsing colon suffixed keywords
2965       2.  Added the ability to shut the error messages up
2966       3.  Reimplemented the pretty printer and fixed up/improved
2967           the pretty printing of handle and -> types.
2968       4.  Fixed up generation of literal symbols in the nowhere tool.
2969       5.  Added some SML keywords to to sml.sty
2970    
2971    ----------------------------------------------------------------------
2972    Name: Matthias Blume
2973    Date: 2002/02/19 16:20:00 EST
2974    Tag: blume-20020219-cmffi
2975    Description:
2976    
2977    A wild mix of changes, some minor, some major:
2978    
2979    * All C FFI-related libraries are now anchored under $c:
2980        $/c.cm      --> $c/c.cm
2981        $/c-int.cm  --> $c/internals/c-int.cm
2982        $/memory.cm --> $c/memory/memory.cm
2983    
2984    * "make" tool (in CM) now treats its argument pathname slightly
2985      differently:
2986        1. If the native expansion is an absolute name, then before invoking
2987           the "make" command on it, CM will apply OS.Path.mkRelative
2988           (with relativeTo = OS.FileSys.getDir()) to it.
2989        2. The argument will be passed through to subsequent phases of CM
2990           processing without "going native".  In particular, if the argument
2991           was an anchored path, then "make" will not lose track of that anchor.
2992    
2993    * Compiler backends now "know" their respective C calling conventions
2994      instead of having to be told about it by ml-nlffigen.  This relieves
2995      ml-nlffigen from one of its burdens.
2996    
2997    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
2998    
2999    * Export C_DEBUG and C_Debug from $c/c.cm.
3000    
3001    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3002      subtyping relationship between incomplete pointers and their complete
3003      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3004      with the convention of instantiating 'o with ('t, 'c) obj whenever
3005      the pointer target type is complete.  In the incomplete case, 'o
3006      will be instantiated with some "'c iobj" -- a type obtained by
3007      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3008    
3009      Operations that work on both incomplete and complete pointer types are
3010      typed as taking an 'o ptr while operations that require the target to
3011      be known are typed as taking some ('t, 'c) obj ptr.
3012    
3013      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3014      where void is an eqtype without any values.  This makes it possible
3015      to work on voidptr values using functions meant to operate on light
3016      incomplete pointers.
3017    
3018    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3019      been vastly simplified.
3020    
3021    ----------------------------------------------------------------------
3022    Name: Matthias Blume
3023    Date: 2002/02/19 10:48:00 EST
3024    Tag: blume-20020219-pqfix
3025    Description:
3026    
3027    Applied Chris Okasaki's bug fix for priority queues.
3028    
3029    ----------------------------------------------------------------------
3030    Name: Matthias Blume
3031    Date: 2002/02/15 17:05:00
3032    Tag: Release_110_39
3033    Description:
3034    
3035    Last-minute retagging is becoming a tradition... :-(
3036    
3037    This is the working release 110.39.
3038    
3039    ----------------------------------------------------------------------
3040    Name: Matthias Blume
3041    Date: 2002/02/15 16:00:00 EST
3042    Tag: Release_110_39-orig
3043    Description:
3044    
3045    Working release 110.39.  New bootfiles.
3046    
3047    (Update: There was a small bug in the installer so it wouldn't work
3048    with all shells.  So I retagged. -Matthias)
3049    
3050    ----------------------------------------------------------------------
3051    Name: Matthias Blume
3052    Date: 2002/02/15 14:17:00 EST
3053    Tag: blume-20020215-showbindings
3054    Description:
3055    
3056    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3057    the latter can be useful for exploring what bindings are available at
3058    the interactive prompt.  (The first function returns only the list
3059    of symbols that are really bound, the second prints those but also the
3060    ones that CM's autoloading mechanism knows about.)
3061    
3062    ----------------------------------------------------------------------
3063    Name: Matthias Blume
3064    Date: 2002/02/15 12:08:00 EST
3065    Tag: blume-20020215-iptrs
3066    Description:
3067    
3068    Two improvements to ml-nlffigen:
3069    
3070      1. Write files only if they do not exist or if their current contents
3071         do not coincide with what's being written.  (That is, avoid messing
3072         with the time stamps unless absolutely necessary.)
3073    
3074      2. Implement a "repository" mechanism for generated files related
3075         to "incomplete pointer types".   See the README file for details.
3076    
3077    ----------------------------------------------------------------------
3078    Name: Matthias Blume
3079    Date: 2002/02/14 11:50:00 EST
3080    Tag: blume-20020214-quote
3081    Description:
3082    
3083    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3084    because of the new and improved tag generation scheme.  (Thanks to Allen
3085    Leung for pointing it out.)
3086    
3087    ----------------------------------------------------------------------
3088    Name: Lal George
3089    Date: 2002/02/14 09:55:27 EST 2002
3090    Tag: george-20020214-isabelle-bug
3091    Description:
3092    
3093    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3094    of Isabelle on the x86.
3095    
3096    From Allen:
3097    -----------
3098     I've found the problem:
3099    
3100         in ra-core.sml, I use the counter "blocked" to keep track of the
3101         true number of elements in the freeze queue.  When the counter goes
3102         to zero, I skip examining the queue.  But I've messed up the
3103         bookkeeping in combine():
3104    
3105             else ();
3106             case !ucol of
3107               PSEUDO => (if !cntv > 0 then
3108                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3109                                        ^^^^^^^^^^^^^^^^^^^^^^^
3110                      moveu := mergeMoveList(!movev, !moveu)
3111                     )
3112                  else ();
3113    
3114         combine() is called to coalesce two nodes u and v.
3115         I think I was thinking that if the move counts of u and v are both
3116         greater than zero then after they are coalesced then one node is
3117         removed from the freeze queue.  Apparently I was thinking that
3118         both u and v are of low degree, but that's clearly not necessarily true.
3119    
3120    
3121    02/12/2002:
3122        Here's the patch.  HOL now compiles.
3123    
3124        I don't know how this impact on performance (compile
3125        time or runtime).  This bug caused the RA (especially on the x86)
3126        to go thru the potential spill phase when there are still nodes on the
3127        freeze queue.
3128    
3129    
3130    
3131    
3132    ----------------------------------------------------------------------
3133    Name: Matthias Blume
3134    Date: 2002/02/13 22:40:00 EST
3135    Tag: blume-20020213-fptr-rtti
3136    Description:
3137    
3138    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3139    updates.
3140    
3141    ----------------------------------------------------------------------
3142    Name: Matthias Blume
3143    Date: 2002/02/13 16:41:00 EST
3144    Tag: blume-20020213-cmlpq
3145    Description:
3146    
3147    Added new priority queue export symbols (which have just been added to
3148    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3149    not compile and the installer would choke.)
3150    
3151    ----------------------------------------------------------------------
3152    Name: Matthias Blume
3153    Date: 2002/02/13 16:15:00 EST
3154    Tag: blume-20020213-various
3155    Description:
3156    
3157    1. More tweaks to ml-nlffigen:
3158    
3159       - better internal datastructures (resulting in slight speedup)
3160       - "-match" option requires exact match
3161       - "localized" gensym counters (untagged structs/unions nested within
3162         other structs/unions or within typedefs get a fresh counter; their
3163         tag will be prefixed by a concatenation of their parents' tags)
3164       - bug fixes (related to calculation of transitive closure of types
3165         to be included in the output)
3166    
3167    2. Minor Basis updates:
3168    
3169       - added implementations for List.collate and Option.app
3170    
3171    ----------------------------------------------------------------------
3172    Name: Matthias Blume
3173    Date: 2002/02/11 15:55:00 EST
3174    Tag: blume-20020211-gensym
3175    Description:
3176    
3177    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3178    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3179    names (ML structure names that correspond to unnamed C structs, unions,
3180    and enums), so that separate runs of ml-nlffigen do not clash.
3181    
3182    ----------------------------------------------------------------------
3183    Name: Matthias Blume
3184    Date: 2002/02/11 12:05:00 EST
3185    Tag: blume-20020211-gensml
3186    Description:
3187    
3188    A quick fix for a problem with GenSML (in the pgraph-util library):
3189    Make generation of toplevel "local" optional.  (Strictly speaking,
3190    signature definitions within "local" are not legal SML.)
3191    
3192    Other than that: updates to INSTALL and cm/TODO.
3193    
3194    ----------------------------------------------------------------------
3195    Name: Matthias Blume
3196    Date: 2002/02/08 15:00:00 EST
3197    Tag: blume-20020208-uniquepid
3198    Description:
3199    
3200    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3201    
3202    1. The installer (config/install.sh) has gotten smarter:
3203    
3204         - Configuration options are a bit easier to specify now
3205           (in config/targets).
3206         - Bug in recognizing .tar.bz2 files fixed.
3207         - Installer automatically resolves dependencies between
3208           configuration options (e.g., if you ask for eXene, you will
3209           also get cml -- regardless whether you asked for it or not).
3210         - Installer can run in "quieter mode" by setting the environment
3211           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3212           "completely silent", though.
3213         - Build HashCons library as part of smlnj-lib.
3214    
3215    2. A new scheme for assigning persistent identifiers to compilation
3216       units (and, by extension, to types etc.) has been put into place.
3217       This fixes a long-standing bug where types and even dynamic values
3218       can get internally confused, thereby compromising type safety
3219       (abstraction) and dynamic correctness.  See
3220    
3221         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3222    
3223       for an example of how things could go wrong until now.
3224    
3225       The downside of the new scheme is that pids are not quite as
3226       persistent as they used to be: CM will generate a fresh pid
3227       for every compilation unit that it thinks it sees for the first
3228       time.  That means that if you compile starting from a clean, fresh
3229       source tree at two different times, you end up with different
3230       binaries.
3231    
3232       Cutoff recompilation, however, has not been compromised because
3233       CM keeps pid information in special caches between runs.
3234    
3235    ----------------------------------------------------------------------
3236    Name: Lal George
3237    Date: 2002/02/07 15:34:13 EST 2002
3238    Tag: <none>
3239    Description:
3240    
3241    Compilers that generate assembly code may produce  global labels
3242    whose value is resolved at link time. The various peephole optimization
3243    modules did not take this in account.
3244    
3245    TODO. The Labels.addrOf function should really return an option
3246    type so that clients are forced to deal with this issue, rather
3247    than an exception being raised.
3248    
3249    ----------------------------------------------------------------------
3250    Name: Lal George
3251    Date: 2002/02/06 13:55:02 EST
3252    Tag: george-20020206-ra-breakup
3253    Description:
3254    
3255    1. A bug fix from Allen.
3256    
3257        A typo causes extra fstp %st(0)'s to be generated at compensation
3258        edges, which might cause stack underflow traps at runtime.  This
3259        occurs in fft where there are extraneous fstps right before the 'into'
3260        trap instruction (in this case they are harmless since none of the
3261        integers overflow.)
3262    
3263    2. Pulled out various utility modules that were embedded in the modules
3264       of the register allocator. I need these modules for other purposes, but
3265       they are not complete enough to put into a library (just yet).
3266    ----------------------------------------------------------------------
3267    Name: Matthias Blume
3268    Date: 2002/01/31 16:05:00 EST
3269    Tag: blume-20020131-sparc-ccalls
3270    Description:
3271    
3272    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3273       of extra stack space by mistake.  Fixed.
3274    
3275    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3276    
3277    ----------------------------------------------------------------------
3278    Name: Allen Leung
3279    Date: 2002/01/30
3280    Tag: leunga-20020130-nowhere-bug-fix
3281    Description:
3282    
3283       MLRISC bug fixes:
3284       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3285       2. MachineInt.fromString was negating its value.
3286    
3287    ----------------------------------------------------------------------
3288    Name: Matthias Blume
3289    Date: 2002/01/29
3290    Tag: blume-20020129-INSTALL
3291    Description:
3292    
3293    - Added somewhat detailed installation instructions (file INSTALL).
3294    - Fixed curl-detection bug in config/install.sh.
3295    - It is now possible to select the URL getter using the URLGETTER
3296      environment variable:
3297    
3298          not set / "unknown"      --> automatic detection (script tries wget,
3299                                       curl, and lynx)
3300          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3301                                       how to properly invoke them)
3302          other                    --> use $URLGETTER directly, it must take
3303                                       precisely two command-line arguments
3304                                       (source URL and destination file name)
3305    
3306    ----------------------------------------------------------------------
3307    Name: Matthias Blume
3308    Date: 2002/01/28
3309    Tag: blume-20020128-sparc-ccalls
3310    Description:
3311    
3312    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3313    - Make use of the allocParam argument in sparc-c-calls.
3314    
3315    ----------------------------------------------------------------------
3316    Name: Matthias Blume
3317    Date: 2002/01/28
3318    Tag: blume-20020128-allocParam
3319    Description:
3320    
3321    John Reppy:  Changes c-calls API to accept client-callback for
3322    allocating extra stack space.
3323    me: Corresponding changes to mlriscGen (using a dummy argument that
3324        does not change the current behavior).
3325    
3326    ----------------------------------------------------------------------
3327    Name: Matthias Blume
3328    Date: 2002/01/28 12:00:00
3329    Tag: Release_110_38
3330    Description:
3331    
3332    This time for real!!!
3333    
3334    ----------------------------------------------------------------------
3335    Name: Matthias Blume
3336    Date: 2002/01/28 10:56:00 EST
3337    Tag: blume-20020128-retraction
3338    Description:
3339    
3340    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3341       with blume-Release_110_38-retracted.)
3342    
3343    1. Fixed a problem with incorrect rounding modes in real64.sml.
3344       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3345    
3346    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3347       and enums fixed.  The naming of corresponding ML identifiers should
3348       now be consistent again.
3349    
3350    ----------------------------------------------------------------------
3351    Name: Allen Leung
3352    Date: 2002/01/27
3353    Tag: leunga-20020127-nowhere
3354    Description:
3355    
3356       Added a target called nowhere in the configuration scripts.
3357       Enabling this will build the MLRISC 'nowhere' tool (for translating
3358       programs with where-clauses into legal SML code) during installation.
3359    
3360    ----------------------------------------------------------------------
3361    Name: Matthias Blume
3362    Date: 2002/01/25 21:27:00 EST
3363    Tag: blume-Release_110_38-retracted
3364    Description:
3365    
3366    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3367    
3368    README will be added later.
3369    
3370    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3371    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3372    
3373    ----------------------------------------------------------------------
3374    Name: Matthias Blume
3375    Date: 2002/01/25
3376    Tag: blume-20020125-ffi
3377    Description:
3378    
3379    A large number of tweaks and improvements to ml-nlffi-lib and
3380    ml-nlffigen:
3381    
3382       - ML represenation types have been streamlined
3383       - getter and setter functions work with concrete values, not abstract
3384         ones where possible
3385       - ml-nlffigen command line more flexible (see README file there)
3386       - some bugs have been fixed (hopefully)
3387    
3388    ----------------------------------------------------------------------
3389    Name: Lal George
3390    Date: 2002/01/24
3391    Tag: george-20020124-risc-ra-interface
3392    Description:
3393    
3394       There is a dramatic simplification in the interface to the
3395       register allocator for RISC architectures as a result of making
3396       parallel copy instructions explicit.
3397    
3398    ----------------------------------------------------------------------
3399    Name: Matthias Blume
3400    Date: 2002/01/22
3401    Tag: blume-20020122-x86-ccalls
3402    Description:
3403    
3404    Bug fix for c-calls on x86 (having to do with how char- and
3405    short-arguments are being handled).
3406    
3407    ----------------------------------------------------------------------
3408    Name: Matthias Blume
3409    Date: 2002/01/21
3410    Tag: blume-20020121-ff
3411    Description:
3412    
3413    Another day of fiddling with the FFI...
3414    
3415    1. Bug fix/workaround:  CKIT does not complain about negative array
3416       dimensions, so ml-nlffigen has to guard itself against this possibility.
3417       (Otherwise a negative dimension would send it into an infinite loop.)
3418    
3419    2. Some of the abstract types (light objects, light pointers, most "base"
3420       types) in structure C are now eqtypes.
3421    
3422    3. Added constructors and test functions for NULL function pointers.
3423    
3424    ----------------------------------------------------------------------
3425    Name: Matthias Blume
3426    Date: 2002/01/18
3427    Tag: blume-20020118-ready-for-new-release
3428    Description:
3429    
3430    Made config/srcarchiveurl point to a new place.  (Will provide boot
3431    files shortly.)
3432    
3433    Maybe we christen this to be 110.38?
3434    
3435    ----------------------------------------------------------------------
3436    Name: Matthias Blume
3437    Date: 2002/01/18
3438    Tag: blume-20020118-more-ffifiddle
3439    Description:
3440    
3441    Today's FFI fiddling:
3442    
3443      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3444        C.Set.* that use concrete (MLRep.*) arguments and results instead
3445        of abstract ones.
3446    
3447      - Provided word-style bit operations etc. for "int" representation
3448        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3449        and Long.
3450    
3451    ----------------------------------------------------------------------
3452    Name: Matthias Blume
3453    Date: 2002/01/18
3454    Tag: blume-20020118-use-x86-fp
3455    Description:
3456    
3457    Now that x86-fast-fp seems to be working, I turned it back on again
3458    by default.  (Seems to work fine now, even with the FFI.)
3459    
3460    Other than that, I added some documentation about the FFI to
3461    src/ml-nlffigen/README and updated the FFI test examples in
3462    src/ml-nlffi-lib/Tests/*.
3463    
3464    ----------------------------------------------------------------------
3465    Name: Allen Leung
3466    Date: 2002/01/17
3467    Tag: leunga-20020117-x86-fast-fp-call
3468    Description:
3469    
3470       1. Fixed a problem with handling return fp values when x86's fast fp
3471          mode is turned on.
3472    
3473       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3474          of %f32.
3475    
3476       3. Added a constructor INT32lit to the ast of MLRISC tools.
3477    
3478    ----------------------------------------------------------------------
3479    Name: Matthias Blume
3480    Date: 2002/01/16
3481    Tag: blume-20020116-ffifiddle
3482    Description:
3483    
3484    More fiddling with the FFI interface:
3485    
3486     - Make constness 'c instead of rw wherever possible.  This eliminates
3487       the need for certain explicit coercions.  (However, due to ML's
3488       value polymorphism, there will still be many cases where explicit
3489       coercions are necessary.  Phantom types are not the whole answer
3490       to modeling a subtyping relationship in ML.)
3491    
3492     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3493    
3494     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3495       clumsy workaround for varargs problem.)
3496    
3497    ----------------------------------------------------------------------
3498    Name: Lal George
3499    Date: 2002/01/15
3500    Tag: <none>
3501    Description:
3502    
3503    1. Since COPY instructions are no longer native to the architecture,
3504       a generic functor can be used to implement the expandCopies function.
3505    
3506    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3507       TEXT segment.
3508    
3509    ----------------------------------------------------------------------
3510    Name: Matthias Blume
3511    Date: 2002/01/15
3512    Tag: blume-20020115-ffiupdates
3513    Description:
3514    
3515    1. Fix for bug resulting in single-precision float values being returned
3516       incorrectly from FFI calls.
3517    
3518    2. Small modifications to C FFI API:
3519    
3520        - memory-allocation routines return straight objects (no options)
3521          and raise an exception in out-of-memory situations
3522        - unsafe extensions to cast between function pointers and pointers
3523          from/to ints
3524        - added structure C_Debug as an alternative to structure C where
3525          pointer-dereferencing (|*| and |*!) always check for null-pointers
3526        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3527          but also takes a (possibly empty) list of existing library handles
3528          that the current library depends on
3529    
3530    ----------------------------------------------------------------------
3531    Name: Matthias Blume
3532    Date: 2002/01/10
3533    Tag: blume-20020110-newffigen
3534    Description:
3535    
3536    1. Updates to portable graph code.
3537    
3538    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3539       more scalable now so that even huge interfaces such as the one
3540       for GTK compile in finite time and space. :-)
3541       See src/ml-nlffigen/README for details on what's new.
3542    
3543    ----------------------------------------------------------------------
3544    Name: Lal George
3545    Date: 2001/01/09 14:31:35 EST 2002
3546    Tag: george-20011206-rm-native-copy
3547    Description:
3548    
3549            Removed the native COPY and FCOPY instructions
3550            from all the architectures and replaced it with the
3551            explicit COPY instruction from the previous commit.
3552    
3553            It is now possible to simplify many of the optimizations
3554            modules that manipulate copies. This has not been
3555            done in this change.
3556    
3557    ----------------------------------------------------------------------
3558    Name: Lal George
3559    Date: 2001/12/06 16:50:13 EST 2001
3560    Tag: george-20011206-mlrisc-instruction
3561    Description:
3562    
3563    Changed the representation of instructions from being fully abstract
3564    to being partially concrete. That is to say:
3565    
3566      from
3567            type instruction
3568    
3569      to
3570            type instr                              (* machine instruction *)
3571    
3572            datatype instruction =
3573                LIVE of {regs: C.cellset, spilled: C.cellset}
3574              | KILL of {regs: C.cellset, spilled: C.cellset}
3575              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3576              | ANNOTATION of {i: instruction, a: Annotations.annotation}
3577              | INSTR of instr
3578    
3579    This makes the handling of certain special instructions that appear on
3580    all architectures easier and uniform.
3581    
3582    LIVE and KILL say that a list of registers are live or killed at the
3583    program point where they appear. No spill code is generated when an
3584    element of the 'regs' field is spilled, but the register is moved to
3585    the 'spilled' (which is present, more for debugging than anything else).
3586    
3587    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3588    We used to generate:
3589    
3590            DEFFREG f1
3591            f1 := f2 + f3
3592            trapb
3593    
3594    but now generate:
3595    
3596            f1 := f2 + f3
3597            trapb
3598            LIVE {regs=[f1,f2,f3], spilled=[]}
3599    
3600    Furthermore, the DEFFREG (hack) required that all floating point instruction
3601    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3602    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3603    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3604    this architecture).
3605    
3606    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3607    all the architectures. This will result in further simplification of the
3608    register allocator that must be aware of them for coalescing purposes, and
3609    will also simplify certain aspects of the machine description that provides
3610    callbacks related to parallel copies.
3611    
3612    ANNOTATION should be obvious, and now INSTR represents the honest to God
3613    machine instruction set!
3614    
3615    The <arch>/instructions/<arch>Instr.sml files define certain utility
3616    functions for making porting easier -- essentially converting upper case
3617    to lower case. All machine instructions (of type instr) are in upper case,
3618    and the lower case form generates an MLRISC instruction. For example on
3619    the alpha we have:
3620    
3621      datatype instr =
3622         LDA of {r:cell, b:cell, d:operand}
3623       | ...
3624    
3625      val lda : {r:cell, b:cell, d:operand} -> instruction
3626        ...
3627    
3628    where lda is just (INSTR o LDA), etc.
3629    
3630    ----------------------------------------------------------------------
3631    Name: Matthias Blume
3632    Date: 2001/11/22 21:40:00 EST
3633    Tag: Release_110_37
3634    Description:
3635    
3636    Release 110.37.  This time for real.
3637    
3638    ----------------------------------------------------------------------
3639    Name: Matthias Blume
3640    Date: 2001/11/21 16:35:00 EST
3641    Tag: blume-20011121-foot-in-mouth
3642    Description:
3643    
3644    Removed the "Release_110_37" tag because of a serious bug.
3645    This will be re-tagged once the bug is fixed.
3646    
3647    ----------------------------------------------------------------------
3648    Name: Matthias Blume
3649    Date: 2001/11/21 16:14:00 EST
3650    Tag: blume-20011121-forgottenfile
3651    Description:
3652    
3653    Forgot to add a file.  (Just a .tex-file -- part of
3654    the CM manual source.)
3655    
3656    ----------------------------------------------------------------------
3657    Name: Matthias Blume
3658    Date: 2001/11/21 16:10:00 EST
3659    Tag: blume-20011121-invalid_110_37
3660    Description:
3661    
3662    Note: I removed the original tag "Release_110_37" from this commit
3663          because we found a serious bug in all non-x86 backends.
3664          - Matthias
3665    
3666    1. Modifications to the SML/NJ code generator and to the runtime system
3667       so that code object name strings are directly inserted into code
3668       objects at code generation time.  The only business the runtime system
3669       has with this is now to read the name strings on occasions.
3670       (The encoding of the name string has also changed somewhat.)
3671    
3672    2. CM now implements a simple "set calculus" for specifying export lists.
3673       In particular, it is now possible to refer to the export lists of
3674       other libraries/groups/sources and form unions as well as differences.
3675       See the latest CM manual for details.
3676    
3677    3. An separate notion of "proxy" libraries has again be eliminated from
3678       CM's model.  (Proxy libraries are now simply a special case of using
3679       the export list calculus.)
3680    
3681    4. Some of the existing libraries now take advantage of the new set
3682       calculus.
3683       (Notice that not all libraries have been converted because some
3684       of the existing .cm-files are supposed to be backward compatible
3685       with 110.0.x.)
3686    
3687    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3688       "General.exnMessage"!  The former relies on a certain hook to be
3689       initialized, and that often does not happen in the stand-alone case.)
3690    
3691    ----------------------------------------------------------------------
3692    Name: Lal George
3693    Date: 2001/11/21  13:56:18 EST
3694    Tag: george-2001121-pseudo-ops
3695    Description:
3696    
3697      Implemented a complete redesign of MLRISC pseudo-ops. Now there
3698      ought to never be any question of incompatabilities with
3699      pseudo-op syntax expected by host assemblers.
3700    
3701      For now, only modules supporting GAS syntax are implemented
3702      but more should follow, such as MASM, and vendor assembler
3703      syntax, e.g. IBM as, Sun as, etc.
3704    
3705    ----------------------------------------------------------------------
3706    Name: Matthias Blume
3707    Date: 2001/11/14 11:52:00 EST
3708    Tag: blume-20011114-srcname
3709    Description:
3710    
3711    1. Routed the name of the current source file to mlriscgen where it
3712       should be directly emitted into the code object.  (This last part
3713       is yet to be done.)
3714    
3715    2. Some cleanup of the pgraph code to make it match the proposal that
3716       I put out the other day.  (The proposal notwithstanding, things are
3717       still in flux here.)
3718    
3719    ----------------------------------------------------------------------
3720    Name: Lal George
3721    Date: 2001/11/14 09:44:04 EST
3722    Tag:
3723    Description:
3724    
3725      Fix for a backpatching bug reported by Allen.
3726    
3727      Because the boundary between short and long span-dependent
3728      instructions is +/- 128, there are an astounding number of
3729      span-dependent instructions whose size is over estimated.
3730    
3731      Allen came up with the idea of letting the size of span
3732      dependent instructions be non-monotonic, for a maxIter
3733      number of times, after which the size must be monotonically
3734      increasing.
3735    
3736      This table shows the number of span-dependent instructions
3737      whose size was over-estimated as a function of maxIter, for the
3738      file Parse/parse/ml.grm.sml:
3739    
3740         maxIter            # of instructions:
3741            10                      687
3742            20                      438
3743            30                      198
3744            40                        0
3745    
3746      In compiling the compiler, there is no significant difference in
3747      compilation speed between maxIter=10 and maxIter=40. Actually,
3748      my measurements showed that maxIter=40 was a tad faster than
3749      maxIter=10! Also 96% of the  files in the compiler reach a fix
3750      point within 13 iterations, so fixing maxIter at 40, while high,
3751      is okay.
3752    
3753    ----------------------------------------------------------------------
3754    Name: Matthias Blume
3755    Date: 2001/10/31 15:25:00 EST
3756    Tag: blume-20011031-pgraph
3757    Description:
3758    
3759    CKIT:
3760    * Changed the "Function" constructor of type Ast.ctype to carry optional
3761      argument identifiers.
3762    * Changed the return type of TypeUtil.getFunction accordingly.
3763    * Type equality ignores the argument names.
3764    * TypeUtil.composite tries to preserve argument names but gives up quickly
3765      if there is a mismatch.
3766    
3767    installation script:
3768    * attempts to use "curl" if available (unless "wget" is available as well)
3769    
3770    CM:
3771    * has an experimental implementation of "portable graphs" which I will
3772      soon propose as an implementation-independent library format
3773    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
3774    
3775    NLFFI-LIB:
3776    * some cleanup (all cosmetic)
3777    
3778    NLFFIGEN:
3779    * temporarily disabled the mechanism that suppresses ML output for
3780      C definitions whose identifiers start with an underscore character
3781    * generate val bindings for enum constants
3782    * user can request that only one style (light or heavy) is being used;
3783      default is to use both (command-line arguments: -heavy and -light)
3784    * fixed bug in handling of function types involving incomplete pointers
3785    * generate ML entry points that take record arguments (i.e., using
3786      named arguments) for C functions that have a prototype with named
3787      arguments
3788      (see changes to CKIT)
3789    
3790    ----------------------------------------------------------------------
3791  Name: Allen Leung  Name: Allen Leung
3792  Date: 2001/10/27 20:34:00 EDT  Date: 2001/10/27 20:34:00 EDT
3793  Tag: leunga-20011027-x86-fast-fp-call  Tag: leunga-20011027-x86-fast-fp-call
# Line 109  Line 3884 
3884    
3885  ----------------------------------------------------------------------  ----------------------------------------------------------------------
3886  Name: Matthias Blume  Name: Matthias Blume
 >>>>>>> 1.169  
3887  Date: 2001/09/18 15:35:00 EDT  Date: 2001/09/18 15:35:00 EDT
3888  Tag: blume-20010918-readme11036  Tag: blume-20010918-readme11036
3889  Description:  Description:

Legend:
Removed from v.972  
changed lines
  Added in v.1672

root@smlnj-gforge.cs.uchicago.edu
ViewVC Help
Powered by ViewVC 1.0.0