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 572, Thu Mar 9 02:43:06 2000 UTC revision 1668, Wed Oct 27 22:22:13 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: 2004/10/27 17:20:00 CDT
18    Tag: blume-20041027-x86-c-calls
19    Description:
20    
21    This is the HISTORY entry for two earlier commits, both concerning
22    the x86 c-calls code in MLRISC:
23    
24      - added a missing LOAD in the code that deals with struct arguments
25      - made sure the caller does not add the wrong number of bytes to the
26        stack pointer after a call of a function returning a struct
27        (the callee already pops the implicit argument which points to
28         the space reserved for the result)
29    
30    ----------------------------------------------------------------------
31    Name: Allen Leung (leunga (at) reservoir (dot) com)
32    Date: 2004/10/24 14:00:00 EST
33    Tag: leunga-20041024-x86-gas-fucomip
34    Description:
35    
36       John discovered a bug in the syntax of fucomip.
37       The opcodes FU?COMIP? have been changed to
38    
39           fu?comip? %st(i), %st
40    
41    ----------------------------------------------------------------------
42    Name: Matthias Blume (blume (at) tti - c (dot) org)
43    Date: 2004/10/20 15:06:00 CDT
44    Tag: blume-20041020-standalone-backtrace
45    Description:
46    
47    Added a mechanism for getting back-trace information from standalone
48    programs.  Here is how it works:
49    
50       1. The part of the program from which you want to get backtrace
51          information (usually the whole program) should be wrapped with
52          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
53          main program could be modified from something like
54    
55             fun main (pgm, args) = ...
56    
57         to
58    
59            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
60    
61       2. To be able to access BackTrace.monitor, you have to add
62          library $smlnj-tdp/plugins.cm to the .cm file that contains your
63          main function.
64    
65       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
66          CM might have created in the past for your project).
67    
68       4. Build the system using this command line:
69    
70             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
71                      myprog.cm MyProg.main myprog
72    
73          instead of the usual
74    
75             ml-build myprog.cm MyProg.main myprog
76    
77    I changed a library name:
78    
79       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
80    
81    New libraries:
82    
83       $smlnj-tdp/back-trace.cm
84          -- when loaded causes the back-trace plugin to be installed
85       $smlnj-tdp/coverage.cm
86          -- when loaded causes the coverage plugin to be installed
87    
88    ----------------------------------------------------------------------
89    Name: Matthias Blume (blume (at) tti - c (dot) org)
90    Date: 2004/10/18 16:45:00 CDT
91    Tag: blume-20041018-groupowner
92    Description:
93    
94    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
95    
96    Eliminated group owner specs from .cm files throughout the source tree.
97    
98    ----------------------------------------------------------------------
99    Name: Matthias Blume (blume (at) tti - c (dot) org)
100    Date: 2004/10/15 15:45:00 CDT
101    Tag: blume-20041015-coverage
102    Description:
103    
104    * Test coverage tool added!
105    * Further reorganization of tracing-, debugging-, and profiling support:
106    
107        - moved original BTImp -- now called BackTrace -- into a separate
108          library called $/trace-debug-profile.cm
109        - eliminated all mentions of BTrace from SMLofNJ.Internals
110        - only the instrumentation mechanism is now left in the compiler proper
111        - BackTrace module is a plugin which is NOT plugged in by default
112        - Coverage module is another such plugin
113    
114    To get the benefits of any of these plugin modules, the code in
115    question must be compiled with tdp instrumentation turned on.  This
116    can be done by setting SMLofNJ.Internals.TDP.mode to true.
117    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
118    
119    Plugins are selected at link time.  (Pre-compiled instrumented code
120    can be re-loaded with different plugins in effect.)  When an
121    instrumented module is linked, whatever plugins are at that time
122    enabled will come into effect for that module.
123    
124    To enable the back-trace plugin, load library $/trace-debug-profile.cm
125    and invoke BackTrace.install() (e.g., from the interactive prompt).
126    To enable the coverage plugin, load the same library and invoke
127    Coverage.install().
128    
129    Back-traces are generated automatically on uncaught exceptions and
130    when the code in question explicitly invokes BackTrace.trigger().
131    
132    Coverage (and execution frequency-) information must be queried
133    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
134    
135    ----------------------------------------------------------------------
136    Name: Matthias Blume (blume (at) tti - c (dot) org)
137    Date: 2004/10/14 17:40:00 CDT
138    Tag: blume-20041014-tdp-core
139    Description:
140    
141    Snapshot of a significant overhaul of how the trace/debug/profile support
142    is hooked into the system (specifically: Core and SMLofNJ.Internals).
143    
144    ----------------------------------------------------------------------
145    Name: Matthias Blume (blume (at) tti - c (dot) org)
146    Date: 2004/10/13 16:34:00 CDT
147    Tag: blume-20041013-tdp
148    Description:
149    
150    Some rationalization of names:
151    
152       structure BTrace -> structure TDPInstrument
153       etc.
154    
155    This is is preparation of using the original back-trace
156    instrumentation for other purposes.  "TDP" stands for
157    Trace/Debug/Profile.
158    
159    The control flag controlling whether instrumentation is on or off is now
160    registered under a different name, so instead of running sml as
161    
162     sml -Cinstrument.btrace-mode=true
163    
164    one has to say
165    
166     sml -Ctdp.instrument=true
167    
168    ----------------------------------------------------------------------
169    Name: Matthias Blume (blume (at) tti - c (dot) org)
170    Date: 2004/10/11 16:37:00 CDT
171    Tag: blume-20041011-regions
172    Description:
173    
174    Made some minor modifications to elabcore.sml to have source regions
175    be propagated more tightly -- resulting in better (i.e., smaller)
176    regions being reported in error- and debug messages.
177    
178    ----------------------------------------------------------------------
179    Name: Matthias Blume (blume (at) tti - c (dot) org)
180    Date: 2004/10/08 22:50:00 CDT
181    Tag: blume-20041008-cmkw
182    Description:
183    
184    Fixed handling of keywords in .cm files: After seeing "is" the lexer
185    treats subsequent occurrences of "group", "library", "source", "is",
186    "*", and "-" as ordinary identifiers rather than keywords.
187    
188    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
189    is supposed to accept a tool argument called "source", but this did
190    not work because of the clash with the keyword.
191    
192    ----------------------------------------------------------------------
193    Name: Matthias Blume (blume (at) tti - c (dot) org)
194    Date: 2004/10/07 16:00:00 CDT
195    Tag: blume-20041007-cleanup
196    Description:
197    
198    Assorted cleanup work:
199    
200      - got rid of intstrmap in favor of using the library's
201        hash table implementation
202      - threw out most of the pathnames stuff, as it was not used anyway
203      - simplified tokentable implementation
204      - fixed some minor spelling errors
205    
206    ----------------------------------------------------------------------
207    Name: Matthias Blume (blume (at) tti - c (dot) org)
208    Date: 2004/10/06 15:15:15 CDT
209    Tag: blume-20041006-handler
210    Description:
211    
212    Cleaned up the absyn to reflect the invariant that HANDLE always
213    carries a FNexp as part of the type definition.  This eliminates some
214    superfluous sanity checks at runtime down the road.
215    
216    Some minor cleanup of the btrace code.
217    
218    ----------------------------------------------------------------------
219    Name: Matthias Blume (blume (at) tti - c (dot) org)
220    Date: 2004/10/01 10:20:30 CDT
221    Tag: blume-20041001-slave
222    Description:
223    
224    Added hack to make slave mode work in the presence of the version
225    tool.  (Still, since the master does two passes over the code for
226    CMB.make, the release number gets bumped twice when slaves are
227    attached.  I don't know if this is worth fixing...)
228    
229    ----------------------------------------------------------------------
230    Name: Matthias Blume (blume (at) tti - c (dot) org)
231    Date: 2004/09/30 10:55:00 CDT
232    Tag: blume-20040930-version
233    Description:
234    
235    * Moved the "version" magic into its own little library under
236    src/system/smlnj/internal.  This avoids expensive reconstruction
237    of a stable src/compiler/core.cm.
238    
239    * At the same time, structure CompilerVersion is now known as structure
240    SMLNJVersion.
241    
242    * Arranged for the version tool to NOT kick in when rebuilding the system
243      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
244      Also, loading the versiontool does not work when rebuilding the system
245      because CM is not properly initialized at that time.
246    
247    ----------------------------------------------------------------------
248    Name: Matthias Blume (blume (at) tti - c (dot) org)
249    Date: 2004/09/29 14:00:00 CDT
250    Tag: blume-20040929-autoversion
251    Description:
252    
253    Implemented some CM magic to have
254        file src/compiler/TopLevel/main/version.sml
255    generated automagically.
256    The version is taken from two files: config/version and config/release.
257    The first is expected to contain a two-part version number such as 110.49.
258    The second should contain a single number, but it may be missing.
259    
260    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
261    time the version tool is loaded (which is the first time you say CMB.make),
262    then the tool will increment the value stored in config/release every
263    time CMB.make is invoked.
264    
265    The binfile format is now insensitive to anything beyond the first
266    two components of a version number, so bumping the release does not render
267    binfiles incompatible.  Auto-bumping can be used to keep track of versions
268    during development without invalidating existing binfiles.
269    
270    In any case, every CMB.make updates the date information in version.sml.
271    (This is the date that is printed in the banner.)
272    
273    ----------------------------------------------------------------------
274    Name: Matthias Blume (blume (at) tti - c (dot) org)
275    Date: 2004/09/28 10:53:00 CDT
276    Tag: blume-20040928-controls
277    Description:
278    
279    Some cleanup of the controls code.
280    
281    ----------------------------------------------------------------------
282    Name: Matthias Blume (blume (at) tti - c (dot) org)
283    Date: 2004/09/27 22:08:00 CDT
284    Tag: blume-20040927-controls
285    Description:
286    
287    Added two pieces of functionality to the Controls interface:
288    
289    1.   val save'restore: 'a control -> unit -> unit
290    
291       grabs the current value of the control in stage 1 and restores it
292       in stage 2.
293    
294    2.   val set' : 'a control * 'a -> unit -> unit
295    
296       stores the given value into the control in stage 2 (i.e., delayed)
297       but does all error checking in stage 1.
298       (This is for string controls that need to do parse their argument --
299        something that might fail.  In some cases, notably in CM, one
300        already knows the intended argument but wants to delay the actual
301        assignment until a time when error recovery would be more difficult.)
302    
303    Changed the handling of controls in tool arguments to classes "sml" and
304    "lazysml":
305       - use Controls.save'restore as a more robust way of restoring the
306         old value (in particular: without having to re-parse the string)
307       - use controls to handle the "overload" keyword in the init group
308         (I believe this change actually fixes a long-standing obscure bug.)
309    
310    ----------------------------------------------------------------------
311    Name: Matthias Blume (blume (at) tti - c (dot) org)
312    Date: 2004/09/27 17:00:00 CDT
313    Tag: blume-20040927-lazysml
314    Description:
315    
316    Added a new tool class called "lazysml" to CM's tool chest.  The only
317    difference to "sml" is that compilation is done with Control.lazysml
318    set to true.  A source of class "lazysml" is automatically recognized
319    by a file name suffix of ".lml".
320    
321    In addition to the above feature, the original class "sml" now also
322    supports a tool argument "lazy" which has the same effect.  As a
323    result, the following three lines are equivalent:
324    
325        foo.sml : lazysml
326        foo.sml : sml (lazy)
327        foo.sml (lazy)
328    
329    The setting goes into effect both during parsing and during
330    compilation.  The original setting is restored right after parsing and
331    after compilation, respectively.
332    
333    In addition to all the above, there is also a general mechanism to set
334    ANY of the "controls" that are available at the command line via
335    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
336    apply as well. (In fact, "lazy" is implemented as a special case of
337    the general mechanism.)
338    
339    The .cm file syntax uses a new keyword tool argument called "with".
340    There are several ways of indicating the desired settings:
341    
342        foo.sml (with:parser.quotations=true)
343        foo.sml (with:(name:parser.quotations value:true))
344        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
345        foo.sml (with:(name1=value1 name2=value2 ...))
346        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
347    
348        etc.
349    
350    Another possible abbreviation is to leave out the =v or value:v part
351    if the name refers to a boolean control (in which case the value is
352    taken to be true).  Thus, one could get lazy sml also by saying:
353    
354        foo.sml (with:parser.lazy-keyword=true)
355        foo.sml (with:parser.lazy-keyword)
356        foo.sml (with:(name:parser.lazy-keyword value:true))
357        foo.sml (with:(name:parser.lazy-keyword))
358    
359    ----------------------------------------------------------------------
360    Name: Matthias Blume (blume (at) tti - c (dot) org)
361    Date: 2004/09/24 16:22:00 CDT
362    Tag: blume-20040924-ppc-long-branch
363    Description:
364    
365    Turned message about "emiting long form of branch" off by default.
366    Added a control flag to turn it back on when desired.
367    
368    ----------------------------------------------------------------------
369    Name: Matthias Blume (blume (at) tti - c (dot) org)
370    Date: 2004/09/24 16:05:00 CDT
371    Tag: blume-20040924-rounding
372    Description:
373    
374    Applied patch for setting rounding modes under Mac OS X.  Thanks to
375    Melissa O'Neill for providing the code!
376    
377    ----------------------------------------------------------------------
378    Name: Matthias Blume (blume (at) tti - c (dot) org)
379    Date: 2004/09/23 17:30:00 CDT
380    Tag: blume-20040923-envvars
381    Description:
382    
383    1. Changed definition of type ControlRegistry.registry_tree to
384       include control_info (i.e., the name of the controlling environment
385       variable).
386    
387    2. Added command-line flags -e and -E to print the names of environment
388       variables that can be used to control internal settings.
389       (This uses the new API mentioned in 1.)
390    
391    ----------------------------------------------------------------------
392    Name: Matthias Blume (blume (at) tti - c (dot) org)
393    Date: 2004/09/13 16:50:00 CDT
394    Tag: Release_110_49
395    Description:
396    
397    New working version (110.49).  NEW BOOTFILES!
398    
399    ----------------------------------------------------------------------
400    Name: Matthias Blume
401    Date: 2004/09/13 16:20:00 CDT
402    Tag: blume-20040913-config-mlrisc
403    Description:
404    
405    Put target "mlrisc" back into the default list.
406    (There is no harm in having it, and some users have expressed their
407    wish to have "mlrisc" included by default.)
408    
409    ----------------------------------------------------------------------
410    Name: John Reppy
411    Date: 2004/09/13
412    Tag: jhr-20040913-signals
413    Description:
414    
415    Fixed the signal masking code to properly nest mask/unmask operations
416    on a per-signal basis.
417    
418    ----------------------------------------------------------------------
419    Name: Matthias Blume (blume (at) tti - c (dot) org)
420    Date: 2004/09/08 13:20:00 CDT
421    Tag: blume-20040908-heap-magic
422    Description:
423    
424    Bumped the heap macig number to 0x09082004 to account for the changed
425    layout of the ML frame under MacOS X.
426    
427    ----------------------------------------------------------------------
428    Name: Allen Leung (leunga (at) reservoir (dot) com)
429    Date: 2004/09/03 11:26:00 EST
430    Tag: leunga-20040903-cygwin-install
431    Description:
432    
433       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
434    The Cygwin runtime is turned on by setting the environment
435    variable SMLNJ_CYGWIN_RUNTIME to 1.
436    
437    ----------------------------------------------------------------------
438    Name: Matthias Blume (blume (at) tti - c (dot) org)
439    Date: 2004/08/31 17:15:00 CDT
440    Tag: blume-20040831-core
441    Description:
442    
443    Added some exports to src/compiler/core.cm upon request by J. Joemann.
444    
445    ----------------------------------------------------------------------
446    Name: Matthias Blume (blume (at) tti - c (dot) org)
447    Date: 2004/08/30 17:55:00 CDT
448    Tag: blume-20040830-installer
449    Description:
450    
451    Upon request by Johannes Joemann:
452    
453    - improved ML code of installer to fall back to coping when renaming
454      fails (i.e., when source and target are on different file systems);
455      the code compiles but has yet to be tested in anger
456    - removed mlrisc from list of default targets (config/targets)
457    
458    ----------------------------------------------------------------------
459    Name: Matthias Blume (blume (at) tti - c (dot) org)
460    Date: 2004/08/27 17:20:00 CDT
461    Tag: blume-20040827-ptreql
462    Description:
463    
464    Added ptreql primop to structure InlineT (upon request from Larry
465    Paulson).
466    
467    ----------------------------------------------------------------------
468    Name: Allen Leung (leunga (at) reservoir (dot) com)
469    Date: 2004/08/15 21:21:00 EST
470    Tag: leunga-110_48-udgraph
471    Description:
472    
473    Another bug fix from Carl Hauser:
474    
475    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
476    > 48c48
477    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
478    > e::L)
479    > ---
480    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
481    > Without this, any deletion of an edge in an undirected graph does severe
482    > violence to the graph.
483    
484    ----------------------------------------------------------------------
485    Name: Allen Leung (leunga (at) reservoir (dot) com)
486    Date: 2004/08/10 23:35:00 EST
487    Tag: leunga-110_48-ppc
488    Description:
489    
490       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
491    
492    ----------------------------------------------------------------------
493    Name: Matthias Blume (blume (at) tti - c (dot) org)
494    Date: 2004/08/10 12:00:00 CDT
495    Tag: Release_110_48
496    Description:
497    
498    New working version (110.48).  NEW BOOTFILES!
499    
500    ----------------------------------------------------------------------
501    Name: Allen Leung (leunga (at) reservoir (dot) com)
502    Date: 2004/08/09 12:21:00 EST
503    Tag: leunga-110_47-dijsktra
504    Description:
505    
506    Bug fix from Carl Hauser:
507    
508    single_source_shortest_paths in dijkstra.sml was observed to get wrong
509    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
510    
511    The problem is that following the expression A.update(dist,s,Num.zero)
512    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
513    
514    ----------------------------------------------------------------------
515    Name: Matthias Blume (blume (at) tti - c (dot) org)
516    Date: 2004/08/06 18:10:00 CDT
517    Tag: blume-20040806-cmdline
518    Description:
519    
520    Fiddled with handling of command-line options:
521    
522      * sml now quits after processing the command line
523        if -H, -S, -h<n>, or -s<n> appears as the last
524        command-line argument
525      * a new option -q terminates the session when encountered on
526        the command line; subsequent arguments will be ignored
527      * bug fixes: short (erroneous) arguments are no longer ignored
528        completely
529    
530    ----------------------------------------------------------------------
531    Name: Allen Leung (leunga (at) reservoir (dot) com)
532    Date: 2004/08/04 18:17:00 EST
533    Tag: leunga-110_47-ppc-ibm-asm
534    Description:
535    
536      - Added minimal IBM assembly syntax support for PowerPC.
537    
538      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
539        been altered in the runtime.  We need an automatic way to keep the file
540        in sync.
541    
542    ----------------------------------------------------------------------
543    Name: Matthias Blume (blume (at) tti - c (dot) org)
544    Date: 2004/08/04 14:00:00 CDT
545    Tag: Release_110_47
546    Description:
547    
548    New working version (110.47).  NEW BOOTFILES!
549    
550    ----------------------------------------------------------------------
551    Name: Matthias Blume (blume (at) tti - c (dot) org)
552    Date: 2004/08/03 14:25:00 CDT
553    Tag: blume-20040803-callingconv
554    Description:
555    
556    Added low-level support for choosing C calling conventions by
557    twiddling the type of rawccall.  (See
558    src/compiler/Semant/types/cproto.sml for details.)
559    
560    ----------------------------------------------------------------------
561    Name: Matthias Blume (blume (at) tti - c (dot) org)
562    Date: 2004/08/02 15:55:00 CDT
563    Tag: blume-20040802-backout
564    Description:
565    
566    Backed out of change to win32-filesys.c.  The earlier patch to
567    get_file_time caused CM to produce files with the wrong time stamp.
568    
569    ----------------------------------------------------------------------
570    Name: Matthias Blume (blume (at) tti - c (dot) org)
571    Date: 2004/08/02 14:45:00 CDT
572    Tag: blume-20040802-nlffi-win32
573    Description:
574    
575    Added NLFFI support for Win32, adapted from a patch provided by David
576    Hansel.  This is currently completely untested.  Also, the issue
577    concerning stdcall vs. ccall is still unresolved.
578    
579    ----------------------------------------------------------------------
580    Name: Matthias Blume (blume (at) tti - c (dot) org)
581    Date: 2004/07/30 17:55:00 CDT
582    Tag: blume-20040730-various
583    Description:
584    
585    Gearing up towards 110.47...
586    
587    - various minor bugfixes to ml-nlffigen
588    - a beginning of a manual for nlffi
589    
590    - eliminated 'export name=value' in config/install.sh as this does
591      not work with certain versions of /bin/sh
592      (Thanks to David King at Motorola for catching this.)
593    
594    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
595      - added a test for tm==NULL to gmtime.c and localtime.c
596      - applied patch for incorrect GetFileTime under win32
597      - toSeconds -> toMilliseconds in Win32/win32-process.sml
598    
599    ----------------------------------------------------------------------
600    Name: Matthias Blume (blume (at) tti - c (dot) org)
601    Date: 2004/07/21 18:20:00 CDT
602    Tag: blume-20040721-nlffigen
603    Description:
604    
605    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
606      for a typedef to an incomplete type, but leave out the "typ" member.
607      (This is just for consistency.)
608    - Started to produce what is supposed to become better (i.e., comprehensive)
609      documentation of what ml-nlffigen does and produces.
610    
611    ----------------------------------------------------------------------
612    Name: Matthias Blume (blume (at) tti - c (dot) org)
613    Date: 2004/07/14 16:25:00 CDT
614    Tag: blume-20040714-union
615    Description:
616    
617    Added C_UNION to c-calls/c-types.sml and updated the machinery
618    (ml-nlffigen, cproto.sml) that conveys C function interface
619    information to the code generator.
620    
621    However, the actual architecture-specific implementation of function
622    arguments and results that are C unions is still not implemented.
623    
624    ----------------------------------------------------------------------
625    Name: Allen Leung (leunga (at) reservoir (dot) com)
626    Date: 2004/07/14 14:38:00 EST
627    Tag: leunga-110_46_1-ppc-lwzu
628    Description:
629    
630        Added these instructions to the PowerPC architecture:
631        LBZU(X), LHZU(X), LWZU(X),
632        STWU(X), STFDU, STFSU
633    
634        etc...
635    
636        Note: I haven't added their instruction encoding into the description.
637    
638    ----------------------------------------------------------------------
639    Name: Allen Leung (leunga (at) reservoir (dot) com)
640    Date: 2004/07/13 15:04:00 EST
641    Tag: leunga-110_46_1-ppc-lwarx
642    Description:
643    
644       Added the two instructions LWARX and STWCX to the PowerPC
645    instruction set.
646    
647       A (untested) rewrite of loop-structure.sml.  The old version
648    is completely broken.
649    
650    ----------------------------------------------------------------------
651    Name: Matthias Blume (blume (at) tti - c (dot) org)
652    Date: 2004/07/13 13:50:00 CDT
653    Tag: blume-20040713-nlffi
654    Description:
655    
656    - use paramAlloc to report c-calls with too many arguments
657      (for PPC version where parameter area is pre-allocated)
658    - added ccall_maxargspace to machspec (to implement the above)
659    - made "make" commend in CM's "make" tool configurable
660    - added option (default: on) for passing the name of the SML/NJ's "bin"
661      directory to "make"; the call looks like this:
662    
663         make <options> SMLNJ_BINDIR=<dir> <target>
664    
665      This can be used by the Makefile to, e.g., pick the "right" version
666      of ml-nlffigen.
667    - minor code tweaks
668    
669    ----------------------------------------------------------------------
670    Name: Matthias Blume (blume (at) tti - c (dot) org)
671    Date: 2004/07/12 22:50:00 CDT
672    Tag: blume-110_46_1-macosx-nlffi
673    Description:
674    
675    NLFFI under Mac OS X now working (sort of).  This is largely untested,
676    though.
677    
678    Note:  1. You have to make a new, clean build of the runtime system.
679           2. There are new BOOTFILES, you have to use them!
680              (Doing the bootstrap process yourself would be *very* painful!
681               If you absolutely have to do it, build the system under
682               a different architecture and then cross-compile.)
683    
684    Version bumped to 110.46.1 to account for runtime data format changes.
685    
686    ----------------------------------------------------------------------
687    Name: Matthias Blume (blume (at) tti - c (dot) org)
688    Date: 2004/06/18 14:30:00 CDT
689    Tag: blume-20040618-unix
690    Description:
691    
692    Changed the implementation of structure Unix so that the same stream
693    is returned every time one of the {text,bin}{In,Out}streamOf functions
694    is invoked on the same proc.  This is not what the spec currently
695    says -- although IMO it arguably should.  (See discussion below.)
696    
697    ----------------------------------------------------------------------
698    Name: Matthias Blume (blume (at) tti - c (dot) org)
699    Date: 2004/06/17 18:15:00 CDT
700    Tag: Release_110_46
701    Description:
702    
703    New working version (110.46).  NEW BOOTFILES!
704    
705    ----------------------------------------------------------------------
706    Name: Matthias Blume (blume (at) tti - c (dot) org)
707    Date: 2004/06/17 17:20:00 CDT
708    Tag: blume-20040617-timer-unix
709    Description:
710    
711    Changed the interface of structures Timer and Unix to match the most
712    recent Basis spec.
713    
714    In the case of Unix there still seems to be an open/weird issue:
715    
716       The {text,bin}{In,Out}streamOf functions are supposed to create
717       fresh streams whenever they are called -- as opposed to have them
718       return the same stream every time.  This design is supposed to
719       prevent space leaks caused by proc values hanging on to streams.
720    
721       The reap function, on the other hand, is supposed to close the
722       streams.  This cannot be done without having a handle on the
723       stream in proc after all...
724    
725    I took the liberty to implement the following stopgap solution:
726    
727      The proc value hangs on to the most recently created stream(s).
728      Reap closes those.  If either or both of the two streams hadn't
729      been created at all yet, then reap will close the corresponding
730      file descriptors directly.
731    
732    PS: I don't understand the original space leak argument anymore.  If
733    a proc hangs on to the imperative stream, then I/O operations on those
734    will advance the state of the cached stream and avoid the space leak.
735    
736    ----------------------------------------------------------------------
737    Name: Matthias Blume (blume (at) tti - c (dot) org)
738    Date: 2004/05/28 16:45:00 CDT
739    Tag: blume-20040528-basis
740    Description:
741    
742    Added signature PACK_REAL and exported functor PrimIO.
743    
744    ----------------------------------------------------------------------
745    Name: Matthias Blume (blume (at) tti - c (dot) org)
746    Date: 2004/05/25 16:00:00 CDT
747    Tag: blume-20040525-group-owner
748    Description:
749    
750    CM now ignores (but still accepts) the "owner" information in group
751    descriptions.  The owner of a group is its next enclosing
752    library. Each group must have a unique owner.  (There is a virtual
753    "toplevel" library that owns groups which are not nested within a real
754    library.)  Previously, each group had to explicitly declare its owner,
755    and CM would check that such a declaration is correct.  The new scheme
756    is to have CM check that for each group there is precisely one owning
757    library.
758    
759    The advantage of the new scheme is that the programmer no longer needs
760    to maintain the somewhat annoying owner information.  The downside is
761    that CM cannot enforce the ownership rule across multiple runs of
762    CM.make.  Fortunately, enclosing the same group in two different
763    libraries A and B which are not part of the same program does not
764    cause real problems.
765    
766    ----------------------------------------------------------------------
767    Name: Matthias Blume (blume (at) tti - c (dot) org)
768    Date: 2004/05/20 16:00:00 CDT
769    Tag: blume-20040520-win32
770    Description:
771    
772    Made the win32 version work again.  (Strangely, a misplaced comma had
773    slipped into win32-process.c which prevented the runtime from being
774    compiled correctly.)
775    
776    Also, included a minor addition to ml-build.bat analogous to what was
777    done in blume-20040519-ml-build.
778    
779    ----------------------------------------------------------------------
780    Name: Matthias Blume (blume (at) tti - c (dot) org)
781    Date: 2004/05/19 22:10:00 CDT
782    Tag: blume-20040519-ml-build
783    Description:
784    
785    Arranged for ml-build to clean up after itself a little bit better.
786    The script generates a temporary SML source file and compiles it using
787    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
788    gets rid of those at the end, so they don't accumulate under .cm.
789    
790    This required a minor change to install.sh because the name of the
791    metadata directory (default: .cm) is actually configurable at
792    installation time.
793    
794    ----------------------------------------------------------------------
795    Name: Matthias Blume (blume (at) tti - c (dot) org)
796    Date: 2004/05/18 15:50:00 CDT
797    Tag: blume-20040518-mkreader
798    Description:
799    
800    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
801    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
802    
803    ----------------------------------------------------------------------
804    Name: Matthias Blume (blume (at) tti - c (dot) org)
805    Date: 2004/05/11 14:35:00 CDT
806    Tag: blume-20040511-win32sock
807    Description:
808    
809    Added previously missing support for many socket-related functions
810    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
811    for the voluminous patch!
812    
813    (I have not tested this patch under win32 yet.)
814    
815    Here is David's e-mail:
816    
817    Hi,
818    
819    Attached to this email you find a diff against sml/nj 110.45
820    that will enable socket support under Windows.
821    
822    To apply the patch (using unix or cygwin)
823    1) gunzip runtime.diff.gz
824    2) "cd" into "src/runtime" in the source tree of a fresh
825       110.45 installation.
826    3) patch -p 1 < [your/path/to]runtime.diff
827    
828    The code compiles fine but has NOT yet been extensively tested.
829    I only ran a few tests for basic socket client functionality
830    (which worked fine).  Especially the functions that use ioctl
831    are not tested at all and might not work (see below).
832    
833    I implemented this since we want to move to a newer version of sml/nj
834    but need socket support in order to use it.  This is the first time I
835    even had a look at the sml/nj source,  so please review my changes
836    before making this part of the distribution!  Here are a few issues
837    that I think might be better for someone to solve who is more
838    familiar with the sml/nj source (and socket programming):
839    
840    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
841      exception since I could not figure out what the windows equivalent
842      of these functions is
843    
844    - In sockets-osdep.h there are a some #include statements that are
845      only used in a few files that include sockets-osdep.h
846    
847    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
848      does not process its return value since I don't know how to
849      report an error upwards.
850    
851    - It would probably be good to have a call to WSACleanup() when
852      the library is unloaded (if there is such a possibility).
853      Otherwise I think Windows will take care of this automatically
854      when the process finishes.
855    
856    - I used ioctlsocket() as a replacement for ioctl() but I have
857      no idea if that is actually the proper replacement on Windows.
858    
859    - All these issues are marked in the code by "FIXME" comments.
860    
861    We use sml/nj extensively in our products and are quite happy
862    with it.  I hope this contribution will help you.
863    
864    Keep up the good work!
865    
866    David
867    
868    ----------------------------------------------------------------------
869    Name: Matthias Blume (blume (at) tti - c (dot) org)
870    Date: 2004/05/11 14:20:00 CDT
871    Tag: blume-20040511-installml
872    Description:
873    
874    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
875    the patch.)
876    
877    ----------------------------------------------------------------------
878    Name: Matthias Blume (blume (at) tti - c (dot) org)
879    Date: 2004/05/11 14:05:00 CDT
880    Tag: blume-20040511-nlffi-netbsd
881    Description:
882    
883    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
884    the patch.)
885    
886    ----------------------------------------------------------------------
887    Name: Matthias Blume (blume (at) tti - c (dot) org)
888    Date: 2004/05/11 12:05:00 CDT
889    Tag: blume-20040511-exports
890    Description:
891    
892    As per request by Adam Chlipala <adam@hcoop.net>, extended various
893    export lists in compiler-related .cm-files.
894    
895    ----------------------------------------------------------------------
896    Name: Matthias Blume (blume (at) tti - c (dot) org)
897    Date: 2004/05/11 11:35:00 CDT
898    Tag: blume-20040511-allsource
899    Description:
900    
901    The installer now honors the "src-smlnj" target again, although its meaning
902    has changed from "all sources required for the compiler" to "all sources
903    the installer knows about".  In other words, if you enable "src-smlnj"
904    in the "targets" file, then the installer will pull in sources for
905    everything.  (Notice that this refers to source code only.  Compiled
906    code is still only installed for modules that were requested explicitly
907    or which are required for other modules that were requested explicitly.)
908    
909    ----------------------------------------------------------------------
910    Name: Matthias Blume (blume (at) tti - c (dot) org)
911    Date: 2004/04/23 17:40:00 CDT
912    Tag: blume-20040423-ieee-scan
913    Description:
914    
915    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
916    in the exponent calculation we get INF or ZERO (depending on the mantissa
917    and the sign of the exponent).
918    
919    ----------------------------------------------------------------------
920    Name: Matthias Blume (blume (at) tti - c (dot) org)
921    Date: 2004/04/23 10:40:00 CDT
922    Tag: blume-20040423-ml-build
923    Description:
924    
925    The ml-build script now terminates with a non-0 status when something
926    goes wrong.
927    
928    ----------------------------------------------------------------------
929    Name: Matthias Blume (blume (at) tti - c (dot) org)
930    Date: 2004/04/22 16:35:00 CDT
931    Tag: blume-20040422-Option
932    Description:
933    
934    Made exception Option to be the same as exception Option.Option
935    (as it should be).
936    
937    ----------------------------------------------------------------------
938    Name: Allen Leung (leunga (at) reservoir (dot) com)
939    Date: 2004/03/19 14:40:00 EST
940    Tag: leunga-20040319-cygwin-nlffi
941    Description:
942    
943    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
944    of SML/NJ.  The problem is that
945    
946       lib = dlopen(NULL, ...)
947       f   = dlsym(lib, "malloc");
948    
949    does not work on Windows unless we explicitly export symbols
950    such as 'malloc' during linking.  We fixed this by explicitly
951    exporting the required symbols with the magic gcc incantation:
952    
953        -Wl,--export-all cygwin.def
954    
955    where cygwin.def is a file containing all the symbols that
956    we wish to export.
957    
958    I suspect this is a Windows problem and we'll have to
959    do the same (somehow with windows compilers) when
960    we build the native win32 version with the system
961    calls LoadLibrary/GetProcAddress.
962    
963    ----------------------------------------------------------------------
964    Name: Matthias Blume (blume (at) tti - c (dot) org)
965    Date: 2004/03/04 16:35:00 CST
966    Tag: blume-20040304-intinf-fmt
967    Description:
968    
969    Fixed problem with IntInf.fmt (sign would show up on the right instead
970    of on the left for BIN, OCT, and HEX).
971    
972    ----------------------------------------------------------------------
973    Name: Matthias Blume (blume (at) tti - c (dot) org)
974    Date: 2004/03/04 11:25:00 CST
975    Tag: blume-20040304-symlinks
976    Description:
977    
978    Fixed problem with installer script (unix only) where bin/ml-yacc and
979    friends pointed (via symlinks) to absolute locations instead of just
980    .run-sml.  This was reported by Vesa A Norrman.
981    
982    ----------------------------------------------------------------------
983    Name: Matthias Blume (blume (at) tti - c (dot) org)
984    Date: 2004/02/13 14:50:00 CST
985    Tag: Release_110_45
986    Description:
987    
988    New working version (110.45). New bootfiles.
989    
990    ----------------------------------------------------------------------
991    Name: Matthias Blume (blume (at) tti - c (dot) org)
992    Date: 2004/01/26 15:15:15 CST
993    Tag: blume-20040126-toplevel
994    Description:
995    
996    Improved handling of exceptions at the interactive toplevel.
997    
998    ----------------------------------------------------------------------
999    Name: Matthias Blume (blume (at) tti - c (dot) org)
1000    Date: 2004/01/26 11:25:00
1001    Tag: blume-20040126-app
1002    Description:
1003    
1004    Type of top-level "app" corrected.
1005    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1006    
1007    ----------------------------------------------------------------------
1008    Name: Matthias Blume (blume (at) tti - c (dot) org)
1009    Date: 2003/11/18 17:10 CST
1010    Tag: blume-20031118-basis-fiddle
1011    Description:
1012    
1013    - changed Timer interface to what might become the spec
1014    - POSIX_FLAGS -> BIT_FLAGS according to spec
1015    - some other minor discrepancies wrt. spec eliminated
1016    
1017    ----------------------------------------------------------------------
1018    Name: Matthias Blume (blume (at) tti - c (dot) org)
1019    Date: 2003/11/06 12:00:00 CST
1020    Tag: Release_110_44
1021    Description:
1022    
1023    New working version (110.44). New bootfiles.
1024    
1025    ----------------------------------------------------------------------
1026    Name: Matthias Blume (blume (at) tti - c (dot) org)
1027    Date: 2003/11/04 11:50:00 CST
1028    Tag: blume-20031104-move-libraries
1029    Description:
1030    
1031    Eliminated the "dont_move_libraries" directive in config/targets.
1032    (The mechanism was broken and could not be fixed easily.  Moreover,
1033    there does not seem to be any reason not to move all libraries into
1034    lib during installation.  I originally implemented this directive as a
1035    backward-compatibility feature when I first introduced the new CM.
1036    Now that things have been stable for a long time and going back to the
1037    old CM is not an option, there is no reason to keep it around.)
1038    
1039    ----------------------------------------------------------------------
1040    Name: Matthias Blume (blume (at) tti - c (dot) org)
1041    Date: 2003/11/03 16:00:00 CST
1042    Tag: blume-20031103-installdir
1043    Description:
1044    
1045    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1046    Richards for pointing out the problem and providing the solution.)
1047    
1048    ----------------------------------------------------------------------
1049    Name: Matthias Blume (blume (at) tti - c (dot) org)
1050    Date: 2003/10/01 17:05:00 CDT
1051    Tag: blume-20031001-lal-mlrisc
1052    Description:
1053    
1054    MLRISC bug fix from Lal.
1055    
1056    ----------------------------------------------------------------------
1057    Name: Matthias Blume (blume (at) tti - c (dot) org)
1058    Date: 2003/09/30 16:10:00 CDT
1059    Tag: blume-20030930-primio-bat
1060    Description:
1061    
1062    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1063    2. Improved .bat files (for Win32 port) to make things work under Win95.
1064       (thanks to Aaron S. Hawley for this one)
1065    
1066    ----------------------------------------------------------------------
1067    Name: Matthias Blume (blume (at) tti - c (dot) org)
1068    Date: 2003/09/26 16:05:00 CDT
1069    Tag: blume-20030926-wrappriv
1070    Description:
1071    
1072    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1073    
1074    ----------------------------------------------------------------------
1075    Name: Matthias Blume (blume (at) tti - c (dot) org)
1076    Date: 2003/09/26 15:00:00 CDT
1077    Tag: blume-20030926-110_43_3
1078    Description:
1079    
1080    - additional cleanup
1081    - version number bump, NEW BOOTFILES
1082    
1083    ----------------------------------------------------------------------
1084    Name: Matthias Blume (blume (at) tti - c (dot) org)
1085    Date: 2003/09/26 12:00:00 CDT
1086    Tag: blume-20030926-ppautoload
1087    Description:
1088    
1089    I modified the read-eval-print loop so that the autoloader gets
1090    invoked whenever the prettyprinter tries to look up a symbol that
1091    is not currently defined in the toplevel environment but which
1092    appears in CM's autoload registry.  As a result, we see far fewer of
1093    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1094    
1095    In addition to this I tried to clean up some pieces of the Basis
1096    implementation (e.g., Socket, Word8Array) in order to prevent other
1097    instances of these ?.Foo.Bar.xxx names from being printed.
1098    
1099    The mechanism that picks names for types still needs some work, though.
1100    (Right now it seems that if there is a type A.t which is defined to
1101    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1102    "?.B.u" although the perhaps more sensible solution would be to use
1103    "A.t" in this case.  In other words, the prettyprinter should follow
1104    a chain of DEFtycs not farther than there are corresponding toplevel
1105    names in the current environment.)
1106    
1107    ----------------------------------------------------------------------
1108    Name: Matthias Blume (blume (at) tti - c (dot) org)
1109    Date: 2003/09/24 16:31:00 CDT
1110    Tag: blume-20030924-installer
1111    Description:
1112    
1113    Another installer tweak:  All the ML code for the installer is now
1114    compiled during CMB.make and put into a little library called
1115    $smlnj/installer.cm.  The installation then simply invokes
1116    
1117       sml -m $smlnj/installer.cm
1118    
1119    and everything happens automagically.
1120    
1121    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1122    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1123            and BINDIR.
1124    
1125    The new scheme guarantees that the ML code responsible for the installation
1126    is in sync with the APIs of the main system.  Also, the installer is
1127    somewhat faster because the installer script is precompiled.
1128    
1129    ----------------------------------------------------------------------
1130    Name: Matthias Blume (blume (at) tti - c (dot) org)
1131    Date: 2003/09/24 15:35:00 CDT
1132    Tag: blume-20030924-synsock
1133    Description:
1134    
1135    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1136    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1137    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1138    implementations of non-blocking operations from CML's Socket
1139    structure.
1140    
1141    ----------------------------------------------------------------------
1142    Name: Matthias Blume (blume (at) tti - c (dot) org)
1143    Date: 2003/09/24 15:10:05 CDT
1144    Tag: blume-20030924-sockets
1145    Description:
1146    
1147    1. Fixed SOCKET API and implementation to match Basis spec.
1148       This required changing the internal representation of sockets to one
1149       that remembers (for each socket file descriptor) whether it is currently
1150       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1151       call is made only if the state actually needs to change).
1152    
1153    2. OS-specific details of sockets were moved into separate files, thus
1154       making it possible to unify the bulk of the socket implementations
1155       between Unix and Win32.
1156    
1157    3. CML's socket API changed accordingly.
1158       (Note that we need to remove non-blocking functions from this API
1159       since they are redundant in the case of CML!)
1160    
1161    4. CML's socket implementation now makes use of non-blocking functions
1162       provided by Basis, thus removing all OS-dependent code from this part
1163       of CML.
1164    
1165    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1166    
1167    ----------------------------------------------------------------------
1168    Name: Matthias Blume (blume (at) tti - c (dot) org)
1169    Date: 2003/09/22 12:10:00 CDT
1170    Tag: blume-20030922-110_43_2
1171    Description:
1172    
1173    Made a new interim version and bootfiles for developer's bootstrapping
1174    convenience.
1175    
1176       110.43.2 -- NEW BOOTFILES
1177    
1178    ----------------------------------------------------------------------
1179    Name: Matthias Blume (blume (at) tti - c (dot) org)
1180    Date: 2003/09/19 15:55:00 CDT
1181    Tag: blume-20030919-cmdir
1182    Description:
1183    
1184    1. new-install.sh -> install.sh
1185    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1186    3. tweaked installer so that another name instead of .cm can be chosen
1187       at install time (by setting the CM_DIR_ARC environment variable
1188       during installation); once installation is complete, the name is
1189       fixed
1190    
1191    ----------------------------------------------------------------------
1192    Name: Matthias Blume (blume (at) tti - c (dot) org)
1193    Date: 2003/09/18 16:00:00 CDT
1194    Tag: blume-20030918-110_43_1
1195    Description:
1196    
1197    Made a new interim version and bootfiles for developer's bootstrapping
1198    convenience.
1199    
1200       110.43.1 -- NEW BOOTFILES
1201    
1202    ----------------------------------------------------------------------
1203    Name: Matthias Blume (blume (at) tti - c (dot) org)
1204    Date: 2003/09/18 15:20:00 CDT
1205    Tag: blume-20030918-misc
1206    Description:
1207    
1208    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1209       this seems to be controversial at the moment) and used those in
1210       Posix.ProcEnv.times.
1211    
1212    2. Added Time.{from,to}Nanoseconds to Time.
1213    
1214    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1215       For example, fromLargeInt never needs to look at more than 3 "big
1216       digits" to get its 53 bits of precision.
1217    
1218    ----------------------------------------------------------------------
1219    Name: Matthias Blume (blume (at) tti - c (dot) org)
1220    Date: 2003/09/17 16:30:00 CDT
1221    Tag: blume-20030917-real32-slices
1222    Description:
1223    
1224    Added an entry to the primitive environment
1225    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1226    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1227    
1228    Removed some of the "magic" constants in real64.sml and replaced them
1229    with code that generates these values from their corresponding
1230    integer counterparts.
1231    
1232    Made all(?) the slice-related changes to the Basis and made everything
1233    compile again...
1234    
1235    ----------------------------------------------------------------------
1236    Name: Matthias Blume (blume (at) tti - c (dot) org)
1237    Date: 2003/09/15 17:45:00 CDT
1238    Tag: blume-20030915-rbase
1239    Description:
1240    
1241    Fixed bug in Real.fromLargeInt.
1242    
1243    ----------------------------------------------------------------------
1244    Name: Matthias Blume (blume (at) tti - c (dot) org)
1245    Date: 2003/09/13 18:11:00 CDT
1246    Tag: blume-20030913-libinstall
1247    Description:
1248    
1249    Minor bugfix in config/libinstall (set anchor with path to
1250    standalone tool after installing it, otherwise libraries that
1251    need ml-lex or ml-yacc won't compile the first time the installer
1252    runs).
1253    
1254    ----------------------------------------------------------------------
1255    Name: Matthias Blume (blume (at) tti - c (dot) org)
1256    Date: 2003/09/12 11:45:00 CDT
1257    Tag: blume-20030912-various
1258    Description:
1259    
1260     - fixed bug in Real.toLargeInt
1261     - fixed bug in Posix.ProcEnv.times
1262     - changed inputLine functions to return an option
1263     - minor installer improvements / bugfixes
1264     - changed default @SMLalloc parameter for x86/celeron to 64k
1265    
1266    ----------------------------------------------------------------------
1267    Name: Matthias Blume (blume (at) tti - c (dot) org)
1268    Date: 2003/09/09 22:00:00 CDT
1269    Tag: Release_110_43
1270    Description:
1271    
1272        New working release 110.43.  New bootfiles.
1273    
1274    ----------------------------------------------------------------------
1275    Name: Matthias Blume (blume (at) tti - c (dot) org)
1276    Date: 2003/09/09 19:20:00 CDT
1277    Tag: blume-20030909-installer
1278    Description:
1279    
1280    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1281    Modified config/install.bat to take advantage of it.  Also modified
1282    config/install.sh (and called it config/new-install.sh) to take advantage
1283    of it on Unix systems.  (The SML code is (supposed to be) platform-
1284    independent.)
1285    
1286    The installer can now install everything under Win32
1287    as well as under *nix as long as it compiles.
1288    
1289    Other changes:
1290    
1291      - made CML compile again under Win32
1292      - made eXene compile under Win32 (by providing a fake structure UnixSock
1293        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1294      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1295        same as type int;  under Win32 it isn't
1296      - fixed some slice-related problems in the win32-specific parts of CML
1297      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1298        (under Win32, the volume name is case-insensitive, and the
1299        OS.Path code compares volume names for equality)
1300    
1301    ----------------------------------------------------------------------
1302    Name: Matthias Blume (blume (at) tti - c (dot) org)
1303    Date: 2003/09/08 11:55:00 CDT
1304    Tag: blume-20030908-fullpath
1305    Description:
1306    
1307    Made Win32 version of OS.FileSys.fullPath return current directory
1308    when given an empty string.  This is what the spec says, and incidentally,
1309    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1310    cases when presented with the name of a non-existing .cm file.)
1311    
1312    ----------------------------------------------------------------------
1313    Name: Matthias Blume (blume (at) tti - c (dot) org)
1314    Date: 2003/09/04 16:30:00 CDT
1315    Tag: blume-20030905-slices-etc
1316    Description:
1317    
1318      1. Changed interface to vectors and arrays in Basis to match
1319         (draft) Basis spec.
1320      2. Added signatures and implementations of slices according to
1321         Basis spec.
1322      3. Edited source code throughout the system to make it compile again
1323         under 1. and 2.  (In some cases code had to be added to have it
1324         match the new signatures.)
1325      4. MLRISC should be backward-compatible: the copies of the originals
1326         of files that needed to change under 3. were retained, the .cm files
1327         check the compiler version number and use old versions when
1328         appropriate.
1329      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1330         dirstream -> string option (in accordance with Basis spec).
1331      6. When generating code that counts lines, ml-lex used function
1332         CharVector.foldli, taking advantage of its old interface.
1333         This has been replaced with the corresponding code from
1334         CharVectorSlice.  (html-lex must be re-lexed!)
1335      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1336         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1337      8. Removed temporary additions (fromInternal, toInternal) from the
1338         (now obsolete) IntInf in smlnj-lib/Util.
1339      9. Cleaned up structure Byte.
1340     10. Added localOffset, scan, and fromString to Date (according to spec).
1341         Cleaned/corrected implementation of Date.
1342         (Still need to check for correctness; implement better canonicalizeDate.)
1343     11. Added "scan" to signature IEEE_REAL.
1344     12. Some improvements to IntInf [in particular: efficiency-hack for
1345         mod and rem when second operand is 2 (for parity checks).]
1346     13. Changed representation of type Time.time, using a single IntInf.int
1347         value counting microseconds.  This considerably simplified the
1348         implementation of structure Time.  We now support negative time
1349         values; scan and fromString handle signs.
1350     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1351         ArraySlice).
1352    
1353    ----------------------------------------------------------------------
1354    Name: Matthias Blume (blume (at) tti - c (dot) org)
1355    Date: 2003/08/28 17:00:00 CDT
1356    Tag: blume-20030828-intinf
1357    Description:
1358    
1359    This is a major update which comes with a version number bump
1360    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1361    and an implementation of IntInf in the Basis.
1362    
1363    There are a fairly large number of related changes and updates throughout
1364    the system:
1365    
1366    Basis:
1367            - Implemented IntInf.
1368            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1369            - Added some missing Real64 operations, most notably Real.toLargeInt.
1370            - Added FixedInt as a synonym for Int32.
1371    
1372    compiler:
1373            * Added support for a built-in intinf type.
1374              - literals
1375              - pattern matching
1376              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1377              - overloading on literals and operations
1378    
1379            This required adding a primitive type intinf, some additional
1380            primops, and implementations for several non-trivial intinf
1381            operations in Core.  (The intinf type is completely abstract
1382            to the compiler; all operations get delegated back to the Core.)
1383    
1384            * Intinf equality is handled by polyequal.  However, the compiler
1385            does not print its usual warning in this case (since polyequal
1386            is the right thing to do there).
1387    
1388            * Improved the organization of structure InlineT.
1389    
1390            * A word about conversion primops:
1391            If conversions involving intinf do not cancel out during
1392            CPS contract, then the compiler must insert calls to Core functions.
1393            Since all core access must be resolved already during the FLINT
1394            translate phase, it would be too late a the time of CPS contract
1395            to add new Core calls.  For this reason, conversion primops
1396            for intinf carry two arguments:  1. the numeric argument that
1397            they are supposed to convert, and 2. the Core function that
1398            can help with this conversion if necessary.  If CPS contract
1399            eliminates a primop, then the associated Core function becomes
1400            dead and goes away.  Intinf conversion primops that do not get
1401            eliminated by CPS contract get rewritten into calls of their
1402            core functions by a separate, new phase.
1403    
1404    interactive system:
1405            - Control.Print.intinfDepth controls max length of intinf constants
1406            being printed. (Analogous to Control.Print.stringDepth.)
1407            - Cleanup in printutil and pputil: got rid of unused stuff and
1408            duplicates; replaced some of the code with code that makes better
1409            use of library functionality.
1410    
1411    CM:
1412            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1413            are no longer silent.
1414    
1415    CKIT:
1416            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1417            in favor of LargeInt -- which is now the same as IntInf.
1418            CKIT-knowledgable people should check whether this is what's
1419            intended and otherwise change things back to using Int32 or
1420            FixedInt.
1421    
1422    Throughout the code:
1423            Started using IntInf.int literals and built-in operations
1424            (e.g., comparison with 0) where this seems appropriate.
1425    
1426    
1427    ----------------------------------------------------------------------
1428    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1429    Date: 2003/08/13 11:36:00 CDT
1430    Tag: dbm-20030813-mcz-merge1
1431    Description:
1432    
1433    Merging changes from the mcz-branch development branch into trunk.
1434    These changes involve replacement of the emulated old prettyprinter
1435    interface with direct use of the SML/NJ Lib PP library, and fixing
1436    of a couple of bugs (895, 1186) relating to error messages.  A new
1437    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1438    has been added.
1439    
1440    ----------------------------------------------------------------------
1441    Name: Matthias Blume (blume (at) tti - c (dot) org)
1442    Date: 2003/08/11 15:45:00 CDT
1443    Tag: blume-20030811-windows
1444    Description:
1445    
1446    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1447    
1448        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1449    
1450    This patch restores SML/NJ's ability to run under win32.  There are a
1451    number of changes, including fixes for several bugs that had gone
1452    unnoticed until now:
1453    
1454      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1455        trying to run the win32 version from within cygwin.)
1456    
1457      - There are a number of simple .bat scripts that substitute for their
1458        corresponding Unix shell-scripts.  (See below.)
1459    
1460      - The internals of ml-build have been modified slightly.  The main
1461        difference is that instead of calling ".link-sml" (or link-sml.bat)
1462        using OS.Process.system, the ML process delegates this task back
1463        to the script.  Otherwise problems arise in mixed environments such
1464        as Cygwin where scripts look and work like Unix scripts, but
1465        where OS.Process.system cannot run them.
1466    
1467      - In CM, the srcpath pickler used native pathname syntax -- which
1468        is incorrect in the case of cross-compilation.  The new pickle format
1469        is independent of platform-specific naming conventions.
1470    
1471      - Path configuration files (such as lib/pathconfig) can now choose
1472        between native and standard syntax.  Placing a line of the form
1473    
1474            standard!
1475    
1476        into the file causes all subsequent paths to be interpreted using
1477        CM standard pathname syntax (= Unix conventions); a line
1478    
1479            native!
1480    
1481        switches back to native style.  This was needed so that
1482        path config files can be written portably, see src/system/pathconfig.
1483    
1484      - Runtime system:
1485    
1486          - win32-filesys.c:  get_file_time and set_file_time now
1487            access modification time, not creation time.
1488    
1489          - I/O code made aware of new array representation.
1490    
1491          - Bug fixes in X86.prim.masm.
1492    
1493       - src/system/makeml made aware of win32. (For use under cygwin
1494            and other Unix-environments for windows.)
1495    
1496       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1497         which caused BinIO.inputAll to fail consistently.
1498    
1499    .bat scripts:
1500    
1501       Windows .bat scripts assume that SMLNJ_HOME is defined.
1502    
1503       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1504         applications (sml, ml-yacc, ml-lex).
1505       - ml-build.bat: analogous to ml-build.
1506       - config\install.bat:  Analogous to config/install.sh.  This requires
1507         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1508         (nmake etc. must be on the path, and vcvars32 must have been run.)
1509         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1510         and the bootfile hierarchy must have been unpacked under
1511         sml.boot.x86-win32.
1512         The script is very primitive and does a poor job at error checking.
1513         It only installs the base system, ml-lex, and ml-yacc.  No other
1514         libraries are being installed (i.e., you get only those that
1515         are part of the compiler.)
1516       - link-sml.bat:  analogous to .link-sml, but not currently used
1517    
1518    Unrelated bug fixes:
1519    
1520      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1521        types.
1522      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1523        otherwise.)
1524    
1525    ----------------------------------------------------------------------
1526    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1527    Date: 2003/06/17
1528    Tag: macqueen-20030617-bug895
1529    Description:
1530    
1531    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1532    Tag will be used for new development branch (mcz-branch) for
1533    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1534    
1535    ----------------------------------------------------------------------
1536    Name: Matthias Blume (blume (at) tti - c (dot) org)
1537    Date: 2003/05/27 16:55:00 CDT
1538    Tag: blume-20030527-polyeq
1539    Description:
1540    
1541    Tried to eliminated most cases of polymorphic equality.
1542    
1543    ----------------------------------------------------------------------
1544    Name: Matthias Blume (blume (at) tti - c (dot) org)
1545    Date: 2003/05/21 17:45:00 CDT
1546    Tag: blume-20030517-complete
1547    Description:
1548    
1549    Two changes:
1550    
1551      1. Added a flag for controlling whether non-exhaustive bindings will
1552         be treated as errors (default is false).
1553      2. Cleaned up the *entire* source tree so that CMB.make goes through
1554         without a single non-exhaustive match- or bind warning.
1555    
1556    ----------------------------------------------------------------------
1557    Name: Matthias Blume (blume (at) tti - c (dot) org)
1558    Date: 2003/05/17 10:20:00 CDT
1559    Tag: blume-20030517-absyn
1560    Description:
1561    
1562    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1563    
1564       This mainly affects the quality of error messages.  However, some
1565       of the code is now more straightforward than before. (Treatment of
1566       the above four constructs in translate.sml is much simpler than
1567       the "macro-expansion" that was going on before.  Plus, the mach-
1568       compiler no longer gets invoked just to be able to compile an
1569       if-expression.)
1570    
1571    2. The ErrorMsg.Error exception is now caught and absorbed by the
1572       interactive loop.
1573    
1574    ----------------------------------------------------------------------
1575    Name: Allen Leung
1576    Date: 2003/05/16 13:05:00 CDT
1577    Tag: leunga-20030516-cygwin-runtime
1578    Description:
1579    
1580        Ported the runtime system to cygwin, which uses the unix
1581    x86-unix bin files.  Missing/buggy features:
1582    
1583       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1584         the Cygwin library.
1585       o Ctrl-C handling may be flaky.
1586       o Windows system calls and Windows I/O are not supported.
1587    
1588       A new set of binfiles is located at:
1589    
1590         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1591    
1592       This is only needed for bootstrapping the cygwin version of smlnj.
1593       Other x86 versions can use the existing binfiles.
1594    
1595    ----------------------------------------------------------------------
1596    Name: Matthias Blume
1597    Date: 2003/04/08 15:42:00 CDT
1598    Tag: blume-20030408-listpair
1599    Description:
1600    
1601    1. Added a target 'mlrisc' to installer.
1602    
1603    2. Added missing elements to structure ListPair.
1604    
1605    ----------------------------------------------------------------------
1606    Name: Allen Leung
1607    Date: 2003/01/07 10:40:00 EST
1608    Tag: leunga-20030107-int-rem
1609    Description:
1610    
1611    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1612    The arguments to the SUBL instruction were swapped.
1613    
1614    ----------------------------------------------------------------------
1615    Name: Matthias Blume
1616    Date: 2002/12/12 16:25:00 EST
1617    Tag: blume-20021212-risc-ra
1618    Description:
1619    
1620    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1621    sent the RA into an infinite loop when floating point registers get
1622    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1623    110.37 and 110.38.)
1624    There was another set of potential problems related to the handling of
1625    MLRISC annotations (but those did not yet cause real problems, apparently).
1626    
1627    ----------------------------------------------------------------------
1628    Name: Matthias Blume
1629    Date: 2002/12/06 22:40:00 EST
1630    Tag: blume-20021206-cm-fileid
1631    Description:
1632    
1633    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1634    This fixes the problem of CM getting confused by files that suddenly
1635    change their identity (e.g., by getting unlinked and recreated by some
1636    text editor such as vi).  There might be a better/cheaper/cleaner way
1637    of doing this, but for now this will have to do.
1638    
1639    ----------------------------------------------------------------------
1640    Name: Matthias Blume
1641    Date: 2002/10/28 09:50:00 EST
1642    Tag: blume-20021028-typecheck
1643    Description:
1644    
1645    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1646    
1647    ----------------------------------------------------------------------
1648    Name: Matthias Blume
1649    Date: 2002/10/17 09:10:00 EDT
1650    Tag: Release_110_42
1651    Description:
1652    
1653    In good old tradition, there has been a slight hiccup so that we have
1654    to patch 110.42 after the fact.   The old release tag has been replaced
1655    (see below).
1656    
1657    The change solves a problem with two competing approaches the
1658    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1659    each other's way.
1660    
1661    This change only affects the runtime system code and the installer script.
1662    (No new bootfiles.)
1663    
1664    ----------------------------------------------------------------------
1665    Name: Matthias Blume
1666    Date: 2002/10/16 12:00:00 EDT
1667    Tag: Release_110_42_removed
1668    Description:
1669    
1670    New working release.  New bootfiles.
1671    
1672    ----------------------------------------------------------------------
1673    Name: Matthias Blume
1674    Date: 2002/10/10 13:10:00 EDT
1675    Tag: blume-20021010-ppc-divs
1676    Description:
1677    
1678    The mltree operator DIVS must be implemented with an overflow check on
1679    the PPC because the hardware indicates divide-by-zero using "overflow" as
1680    well.
1681    
1682    ----------------------------------------------------------------------
1683    Name: Matthias Blume
1684    Date: 2002/07/23 11:45:00
1685    Tag: blume-20020723-smlnj-home
1686    Description:
1687    
1688    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1689    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1690    CM_PATHCONFIG is also set), the path configuration file is assumed
1691    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1692    the entire tree to some other place and everything will "just work".
1693    
1694    (Companion commands such as ml-build and ml-makedepend also sense this
1695    variable.)
1696    
1697    ----------------------------------------------------------------------
1698    Name: Matthias Blume
1699    Date: 2002/07/12 21:19:00 EDT
1700    Tag: blume-20020712-liveness
1701    Description:
1702    
1703    Exported two useful "step" functions from liveness module (MLRISC).
1704    
1705    ----------------------------------------------------------------------
1706    Name: Matthias Blume
1707    Date: 2002/07/05 16:00 EDT
1708    Tag: Release_110_41
1709    Description:
1710    
1711    New working release.  New bootfiles.
1712    
1713    ----------------------------------------------------------------------
1714    Name: Matthias Blume
1715    Date: 2002/07/05 10:25:00 EDT
1716    Tag: blume-20020705-btimp
1717    Description:
1718    
1719    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1720    other clients can set up backtracing support.
1721    
1722    ----------------------------------------------------------------------
1723    Name: Matthias Blume
1724    Date: 2002/06/25 17:23:00 EDT
1725    Tag: blume-20020625-fpmax
1726    Description:
1727    
1728    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1729    case.  (The sense of the isNaN test was reversed -- which made min and
1730    max always return their first argument.)
1731    
1732    ----------------------------------------------------------------------
1733    Name: Matthias Blume
1734    Date: 2002/06/11
1735    Tag: blume-20020611-unixpath
1736    Description:
1737    
1738    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1739    
1740    ----------------------------------------------------------------------
1741    Name: Matthias Blume
1742    Date: 2002/06/10 16:35:00 EDT
1743    Tag: blume-20020610-ieeereal
1744    Description:
1745    
1746    I back-ported my implementation of IEEEReal.fromString from the
1747    idlbasis-devel branch so that we can test it.
1748    
1749    Another small change is that ppDec tries to give more information
1750    than just "<sig>" in the case of functors.  However, this code is
1751    broken in some mysterious way if the functor's body's signature
1752    has not been declared by ascription but gets inferred from the
1753    implementation.  This needs fixing...
1754    
1755    ----------------------------------------------------------------------
1756    Name: Matthias Blume
1757    Date: 2002/05/31
1758    Tag: blume-20020531-btrace-mode
1759    Description:
1760    
1761    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1762    the wayside when I switched over to using Controls everywhere.)
1763    
1764    ----------------------------------------------------------------------
1765    Name: Lal George
1766    Date: 2002/05/23 12:21:40 EDT
1767    Tag: george-20020523-visual-labels
1768    Description:
1769    
1770    Labels are now displayed in the graphical output to make
1771    the fall-through and target blocks obvious.
1772    
1773    ----------------------------------------------------------------------
1774    Name: Matthias Blume
1775    Date: 2002/05/22 11:03:00 EDT
1776    Tag: blume-20020522-shrink
1777    Description:
1778    
1779    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1780    situation that comes up when allocating huge arrays.
1781    
1782    ----------------------------------------------------------------------
1783    Name: Matthias Blume
1784    Date: 2002/05/21 16:00:00 EDT
1785    Tag: Release_110_40
1786    Description:
1787    
1788    New working release (110.40).  New bootfiles.
1789    
1790    [Also: John Reppy fixed GC bug 1131.]
1791    
1792    ----------------------------------------------------------------------
1793    Name: Matthias Blume
1794    Date: 2002/05/21 12:35:00 EDT
1795    Tag: blume-20020521-cmdoc
1796    Description:
1797    
1798    CM documentation update.
1799    
1800    ----------------------------------------------------------------------
1801    Name: Matthias Blume
1802    Date: 2002/05/21 10:55:00 EDT
1803    Tag: blume-20020521-misc
1804    Description:
1805    
1806    - John tweaked runtime to be silent on heap export (except when
1807      GC messages are on).
1808    - I added a few more things (cross-compiling versions of CMB) to
1809      config/preloads (as suggestions).
1810    
1811    ----------------------------------------------------------------------
1812    Name: Matthias Blume
1813    Date: 2002/05/20 22:25:00 EDT
1814    Tag: blume-20020520-controls
1815    Description:
1816    
1817    - Added ControlUtil structure to control-lib.cm.
1818    - Use it throughout.
1819    - Used Controls facility to define MLRISC controls (as opposed to
1820      registering MLRISC control ref cells with Controls after the fact)
1821    - Fixed messed-up controls priorities.
1822    
1823    * Removed again all the stuff from config/preloads that one wouldn't
1824      be able to preload at the time the initial heap image is built.
1825      (Many libraries, e.g., CML, do not exist yet at this time.  The
1826       only libraries that can be preloaded via config/preloads are those
1827       that come bundled with the bootfiles.)
1828    
1829    ----------------------------------------------------------------------
1830    Name: Matthias Blume
1831    Date: 2002/05/20 10:59:00 EDT
1832    Tag: blume-20020520-preloads
1833    Description:
1834    
1835    Added a lot of commented-out suggestions for things to be included
1836    in config/preloads.
1837    
1838    ----------------------------------------------------------------------
1839    Name: Allen Leung
1840    Date: 2002/05/18 14:20:00 EDT
1841    Tag: leunga-20020518-mdl
1842    Description:
1843    
1844    o Made the mdl tool stuff compile and run again.
1845    o I've disabled all the stuff that depends on RTL specifications; they
1846      are all badly broken anyway.
1847    
1848    ----------------------------------------------------------------------
1849    Name: Matthias Blume
1850    Date: 2002/05/17 16:49:00 EDT
1851    Tag: blume-20020517-controls
1852    Description:
1853    
1854    0. John Reppy made several modifications to the SML/NJ library.
1855       In particular, there is a shiny new controls-lib.cm.
1856    
1857    1. Pushed new controls interface through compiler so that everything
1858       compiles again.
1859    
1860    2. Added FormatComb and FORMAT_COMB to the CML version of the
1861       SML/NJ library (so that CML compiles again).
1862    
1863    3. Modified init scripts because XXX_DEFAULT environment variables
1864       are no longer with us.  (Boot-time initialization is now done
1865       using the same environment variables that are also used for
1866       startup-time initialization of controls.)
1867    
1868    ----------------------------------------------------------------------
1869    Name: Lal George
1870    Date: 2002/05/15 09:20:10 EDT
1871    Tag: george-20020515-pseudo-op-decls
1872    Description:
1873    
1874            All pseudo-ops emitted before the first segment declaration
1875    such as TEXT, DATA, and BSS directives are assumed to be global
1876    declarations and are emitted first in the assembly file. This is
1877    useful in a number of situations where one has pseudo-ops that are not
1878    specific to any segment, and also works around the constraint that one
1879    cannot have client pseudo-ops in the TEXT segment.
1880    
1881            Because no segment is associated with these declarations it is
1882    an error to allocate any space or objects before the first segment
1883    directive and an exception will be raised. However, we cannot make
1884    this check for client pseudo-ops.
1885    
1886            These top level declarations are a field in the CFG graph_info.
1887    In theory you can continue to add to this field after the CFG has been
1888    built -- provided you know what you are doing;-)
1889    
1890    ----------------------------------------------------------------------
1891    Name: Matthias Blume
1892    Date: 2002/05/13 16:40:00 EDT
1893    Tag: blume-20020513-pp-etc
1894    Description:
1895    
1896    A few minor bugfixes:
1897    
1898      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1899        (Bogus printouts for redefined bindings still occur.  Compiler
1900        bug should no longer occur now.  We need to redo the prettyprinter
1901        from scratch.)
1902    
1903      - CM pathname printer now also adds escape sequences for ( and )
1904    
1905      - commend and docu fixes for ml-nlffi
1906    
1907    ----------------------------------------------------------------------
1908    Name: Matthias Blume
1909    Date: 2002/05/10 16:40:00 EDT
1910    Tag: blume-20020510-erg-textio
1911    Description:
1912    
1913    Applied the following bugfix provided by Emden Gansner:
1914    
1915        Output is corrupted when outputSubstr is used rather than output.
1916    
1917        The problem occurs when a substring
1918    
1919            ss = (s, dataStart, dataLen)
1920    
1921        where dataStart > 0, fills a stream buffer with avail bytes left.
1922        avail bytes of s, starting at index dataStart, are copied into the
1923        buffer, the buffer is flushed, and then the remaining dataLen-avail
1924        bytes of ss are copied into the beginning of the buffer. Instead of
1925        starting this copy at index dataStart+avail in s, the current code
1926        starts the copy at index avail.
1927    
1928        Fix:
1929        In text-io-fn.sml, change line 695 from
1930             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1931        to
1932             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1933    
1934    ----------------------------------------------------------------------
1935    Name: Matthias Blume
1936    Date: 2002/04/12 13:55:00 EDT
1937    Tag: blume-20020412-assyntax
1938    Description:
1939    
1940    1. Grabbed newer assyntax.h from the XFree86 project.
1941    2. Fiddled with how to compile X86.prim.asm without warnings.
1942    3. (Very) Minor cleanup in CM.
1943    
1944    ----------------------------------------------------------------------
1945    Name: Matthias Blume
1946    Date: 2002/04/01 (no joke!) 17:07:00 EST
1947    Tag: blume-20020401-x86div
1948    Description:
1949    
1950    Added full support for div/mod/rem/quot on the x86, using the machine
1951    instruction's two results (without clumsily recomputing the remainder)
1952    directly where appropriate.
1953    
1954    Some more extensive power-of-two support was added to the x86 instruction
1955    selector (avoiding expensive divs, mods, and muls where they can be
1956    replaced with cheaper shifts and masks).  However, this sort of thing
1957    ought to be done earlier, e.g., within the CPS optimizer so that
1958    all architectures benefit from it.
1959    
1960    The compiler compiles to a fixed point, but changes might be somewhat
1961    fragile nevertheless.  Please, report any strange things that you might
1962    see wrt. div/mod/quot/rem...
1963    
1964    ----------------------------------------------------------------------
1965    Name: Matthias Blume
1966    Date: 2002/03/29 17:22:00
1967    Tag: blume-20020329-div
1968    Description:
1969    
1970    Fixed my broken div/mod logic.  Unfortunately, this means that the
1971    inline code for div/mod now has one more comparison than before.
1972    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
1973    The problem was with quotient = 0, because that alone does not tell
1974    us which way the rounding went.  One then has to look at whether
1975    remainder and divisor have the same sign...  :(
1976    
1977    Anyway, I replaced the bootfiles with fresh ones...
1978    
1979    ----------------------------------------------------------------------
1980    Name: Matthias Blume
1981    Date: 2002/03/29 14:10:00 EST
1982    Tag: blume-20020329-inlprims
1983    Description:
1984    
1985    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
1986    
1987    Primops have changed. This means that the bin/boot-file formats have
1988    changed as well.
1989    
1990    To make sure that there is no confusion, I made a new version.
1991    
1992    
1993    CHANGES:
1994    
1995    * removed REMT from mltree (remainder should never overflow).
1996    
1997    * added primops to deal with divisions of all flavors to the frontend
1998    
1999    * handled these primops all the way through so they map to their respective
2000      MLRISC support
2001    
2002    * used these primops in the implementation of Int, Int32, Word, Word32
2003    
2004    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2005    
2006    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2007    
2008    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2009      floating point
2010    
2011    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2012      and Real (but Real.abs maps to a separate floating-point-only primop)
2013    
2014    
2015    TODO items:
2016    
2017    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2018      instructions because the machine instruction encoder cannot handle
2019      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2020      REM in the first place, and REM is supported by the x86 machine coder.)
2021    
2022    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2023      selection phase.  (The two can be streamlined because the hardware
2024      delivers both quotient and remainder at the same time anyway.)
2025    
2026    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2027      (Currently the behavior is inconsistent both across architectures and
2028      wrt. the draft Basis spec.)
2029    
2030    * Word8 should eventually be handled natively, too.
2031    
2032    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2033      as if there currently is no execution path that could trigger it in
2034      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2035      hold for things like multiplication and division.)
2036    
2037    ----------------------------------------------------------------------
2038    Name: Matthias Blume
2039    Date: 2002/03/27 16:27:00 EST
2040    Tag: blume-20020327-mlrisc-divisions
2041    Description:
2042    
2043    Added support for all four division operations (ML's div, mod, quot,
2044    and rem) to MLRISC.  In the course of doing so, I also rationalized
2045    the naming (no more annoying switch-around of DIV and QUOT), by
2046    parameterizing the operation by div_rounding_mode (which can be either
2047    DIV_TO_ZERO or DIV_TO_NEGINF).
2048    
2049    The generic MLTreeGen functor takes care of compiling all four
2050    operations down to only round-to-zero div.
2051    
2052    Missing pieces:
2053    
2054      * Doing something smarter than relying on MLTreeGen on architectures
2055        like, e.g., the x86 where hardware division delivers both quotient and
2056        remainder at the same time.  With this, the implementation of the
2057        round-to-neginf operations could be further streamlined.
2058    
2059      * Remove inlining support for div/mod/rem from the frontend and replace it
2060        with primops that get carried through to the backend.  Do this for all
2061        int and word types.
2062    
2063    ----------------------------------------------------------------------
2064    Name: Matthias Blume
2065    Date: 2002/03/25 17:25:00 EST
2066    Tag: blume-20020325-divmod
2067    Description:
2068    
2069    I improved (hopefully without breaking them) the implementation of Int.div,
2070    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2071    advantage of the following observations:
2072    
2073      Let  q = x quot y      r = x rem y
2074           d = x div  y      m = x mod y
2075    
2076    where "quot" is the round-to-zero version of integer division that
2077    hardware usually provides.  Then we have:
2078    
2079         r = x - q * y        where neither the * nor the - will overflow
2080         d = if q >= 0 orelse x = q * y then q else q - 1
2081                              where neither the * nor the - will overflow
2082         m = if q >= 0 orelse r = 0 then r else r + y
2083                              where the + will not overflow
2084    
2085    This results in substantial simplification of the generated code.
2086    The following table shows the number of CFG nodes and edges generated
2087    for
2088            fun f (x, y) = x OPER y
2089            (* with OPER \in div, mod, quot, rem *)
2090    
2091    
2092        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2093        --------------------------------------------------------
2094         div |         24 |         39 |         12 |         16
2095         mod |         41 |         71 |         12 |         16
2096        quot |          8 |         10 |          8 |         10
2097         rem |         10 |         14 |          8 |         10
2098    
2099    
2100    ----------------------------------------------------------------------
2101    Name: Matthias Blume
2102    Date: 2002/03/25 22:06:00 EST
2103    Tag: blume-20020325-cprotobug
2104    Description:
2105    
2106    Fixed a bug in cproto (c prototype decoder).
2107    
2108    ----------------------------------------------------------------------
2109    Name: Matthias Blume
2110    Date: 2002/03/25 16:00:00 EST
2111    Tag: blume-20020325-raw-primops
2112    Description:
2113    
2114    I did some cleanup to Allen's new primop code and
2115    replaced yesterday's bootfiles with new ones.
2116    (But they are stored in the same place.)
2117    
2118    ----------------------------------------------------------------------
2119    Name: Matthias Blume
2120    Date: 2002/03/24 22:40:00 EST
2121    Tag: blume-20020324-bootfiles
2122    Description:
2123    
2124    Made the bootfiles that Allen asked for.
2125    
2126    ----------------------------------------------------------------------
2127    Name: Allen Leung
2128    Date: 2002/03/23 15:50:00 EST
2129    Tag: leunga-20020323-flint-cps-rcc-primops
2130    Description:
2131    
2132      1. Changes to FLINT primops:
2133    
2134        (* make a call to a C-function;
2135         * The primop carries C function prototype information and specifies
2136         * which of its (ML-) arguments are floating point. C prototype
2137         * information is for use by the backend, ML information is for
2138         * use by the CPS converter. *)
2139      | RAW_CCALL of { c_proto: CTypes.c_proto,
2140                       ml_args: ccall_type list,
2141                       ml_res_opt: ccall_type option,
2142                       reentrant : bool
2143                     } option
2144       (* Allocate uninitialized storage on the heap.
2145        * The record is meant to hold short-lived C objects, i.e., they
2146        * are not ML pointers.  With the tag, the representation is
2147        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2148        *)
2149      | RAW_RECORD of {tag:bool,sz:int}
2150      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2151    
2152      2.  These CPS primops are now overloaded:
2153    
2154           rawload of {kind:numkind}
2155           rawstore of {kind:numkind}
2156    
2157          The one argument form is:
2158    
2159             rawload {kind} address
2160    
2161          The two argument form is:
2162    
2163             rawload {kind} [ml object, byte-offset]
2164    
2165      3. RAW_CCALL/RCC now takes two extra arguments:
2166    
2167         a. The first is whether the C call is reentrant, i.e., whether
2168            ML state should be saved and restored.
2169         b. The second argument is a string argument specifying the name of
2170            library and the C function.
2171    
2172         These things are currently not handled in the code generator, yet.
2173    
2174      4. In CProto,
2175    
2176         An encoding type of "bool" means "ml object" and is mapped into
2177         C prototype of PTR.  Note that "bool" is different than "string",
2178         even though "string" is also mapped into PTR, because "bool"
2179         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2180    
2181      5. Pickler/unpicker
2182    
2183         Changed to handle RAW_RECORD and newest RAW_CCALL
2184    
2185      6. MLRiscGen,
2186    
2187         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2188         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2189            in the file CodeGen/cpscompile/cps-c-calls.sml
2190    
2191      7. Added the conditional move operator
2192    
2193             condmove of branch
2194    
2195         to cps.  Generation of this is still buggy so it is currently
2196         disabled.
2197    
2198    ----------------------------------------------------------------------
2199    Name: Lal George
2200    Date: 2002/03/22 14:18:25 EST
2201    Tag: george-20020322-cps-branch-prob
2202    Description:
2203    
2204    Implemented the Ball-Larus branch prediction-heuristics, and
2205    incorporated graphical viewers for control flow graphs.
2206    
2207    Ball-Larus Heuristics:
2208    ---------------------
2209    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2210    
2211    By design it uses the Dempster-Shafer theory for combining
2212    probabilities.  For example, in the function:
2213    
2214        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2215    
2216    the ball-larus heuristics predicts that the n=0 is unlikely
2217    (OH-heuristic), and the 'then' branch is unlikely because of the
2218    RH-heuristic -- giving the 'then' branch an even lower combined
2219    probability using the Dempster-Shafer theory.
2220    
2221    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2222    probability of the 'then' branch because of the loop in the else
2223    branch.
2224    
2225    
2226    Graphical Viewing:
2227    ------------------
2228    I merely plugged in Allen's graphical viewers into the compiler. The
2229    additional code is not much. At the top level, saying:
2230    
2231            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2232    
2233    will display the graphical view of the control flow graph just before
2234    back-patching.  daVinci must be in your path for this to work. If
2235    daVinci is not available, then the default viewer can be changed
2236    using:
2237    
2238            Control.MLRISC.getString "viewer"
2239    
2240    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2241    course, these viewers must be in your path.
2242    
2243    The above will display the compilation unit at the level of clusters,
2244    many of which are small, boring, and un-interesting. Also setting:
2245    
2246            Control.MLRISC.getInt "cfg-graphical-view_size"
2247    
2248    will display clusters that are larger than the value set by the above.
2249    
2250    
2251    ----------------------------------------------------------------------
2252    Name: Matthias Blume
2253    Date: 2002/03/21 22:20:00 EST
2254    Tag: blume-20020321-kmp-bugfix
2255    Description:
2256    
2257    Changed the interface to the KMP routine in PreString and fixed
2258    a minor bug in one place where it was used.
2259    
2260    ----------------------------------------------------------------------
2261    Name: Allen Leung
2262    Date: 2002/03/21 20:30:00 EST
2263    Tag: leunga-20020321-cfg
2264    Description:
2265    
2266      Fixed a potential problem in cfg edge splitting.
2267    
2268    ----------------------------------------------------------------------
2269    Name: Allen Leung
2270    Date: 2002/03/21 17:15:00 EST
2271    Tag: leunga-20020321-x86-fp-cfg
2272    Description:
2273    
2274      1. Recoded the buggy parts of x86-fp.
2275    
2276         a. All the block reordering code has been removed.
2277            We now depend on the block placement phases to do this work.
2278    
2279         b. Critical edge splitting code has been simplified and moved into the
2280            CFG modules, as where they belong.
2281    
2282         Both of these were quite buggy and complex.  The code is now much, much
2283         simpler.
2284    
2285      2. X86 backend.
2286    
2287         a. Added instructions for 64-bit support.  Instruction selection for
2288            64-bit has not been committed, however, since that
2289            requires changes to MLTREE which haven't been approved by
2290            Lal and John.
2291    
2292         b. Added support for FUCOMI and FUCOMIP when generating code for
2293            PentiumPro and above.  We only generate these instructions in
2294            the fast-fp mode.
2295    
2296         c. Added cases for JP and JNP in X86FreqProps.
2297    
2298      3. CFG
2299    
2300         CFG now has a bunch of methods for edge splitting and merging.
2301    
2302      4. Machine description.
2303    
2304         John's simplification of MLTREE_BASIS.fcond broke a few machine
2305         description things:
2306    
2307         rtl-build.{sig,sml} and hppa.mdl fixed.
2308    
2309         NOTE: the machine description stuff in the repository is still broken.
2310               Again, I can't put my fixes in because that involves
2311               changes to MLTREE.
2312    
2313    ----------------------------------------------------------------------
2314    Name: Matthias Blume
2315    Date: 2002/03/20 15:55:00 EST
2316    Tag: blume-20020320-kmp
2317    Description:
2318    
2319    Implemented Knuth-Morris-Pratt string matching in PreString and used
2320    it for String.isSubstring, Substring.isSubstring, and
2321    Substring.position.
2322    
2323    (Might need some stress-testing.  Simple examples worked fine.)
2324    
2325    ----------------------------------------------------------------------
2326    Name: Matthias Blume
2327    Date: 2002/03/19 16:37:00 EST
2328    Tag: blume-20020319-witnesses
2329    Description:
2330    
2331    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2332    
2333    This implements a generic mechanism for changing constness qualifiers
2334    anywhere within big C types without resorting to outright "casts".
2335    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2336    modify the constness at the outermost level.)
2337    The implementation of "convert" is based on the idea of "witness"
2338    values -- values that are not used by the operation but whose types
2339    "testify" to their applicability.  On the implementation side, "convert"
2340    is simply a projection (returning its second curried argument).  With
2341    cross-module inlining, it should not result in any machine code being
2342    generated.
2343    
2344    ----------------------------------------------------------------------
2345    Name: Matthias Blume
2346    Date: 2002/03/15 16:40:00 EST
2347    Tag: blume-20020315-basis
2348    Description:
2349    
2350    Provided (preliminary?) implementations for
2351    
2352      {String,Substring}.{concatWith,isSuffix,isSubstring}
2353    
2354    and
2355    
2356      Substring.full
2357    
2358    Those are in the Basis spec but they were missing in SML/NJ.
2359    
2360    ----------------------------------------------------------------------
2361    Name: Matthias Blume
2362    Date: 2002/03/14 21:30:00 EST
2363    Tag: blume-20020314-controls
2364    Description:
2365    
2366    Controls:
2367    ---------
2368    
2369    1. Factored out the recently-added Controls : CONTROLS stuff and put
2370       it into its own library $/controls-lib.cm.  The source tree for
2371       this is under src/smlnj-lib/Controls.
2372    
2373    2. Changed the names of types and functions in this interface, so they
2374       make a bit more "sense":
2375    
2376          module -> registry
2377          'a registry -> 'a group
2378    
2379    3. The interface now deals in ref cells only.  The getter/setter interface
2380       is (mostly) gone.
2381    
2382    4. Added a function that lets one register an already-existing ref cell.
2383    
2384    5. Made the corresponding modifications to the rest of the code so that
2385       everything compiles again.
2386    
2387    6. Changed the implementation of Controls.MLRISC back to something closer
2388       to the original.  In particular, this module (and therefore MLRISC)
2389       does not depend on Controls.  There now is some link-time code in
2390       int-sys.sml that registers the MLRISC controls with the Controls
2391       module.
2392    
2393    CM:
2394    ---
2395    
2396      * One can now specify the lambda-split aggressiveness in init.cmi.
2397    
2398    ----------------------------------------------------------------------
2399    Name: Allen Leung
2400    Date: 2002/03/13 17:30:00 EST
2401    Tag: leunga-20020313-x86-fp-unary
2402    Description:
2403    
2404    Bug fix for:
2405    
2406    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2407    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2408    > - fun f(x,(y,z)) = Real.~ y;
2409    > [autoloading]
2410    > [autoloading done]
2411    >       fchsl   (%eax), 184(%esp)
2412    > Error: MLRisc bug: X86MCEmitter.emitInstr
2413    >
2414    > uncaught exception Error
2415    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2416    
2417    The problem was that the code generator did not generate any fp registers
2418    in this case, and the ra didn't know that it needed to run the X86FP phase to
2419    translate the pseudo fp instruction.   This only happened with unary fp
2420    operators in certain situations.
2421    
2422    ----------------------------------------------------------------------
2423    Name: Matthias Blume
2424    Date: 2002/03/13 14:00:00 EST
2425    Tag: blume-20020313-overload-etc
2426    Description:
2427    
2428    1. Added _overload as a synonym for overload for backward compatibility.
2429       (Control.overloadKW must be true for either version to be accepted.)
2430    
2431    2. Fixed bug in install script that caused more things to be installed
2432       than what was requested in config/targets.
2433    
2434    3. Made CM aware of the (_)overload construct so that autoloading
2435       works.
2436    
2437    ----------------------------------------------------------------------
2438    Name: Matthias Blume
2439    Date: 2002/03/12 22:03:00 EST
2440    Tag: blume-20020312-url
2441    Description:
2442    
2443    Forgot to update BOOT and srcarchiveurl.
2444    
2445    ----------------------------------------------------------------------
2446    Name: Matthias Blume
2447    Date: 2002/03/12 17:30:00 EST
2448    Tag: blume-20020312-version110392
2449    Description:
2450    
2451    Yet another version number bump (because of small changes to the
2452    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2453    
2454    Changes:
2455    
2456      The new pid generation scheme described a few weeks ago was overly
2457      complicated.  I implemented a new mechanism that is simpler and
2458      provides a bit more "stability":  Once CM has seen a compilation
2459      unit, it keeps its identity constant (as long as you do not delete
2460      those crucial CM/GUID/* files).  This means that when you change
2461      an interface, compile, then go back to the old interface, and
2462      compile again, you arrive at the original pid.
2463    
2464      There now also is a mechanism that instructs CM to use the plain
2465      environment hash as a module's pid (effectively making its GUID
2466      the empty string).  For this, "noguid" must be specified as an
2467      option to the .sml file in question within its .cm file.
2468      This is most useful for code that is being generated by tools such
2469      as ml-nlffigen (because during development programmers tend to
2470      erase the tool's entire output directory tree including CM's cached
2471      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2472      revert to the old, broken behavior of SML/NJ, but in specific cases
2473      where there is no danger of interface confusion, its use is ok
2474      (I think).
2475    
2476      ml-nlffigen by default generates "noguid" annotations.  They can be
2477      turned off by specifying -guid in its command line.
2478    
2479    ----------------------------------------------------------------------
2480    Name: Lal George
2481    Date: 2002/03/12 12 14:42:36 EST
2482    Tag: george-20020312-frequency-computation
2483    Description:
2484    
2485    Integrated jump chaining and static block frequency into the
2486    compiler. More details and numbers later.
2487    
2488    ----------------------------------------------------------------------
2489    Name: Lal George
2490    Date: 2002/03/11 11 22:38:53 EST
2491    Tag: george-20020311-jump-chain-elim
2492    Description:
2493    
2494    Tested the jump chain elimination on all architectures (except the
2495    hppa).  This is on by default right now and is profitable for the
2496    alpha and x86, however, it may not be profitable for the sparc and ppc
2497    when compiling the compiler.
2498    
2499    The gc test will typically jump to a label at the end of the cluster,
2500    where there is another jump to an external cluster containing the actual
2501    code to invoke gc. This is to allow factoring of common gc invocation
2502    sequences. That is to say, we generate:
2503    
2504            f:
2505               testgc
2506               ja   L1      % jump if above to L1
2507    
2508            L1:
2509               jmp L2
2510    
2511    
2512    After jump chain elimination the 'ja L1' instructions is converted to
2513    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2514    up being implemented in their long form (if L2 is far away) using:
2515    
2516            jbe     L3      % jump if below or equal to L3
2517            jmp     L2
2518         L3:
2519            ...
2520    
2521    
2522    For large compilation units L2  may be far away.
2523    
2524    
2525    ----------------------------------------------------------------------
2526    Name: Matthias Blume
2527    Date: 2002/03/11 13:30:00 EST
2528    Tag: blume-20020311-mltreeeval
2529    Description:
2530    
2531    A functor parameter was missing.
2532    
2533    ----------------------------------------------------------------------
2534    Name: Allen Leung
2535    Date: 2002/03/11 10:30:00 EST
2536    Tag: leunga-20020311-runtime-string0
2537    Description:
2538    
2539       The representation of the empty string now points to a
2540    legal null terminated C string instead of unit.  It is now possible
2541    to convert an ML string into C string with InlineT.CharVector.getData.
2542    This compiles into one single machine instruction.
2543    
2544    ----------------------------------------------------------------------
2545    Name: Allen Leung
2546    Date: 2002/03/10 23:55:00 EST
2547    Tag: leunga-20020310-x86-call
2548    Description:
2549    
2550       Added machine generation for CALL instruction (relative displacement mode)
2551    
2552    ----------------------------------------------------------------------
2553    Name: Matthias Blume
2554    Date: 2002/03/08 16:05:00
2555    Tag: blume-20020308-entrypoints
2556    Description:
2557    
2558    Version number bumped to 110.39.1.  NEW BOOTFILES!
2559    
2560    Entrypoints: non-zero offset into a code object where execution should begin.
2561    
2562    - Added the notion of an entrypoint to CodeObj.
2563    - Added reading/writing of entrypoint info to Binfile.
2564    - Made runtime system bootloader aware of entrypoints.
2565    - Use the address of the label of the first function given to mlriscGen
2566      as the entrypoint.  This address is currently always 0, but it will
2567      not be 0 once we turn on block placement.
2568    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2569      with entry points) from mlriscGen.
2570    
2571    ----------------------------------------------------------------------
2572    Name: Allen Leung
2573    Date: 2002/03/07 20:45:00 EST
2574    Tag: leunga-20020307-x86-cmov
2575    Description:
2576    
2577       Bug fixes for CMOVcc on x86.
2578    
2579       1. Added machine code generation for CMOVcc
2580       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2581       3. CMOVcc cannot have an immediate operand as argument.
2582    
2583    ----------------------------------------------------------------------
2584    Name: Matthias Blume
2585    Date: 2002/03/07 16:15:00 EST
2586    Tag: blume-20020307-controls
2587    Description:
2588    
2589    This is a very large but mostly boring patch which makes (almost)
2590    every tuneable compiler knob (i.e., pretty much everything under
2591    Control.* plus a few other things) configurable via both the command
2592    line and environment variables in the style CM did its configuration
2593    until now.
2594    
2595    Try starting sml with '-h' (or, if you are brave, '-H')
2596    
2597    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2598    implements the underlying generic mechanism.
2599    
2600    The interface to some of the existing such facilities has changed somewhat.
2601    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2602    (The getFoo interface is still there for backward-compatibility, but its
2603    use is deprecated.)
2604    
2605    The ml-build script passes -Cxxx=yyy command-line arguments through so
2606    that one can now twiddle the compiler settings when using this "batch"
2607    compiler.
2608    
2609    TODO items:
2610    
2611    We should go through and throw out all controls that are no longer
2612    connected to anything.  Moreover, we should go through and provide
2613    meaningful (and correct!) documentation strings for those controls
2614    that still are connected.
2615    
2616    Currently, multiple calls to Controls.new are accepted (only the first
2617    has any effect).  Eventually we should make sure that every control
2618    is being made (via Controls.new) exactly once.  Future access can then
2619    be done using Controls.acc.
2620    
2621    Finally, it would probably be a good idea to use the getter-setter
2622    interface to controls rather than ref cells.  For the time being, both
2623    styles are provided by the Controls module, but getter-setter pairs are
2624    better if thread-safety is of any concern because they can be wrapped.
2625    
2626    *****************************************
2627    
2628    One bug fix: The function blockPlacement in three of the MLRISC
2629    backpatch files used to be hard-wired to one of two possibilities at
2630    link time (according to the value of the placementFlag).  But (I
2631    think) it should rather sense the flag every time.
2632    
2633    *****************************************
2634    
2635    Other assorted changes (by other people who did not supply a HISTORY entry):
2636    
2637    1. the cross-module inliner now works much better (Monnier)
2638    2. representation of weights, frequencies, and probabilities in MLRISC
2639       changed in preparation of using those for weighted block placement
2640       (Reppy, George)
2641    
2642    ----------------------------------------------------------------------
2643    Name: Lal George
2644    Date: 2002/03/07 14:44:24 EST 2002
2645    Tag: george-20020307-weighted-block-placement
2646    
2647    Tested the weighted block placement optimization on all architectures
2648    (except the hppa) using AMPL to generate the block and edge frequencies.
2649    Changes were required in the machine properties to correctly
2650    categorize trap instructions. There is an MLRISC flag
2651    "weighted-block-placement" that can be used to enable weighted block
2652    placement, but this will be ineffective without block/edge
2653    frequencies (coming soon).
2654    
2655    
2656    ----------------------------------------------------------------------
2657    Name: Lal George
2658    Date: 2002/03/05 17:24:48 EST
2659    Tag: george-20020305-linkage-cluster
2660    
2661    In order to support the block placement optimization, a new cluster
2662    is generated as the very first cluster (called the linkage cluster).
2663    It contains a single jump to the 'real' entry point for the compilation
2664    unit. Block placement has no effect on the linkage cluster itself, but
2665    all the other clusters  have full freedom in the manner in which they
2666    reorder blocks or functions.
2667    
2668    On the x86 the typical linkage code that is generated is:
2669       ----------------------
2670            .align 2
2671       L0:
2672            addl    $L1-L0, 72(%esp)
2673            jmp     L1
2674    
2675    
2676            .align  2
2677       L1:
2678       ----------------------
2679    
2680    72(%esp) is the memory location for the stdlink register. This
2681    must contain the address of the CPS function being called. In the
2682    above example, it contains the address of  L0; before
2683    calling L1 (the real entry point for the compilation unit), it
2684    must contain the address for L1, and hence
2685    
2686            addl $L1-L0, 72(%esp)
2687    
2688    I have tested this on all architectures except the hppa.The increase
2689    in code size is of course negligible
2690    
2691    ----------------------------------------------------------------------
2692    Name: Allen Leung
2693    Date: 2002/03/03 13:20:00 EST
2694    Tag: leunga-20020303-mlrisc-tools
2695    
2696      Added #[ ... ] expressions to mlrisc tools
2697    
2698    ----------------------------------------------------------------------
2699    Name: Matthias Blume
2700    Date: 2002/02/27 12:29:00 EST
2701    Tag: blume-20020227-cdebug
2702    Description:
2703    
2704    - made types in structure C and C_Debug to be equal
2705    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2706    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2707    
2708    ----------------------------------------------------------------------
2709    Name: Matthias Blume
2710    Date: 2002/02/26 12:00:00 EST
2711    Tag: blume-20020226-ffi
2712    Description:
2713    
2714    1. Fixed a minor bug in CM's "noweb" tool:
2715       If numbering is turned off, then truly don't number (i.e., do not
2716       supply the -L option to noweb).  The previous behavior was to supply
2717       -L'' -- which caused noweb to use the "default" line numbering scheme.
2718       Thanks to Chris Richards for pointing this out (and supplying the fix).
2719    
2720    2. Once again, I reworked some aspects of the FFI:
2721    
2722       A. The incomplete/complete type business:
2723    
2724       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2725         gone!
2726       - ML types representing an incomplete type are now *equal* to
2727         ML types representing their corresponding complete types (just like
2728         in C).  This is still safe because ml-nlffigen will not generate
2729         RTTI for incomplete types, nor will it generate functions that
2730         require access to such RTTI.   But when ML code generated from both
2731         incomplete and complete versions of the C type meet, the ML types
2732         are trivially interoperable.
2733    
2734         NOTE:  These changes restore the full generality of the translation
2735         (which was previously lost when I eliminated functorization)!
2736    
2737       B. Enum types:
2738    
2739       - Structure C now has a type constructor "enum" that is similar to
2740         how the "su" constructor works.  However, "enum" is not a phantom
2741         type because each "T enum" has values (and is isomorphic to
2742         MLRep.Signed.int).
2743       - There are generic access operations for enum objects (using
2744         MLRep.Signed.int).
2745       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2746         * The structure contains the definition of type "mlrep" (the ML-side
2747         representation type of the enum).  Normally, mlrep is the same
2748         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2749         then mlrep will be defined as a datatype -- thus facilitating
2750         pattern matching on mlrep values.
2751         ("-ec" will be suppressed if there are duplicate values in an
2752          enumeration.)
2753         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2754         will be generated for each C enum constant xxx.
2755         * Conversion functions m2i and i2m convert between mlrep and
2756         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2757         * Coversion functions c and ml convert between mlrep and "tag enum".
2758         * Access functions (get/set) fetch and store mlrep values.
2759       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2760         enumerations are merged into one single enumeration represented by
2761         structure E_'.
2762    
2763    ----------------------------------------------------------------------
2764    Name: Allen Leung
2765    Date: 2002/02/25 04:45:00 EST
2766    Tag: leunga-20020225-cps-spill
2767    
2768    This is a new implementation of the CPS spill phase.
2769    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2770    In case of problems, replace it with the old file spill.sml
2771    
2772    The current compiler runs into some serious performance problems when
2773    constructing a large record.  This can happen when we try to compile a
2774    structure with many items.  Even a very simple structure like the following
2775    makes the compiler slow down.
2776    
2777        structure Foo = struct
2778           val x_1 = 0w1 : Word32.int
2779           val x_2 = 0w2 : Word32.int
2780           val x_3 = 0w3 : Word32.int
2781           ...
2782           val x_N = 0wN : Word32.int
2783        end
2784    
2785    The following table shows the compile time, from N=1000 to N=4000,
2786    with the old compiler:
2787    
2788    N
2789    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2790           MLRISC ra                               0.06u  0.00s  0.05g
2791              (spills = 0 reloads = 0)
2792           TOTAL                                   0.63u  0.07s  0.21g
2793    
2794    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2795           MLRISC ra                               5.68u  0.59s  3.93g
2796              (spills = 0 reloads = 0)
2797           TOTAL                                   14.71u  0.99s  4.81g
2798    
2799    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2800           MLRISC ra                               5.54u  0.65s  3.91g
2801              (spills = 543 reloads = 1082)
2802           TOTAL                                   65.40u  3.13s  6.00g
2803    
2804    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2805           MLRISC ra                               0.80u  0.10s  0.55g
2806              (spills = 42 reloads = 84)
2807           TOTAL                                   129.42u  5.10s  4.13g
2808    
2809    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2810           MLRISC ra                               2.69u  0.27s  1.38g
2811              (spills = 62 reloads = 124)
2812           TOTAL                                   682.48u  19.61s  13.99g
2813    
2814    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2815           MLRISC ra                               4.96u  0.27s  2.72g
2816              (spills = 85 reloads = 170)
2817           TOTAL                                   2375.26u  57.21s  48.00g
2818    
2819    As you can see the old cps spill module suffers from some serious
2820    performance problem.  But since I cannot decipher the old code fully,
2821    instead of patching the problems up, I'm reimplementing it
2822    with a different algorithm.  The new code is more modular,
2823    smaller when compiled, and substantially faster
2824    (O(n log n) time and O(n) space).  Timing of the new spill module:
2825    
2826    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2827          MLRISC ra                               0.25u  0.02s  0.15g
2828             (spills=1 reloads=3)
2829          TOTAL                                   7.74u  0.34s  1.62g
2830    
2831    Implementation details:
2832    
2833    As far as I can tell, the purpose of the CPS spill module is to make sure the
2834    number of live variables at any program point (the bandwidth)
2835    does not exceed a certain limit, which is determined by the
2836    size of the spill area.
2837    
2838    When the bandwidth is too large, we decrease the register pressure by
2839    packing live variables into spill records.  How we achieve this is
2840    completely different than what we did in the old code.
2841    
2842    First, there is something about the MLRiscGen code generator
2843    that we should be aware of:
2844    
2845    o MLRiscGen performs code motion!
2846    
2847       In particular, it will move floating point computations and
2848       address computations involving only the heap pointer to
2849       their use sites (if there is only a single use).
2850       What this means is that if we have a CPS record construction
2851       statement
2852    
2853           RECORD(k,vl,w,e)
2854    
2855       we should never count the new record address w as live if w
2856       has only one use (which is often the case).
2857    
2858       We should do something similar to floating point, but the transformation
2859       there is much more complex, so I won't deal with that.
2860    
2861    Secondly, there are now two new cps primops at our disposal:
2862    
2863     1. rawrecord of record_kind option
2864        This pure operator allocates some uninitialized storage from the heap.
2865        There are two forms:
2866    
2867         rawrecord NONE [INT n]  allocates a tagless record of length n
2868         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2869                                     and initializes the tag.
2870    
2871     2. rawupdate of cty
2872          rawupdate cty (v,i,x)
2873          Assigns to x to the ith component of record v.
2874          The storelist is not updated.
2875    
2876    We use these new primops for both spilling and increment record construction.
2877    
2878     1. Spilling.
2879    
2880        This is implemented with a linear scan algorithm (but generalized
2881        to trees).  The algorithm will create a single spill record at the
2882        beginning of the cps function and use rawupdate to spill to it,
2883        and SELECT or SELp to reload from it.  So both spills and reloads
2884        are fine-grain operations.  In contrast, in the old algorithm
2885        "spills" have to be bundled together in records.
2886    
2887        Ideally, we should sink the spill record construction to where
2888        it is needed.  We can even split the spill record into multiple ones
2889        at the places where they are needed.  But CPS is not a good
2890        representation for global code motion, so I'll keep it simple and
2891        am not attempting this.
2892    
2893     2. Incremental record construction (aka record splitting).
2894    
2895        Long records with many component values which are simulatenously live
2896        (recall that single use record addresses are not considered to
2897         be live) are constructed with rawrecord and rawupdate.
2898        We allocate space on the heap with rawrecord first, then gradually
2899        fill it in with rawupdate.  This is the technique suggested to me
2900        by Matthias.
2901    
2902        Some restrictions on when this is applicable:
2903        1. It is not a VECTOR record.  The code generator currently does not handle
2904           this case. VECTOR record uses double indirection like arrays.
2905        2. All the record component values are defined in the same "basic block"
2906           as the record constructor.  This is to prevent speculative
2907           record construction.
2908    
2909    ----------------------------------------------------------------------
2910    Name: Allen Leung
2911    Date: 2002/02/22 01:02:00 EST
2912    Tag: leunga-20020222-mlrisc-tools
2913    
2914    Minor bug fixes in the parser and rewriter
2915    
2916    ----------------------------------------------------------------------
2917    Name: Allen Leung
2918    Date: 2002/02/21 20:20:00 EST
2919    Tag: leunga-20020221-peephole
2920    
2921    Regenerated the peephole files.  Some contained typos in the specification
2922    and some didn't compile because of pretty printing bugs in the old version
2923    of 'nowhere'.
2924    
2925    ----------------------------------------------------------------------
2926    Name: Allen Leung
2927    Date: 2002/02/19 20:20:00 EST
2928    Tag: leunga-20020219-mlrisc-tools
2929    Description:
2930    
2931       Minor bug fixes to the mlrisc-tools library:
2932    
2933       1.  Fixed up parsing colon suffixed keywords
2934       2.  Added the ability to shut the error messages up
2935       3.  Reimplemented the pretty printer and fixed up/improved
2936           the pretty printing of handle and -> types.
2937       4.  Fixed up generation of literal symbols in the nowhere tool.
2938       5.  Added some SML keywords to to sml.sty
2939    
2940    ----------------------------------------------------------------------
2941    Name: Matthias Blume
2942    Date: 2002/02/19 16:20:00 EST
2943    Tag: blume-20020219-cmffi
2944    Description:
2945    
2946    A wild mix of changes, some minor, some major:
2947    
2948    * All C FFI-related libraries are now anchored under $c:
2949        $/c.cm      --> $c/c.cm
2950        $/c-int.cm  --> $c/internals/c-int.cm
2951        $/memory.cm --> $c/memory/memory.cm
2952    
2953    * "make" tool (in CM) now treats its argument pathname slightly
2954      differently:
2955        1. If the native expansion is an absolute name, then before invoking
2956           the "make" command on it, CM will apply OS.Path.mkRelative
2957           (with relativeTo = OS.FileSys.getDir()) to it.
2958        2. The argument will be passed through to subsequent phases of CM
2959           processing without "going native".  In particular, if the argument
2960           was an anchored path, then "make" will not lose track of that anchor.
2961    
2962    * Compiler backends now "know" their respective C calling conventions
2963      instead of having to be told about it by ml-nlffigen.  This relieves
2964      ml-nlffigen from one of its burdens.
2965    
2966    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
2967    
2968    * Export C_DEBUG and C_Debug from $c/c.cm.
2969    
2970    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
2971      subtyping relationship between incomplete pointers and their complete
2972      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
2973      with the convention of instantiating 'o with ('t, 'c) obj whenever
2974      the pointer target type is complete.  In the incomplete case, 'o
2975      will be instantiated with some "'c iobj" -- a type obtained by
2976      using one of the functors PointerToIncompleteType or PointerToCompleteType.
2977    
2978      Operations that work on both incomplete and complete pointer types are
2979      typed as taking an 'o ptr while operations that require the target to
2980      be known are typed as taking some ('t, 'c) obj ptr.
2981    
2982      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
2983      where void is an eqtype without any values.  This makes it possible
2984      to work on voidptr values using functions meant to operate on light
2985      incomplete pointers.
2986    
2987    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
2988      been vastly simplified.
2989    
2990    ----------------------------------------------------------------------
2991    Name: Matthias Blume
2992    Date: 2002/02/19 10:48:00 EST
2993    Tag: blume-20020219-pqfix
2994    Description:
2995    
2996    Applied Chris Okasaki's bug fix for priority queues.
2997    
2998    ----------------------------------------------------------------------
2999    Name: Matthias Blume
3000    Date: 2002/02/15 17:05:00
3001    Tag: Release_110_39
3002    Description:
3003    
3004    Last-minute retagging is becoming a tradition... :-(
3005    
3006    This is the working release 110.39.
3007    
3008    ----------------------------------------------------------------------
3009    Name: Matthias Blume
3010    Date: 2002/02/15 16:00:00 EST
3011    Tag: Release_110_39-orig
3012    Description:
3013    
3014    Working release 110.39.  New bootfiles.
3015    
3016    (Update: There was a small bug in the installer so it wouldn't work
3017    with all shells.  So I retagged. -Matthias)
3018    
3019    ----------------------------------------------------------------------
3020    Name: Matthias Blume
3021    Date: 2002/02/15 14:17:00 EST
3022    Tag: blume-20020215-showbindings
3023    Description:
3024    
3025    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3026    the latter can be useful for exploring what bindings are available at
3027    the interactive prompt.  (The first function returns only the list
3028    of symbols that are really bound, the second prints those but also the
3029    ones that CM's autoloading mechanism knows about.)
3030    
3031    ----------------------------------------------------------------------
3032    Name: Matthias Blume
3033    Date: 2002/02/15 12:08:00 EST
3034    Tag: blume-20020215-iptrs
3035    Description:
3036    
3037    Two improvements to ml-nlffigen:
3038    
3039      1. Write files only if they do not exist or if their current contents
3040         do not coincide with what's being written.  (That is, avoid messing
3041         with the time stamps unless absolutely necessary.)
3042    
3043      2. Implement a "repository" mechanism for generated files related
3044         to "incomplete pointer types".   See the README file for details.
3045    
3046    ----------------------------------------------------------------------
3047    Name: Matthias Blume
3048    Date: 2002/02/14 11:50:00 EST
3049    Tag: blume-20020214-quote
3050    Description:
3051    
3052    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3053    because of the new and improved tag generation scheme.  (Thanks to Allen
3054    Leung for pointing it out.)
3055    
3056    ----------------------------------------------------------------------
3057    Name: Lal George
3058    Date: 2002/02/14 09:55:27 EST 2002
3059    Tag: george-20020214-isabelle-bug
3060    Description:
3061    
3062    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3063    of Isabelle on the x86.
3064    
3065    From Allen:
3066    -----------
3067     I've found the problem:
3068    
3069         in ra-core.sml, I use the counter "blocked" to keep track of the
3070         true number of elements in the freeze queue.  When the counter goes
3071         to zero, I skip examining the queue.  But I've messed up the
3072         bookkeeping in combine():
3073    
3074             else ();
3075             case !ucol of
3076               PSEUDO => (if !cntv > 0 then
3077                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3078                                        ^^^^^^^^^^^^^^^^^^^^^^^
3079                      moveu := mergeMoveList(!movev, !moveu)
3080                     )
3081                  else ();
3082    
3083         combine() is called to coalesce two nodes u and v.
3084         I think I was thinking that if the move counts of u and v are both
3085         greater than zero then after they are coalesced then one node is
3086         removed from the freeze queue.  Apparently I was thinking that
3087         both u and v are of low degree, but that's clearly not necessarily true.
3088    
3089    
3090    02/12/2002:
3091        Here's the patch.  HOL now compiles.
3092    
3093        I don't know how this impact on performance (compile
3094        time or runtime).  This bug caused the RA (especially on the x86)
3095        to go thru the potential spill phase when there are still nodes on the
3096        freeze queue.
3097    
3098    
3099    
3100    
3101    ----------------------------------------------------------------------
3102    Name: Matthias Blume
3103    Date: 2002/02/13 22:40:00 EST
3104    Tag: blume-20020213-fptr-rtti
3105    Description:
3106    
3107    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3108    updates.
3109    
3110    ----------------------------------------------------------------------
3111    Name: Matthias Blume
3112    Date: 2002/02/13 16:41:00 EST
3113    Tag: blume-20020213-cmlpq
3114    Description:
3115    
3116    Added new priority queue export symbols (which have just been added to
3117    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3118    not compile and the installer would choke.)
3119    
3120    ----------------------------------------------------------------------
3121    Name: Matthias Blume
3122    Date: 2002/02/13 16:15:00 EST
3123    Tag: blume-20020213-various
3124    Description:
3125    
3126    1. More tweaks to ml-nlffigen:
3127    
3128       - better internal datastructures (resulting in slight speedup)
3129       - "-match" option requires exact match
3130       - "localized" gensym counters (untagged structs/unions nested within
3131         other structs/unions or within typedefs get a fresh counter; their
3132         tag will be prefixed by a concatenation of their parents' tags)
3133       - bug fixes (related to calculation of transitive closure of types
3134         to be included in the output)
3135    
3136    2. Minor Basis updates:
3137    
3138       - added implementations for List.collate and Option.app
3139    
3140    ----------------------------------------------------------------------
3141    Name: Matthias Blume
3142    Date: 2002/02/11 15:55:00 EST
3143    Tag: blume-20020211-gensym
3144    Description:
3145    
3146    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3147    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3148    names (ML structure names that correspond to unnamed C structs, unions,
3149    and enums), so that separate runs of ml-nlffigen do not clash.
3150    
3151    ----------------------------------------------------------------------
3152    Name: Matthias Blume
3153    Date: 2002/02/11 12:05:00 EST
3154    Tag: blume-20020211-gensml
3155    Description:
3156    
3157    A quick fix for a problem with GenSML (in the pgraph-util library):
3158    Make generation of toplevel "local" optional.  (Strictly speaking,
3159    signature definitions within "local" are not legal SML.)
3160    
3161    Other than that: updates to INSTALL and cm/TODO.
3162    
3163    ----------------------------------------------------------------------
3164    Name: Matthias Blume
3165    Date: 2002/02/08 15:00:00 EST
3166    Tag: blume-20020208-uniquepid
3167    Description:
3168    
3169    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3170    
3171    1. The installer (config/install.sh) has gotten smarter:
3172    
3173         - Configuration options are a bit easier to specify now
3174           (in config/targets).
3175         - Bug in recognizing .tar.bz2 files fixed.
3176         - Installer automatically resolves dependencies between
3177           configuration options (e.g., if you ask for eXene, you will
3178           also get cml -- regardless whether you asked for it or not).
3179         - Installer can run in "quieter mode" by setting the environment
3180           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3181           "completely silent", though.
3182         - Build HashCons library as part of smlnj-lib.
3183    
3184    2. A new scheme for assigning persistent identifiers to compilation
3185       units (and, by extension, to types etc.) has been put into place.
3186       This fixes a long-standing bug where types and even dynamic values
3187       can get internally confused, thereby compromising type safety
3188       (abstraction) and dynamic correctness.  See
3189    
3190         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3191    
3192       for an example of how things could go wrong until now.
3193    
3194       The downside of the new scheme is that pids are not quite as
3195       persistent as they used to be: CM will generate a fresh pid
3196       for every compilation unit that it thinks it sees for the first
3197       time.  That means that if you compile starting from a clean, fresh
3198       source tree at two different times, you end up with different
3199       binaries.
3200    
3201       Cutoff recompilation, however, has not been compromised because
3202       CM keeps pid information in special caches between runs.
3203    
3204    ----------------------------------------------------------------------
3205    Name: Lal George
3206    Date: 2002/02/07 15:34:13 EST 2002
3207    Tag: <none>
3208    Description:
3209    
3210    Compilers that generate assembly code may produce  global labels
3211    whose value is resolved at link time. The various peephole optimization
3212    modules did not take this in account.
3213    
3214    TODO. The Labels.addrOf function should really return an option
3215    type so that clients are forced to deal with this issue, rather
3216    than an exception being raised.
3217    
3218    ----------------------------------------------------------------------
3219    Name: Lal George
3220    Date: 2002/02/06 13:55:02 EST
3221    Tag: george-20020206-ra-breakup
3222    Description:
3223    
3224    1. A bug fix from Allen.
3225    
3226        A typo causes extra fstp %st(0)'s to be generated at compensation
3227        edges, which might cause stack underflow traps at runtime.  This
3228        occurs in fft where there are extraneous fstps right before the 'into'
3229        trap instruction (in this case they are harmless since none of the
3230        integers overflow.)
3231    
3232    2. Pulled out various utility modules that were embedded in the modules
3233       of the register allocator. I need these modules for other purposes, but
3234       they are not complete enough to put into a library (just yet).
3235    ----------------------------------------------------------------------
3236    Name: Matthias Blume
3237    Date: 2002/01/31 16:05:00 EST
3238    Tag: blume-20020131-sparc-ccalls
3239    Description:
3240    
3241    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3242       of extra stack space by mistake.  Fixed.
3243    
3244    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3245    
3246    ----------------------------------------------------------------------
3247    Name: Allen Leung
3248    Date: 2002/01/30
3249    Tag: leunga-20020130-nowhere-bug-fix
3250    Description:
3251    
3252       MLRISC bug fixes:
3253       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3254       2. MachineInt.fromString was negating its value.
3255    
3256    ----------------------------------------------------------------------
3257    Name: Matthias Blume
3258    Date: 2002/01/29
3259    Tag: blume-20020129-INSTALL
3260    Description:
3261    
3262    - Added somewhat detailed installation instructions (file INSTALL).
3263    - Fixed curl-detection bug in config/install.sh.
3264    - It is now possible to select the URL getter using the URLGETTER
3265      environment variable:
3266    
3267          not set / "unknown"      --> automatic detection (script tries wget,
3268                                       curl, and lynx)
3269          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3270                                       how to properly invoke them)
3271          other                    --> use $URLGETTER directly, it must take
3272                                       precisely two command-line arguments
3273                                       (source URL and destination file name)
3274    
3275    ----------------------------------------------------------------------
3276    Name: Matthias Blume
3277    Date: 2002/01/28
3278    Tag: blume-20020128-sparc-ccalls
3279    Description:
3280    
3281    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3282    - Make use of the allocParam argument in sparc-c-calls.
3283    
3284    ----------------------------------------------------------------------
3285    Name: Matthias Blume
3286    Date: 2002/01/28
3287    Tag: blume-20020128-allocParam
3288    Description:
3289    
3290    John Reppy:  Changes c-calls API to accept client-callback for
3291    allocating extra stack space.
3292    me: Corresponding changes to mlriscGen (using a dummy argument that
3293        does not change the current behavior).
3294    
3295    ----------------------------------------------------------------------
3296    Name: Matthias Blume
3297    Date: 2002/01/28 12:00:00
3298    Tag: Release_110_38
3299    Description:
3300    
3301    This time for real!!!
3302    
3303    ----------------------------------------------------------------------
3304    Name: Matthias Blume
3305    Date: 2002/01/28 10:56:00 EST
3306    Tag: blume-20020128-retraction
3307    Description:
3308    
3309    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3310       with blume-Release_110_38-retracted.)
3311    
3312    1. Fixed a problem with incorrect rounding modes in real64.sml.
3313       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3314    
3315    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3316       and enums fixed.  The naming of corresponding ML identifiers should
3317       now be consistent again.
3318    
3319    ----------------------------------------------------------------------
3320    Name: Allen Leung
3321    Date: 2002/01/27
3322    Tag: leunga-20020127-nowhere
3323    Description:
3324    
3325       Added a target called nowhere in the configuration scripts.
3326       Enabling this will build the MLRISC 'nowhere' tool (for translating
3327       programs with where-clauses into legal SML code) during installation.
3328    
3329    ----------------------------------------------------------------------
3330    Name: Matthias Blume
3331    Date: 2002/01/25 21:27:00 EST
3332    Tag: blume-Release_110_38-retracted
3333    Description:
3334    
3335    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3336    
3337    README will be added later.
3338    
3339    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3340    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3341    
3342    ----------------------------------------------------------------------
3343    Name: Matthias Blume
3344    Date: 2002/01/25
3345    Tag: blume-20020125-ffi
3346    Description:
3347    
3348    A large number of tweaks and improvements to ml-nlffi-lib and
3349    ml-nlffigen:
3350    
3351       - ML represenation types have been streamlined
3352       - getter and setter functions work with concrete values, not abstract
3353         ones where possible
3354       - ml-nlffigen command line more flexible (see README file there)
3355       - some bugs have been fixed (hopefully)
3356    
3357    ----------------------------------------------------------------------
3358    Name: Lal George
3359    Date: 2002/01/24
3360    Tag: george-20020124-risc-ra-interface
3361    Description:
3362    
3363       There is a dramatic simplification in the interface to the
3364       register allocator for RISC architectures as a result of making
3365       parallel copy instructions explicit.
3366    
3367    ----------------------------------------------------------------------
3368    Name: Matthias Blume
3369    Date: 2002/01/22
3370    Tag: blume-20020122-x86-ccalls
3371    Description:
3372    
3373    Bug fix for c-calls on x86 (having to do with how char- and
3374    short-arguments are being handled).
3375    
3376    ----------------------------------------------------------------------
3377    Name: Matthias Blume
3378    Date: 2002/01/21
3379    Tag: blume-20020121-ff
3380    Description:
3381    
3382    Another day of fiddling with the FFI...
3383    
3384    1. Bug fix/workaround:  CKIT does not complain about negative array
3385       dimensions, so ml-nlffigen has to guard itself against this possibility.
3386       (Otherwise a negative dimension would send it into an infinite loop.)
3387    
3388    2. Some of the abstract types (light objects, light pointers, most "base"
3389       types) in structure C are now eqtypes.
3390    
3391    3. Added constructors and test functions for NULL function pointers.
3392    
3393    ----------------------------------------------------------------------
3394    Name: Matthias Blume
3395    Date: 2002/01/18
3396    Tag: blume-20020118-ready-for-new-release
3397    Description:
3398    
3399    Made config/srcarchiveurl point to a new place.  (Will provide boot
3400    files shortly.)
3401    
3402    Maybe we christen this to be 110.38?
3403    
3404    ----------------------------------------------------------------------
3405    Name: Matthias Blume
3406    Date: 2002/01/18
3407    Tag: blume-20020118-more-ffifiddle
3408    Description:
3409    
3410    Today's FFI fiddling:
3411    
3412      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3413        C.Set.* that use concrete (MLRep.*) arguments and results instead
3414        of abstract ones.
3415    
3416      - Provided word-style bit operations etc. for "int" representation
3417        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3418        and Long.
3419    
3420    ----------------------------------------------------------------------
3421    Name: Matthias Blume
3422    Date: 2002/01/18
3423    Tag: blume-20020118-use-x86-fp
3424    Description:
3425    
3426    Now that x86-fast-fp seems to be working, I turned it back on again
3427    by default.  (Seems to work fine now, even with the FFI.)
3428    
3429    Other than that, I added some documentation about the FFI to
3430    src/ml-nlffigen/README and updated the FFI test examples in
3431    src/ml-nlffi-lib/Tests/*.
3432    
3433    ----------------------------------------------------------------------
3434    Name: Allen Leung
3435    Date: 2002/01/17
3436    Tag: leunga-20020117-x86-fast-fp-call
3437    Description:
3438    
3439       1. Fixed a problem with handling return fp values when x86's fast fp
3440          mode is turned on.
3441    
3442       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3443          of %f32.
3444    
3445       3. Added a constructor INT32lit to the ast of MLRISC tools.
3446    
3447    ----------------------------------------------------------------------
3448    Name: Matthias Blume
3449    Date: 2002/01/16
3450    Tag: blume-20020116-ffifiddle
3451    Description:
3452    
3453    More fiddling with the FFI interface:
3454    
3455     - Make constness 'c instead of rw wherever possible.  This eliminates
3456       the need for certain explicit coercions.  (However, due to ML's
3457       value polymorphism, there will still be many cases where explicit
3458       coercions are necessary.  Phantom types are not the whole answer
3459       to modeling a subtyping relationship in ML.)
3460    
3461     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3462    
3463     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3464       clumsy workaround for varargs problem.)
3465    
3466    ----------------------------------------------------------------------
3467    Name: Lal George
3468    Date: 2002/01/15
3469    Tag: <none>
3470    Description:
3471    
3472    1. Since COPY instructions are no longer native to the architecture,
3473       a generic functor can be used to implement the expandCopies function.
3474    
3475    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3476       TEXT segment.
3477    
3478    ----------------------------------------------------------------------
3479    Name: Matthias Blume
3480    Date: 2002/01/15
3481    Tag: blume-20020115-ffiupdates
3482    Description:
3483    
3484    1. Fix for bug resulting in single-precision float values being returned
3485       incorrectly from FFI calls.
3486    
3487    2. Small modifications to C FFI API:
3488    
3489        - memory-allocation routines return straight objects (no options)
3490          and raise an exception in out-of-memory situations
3491        - unsafe extensions to cast between function pointers and pointers
3492          from/to ints
3493        - added structure C_Debug as an alternative to structure C where
3494          pointer-dereferencing (|*| and |*!) always check for null-pointers
3495        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3496          but also takes a (possibly empty) list of existing library handles
3497          that the current library depends on
3498    
3499    ----------------------------------------------------------------------
3500    Name: Matthias Blume
3501    Date: 2002/01/10
3502    Tag: blume-20020110-newffigen
3503    Description:
3504    
3505    1. Updates to portable graph code.
3506    
3507    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3508       more scalable now so that even huge interfaces such as the one
3509       for GTK compile in finite time and space. :-)
3510       See src/ml-nlffigen/README for details on what's new.
3511    
3512    ----------------------------------------------------------------------
3513    Name: Lal George
3514    Date: 2001/01/09 14:31:35 EST 2002
3515    Tag: george-20011206-rm-native-copy
3516    Description:
3517    
3518            Removed the native COPY and FCOPY instructions
3519            from all the architectures and replaced it with the
3520            explicit COPY instruction from the previous commit.
3521    
3522            It is now possible to simplify many of the optimizations
3523            modules that manipulate copies. This has not been
3524            done in this change.
3525    
3526    ----------------------------------------------------------------------
3527    Name: Lal George
3528    Date: 2001/12/06 16:50:13 EST 2001
3529    Tag: george-20011206-mlrisc-instruction
3530    Description:
3531    
3532    Changed the representation of instructions from being fully abstract
3533    to being partially concrete. That is to say:
3534    
3535      from
3536            type instruction
3537    
3538      to
3539            type instr                              (* machine instruction *)
3540    
3541            datatype instruction =
3542                LIVE of {regs: C.cellset, spilled: C.cellset}
3543              | KILL of {regs: C.cellset, spilled: C.cellset}
3544              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3545              | ANNOTATION of {i: instruction, a: Annotations.annotation}
3546              | INSTR of instr
3547    
3548    This makes the handling of certain special instructions that appear on
3549    all architectures easier and uniform.
3550    
3551    LIVE and KILL say that a list of registers are live or killed at the
3552    program point where they appear. No spill code is generated when an
3553    element of the 'regs' field is spilled, but the register is moved to
3554    the 'spilled' (which is present, more for debugging than anything else).
3555    
3556    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3557    We used to generate:
3558    
3559            DEFFREG f1
3560            f1 := f2 + f3
3561            trapb
3562    
3563    but now generate:
3564    
3565            f1 := f2 + f3
3566            trapb
3567            LIVE {regs=[f1,f2,f3], spilled=[]}
3568    
3569    Furthermore, the DEFFREG (hack) required that all floating point instruction
3570    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3571    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3572    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3573    this architecture).
3574    
3575    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3576    all the architectures. This will result in further simplification of the
3577    register allocator that must be aware of them for coalescing purposes, and
3578    will also simplify certain aspects of the machine description that provides
3579    callbacks related to parallel copies.
3580    
3581    ANNOTATION should be obvious, and now INSTR represents the honest to God
3582    machine instruction set!
3583    
3584    The <arch>/instructions/<arch>Instr.sml files define certain utility
3585    functions for making porting easier -- essentially converting upper case
3586    to lower case. All machine instructions (of type instr) are in upper case,
3587    and the lower case form generates an MLRISC instruction. For example on
3588    the alpha we have:
3589    
3590      datatype instr =
3591         LDA of {r:cell, b:cell, d:operand}
3592       | ...
3593    
3594      val lda : {r:cell, b:cell, d:operand} -> instruction
3595        ...
3596    
3597    where lda is just (INSTR o LDA), etc.
3598    
3599    ----------------------------------------------------------------------
3600    Name: Matthias Blume
3601    Date: 2001/11/22 21:40:00 EST
3602    Tag: Release_110_37
3603    Description:
3604    
3605    Release 110.37.  This time for real.
3606    
3607    ----------------------------------------------------------------------
3608    Name: Matthias Blume
3609    Date: 2001/11/21 16:35:00 EST
3610    Tag: blume-20011121-foot-in-mouth
3611    Description:
3612    
3613    Removed the "Release_110_37" tag because of a serious bug.
3614    This will be re-tagged once the bug is fixed.
3615    
3616    ----------------------------------------------------------------------
3617    Name: Matthias Blume
3618    Date: 2001/11/21 16:14:00 EST
3619    Tag: blume-20011121-forgottenfile
3620    Description:
3621    
3622    Forgot to add a file.  (Just a .tex-file -- part of
3623    the CM manual source.)
3624    
3625    ----------------------------------------------------------------------
3626    Name: Matthias Blume
3627    Date: 2001/11/21 16:10:00 EST
3628    Tag: blume-20011121-invalid_110_37
3629    Description:
3630    
3631    Note: I removed the original tag "Release_110_37" from this commit
3632          because we found a serious bug in all non-x86 backends.
3633          - Matthias
3634    
3635    1. Modifications to the SML/NJ code generator and to the runtime system
3636       so that code object name strings are directly inserted into code
3637       objects at code generation time.  The only business the runtime system
3638       has with this is now to read the name strings on occasions.
3639       (The encoding of the name string has also changed somewhat.)
3640    
3641    2. CM now implements a simple "set calculus" for specifying export lists.
3642       In particular, it is now possible to refer to the export lists of
3643       other libraries/groups/sources and form unions as well as differences.
3644       See the latest CM manual for details.
3645    
3646    3. An separate notion of "proxy" libraries has again be eliminated from
3647       CM's model.  (Proxy libraries are now simply a special case of using
3648       the export list calculus.)
3649    
3650    4. Some of the existing libraries now take advantage of the new set
3651       calculus.
3652       (Notice that not all libraries have been converted because some
3653       of the existing .cm-files are supposed to be backward compatible
3654       with 110.0.x.)
3655    
3656    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3657       "General.exnMessage"!  The former relies on a certain hook to be
3658       initialized, and that often does not happen in the stand-alone case.)
3659    
3660    ----------------------------------------------------------------------
3661    Name: Lal George
3662    Date: 2001/11/21  13:56:18 EST
3663    Tag: george-2001121-pseudo-ops
3664    Description:
3665    
3666      Implemented a complete redesign of MLRISC pseudo-ops. Now there
3667      ought to never be any question of incompatabilities with
3668      pseudo-op syntax expected by host assemblers.
3669    
3670      For now, only modules supporting GAS syntax are implemented
3671      but more should follow, such as MASM, and vendor assembler
3672      syntax, e.g. IBM as, Sun as, etc.
3673    
3674    ----------------------------------------------------------------------
3675    Name: Matthias Blume
3676    Date: 2001/11/14 11:52:00 EST
3677    Tag: blume-20011114-srcname
3678    Description:
3679    
3680    1. Routed the name of the current source file to mlriscgen where it
3681       should be directly emitted into the code object.  (This last part
3682       is yet to be done.)
3683    
3684    2. Some cleanup of the pgraph code to make it match the proposal that
3685       I put out the other day.  (The proposal notwithstanding, things are
3686       still in flux here.)
3687    
3688    ----------------------------------------------------------------------
3689    Name: Lal George
3690    Date: 2001/11/14 09:44:04 EST
3691    Tag:
3692    Description:
3693    
3694      Fix for a backpatching bug reported by Allen.
3695    
3696      Because the boundary between short and long span-dependent
3697      instructions is +/- 128, there are an astounding number of
3698      span-dependent instructions whose size is over estimated.
3699    
3700      Allen came up with the idea of letting the size of span
3701      dependent instructions be non-monotonic, for a maxIter
3702      number of times, after which the size must be monotonically
3703      increasing.
3704    
3705      This table shows the number of span-dependent instructions
3706      whose size was over-estimated as a function of maxIter, for the
3707      file Parse/parse/ml.grm.sml:
3708    
3709         maxIter            # of instructions:
3710            10                      687
3711            20                      438
3712            30                      198
3713            40                        0
3714    
3715      In compiling the compiler, there is no significant difference in
3716      compilation speed between maxIter=10 and maxIter=40. Actually,
3717      my measurements showed that maxIter=40 was a tad faster than
3718      maxIter=10! Also 96% of the  files in the compiler reach a fix
3719      point within 13 iterations, so fixing maxIter at 40, while high,
3720      is okay.
3721    
3722    ----------------------------------------------------------------------
3723    Name: Matthias Blume
3724    Date: 2001/10/31 15:25:00 EST
3725    Tag: blume-20011031-pgraph
3726    Description:
3727    
3728    CKIT:
3729    * Changed the "Function" constructor of type Ast.ctype to carry optional
3730      argument identifiers.
3731    * Changed the return type of TypeUtil.getFunction accordingly.
3732    * Type equality ignores the argument names.
3733    * TypeUtil.composite tries to preserve argument names but gives up quickly
3734      if there is a mismatch.
3735    
3736    installation script:
3737    * attempts to use "curl" if available (unless "wget" is available as well)
3738    
3739    CM:
3740    * has an experimental implementation of "portable graphs" which I will
3741      soon propose as an implementation-independent library format
3742    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
3743    
3744    NLFFI-LIB:
3745    * some cleanup (all cosmetic)
3746    
3747    NLFFIGEN:
3748    * temporarily disabled the mechanism that suppresses ML output for
3749      C definitions whose identifiers start with an underscore character
3750    * generate val bindings for enum constants
3751    * user can request that only one style (light or heavy) is being used;
3752      default is to use both (command-line arguments: -heavy and -light)
3753    * fixed bug in handling of function types involving incomplete pointers
3754    * generate ML entry points that take record arguments (i.e., using
3755      named arguments) for C functions that have a prototype with named
3756      arguments
3757      (see changes to CKIT)
3758    
3759    ----------------------------------------------------------------------
3760    Name: Allen Leung
3761    Date: 2001/10/27 20:34:00 EDT
3762    Tag: leunga-20011027-x86-fast-fp-call
3763    Description:
3764    
3765       Fixed the bug described in blume-20010920-slowfp.
3766    
3767       The fix involves
3768          1. generating FCOPYs in FSTP in ia32-svid
3769          2. marking a CALL with the appropriate annotation
3770    
3771    ----------------------------------------------------------------------
3772    Name: Matthias Blume
3773    Date: 2001/10/16 11:32:00 EDT
3774    Tag: blume-20011016-netbsd
3775    Description:
3776    
3777    Underscore patch from Chris Richards (fixing problem with compiling
3778    runtime system under recent NetBSD).
3779    
3780    ----------------------------------------------------------------------
3781    Name: Allen Leung
3782    Date: 2001/10/12 17:18:32 EDT 2001
3783    Tag: leung-20011012-x86-printflowgraph
3784    Description:
3785    
3786    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
3787    
3788    ----------------------------------------------------------------------
3789    Name: Lal George
3790    Date: 2001/10/11 23:51:34 EDT
3791    Tag: george-20011011-too-many-instrs
3792    Description:
3793    
3794    The representation of a program point never expected to see more
3795    than 65536 instructions in a basic block!
3796    
3797    ----------------------------------------------------------------------
3798    Name: Lal George
3799    Date: 2001/10/09 09:41:37 EDT
3800    Tag: george-20011008-mlrisc-labels
3801    Description:
3802    
3803    Changed the machine description files to support printing of
3804    local and global labels in assembly code, based on host assembler
3805    conventions.
3806    
3807    ----------------------------------------------------------------------
3808    Name: Matthias Blume
3809    Date: 2001/09/25 15:25:00 EDT
3810    Tag: blume-20010925-exninfo
3811    Description:
3812    
3813    I provided a non-hook implementation of exnName (at the toplevel) and
3814    made the "dummy" implementation of exnMessage (at the toplevel) more
3815    useful: if nothing gets "hooked in", then at least you are going to
3816    see the exception name and a message indicating why you don't see more.
3817    
3818    [For the time being, programs that need exnMessage and want to use
3819    ml-build should either use General.exnMessage (strongly recommended) or
3820    refer to structure General at some other point so that CM sees a
3821    static dependency.]
3822    
3823    [Similar remarks go for "print" and "use":  If you want to use their
3824    functionality in stand-alone programs generated by ml-build, then use
3825    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
3826    
3827    ----------------------------------------------------------------------
3828    Name: Matthias Blume
3829    Date: 2001/09/20 17:28:00 EDT
3830    Tag: blume-20010920-slowfp
3831    Description:
3832    
3833    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
3834    
3835    ----------------------------------------------------------------------
3836    Name: Matthias Blume
3837    Date: 2001/09/20 17:20:00 EDT
3838    Tag: blume-20010920-canonicalpaths
3839    Description:
3840    
3841    0. Updated the BOOT file (something that I forgot to do earlier).
3842    
3843    1. Small internal change to CM so that it avoids "/../" in filenames
3844       as much as possible (but only where it is safe).
3845    
3846    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
3847       that arguments that contain delimiters are passed through correctly.
3848       This change also means that all "special" arguments of the form