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 959, Fri Oct 12 21:21:41 2001 UTC revision 1682, Tue Nov 9 23:48:09 2004 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/11/09 17:50:00 CST
18    Tag: blume-20041109-64bit
19    Description:
20    
21    Started adding some infrastructure for supporting 64-bit int- and
22    word-types.  (Still in its very early stages.)
23    
24    ----------------------------------------------------------------------
25    Name: Matthias Blume (blume (at) tti - c (dot) org)
26    Date: 2004/10/28 10:45:00 CDT
27    Tag: Release_110_50
28    Description:
29    
30    New working version (110.50).  NEW BOOTFILES!
31    
32                 =====================
33    
34    Also:
35    
36      - Changend config/srcarchiveurl from a file just
37        containing the URL string into a file containing
38        shell script code.  The code has access to the $VERSION variable.
39      - Made corresponding changes to config/install.sh and config/unpack.
40      - Default contents of config/srcarchiveurl uses $VERSION and
41        normally does not have to be edited to reflect a version change.
42    
43      (As a result, a version change can be done by just editing
44       config/version, the rest is now automatic.)
45    
46    ----------------------------------------------------------------------
47    Name: Matthias Blume (blume (at) tti - c (dot) org)
48    Date: 2004/10/27 17:50:00 CDT
49    Tag: blume-20041027-btrace-msg
50    Description:
51    
52    BackTrace.monitor now also reports the source of the exception that
53    triggered the trace.
54    
55    ----------------------------------------------------------------------
56    Name: Matthias Blume (blume (at) tti - c (dot) org)
57    Date: 2004/10/27 17:20:00 CDT
58    Tag: blume-20041027-x86-c-calls
59    Description:
60    
61    This is the HISTORY entry for two earlier commits, both concerning
62    the x86 c-calls code in MLRISC:
63    
64      - added a missing LOAD in the code that deals with struct arguments
65      - made sure the caller does not add the wrong number of bytes to the
66        stack pointer after a call of a function returning a struct
67        (the callee already pops the implicit argument which points to
68         the space reserved for the result)
69    
70    ----------------------------------------------------------------------
71    Name: Allen Leung (leunga (at) reservoir (dot) com)
72    Date: 2004/10/24 14:00:00 EST
73    Tag: leunga-20041024-x86-gas-fucomip
74    Description:
75    
76       John discovered a bug in the syntax of fucomip.
77       The opcodes FU?COMIP? have been changed to
78    
79           fu?comip? %st(i), %st
80    
81    ----------------------------------------------------------------------
82    Name: Matthias Blume (blume (at) tti - c (dot) org)
83    Date: 2004/10/20 15:06:00 CDT
84    Tag: blume-20041020-standalone-backtrace
85    Description:
86    
87    Added a mechanism for getting back-trace information from standalone
88    programs.  Here is how it works:
89    
90       1. The part of the program from which you want to get backtrace
91          information (usually the whole program) should be wrapped with
92          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
93          main program could be modified from something like
94    
95             fun main (pgm, args) = ...
96    
97         to
98    
99            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
100    
101       2. To be able to access BackTrace.monitor, you have to add
102          library $smlnj-tdp/plugins.cm to the .cm file that contains your
103          main function.
104    
105       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
106          CM might have created in the past for your project).
107    
108       4. Build the system using this command line:
109    
110             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
111                      myprog.cm MyProg.main myprog
112    
113          instead of the usual
114    
115             ml-build myprog.cm MyProg.main myprog
116    
117    I changed a library name:
118    
119       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
120    
121    New libraries:
122    
123       $smlnj-tdp/back-trace.cm
124          -- when loaded causes the back-trace plugin to be installed
125       $smlnj-tdp/coverage.cm
126          -- when loaded causes the coverage plugin to be installed
127    
128    ----------------------------------------------------------------------
129    Name: Matthias Blume (blume (at) tti - c (dot) org)
130    Date: 2004/10/18 16:45:00 CDT
131    Tag: blume-20041018-groupowner
132    Description:
133    
134    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
135    
136    Eliminated group owner specs from .cm files throughout the source tree.
137    
138    ----------------------------------------------------------------------
139    Name: Matthias Blume (blume (at) tti - c (dot) org)
140    Date: 2004/10/15 15:45:00 CDT
141    Tag: blume-20041015-coverage
142    Description:
143    
144    * Test coverage tool added!
145    * Further reorganization of tracing-, debugging-, and profiling support:
146    
147        - moved original BTImp -- now called BackTrace -- into a separate
148          library called $/trace-debug-profile.cm
149        - eliminated all mentions of BTrace from SMLofNJ.Internals
150        - only the instrumentation mechanism is now left in the compiler proper
151        - BackTrace module is a plugin which is NOT plugged in by default
152        - Coverage module is another such plugin
153    
154    To get the benefits of any of these plugin modules, the code in
155    question must be compiled with tdp instrumentation turned on.  This
156    can be done by setting SMLofNJ.Internals.TDP.mode to true.
157    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
158    
159    Plugins are selected at link time.  (Pre-compiled instrumented code
160    can be re-loaded with different plugins in effect.)  When an
161    instrumented module is linked, whatever plugins are at that time
162    enabled will come into effect for that module.
163    
164    To enable the back-trace plugin, load library $/trace-debug-profile.cm
165    and invoke BackTrace.install() (e.g., from the interactive prompt).
166    To enable the coverage plugin, load the same library and invoke
167    Coverage.install().
168    
169    Back-traces are generated automatically on uncaught exceptions and
170    when the code in question explicitly invokes BackTrace.trigger().
171    
172    Coverage (and execution frequency-) information must be queried
173    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
174    
175    ----------------------------------------------------------------------
176    Name: Matthias Blume (blume (at) tti - c (dot) org)
177    Date: 2004/10/14 17:40:00 CDT
178    Tag: blume-20041014-tdp-core
179    Description:
180    
181    Snapshot of a significant overhaul of how the trace/debug/profile support
182    is hooked into the system (specifically: Core and SMLofNJ.Internals).
183    
184    ----------------------------------------------------------------------
185    Name: Matthias Blume (blume (at) tti - c (dot) org)
186    Date: 2004/10/13 16:34:00 CDT
187    Tag: blume-20041013-tdp
188    Description:
189    
190    Some rationalization of names:
191    
192       structure BTrace -> structure TDPInstrument
193       etc.
194    
195    This is is preparation of using the original back-trace
196    instrumentation for other purposes.  "TDP" stands for
197    Trace/Debug/Profile.
198    
199    The control flag controlling whether instrumentation is on or off is now
200    registered under a different name, so instead of running sml as
201    
202     sml -Cinstrument.btrace-mode=true
203    
204    one has to say
205    
206     sml -Ctdp.instrument=true
207    
208    ----------------------------------------------------------------------
209    Name: Matthias Blume (blume (at) tti - c (dot) org)
210    Date: 2004/10/11 16:37:00 CDT
211    Tag: blume-20041011-regions
212    Description:
213    
214    Made some minor modifications to elabcore.sml to have source regions
215    be propagated more tightly -- resulting in better (i.e., smaller)
216    regions being reported in error- and debug messages.
217    
218    ----------------------------------------------------------------------
219    Name: Matthias Blume (blume (at) tti - c (dot) org)
220    Date: 2004/10/08 22:50:00 CDT
221    Tag: blume-20041008-cmkw
222    Description:
223    
224    Fixed handling of keywords in .cm files: After seeing "is" the lexer
225    treats subsequent occurrences of "group", "library", "source", "is",
226    "*", and "-" as ordinary identifiers rather than keywords.
227    
228    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
229    is supposed to accept a tool argument called "source", but this did
230    not work because of the clash with the keyword.
231    
232    ----------------------------------------------------------------------
233    Name: Matthias Blume (blume (at) tti - c (dot) org)
234    Date: 2004/10/07 16:00:00 CDT
235    Tag: blume-20041007-cleanup
236    Description:
237    
238    Assorted cleanup work:
239    
240      - got rid of intstrmap in favor of using the library's
241        hash table implementation
242      - threw out most of the pathnames stuff, as it was not used anyway
243      - simplified tokentable implementation
244      - fixed some minor spelling errors
245    
246    ----------------------------------------------------------------------
247    Name: Matthias Blume (blume (at) tti - c (dot) org)
248    Date: 2004/10/06 15:15:15 CDT
249    Tag: blume-20041006-handler
250    Description:
251    
252    Cleaned up the absyn to reflect the invariant that HANDLE always
253    carries a FNexp as part of the type definition.  This eliminates some
254    superfluous sanity checks at runtime down the road.
255    
256    Some minor cleanup of the btrace code.
257    
258    ----------------------------------------------------------------------
259    Name: Matthias Blume (blume (at) tti - c (dot) org)
260    Date: 2004/10/01 10:20:30 CDT
261    Tag: blume-20041001-slave
262    Description:
263    
264    Added hack to make slave mode work in the presence of the version
265    tool.  (Still, since the master does two passes over the code for
266    CMB.make, the release number gets bumped twice when slaves are
267    attached.  I don't know if this is worth fixing...)
268    
269    ----------------------------------------------------------------------
270    Name: Matthias Blume (blume (at) tti - c (dot) org)
271    Date: 2004/09/30 10:55:00 CDT
272    Tag: blume-20040930-version
273    Description:
274    
275    * Moved the "version" magic into its own little library under
276      src/system/smlnj/internal.  This avoids expensive reconstruction of
277      a stable src/compiler/core.cm.
278    
279    * At the same time, structure CompilerVersion is now known as
280      structure SMLNJVersion.
281    
282    * Arranged for the version tool to NOT kick in when rebuilding the system
283      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
284      Also, loading the versiontool does not work when rebuilding the system
285      because CM is not properly initialized at that time.
286    
287    ----------------------------------------------------------------------
288    Name: Matthias Blume (blume (at) tti - c (dot) org)
289    Date: 2004/09/29 14:00:00 CDT
290    Tag: blume-20040929-autoversion
291    Description:
292    
293    Implemented some CM magic to have
294        file src/compiler/TopLevel/main/version.sml
295    generated automagically.
296    The version is taken from two files: config/version and config/release.
297    The first is expected to contain a two-part version number such as 110.49.
298    The second should contain a single number, but it may be missing.
299    
300    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
301    time the version tool is loaded (which is the first time you say CMB.make),
302    then the tool will increment the value stored in config/release every
303    time CMB.make is invoked.
304    
305    The binfile format is now insensitive to anything beyond the first
306    two components of a version number, so bumping the release does not render
307    binfiles incompatible.  Auto-bumping can be used to keep track of versions
308    during development without invalidating existing binfiles.
309    
310    In any case, every CMB.make updates the date information in version.sml.
311    (This is the date that is printed in the banner.)
312    
313    ----------------------------------------------------------------------
314    Name: Matthias Blume (blume (at) tti - c (dot) org)
315    Date: 2004/09/28 10:53:00 CDT
316    Tag: blume-20040928-controls
317    Description:
318    
319    Some cleanup of the controls code.
320    
321    ----------------------------------------------------------------------
322    Name: Matthias Blume (blume (at) tti - c (dot) org)
323    Date: 2004/09/27 22:08:00 CDT
324    Tag: blume-20040927-controls
325    Description:
326    
327    Added two pieces of functionality to the Controls interface:
328    
329    1.   val save'restore: 'a control -> unit -> unit
330    
331       grabs the current value of the control in stage 1 and restores it
332       in stage 2.
333    
334    2.   val set' : 'a control * 'a -> unit -> unit
335    
336       stores the given value into the control in stage 2 (i.e., delayed)
337       but does all error checking in stage 1.
338       (This is for string controls that need to do parse their argument --
339        something that might fail.  In some cases, notably in CM, one
340        already knows the intended argument but wants to delay the actual
341        assignment until a time when error recovery would be more difficult.)
342    
343    Changed the handling of controls in tool arguments to classes "sml" and
344    "lazysml":
345       - use Controls.save'restore as a more robust way of restoring the
346         old value (in particular: without having to re-parse the string)
347       - use controls to handle the "overload" keyword in the init group
348         (I believe this change actually fixes a long-standing obscure bug.)
349    
350    ----------------------------------------------------------------------
351    Name: Matthias Blume (blume (at) tti - c (dot) org)
352    Date: 2004/09/27 17:00:00 CDT
353    Tag: blume-20040927-lazysml
354    Description:
355    
356    Added a new tool class called "lazysml" to CM's tool chest.  The only
357    difference to "sml" is that compilation is done with Control.lazysml
358    set to true.  A source of class "lazysml" is automatically recognized
359    by a file name suffix of ".lml".
360    
361    In addition to the above feature, the original class "sml" now also
362    supports a tool argument "lazy" which has the same effect.  As a
363    result, the following three lines are equivalent:
364    
365        foo.sml : lazysml
366        foo.sml : sml (lazy)
367        foo.sml (lazy)
368    
369    The setting goes into effect both during parsing and during
370    compilation.  The original setting is restored right after parsing and
371    after compilation, respectively.
372    
373    In addition to all the above, there is also a general mechanism to set
374    ANY of the "controls" that are available at the command line via
375    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
376    apply as well. (In fact, "lazy" is implemented as a special case of
377    the general mechanism.)
378    
379    The .cm file syntax uses a new keyword tool argument called "with".
380    There are several ways of indicating the desired settings:
381    
382        foo.sml (with:parser.quotations=true)
383        foo.sml (with:(name:parser.quotations value:true))
384        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
385        foo.sml (with:(name1=value1 name2=value2 ...))
386        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
387    
388        etc.
389    
390    Another possible abbreviation is to leave out the =v or value:v part
391    if the name refers to a boolean control (in which case the value is
392    taken to be true).  Thus, one could get lazy sml also by saying:
393    
394        foo.sml (with:parser.lazy-keyword=true)
395        foo.sml (with:parser.lazy-keyword)
396        foo.sml (with:(name:parser.lazy-keyword value:true))
397        foo.sml (with:(name:parser.lazy-keyword))
398    
399    ----------------------------------------------------------------------
400    Name: Matthias Blume (blume (at) tti - c (dot) org)
401    Date: 2004/09/24 16:22:00 CDT
402    Tag: blume-20040924-ppc-long-branch
403    Description:
404    
405    Turned message about "emiting long form of branch" off by default.
406    Added a control flag to turn it back on when desired.
407    
408    ----------------------------------------------------------------------
409    Name: Matthias Blume (blume (at) tti - c (dot) org)
410    Date: 2004/09/24 16:05:00 CDT
411    Tag: blume-20040924-rounding
412    Description:
413    
414    Applied patch for setting rounding modes under Mac OS X.  Thanks to
415    Melissa O'Neill for providing the code!
416    
417    ----------------------------------------------------------------------
418    Name: Matthias Blume (blume (at) tti - c (dot) org)
419    Date: 2004/09/23 17:30:00 CDT
420    Tag: blume-20040923-envvars
421    Description:
422    
423    1. Changed definition of type ControlRegistry.registry_tree to
424       include control_info (i.e., the name of the controlling environment
425       variable).
426    
427    2. Added command-line flags -e and -E to print the names of environment
428       variables that can be used to control internal settings.
429       (This uses the new API mentioned in 1.)
430    
431    ----------------------------------------------------------------------
432    Name: Matthias Blume (blume (at) tti - c (dot) org)
433    Date: 2004/09/13 16:50:00 CDT
434    Tag: Release_110_49
435    Description:
436    
437    New working version (110.49).  NEW BOOTFILES!
438    
439    ----------------------------------------------------------------------
440    Name: Matthias Blume
441    Date: 2004/09/13 16:20:00 CDT
442    Tag: blume-20040913-config-mlrisc
443    Description:
444    
445    Put target "mlrisc" back into the default list.
446    (There is no harm in having it, and some users have expressed their
447    wish to have "mlrisc" included by default.)
448    
449    ----------------------------------------------------------------------
450    Name: John Reppy
451    Date: 2004/09/13
452    Tag: jhr-20040913-signals
453    Description:
454    
455    Fixed the signal masking code to properly nest mask/unmask operations
456    on a per-signal basis.
457    
458    ----------------------------------------------------------------------
459    Name: Matthias Blume (blume (at) tti - c (dot) org)
460    Date: 2004/09/08 13:20:00 CDT
461    Tag: blume-20040908-heap-magic
462    Description:
463    
464    Bumped the heap macig number to 0x09082004 to account for the changed
465    layout of the ML frame under MacOS X.
466    
467    ----------------------------------------------------------------------
468    Name: Allen Leung (leunga (at) reservoir (dot) com)
469    Date: 2004/09/03 11:26:00 EST
470    Tag: leunga-20040903-cygwin-install
471    Description:
472    
473       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
474    The Cygwin runtime is turned on by setting the environment
475    variable SMLNJ_CYGWIN_RUNTIME to 1.
476    
477    ----------------------------------------------------------------------
478    Name: Matthias Blume (blume (at) tti - c (dot) org)
479    Date: 2004/08/31 17:15:00 CDT
480    Tag: blume-20040831-core
481    Description:
482    
483    Added some exports to src/compiler/core.cm upon request by J. Joemann.
484    
485    ----------------------------------------------------------------------
486    Name: Matthias Blume (blume (at) tti - c (dot) org)
487    Date: 2004/08/30 17:55:00 CDT
488    Tag: blume-20040830-installer
489    Description:
490    
491    Upon request by Johannes Joemann:
492    
493    - improved ML code of installer to fall back to coping when renaming
494      fails (i.e., when source and target are on different file systems);
495      the code compiles but has yet to be tested in anger
496    - removed mlrisc from list of default targets (config/targets)
497    
498    ----------------------------------------------------------------------
499    Name: Matthias Blume (blume (at) tti - c (dot) org)
500    Date: 2004/08/27 17:20:00 CDT
501    Tag: blume-20040827-ptreql
502    Description:
503    
504    Added ptreql primop to structure InlineT (upon request from Larry
505    Paulson).
506    
507    ----------------------------------------------------------------------
508    Name: Allen Leung (leunga (at) reservoir (dot) com)
509    Date: 2004/08/15 21:21:00 EST
510    Tag: leunga-110_48-udgraph
511    Description:
512    
513    Another bug fix from Carl Hauser:
514    
515    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
516    > 48c48
517    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
518    > e::L)
519    > ---
520    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
521    > Without this, any deletion of an edge in an undirected graph does severe
522    > violence to the graph.
523    
524    ----------------------------------------------------------------------
525    Name: Allen Leung (leunga (at) reservoir (dot) com)
526    Date: 2004/08/10 23:35:00 EST
527    Tag: leunga-110_48-ppc
528    Description:
529    
530       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
531    
532    ----------------------------------------------------------------------
533    Name: Matthias Blume (blume (at) tti - c (dot) org)
534    Date: 2004/08/10 12:00:00 CDT
535    Tag: Release_110_48
536    Description:
537    
538    New working version (110.48).  NEW BOOTFILES!
539    
540    ----------------------------------------------------------------------
541    Name: Allen Leung (leunga (at) reservoir (dot) com)
542    Date: 2004/08/09 12:21:00 EST
543    Tag: leunga-110_47-dijsktra
544    Description:
545    
546    Bug fix from Carl Hauser:
547    
548    single_source_shortest_paths in dijkstra.sml was observed to get wrong
549    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
550    
551    The problem is that following the expression A.update(dist,s,Num.zero)
552    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
553    
554    ----------------------------------------------------------------------
555    Name: Matthias Blume (blume (at) tti - c (dot) org)
556    Date: 2004/08/06 18:10:00 CDT
557    Tag: blume-20040806-cmdline
558    Description:
559    
560    Fiddled with handling of command-line options:
561    
562      * sml now quits after processing the command line
563        if -H, -S, -h<n>, or -s<n> appears as the last
564        command-line argument
565      * a new option -q terminates the session when encountered on
566        the command line; subsequent arguments will be ignored
567      * bug fixes: short (erroneous) arguments are no longer ignored
568        completely
569    
570    ----------------------------------------------------------------------
571    Name: Allen Leung (leunga (at) reservoir (dot) com)
572    Date: 2004/08/04 18:17:00 EST
573    Tag: leunga-110_47-ppc-ibm-asm
574    Description:
575    
576      - Added minimal IBM assembly syntax support for PowerPC.
577    
578      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
579        been altered in the runtime.  We need an automatic way to keep the file
580        in sync.
581    
582    ----------------------------------------------------------------------
583    Name: Matthias Blume (blume (at) tti - c (dot) org)
584    Date: 2004/08/04 14:00:00 CDT
585    Tag: Release_110_47
586    Description:
587    
588    New working version (110.47).  NEW BOOTFILES!
589    
590    ----------------------------------------------------------------------
591    Name: Matthias Blume (blume (at) tti - c (dot) org)
592    Date: 2004/08/03 14:25:00 CDT
593    Tag: blume-20040803-callingconv
594    Description:
595    
596    Added low-level support for choosing C calling conventions by
597    twiddling the type of rawccall.  (See
598    src/compiler/Semant/types/cproto.sml for details.)
599    
600    ----------------------------------------------------------------------
601    Name: Matthias Blume (blume (at) tti - c (dot) org)
602    Date: 2004/08/02 15:55:00 CDT
603    Tag: blume-20040802-backout
604    Description:
605    
606    Backed out of change to win32-filesys.c.  The earlier patch to
607    get_file_time caused CM to produce files with the wrong time stamp.
608    
609    ----------------------------------------------------------------------
610    Name: Matthias Blume (blume (at) tti - c (dot) org)
611    Date: 2004/08/02 14:45:00 CDT
612    Tag: blume-20040802-nlffi-win32
613    Description:
614    
615    Added NLFFI support for Win32, adapted from a patch provided by David
616    Hansel.  This is currently completely untested.  Also, the issue
617    concerning stdcall vs. ccall is still unresolved.
618    
619    ----------------------------------------------------------------------
620    Name: Matthias Blume (blume (at) tti - c (dot) org)
621    Date: 2004/07/30 17:55:00 CDT
622    Tag: blume-20040730-various
623    Description:
624    
625    Gearing up towards 110.47...
626    
627    - various minor bugfixes to ml-nlffigen
628    - a beginning of a manual for nlffi
629    
630    - eliminated 'export name=value' in config/install.sh as this does
631      not work with certain versions of /bin/sh
632      (Thanks to David King at Motorola for catching this.)
633    
634    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
635      - added a test for tm==NULL to gmtime.c and localtime.c
636      - applied patch for incorrect GetFileTime under win32
637      - toSeconds -> toMilliseconds in Win32/win32-process.sml
638    
639    ----------------------------------------------------------------------
640    Name: Matthias Blume (blume (at) tti - c (dot) org)
641    Date: 2004/07/21 18:20:00 CDT
642    Tag: blume-20040721-nlffigen
643    Description:
644    
645    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
646      for a typedef to an incomplete type, but leave out the "typ" member.
647      (This is just for consistency.)
648    - Started to produce what is supposed to become better (i.e., comprehensive)
649      documentation of what ml-nlffigen does and produces.
650    
651    ----------------------------------------------------------------------
652    Name: Matthias Blume (blume (at) tti - c (dot) org)
653    Date: 2004/07/14 16:25:00 CDT
654    Tag: blume-20040714-union
655    Description:
656    
657    Added C_UNION to c-calls/c-types.sml and updated the machinery
658    (ml-nlffigen, cproto.sml) that conveys C function interface
659    information to the code generator.
660    
661    However, the actual architecture-specific implementation of function
662    arguments and results that are C unions is still not implemented.
663    
664    ----------------------------------------------------------------------
665    Name: Allen Leung (leunga (at) reservoir (dot) com)
666    Date: 2004/07/14 14:38:00 EST
667    Tag: leunga-110_46_1-ppc-lwzu
668    Description:
669    
670        Added these instructions to the PowerPC architecture:
671        LBZU(X), LHZU(X), LWZU(X),
672        STWU(X), STFDU, STFSU
673    
674        etc...
675    
676        Note: I haven't added their instruction encoding into the description.
677    
678    ----------------------------------------------------------------------
679    Name: Allen Leung (leunga (at) reservoir (dot) com)
680    Date: 2004/07/13 15:04:00 EST
681    Tag: leunga-110_46_1-ppc-lwarx
682    Description:
683    
684       Added the two instructions LWARX and STWCX to the PowerPC
685    instruction set.
686    
687       A (untested) rewrite of loop-structure.sml.  The old version
688    is completely broken.
689    
690    ----------------------------------------------------------------------
691    Name: Matthias Blume (blume (at) tti - c (dot) org)
692    Date: 2004/07/13 13:50:00 CDT
693    Tag: blume-20040713-nlffi
694    Description:
695    
696    - use paramAlloc to report c-calls with too many arguments
697      (for PPC version where parameter area is pre-allocated)
698    - added ccall_maxargspace to machspec (to implement the above)
699    - made "make" commend in CM's "make" tool configurable
700    - added option (default: on) for passing the name of the SML/NJ's "bin"
701      directory to "make"; the call looks like this:
702    
703         make <options> SMLNJ_BINDIR=<dir> <target>
704    
705      This can be used by the Makefile to, e.g., pick the "right" version
706      of ml-nlffigen.
707    - minor code tweaks
708    
709    ----------------------------------------------------------------------
710    Name: Matthias Blume (blume (at) tti - c (dot) org)
711    Date: 2004/07/12 22:50:00 CDT
712    Tag: blume-110_46_1-macosx-nlffi
713    Description:
714    
715    NLFFI under Mac OS X now working (sort of).  This is largely untested,
716    though.
717    
718    Note:  1. You have to make a new, clean build of the runtime system.
719           2. There are new BOOTFILES, you have to use them!
720              (Doing the bootstrap process yourself would be *very* painful!
721               If you absolutely have to do it, build the system under
722               a different architecture and then cross-compile.)
723    
724    Version bumped to 110.46.1 to account for runtime data format changes.
725    
726    ----------------------------------------------------------------------
727    Name: Matthias Blume (blume (at) tti - c (dot) org)
728    Date: 2004/06/18 14:30:00 CDT
729    Tag: blume-20040618-unix
730    Description:
731    
732    Changed the implementation of structure Unix so that the same stream
733    is returned every time one of the {text,bin}{In,Out}streamOf functions
734    is invoked on the same proc.  This is not what the spec currently
735    says -- although IMO it arguably should.  (See discussion below.)
736    
737    ----------------------------------------------------------------------
738    Name: Matthias Blume (blume (at) tti - c (dot) org)
739    Date: 2004/06/17 18:15:00 CDT
740    Tag: Release_110_46
741    Description:
742    
743    New working version (110.46).  NEW BOOTFILES!
744    
745    ----------------------------------------------------------------------
746    Name: Matthias Blume (blume (at) tti - c (dot) org)
747    Date: 2004/06/17 17:20:00 CDT
748    Tag: blume-20040617-timer-unix
749    Description:
750    
751    Changed the interface of structures Timer and Unix to match the most
752    recent Basis spec.
753    
754    In the case of Unix there still seems to be an open/weird issue:
755    
756       The {text,bin}{In,Out}streamOf functions are supposed to create
757       fresh streams whenever they are called -- as opposed to have them
758       return the same stream every time.  This design is supposed to
759       prevent space leaks caused by proc values hanging on to streams.
760    
761       The reap function, on the other hand, is supposed to close the
762       streams.  This cannot be done without having a handle on the
763       stream in proc after all...
764    
765    I took the liberty to implement the following stopgap solution:
766    
767      The proc value hangs on to the most recently created stream(s).
768      Reap closes those.  If either or both of the two streams hadn't
769      been created at all yet, then reap will close the corresponding
770      file descriptors directly.
771    
772    PS: I don't understand the original space leak argument anymore.  If
773    a proc hangs on to the imperative stream, then I/O operations on those
774    will advance the state of the cached stream and avoid the space leak.
775    
776    ----------------------------------------------------------------------
777    Name: Matthias Blume (blume (at) tti - c (dot) org)
778    Date: 2004/05/28 16:45:00 CDT
779    Tag: blume-20040528-basis
780    Description:
781    
782    Added signature PACK_REAL and exported functor PrimIO.
783    
784    ----------------------------------------------------------------------
785    Name: Matthias Blume (blume (at) tti - c (dot) org)
786    Date: 2004/05/25 16:00:00 CDT
787    Tag: blume-20040525-group-owner
788    Description:
789    
790    CM now ignores (but still accepts) the "owner" information in group
791    descriptions.  The owner of a group is its next enclosing
792    library. Each group must have a unique owner.  (There is a virtual
793    "toplevel" library that owns groups which are not nested within a real
794    library.)  Previously, each group had to explicitly declare its owner,
795    and CM would check that such a declaration is correct.  The new scheme
796    is to have CM check that for each group there is precisely one owning
797    library.
798    
799    The advantage of the new scheme is that the programmer no longer needs
800    to maintain the somewhat annoying owner information.  The downside is
801    that CM cannot enforce the ownership rule across multiple runs of
802    CM.make.  Fortunately, enclosing the same group in two different
803    libraries A and B which are not part of the same program does not
804    cause real problems.
805    
806    ----------------------------------------------------------------------
807    Name: Matthias Blume (blume (at) tti - c (dot) org)
808    Date: 2004/05/20 16:00:00 CDT
809    Tag: blume-20040520-win32
810    Description:
811    
812    Made the win32 version work again.  (Strangely, a misplaced comma had
813    slipped into win32-process.c which prevented the runtime from being
814    compiled correctly.)
815    
816    Also, included a minor addition to ml-build.bat analogous to what was
817    done in blume-20040519-ml-build.
818    
819    ----------------------------------------------------------------------
820    Name: Matthias Blume (blume (at) tti - c (dot) org)
821    Date: 2004/05/19 22:10:00 CDT
822    Tag: blume-20040519-ml-build
823    Description:
824    
825    Arranged for ml-build to clean up after itself a little bit better.
826    The script generates a temporary SML source file and compiles it using
827    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
828    gets rid of those at the end, so they don't accumulate under .cm.
829    
830    This required a minor change to install.sh because the name of the
831    metadata directory (default: .cm) is actually configurable at
832    installation time.
833    
834    ----------------------------------------------------------------------
835    Name: Matthias Blume (blume (at) tti - c (dot) org)
836    Date: 2004/05/18 15:50:00 CDT
837    Tag: blume-20040518-mkreader
838    Description:
839    
840    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
841    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
842    
843    ----------------------------------------------------------------------
844    Name: Matthias Blume (blume (at) tti - c (dot) org)
845    Date: 2004/05/11 14:35:00 CDT
846    Tag: blume-20040511-win32sock
847    Description:
848    
849    Added previously missing support for many socket-related functions
850    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
851    for the voluminous patch!
852    
853    (I have not tested this patch under win32 yet.)
854    
855    Here is David's e-mail:
856    
857    Hi,
858    
859    Attached to this email you find a diff against sml/nj 110.45
860    that will enable socket support under Windows.
861    
862    To apply the patch (using unix or cygwin)
863    1) gunzip runtime.diff.gz
864    2) "cd" into "src/runtime" in the source tree of a fresh
865       110.45 installation.
866    3) patch -p 1 < [your/path/to]runtime.diff
867    
868    The code compiles fine but has NOT yet been extensively tested.
869    I only ran a few tests for basic socket client functionality
870    (which worked fine).  Especially the functions that use ioctl
871    are not tested at all and might not work (see below).
872    
873    I implemented this since we want to move to a newer version of sml/nj
874    but need socket support in order to use it.  This is the first time I
875    even had a look at the sml/nj source,  so please review my changes
876    before making this part of the distribution!  Here are a few issues
877    that I think might be better for someone to solve who is more
878    familiar with the sml/nj source (and socket programming):
879    
880    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
881      exception since I could not figure out what the windows equivalent
882      of these functions is
883    
884    - In sockets-osdep.h there are a some #include statements that are
885      only used in a few files that include sockets-osdep.h
886    
887    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
888      does not process its return value since I don't know how to
889      report an error upwards.
890    
891    - It would probably be good to have a call to WSACleanup() when
892      the library is unloaded (if there is such a possibility).
893      Otherwise I think Windows will take care of this automatically
894      when the process finishes.
895    
896    - I used ioctlsocket() as a replacement for ioctl() but I have
897      no idea if that is actually the proper replacement on Windows.
898    
899    - All these issues are marked in the code by "FIXME" comments.
900    
901    We use sml/nj extensively in our products and are quite happy
902    with it.  I hope this contribution will help you.
903    
904    Keep up the good work!
905    
906    David
907    
908    ----------------------------------------------------------------------
909    Name: Matthias Blume (blume (at) tti - c (dot) org)
910    Date: 2004/05/11 14:20:00 CDT
911    Tag: blume-20040511-installml
912    Description:
913    
914    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
915    the patch.)
916    
917    ----------------------------------------------------------------------
918    Name: Matthias Blume (blume (at) tti - c (dot) org)
919    Date: 2004/05/11 14:05:00 CDT
920    Tag: blume-20040511-nlffi-netbsd
921    Description:
922    
923    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
924    the patch.)
925    
926    ----------------------------------------------------------------------
927    Name: Matthias Blume (blume (at) tti - c (dot) org)
928    Date: 2004/05/11 12:05:00 CDT
929    Tag: blume-20040511-exports
930    Description:
931    
932    As per request by Adam Chlipala <adam@hcoop.net>, extended various
933    export lists in compiler-related .cm-files.
934    
935    ----------------------------------------------------------------------
936    Name: Matthias Blume (blume (at) tti - c (dot) org)
937    Date: 2004/05/11 11:35:00 CDT
938    Tag: blume-20040511-allsource
939    Description:
940    
941    The installer now honors the "src-smlnj" target again, although its meaning
942    has changed from "all sources required for the compiler" to "all sources
943    the installer knows about".  In other words, if you enable "src-smlnj"
944    in the "targets" file, then the installer will pull in sources for
945    everything.  (Notice that this refers to source code only.  Compiled
946    code is still only installed for modules that were requested explicitly
947    or which are required for other modules that were requested explicitly.)
948    
949    ----------------------------------------------------------------------
950    Name: Matthias Blume (blume (at) tti - c (dot) org)
951    Date: 2004/04/23 17:40:00 CDT
952    Tag: blume-20040423-ieee-scan
953    Description:
954    
955    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
956    in the exponent calculation we get INF or ZERO (depending on the mantissa
957    and the sign of the exponent).
958    
959    ----------------------------------------------------------------------
960    Name: Matthias Blume (blume (at) tti - c (dot) org)
961    Date: 2004/04/23 10:40:00 CDT
962    Tag: blume-20040423-ml-build
963    Description:
964    
965    The ml-build script now terminates with a non-0 status when something
966    goes wrong.
967    
968    ----------------------------------------------------------------------
969    Name: Matthias Blume (blume (at) tti - c (dot) org)
970    Date: 2004/04/22 16:35:00 CDT
971    Tag: blume-20040422-Option
972    Description:
973    
974    Made exception Option to be the same as exception Option.Option
975    (as it should be).
976    
977    ----------------------------------------------------------------------
978    Name: Allen Leung (leunga (at) reservoir (dot) com)
979    Date: 2004/03/19 14:40:00 EST
980    Tag: leunga-20040319-cygwin-nlffi
981    Description:
982    
983    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
984    of SML/NJ.  The problem is that
985    
986       lib = dlopen(NULL, ...)
987       f   = dlsym(lib, "malloc");
988    
989    does not work on Windows unless we explicitly export symbols
990    such as 'malloc' during linking.  We fixed this by explicitly
991    exporting the required symbols with the magic gcc incantation:
992    
993        -Wl,--export-all cygwin.def
994    
995    where cygwin.def is a file containing all the symbols that
996    we wish to export.
997    
998    I suspect this is a Windows problem and we'll have to
999    do the same (somehow with windows compilers) when
1000    we build the native win32 version with the system
1001    calls LoadLibrary/GetProcAddress.
1002    
1003    ----------------------------------------------------------------------
1004    Name: Matthias Blume (blume (at) tti - c (dot) org)
1005    Date: 2004/03/04 16:35:00 CST
1006    Tag: blume-20040304-intinf-fmt
1007    Description:
1008    
1009    Fixed problem with IntInf.fmt (sign would show up on the right instead
1010    of on the left for BIN, OCT, and HEX).
1011    
1012    ----------------------------------------------------------------------
1013    Name: Matthias Blume (blume (at) tti - c (dot) org)
1014    Date: 2004/03/04 11:25:00 CST
1015    Tag: blume-20040304-symlinks
1016    Description:
1017    
1018    Fixed problem with installer script (unix only) where bin/ml-yacc and
1019    friends pointed (via symlinks) to absolute locations instead of just
1020    .run-sml.  This was reported by Vesa A Norrman.
1021    
1022    ----------------------------------------------------------------------
1023    Name: Matthias Blume (blume (at) tti - c (dot) org)
1024    Date: 2004/02/13 14:50:00 CST
1025    Tag: Release_110_45
1026    Description:
1027    
1028    New working version (110.45). New bootfiles.
1029    
1030    ----------------------------------------------------------------------
1031    Name: Matthias Blume (blume (at) tti - c (dot) org)
1032    Date: 2004/01/26 15:15:15 CST
1033    Tag: blume-20040126-toplevel
1034    Description:
1035    
1036    Improved handling of exceptions at the interactive toplevel.
1037    
1038    ----------------------------------------------------------------------
1039    Name: Matthias Blume (blume (at) tti - c (dot) org)
1040    Date: 2004/01/26 11:25:00
1041    Tag: blume-20040126-app
1042    Description:
1043    
1044    Type of top-level "app" corrected.
1045    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1046    
1047    ----------------------------------------------------------------------
1048    Name: Matthias Blume (blume (at) tti - c (dot) org)
1049    Date: 2003/11/18 17:10 CST
1050    Tag: blume-20031118-basis-fiddle
1051    Description:
1052    
1053    - changed Timer interface to what might become the spec
1054    - POSIX_FLAGS -> BIT_FLAGS according to spec
1055    - some other minor discrepancies wrt. spec eliminated
1056    
1057    ----------------------------------------------------------------------
1058    Name: Matthias Blume (blume (at) tti - c (dot) org)
1059    Date: 2003/11/06 12:00:00 CST
1060    Tag: Release_110_44
1061    Description:
1062    
1063    New working version (110.44). New bootfiles.
1064    
1065    ----------------------------------------------------------------------
1066    Name: Matthias Blume (blume (at) tti - c (dot) org)
1067    Date: 2003/11/04 11:50:00 CST
1068    Tag: blume-20031104-move-libraries
1069    Description:
1070    
1071    Eliminated the "dont_move_libraries" directive in config/targets.
1072    (The mechanism was broken and could not be fixed easily.  Moreover,
1073    there does not seem to be any reason not to move all libraries into
1074    lib during installation.  I originally implemented this directive as a
1075    backward-compatibility feature when I first introduced the new CM.
1076    Now that things have been stable for a long time and going back to the
1077    old CM is not an option, there is no reason to keep it around.)
1078    
1079    ----------------------------------------------------------------------
1080    Name: Matthias Blume (blume (at) tti - c (dot) org)
1081    Date: 2003/11/03 16:00:00 CST
1082    Tag: blume-20031103-installdir
1083    Description:
1084    
1085    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1086    Richards for pointing out the problem and providing the solution.)
1087    
1088    ----------------------------------------------------------------------
1089    Name: Matthias Blume (blume (at) tti - c (dot) org)
1090    Date: 2003/10/01 17:05:00 CDT
1091    Tag: blume-20031001-lal-mlrisc
1092    Description:
1093    
1094    MLRISC bug fix from Lal.
1095    
1096    ----------------------------------------------------------------------
1097    Name: Matthias Blume (blume (at) tti - c (dot) org)
1098    Date: 2003/09/30 16:10:00 CDT
1099    Tag: blume-20030930-primio-bat
1100    Description:
1101    
1102    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1103    2. Improved .bat files (for Win32 port) to make things work under Win95.
1104       (thanks to Aaron S. Hawley for this one)
1105    
1106    ----------------------------------------------------------------------
1107    Name: Matthias Blume (blume (at) tti - c (dot) org)
1108    Date: 2003/09/26 16:05:00 CDT
1109    Tag: blume-20030926-wrappriv
1110    Description:
1111    
1112    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1113    
1114    ----------------------------------------------------------------------
1115    Name: Matthias Blume (blume (at) tti - c (dot) org)
1116    Date: 2003/09/26 15:00:00 CDT
1117    Tag: blume-20030926-110_43_3
1118    Description:
1119    
1120    - additional cleanup
1121    - version number bump, NEW BOOTFILES
1122    
1123    ----------------------------------------------------------------------
1124    Name: Matthias Blume (blume (at) tti - c (dot) org)
1125    Date: 2003/09/26 12:00:00 CDT
1126    Tag: blume-20030926-ppautoload
1127    Description:
1128    
1129    I modified the read-eval-print loop so that the autoloader gets
1130    invoked whenever the prettyprinter tries to look up a symbol that
1131    is not currently defined in the toplevel environment but which
1132    appears in CM's autoload registry.  As a result, we see far fewer of
1133    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1134    
1135    In addition to this I tried to clean up some pieces of the Basis
1136    implementation (e.g., Socket, Word8Array) in order to prevent other
1137    instances of these ?.Foo.Bar.xxx names from being printed.
1138    
1139    The mechanism that picks names for types still needs some work, though.
1140    (Right now it seems that if there is a type A.t which is defined to
1141    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1142    "?.B.u" although the perhaps more sensible solution would be to use
1143    "A.t" in this case.  In other words, the prettyprinter should follow
1144    a chain of DEFtycs not farther than there are corresponding toplevel
1145    names in the current environment.)
1146    
1147    ----------------------------------------------------------------------
1148    Name: Matthias Blume (blume (at) tti - c (dot) org)
1149    Date: 2003/09/24 16:31:00 CDT
1150    Tag: blume-20030924-installer
1151    Description:
1152    
1153    Another installer tweak:  All the ML code for the installer is now
1154    compiled during CMB.make and put into a little library called
1155    $smlnj/installer.cm.  The installation then simply invokes
1156    
1157       sml -m $smlnj/installer.cm
1158    
1159    and everything happens automagically.
1160    
1161    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1162    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1163            and BINDIR.
1164    
1165    The new scheme guarantees that the ML code responsible for the installation
1166    is in sync with the APIs of the main system.  Also, the installer is
1167    somewhat faster because the installer script is precompiled.
1168    
1169    ----------------------------------------------------------------------
1170    Name: Matthias Blume (blume (at) tti - c (dot) org)
1171    Date: 2003/09/24 15:35:00 CDT
1172    Tag: blume-20030924-synsock
1173    Description:
1174    
1175    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1176    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1177    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1178    implementations of non-blocking operations from CML's Socket
1179    structure.
1180    
1181    ----------------------------------------------------------------------
1182    Name: Matthias Blume (blume (at) tti - c (dot) org)
1183    Date: 2003/09/24 15:10:05 CDT
1184    Tag: blume-20030924-sockets
1185    Description:
1186    
1187    1. Fixed SOCKET API and implementation to match Basis spec.
1188       This required changing the internal representation of sockets to one
1189       that remembers (for each socket file descriptor) whether it is currently
1190       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1191       call is made only if the state actually needs to change).
1192    
1193    2. OS-specific details of sockets were moved into separate files, thus
1194       making it possible to unify the bulk of the socket implementations
1195       between Unix and Win32.
1196    
1197    3. CML's socket API changed accordingly.
1198       (Note that we need to remove non-blocking functions from this API
1199       since they are redundant in the case of CML!)
1200    
1201    4. CML's socket implementation now makes use of non-blocking functions
1202       provided by Basis, thus removing all OS-dependent code from this part
1203       of CML.
1204    
1205    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1206    
1207    ----------------------------------------------------------------------
1208    Name: Matthias Blume (blume (at) tti - c (dot) org)
1209    Date: 2003/09/22 12:10:00 CDT
1210    Tag: blume-20030922-110_43_2
1211    Description:
1212    
1213    Made a new interim version and bootfiles for developer's bootstrapping
1214    convenience.
1215    
1216       110.43.2 -- NEW BOOTFILES
1217    
1218    ----------------------------------------------------------------------
1219    Name: Matthias Blume (blume (at) tti - c (dot) org)
1220    Date: 2003/09/19 15:55:00 CDT
1221    Tag: blume-20030919-cmdir
1222    Description:
1223    
1224    1. new-install.sh -> install.sh
1225    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1226    3. tweaked installer so that another name instead of .cm can be chosen
1227       at install time (by setting the CM_DIR_ARC environment variable
1228       during installation); once installation is complete, the name is
1229       fixed
1230    
1231    ----------------------------------------------------------------------
1232    Name: Matthias Blume (blume (at) tti - c (dot) org)
1233    Date: 2003/09/18 16:00:00 CDT
1234    Tag: blume-20030918-110_43_1
1235    Description:
1236    
1237    Made a new interim version and bootfiles for developer's bootstrapping
1238    convenience.
1239    
1240       110.43.1 -- NEW BOOTFILES
1241    
1242    ----------------------------------------------------------------------
1243    Name: Matthias Blume (blume (at) tti - c (dot) org)
1244    Date: 2003/09/18 15:20:00 CDT
1245    Tag: blume-20030918-misc
1246    Description:
1247    
1248    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1249       this seems to be controversial at the moment) and used those in
1250       Posix.ProcEnv.times.
1251    
1252    2. Added Time.{from,to}Nanoseconds to Time.
1253    
1254    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1255       For example, fromLargeInt never needs to look at more than 3 "big
1256       digits" to get its 53 bits of precision.
1257    
1258    ----------------------------------------------------------------------
1259    Name: Matthias Blume (blume (at) tti - c (dot) org)
1260    Date: 2003/09/17 16:30:00 CDT
1261    Tag: blume-20030917-real32-slices
1262    Description:
1263    
1264    Added an entry to the primitive environment
1265    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1266    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1267    
1268    Removed some of the "magic" constants in real64.sml and replaced them
1269    with code that generates these values from their corresponding
1270    integer counterparts.
1271    
1272    Made all(?) the slice-related changes to the Basis and made everything
1273    compile again...
1274    
1275    ----------------------------------------------------------------------
1276    Name: Matthias Blume (blume (at) tti - c (dot) org)
1277    Date: 2003/09/15 17:45:00 CDT
1278    Tag: blume-20030915-rbase
1279    Description:
1280    
1281    Fixed bug in Real.fromLargeInt.
1282    
1283    ----------------------------------------------------------------------
1284    Name: Matthias Blume (blume (at) tti - c (dot) org)
1285    Date: 2003/09/13 18:11:00 CDT
1286    Tag: blume-20030913-libinstall
1287    Description:
1288    
1289    Minor bugfix in config/libinstall (set anchor with path to
1290    standalone tool after installing it, otherwise libraries that
1291    need ml-lex or ml-yacc won't compile the first time the installer
1292    runs).
1293    
1294    ----------------------------------------------------------------------
1295    Name: Matthias Blume (blume (at) tti - c (dot) org)
1296    Date: 2003/09/12 11:45:00 CDT
1297    Tag: blume-20030912-various
1298    Description:
1299    
1300     - fixed bug in Real.toLargeInt
1301     - fixed bug in Posix.ProcEnv.times
1302     - changed inputLine functions to return an option
1303     - minor installer improvements / bugfixes
1304     - changed default @SMLalloc parameter for x86/celeron to 64k
1305    
1306    ----------------------------------------------------------------------
1307    Name: Matthias Blume (blume (at) tti - c (dot) org)
1308    Date: 2003/09/09 22:00:00 CDT
1309    Tag: Release_110_43
1310    Description:
1311    
1312        New working release 110.43.  New bootfiles.
1313    
1314    ----------------------------------------------------------------------
1315    Name: Matthias Blume (blume (at) tti - c (dot) org)
1316    Date: 2003/09/09 19:20:00 CDT
1317    Tag: blume-20030909-installer
1318    Description:
1319    
1320    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1321    Modified config/install.bat to take advantage of it.  Also modified
1322    config/install.sh (and called it config/new-install.sh) to take advantage
1323    of it on Unix systems.  (The SML code is (supposed to be) platform-
1324    independent.)
1325    
1326    The installer can now install everything under Win32
1327    as well as under *nix as long as it compiles.
1328    
1329    Other changes:
1330    
1331      - made CML compile again under Win32
1332      - made eXene compile under Win32 (by providing a fake structure UnixSock
1333        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1334      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1335        same as type int;  under Win32 it isn't
1336      - fixed some slice-related problems in the win32-specific parts of CML
1337      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1338        (under Win32, the volume name is case-insensitive, and the
1339        OS.Path code compares volume names for equality)
1340    
1341    ----------------------------------------------------------------------
1342    Name: Matthias Blume (blume (at) tti - c (dot) org)
1343    Date: 2003/09/08 11:55:00 CDT
1344    Tag: blume-20030908-fullpath
1345    Description:
1346    
1347    Made Win32 version of OS.FileSys.fullPath return current directory
1348    when given an empty string.  This is what the spec says, and incidentally,
1349    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1350    cases when presented with the name of a non-existing .cm file.)
1351    
1352    ----------------------------------------------------------------------
1353    Name: Matthias Blume (blume (at) tti - c (dot) org)
1354    Date: 2003/09/04 16:30:00 CDT
1355    Tag: blume-20030905-slices-etc
1356    Description:
1357    
1358      1. Changed interface to vectors and arrays in Basis to match
1359         (draft) Basis spec.
1360      2. Added signatures and implementations of slices according to
1361         Basis spec.
1362      3. Edited source code throughout the system to make it compile again
1363         under 1. and 2.  (In some cases code had to be added to have it
1364         match the new signatures.)
1365      4. MLRISC should be backward-compatible: the copies of the originals
1366         of files that needed to change under 3. were retained, the .cm files
1367         check the compiler version number and use old versions when
1368         appropriate.
1369      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1370         dirstream -> string option (in accordance with Basis spec).
1371      6. When generating code that counts lines, ml-lex used function
1372         CharVector.foldli, taking advantage of its old interface.
1373         This has been replaced with the corresponding code from
1374         CharVectorSlice.  (html-lex must be re-lexed!)
1375      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1376         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1377      8. Removed temporary additions (fromInternal, toInternal) from the
1378         (now obsolete) IntInf in smlnj-lib/Util.
1379      9. Cleaned up structure Byte.
1380     10. Added localOffset, scan, and fromString to Date (according to spec).
1381         Cleaned/corrected implementation of Date.
1382         (Still need to check for correctness; implement better canonicalizeDate.)
1383     11. Added "scan" to signature IEEE_REAL.
1384     12. Some improvements to IntInf [in particular: efficiency-hack for
1385         mod and rem when second operand is 2 (for parity checks).]
1386     13. Changed representation of type Time.time, using a single IntInf.int
1387         value counting microseconds.  This considerably simplified the
1388         implementation of structure Time.  We now support negative time
1389         values; scan and fromString handle signs.
1390     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1391         ArraySlice).
1392    
1393    ----------------------------------------------------------------------
1394    Name: Matthias Blume (blume (at) tti - c (dot) org)
1395    Date: 2003/08/28 17:00:00 CDT
1396    Tag: blume-20030828-intinf
1397    Description:
1398    
1399    This is a major update which comes with a version number bump
1400    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1401    and an implementation of IntInf in the Basis.
1402    
1403    There are a fairly large number of related changes and updates throughout
1404    the system:
1405    
1406    Basis:
1407            - Implemented IntInf.
1408            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1409            - Added some missing Real64 operations, most notably Real.toLargeInt.
1410            - Added FixedInt as a synonym for Int32.
1411    
1412    compiler:
1413            * Added support for a built-in intinf type.
1414              - literals
1415              - pattern matching
1416              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1417              - overloading on literals and operations
1418    
1419            This required adding a primitive type intinf, some additional
1420            primops, and implementations for several non-trivial intinf
1421            operations in Core.  (The intinf type is completely abstract
1422            to the compiler; all operations get delegated back to the Core.)
1423    
1424            * Intinf equality is handled by polyequal.  However, the compiler
1425            does not print its usual warning in this case (since polyequal
1426            is the right thing to do there).
1427    
1428            * Improved the organization of structure InlineT.
1429    
1430            * A word about conversion primops:
1431            If conversions involving intinf do not cancel out during
1432            CPS contract, then the compiler must insert calls to Core functions.
1433            Since all core access must be resolved already during the FLINT
1434            translate phase, it would be too late a the time of CPS contract
1435            to add new Core calls.  For this reason, conversion primops
1436            for intinf carry two arguments:  1. the numeric argument that
1437            they are supposed to convert, and 2. the Core function that
1438            can help with this conversion if necessary.  If CPS contract
1439            eliminates a primop, then the associated Core function becomes
1440            dead and goes away.  Intinf conversion primops that do not get
1441            eliminated by CPS contract get rewritten into calls of their
1442            core functions by a separate, new phase.
1443    
1444    interactive system:
1445            - Control.Print.intinfDepth controls max length of intinf constants
1446            being printed. (Analogous to Control.Print.stringDepth.)
1447            - Cleanup in printutil and pputil: got rid of unused stuff and
1448            duplicates; replaced some of the code with code that makes better
1449            use of library functionality.
1450    
1451    CM:
1452            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1453            are no longer silent.
1454    
1455    CKIT:
1456            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1457            in favor of LargeInt -- which is now the same as IntInf.
1458            CKIT-knowledgable people should check whether this is what's
1459            intended and otherwise change things back to using Int32 or
1460            FixedInt.
1461    
1462    Throughout the code:
1463            Started using IntInf.int literals and built-in operations
1464            (e.g., comparison with 0) where this seems appropriate.
1465    
1466    
1467    ----------------------------------------------------------------------
1468    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1469    Date: 2003/08/13 11:36:00 CDT
1470    Tag: dbm-20030813-mcz-merge1
1471    Description:
1472    
1473    Merging changes from the mcz-branch development branch into trunk.
1474    These changes involve replacement of the emulated old prettyprinter
1475    interface with direct use of the SML/NJ Lib PP library, and fixing
1476    of a couple of bugs (895, 1186) relating to error messages.  A new
1477    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1478    has been added.
1479    
1480    ----------------------------------------------------------------------
1481    Name: Matthias Blume (blume (at) tti - c (dot) org)
1482    Date: 2003/08/11 15:45:00 CDT
1483    Tag: blume-20030811-windows
1484    Description:
1485    
1486    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1487    
1488        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1489    
1490    This patch restores SML/NJ's ability to run under win32.  There are a
1491    number of changes, including fixes for several bugs that had gone
1492    unnoticed until now:
1493    
1494      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1495        trying to run the win32 version from within cygwin.)
1496    
1497      - There are a number of simple .bat scripts that substitute for their
1498        corresponding Unix shell-scripts.  (See below.)
1499    
1500      - The internals of ml-build have been modified slightly.  The main
1501        difference is that instead of calling ".link-sml" (or link-sml.bat)
1502        using OS.Process.system, the ML process delegates this task back
1503        to the script.  Otherwise problems arise in mixed environments such
1504        as Cygwin where scripts look and work like Unix scripts, but
1505        where OS.Process.system cannot run them.
1506    
1507      - In CM, the srcpath pickler used native pathname syntax -- which
1508        is incorrect in the case of cross-compilation.  The new pickle format
1509        is independent of platform-specific naming conventions.
1510    
1511      - Path configuration files (such as lib/pathconfig) can now choose
1512        between native and standard syntax.  Placing a line of the form
1513    
1514            standard!
1515    
1516        into the file causes all subsequent paths to be interpreted using
1517        CM standard pathname syntax (= Unix conventions); a line
1518    
1519            native!
1520    
1521        switches back to native style.  This was needed so that
1522        path config files can be written portably, see src/system/pathconfig.
1523    
1524      - Runtime system:
1525    
1526          - win32-filesys.c:  get_file_time and set_file_time now
1527            access modification time, not creation time.
1528    
1529          - I/O code made aware of new array representation.
1530    
1531          - Bug fixes in X86.prim.masm.
1532    
1533       - src/system/makeml made aware of win32. (For use under cygwin
1534            and other Unix-environments for windows.)
1535    
1536       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1537         which caused BinIO.inputAll to fail consistently.
1538    
1539    .bat scripts:
1540    
1541       Windows .bat scripts assume that SMLNJ_HOME is defined.
1542    
1543       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1544         applications (sml, ml-yacc, ml-lex).
1545       - ml-build.bat: analogous to ml-build.
1546       - config\install.bat:  Analogous to config/install.sh.  This requires
1547         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1548         (nmake etc. must be on the path, and vcvars32 must have been run.)
1549         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1550         and the bootfile hierarchy must have been unpacked under
1551         sml.boot.x86-win32.
1552         The script is very primitive and does a poor job at error checking.
1553         It only installs the base system, ml-lex, and ml-yacc.  No other
1554         libraries are being installed (i.e., you get only those that
1555         are part of the compiler.)
1556       - link-sml.bat:  analogous to .link-sml, but not currently used
1557    
1558    Unrelated bug fixes:
1559    
1560      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1561        types.
1562      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1563        otherwise.)
1564    
1565    ----------------------------------------------------------------------
1566    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1567    Date: 2003/06/17
1568    Tag: macqueen-20030617-bug895
1569    Description:
1570    
1571    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1572    Tag will be used for new development branch (mcz-branch) for
1573    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1574    
1575    ----------------------------------------------------------------------
1576    Name: Matthias Blume (blume (at) tti - c (dot) org)
1577    Date: 2003/05/27 16:55:00 CDT
1578    Tag: blume-20030527-polyeq
1579    Description:
1580    
1581    Tried to eliminated most cases of polymorphic equality.
1582    
1583    ----------------------------------------------------------------------
1584    Name: Matthias Blume (blume (at) tti - c (dot) org)
1585    Date: 2003/05/21 17:45:00 CDT
1586    Tag: blume-20030517-complete
1587    Description:
1588    
1589    Two changes:
1590    
1591      1. Added a flag for controlling whether non-exhaustive bindings will
1592         be treated as errors (default is false).
1593      2. Cleaned up the *entire* source tree so that CMB.make goes through
1594         without a single non-exhaustive match- or bind warning.
1595    
1596    ----------------------------------------------------------------------
1597    Name: Matthias Blume (blume (at) tti - c (dot) org)
1598    Date: 2003/05/17 10:20:00 CDT
1599    Tag: blume-20030517-absyn
1600    Description:
1601    
1602    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1603    
1604       This mainly affects the quality of error messages.  However, some
1605       of the code is now more straightforward than before. (Treatment of
1606       the above four constructs in translate.sml is much simpler than
1607       the "macro-expansion" that was going on before.  Plus, the mach-
1608       compiler no longer gets invoked just to be able to compile an
1609       if-expression.)
1610    
1611    2. The ErrorMsg.Error exception is now caught and absorbed by the
1612       interactive loop.
1613    
1614    ----------------------------------------------------------------------
1615    Name: Allen Leung
1616    Date: 2003/05/16 13:05:00 CDT
1617    Tag: leunga-20030516-cygwin-runtime
1618    Description:
1619    
1620        Ported the runtime system to cygwin, which uses the unix
1621    x86-unix bin files.  Missing/buggy features:
1622    
1623       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1624         the Cygwin library.
1625       o Ctrl-C handling may be flaky.
1626       o Windows system calls and Windows I/O are not supported.
1627    
1628       A new set of binfiles is located at:
1629    
1630         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1631    
1632       This is only needed for bootstrapping the cygwin version of smlnj.
1633       Other x86 versions can use the existing binfiles.
1634    
1635    ----------------------------------------------------------------------
1636    Name: Matthias Blume
1637    Date: 2003/04/08 15:42:00 CDT
1638    Tag: blume-20030408-listpair
1639    Description:
1640    
1641    1. Added a target 'mlrisc' to installer.
1642    
1643    2. Added missing elements to structure ListPair.
1644    
1645    ----------------------------------------------------------------------
1646    Name: Allen Leung
1647    Date: 2003/01/07 10:40:00 EST
1648    Tag: leunga-20030107-int-rem
1649    Description:
1650    
1651    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1652    The arguments to the SUBL instruction were swapped.
1653    
1654    ----------------------------------------------------------------------
1655    Name: Matthias Blume
1656    Date: 2002/12/12 16:25:00 EST
1657    Tag: blume-20021212-risc-ra
1658    Description:
1659    
1660    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1661    sent the RA into an infinite loop when floating point registers get
1662    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1663    110.37 and 110.38.)
1664    There was another set of potential problems related to the handling of
1665    MLRISC annotations (but those did not yet cause real problems, apparently).
1666    
1667    ----------------------------------------------------------------------
1668    Name: Matthias Blume
1669    Date: 2002/12/06 22:40:00 EST
1670    Tag: blume-20021206-cm-fileid
1671    Description:
1672    
1673    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1674    This fixes the problem of CM getting confused by files that suddenly
1675    change their identity (e.g., by getting unlinked and recreated by some
1676    text editor such as vi).  There might be a better/cheaper/cleaner way
1677    of doing this, but for now this will have to do.
1678    
1679    ----------------------------------------------------------------------
1680    Name: Matthias Blume
1681    Date: 2002/10/28 09:50:00 EST
1682    Tag: blume-20021028-typecheck
1683    Description:
1684    
1685    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1686    
1687    ----------------------------------------------------------------------
1688    Name: Matthias Blume
1689    Date: 2002/10/17 09:10:00 EDT
1690    Tag: Release_110_42
1691    Description:
1692    
1693    In good old tradition, there has been a slight hiccup so that we have
1694    to patch 110.42 after the fact.   The old release tag has been replaced
1695    (see below).
1696    
1697    The change solves a problem with two competing approaches the
1698    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1699    each other's way.
1700    
1701    This change only affects the runtime system code and the installer script.
1702    (No new bootfiles.)
1703    
1704    ----------------------------------------------------------------------
1705    Name: Matthias Blume
1706    Date: 2002/10/16 12:00:00 EDT
1707    Tag: Release_110_42_removed
1708    Description:
1709    
1710    New working release.  New bootfiles.
1711    
1712    ----------------------------------------------------------------------
1713    Name: Matthias Blume
1714    Date: 2002/10/10 13:10:00 EDT
1715    Tag: blume-20021010-ppc-divs
1716    Description:
1717    
1718    The mltree operator DIVS must be implemented with an overflow check on
1719    the PPC because the hardware indicates divide-by-zero using "overflow" as
1720    well.
1721    
1722    ----------------------------------------------------------------------
1723    Name: Matthias Blume
1724    Date: 2002/07/23 11:45:00
1725    Tag: blume-20020723-smlnj-home
1726    Description:
1727    
1728    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1729    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1730    CM_PATHCONFIG is also set), the path configuration file is assumed
1731    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1732    the entire tree to some other place and everything will "just work".
1733    
1734    (Companion commands such as ml-build and ml-makedepend also sense this
1735    variable.)
1736    
1737    ----------------------------------------------------------------------
1738    Name: Matthias Blume
1739    Date: 2002/07/12 21:19:00 EDT
1740    Tag: blume-20020712-liveness
1741    Description:
1742    
1743    Exported two useful "step" functions from liveness module (MLRISC).
1744    
1745    ----------------------------------------------------------------------
1746    Name: Matthias Blume
1747    Date: 2002/07/05 16:00 EDT
1748    Tag: Release_110_41
1749    Description:
1750    
1751    New working release.  New bootfiles.
1752    
1753    ----------------------------------------------------------------------
1754    Name: Matthias Blume
1755    Date: 2002/07/05 10:25:00 EDT
1756    Tag: blume-20020705-btimp
1757    Description:
1758    
1759    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1760    other clients can set up backtracing support.
1761    
1762    ----------------------------------------------------------------------
1763    Name: Matthias Blume
1764    Date: 2002/06/25 17:23:00 EDT
1765    Tag: blume-20020625-fpmax
1766    Description:
1767    
1768    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1769    case.  (The sense of the isNaN test was reversed -- which made min and
1770    max always return their first argument.)
1771    
1772    ----------------------------------------------------------------------
1773    Name: Matthias Blume
1774    Date: 2002/06/11
1775    Tag: blume-20020611-unixpath
1776    Description:
1777    
1778    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1779    
1780    ----------------------------------------------------------------------
1781    Name: Matthias Blume
1782    Date: 2002/06/10 16:35:00 EDT
1783    Tag: blume-20020610-ieeereal
1784    Description:
1785    
1786    I back-ported my implementation of IEEEReal.fromString from the
1787    idlbasis-devel branch so that we can test it.
1788    
1789    Another small change is that ppDec tries to give more information
1790    than just "<sig>" in the case of functors.  However, this code is
1791    broken in some mysterious way if the functor's body's signature
1792    has not been declared by ascription but gets inferred from the
1793    implementation.  This needs fixing...
1794    
1795    ----------------------------------------------------------------------
1796    Name: Matthias Blume
1797    Date: 2002/05/31
1798    Tag: blume-20020531-btrace-mode
1799    Description:
1800    
1801    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1802    the wayside when I switched over to using Controls everywhere.)
1803    
1804    ----------------------------------------------------------------------
1805    Name: Lal George
1806    Date: 2002/05/23 12:21:40 EDT
1807    Tag: george-20020523-visual-labels
1808    Description:
1809    
1810    Labels are now displayed in the graphical output to make
1811    the fall-through and target blocks obvious.
1812    
1813    ----------------------------------------------------------------------
1814    Name: Matthias Blume
1815    Date: 2002/05/22 11:03:00 EDT
1816    Tag: blume-20020522-shrink
1817    Description:
1818    
1819    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1820    situation that comes up when allocating huge arrays.
1821    
1822    ----------------------------------------------------------------------
1823    Name: Matthias Blume
1824    Date: 2002/05/21 16:00:00 EDT
1825    Tag: Release_110_40
1826    Description:
1827    
1828    New working release (110.40).  New bootfiles.
1829    
1830    [Also: John Reppy fixed GC bug 1131.]
1831    
1832    ----------------------------------------------------------------------
1833    Name: Matthias Blume
1834    Date: 2002/05/21 12:35:00 EDT
1835    Tag: blume-20020521-cmdoc
1836    Description:
1837    
1838    CM documentation update.
1839    
1840    ----------------------------------------------------------------------
1841    Name: Matthias Blume
1842    Date: 2002/05/21 10:55:00 EDT
1843    Tag: blume-20020521-misc
1844    Description:
1845    
1846    - John tweaked runtime to be silent on heap export (except when
1847      GC messages are on).
1848    - I added a few more things (cross-compiling versions of CMB) to
1849      config/preloads (as suggestions).
1850    
1851    ----------------------------------------------------------------------
1852    Name: Matthias Blume
1853    Date: 2002/05/20 22:25:00 EDT
1854    Tag: blume-20020520-controls
1855    Description:
1856    
1857    - Added ControlUtil structure to control-lib.cm.
1858    - Use it throughout.
1859    - Used Controls facility to define MLRISC controls (as opposed to
1860      registering MLRISC control ref cells with Controls after the fact)
1861    - Fixed messed-up controls priorities.
1862    
1863    * Removed again all the stuff from config/preloads that one wouldn't
1864      be able to preload at the time the initial heap image is built.
1865      (Many libraries, e.g., CML, do not exist yet at this time.  The
1866       only libraries that can be preloaded via config/preloads are those
1867       that come bundled with the bootfiles.)
1868    
1869    ----------------------------------------------------------------------
1870    Name: Matthias Blume
1871    Date: 2002/05/20 10:59:00 EDT
1872    Tag: blume-20020520-preloads
1873    Description:
1874    
1875    Added a lot of commented-out suggestions for things to be included
1876    in config/preloads.
1877    
1878    ----------------------------------------------------------------------
1879    Name: Allen Leung
1880    Date: 2002/05/18 14:20:00 EDT
1881    Tag: leunga-20020518-mdl
1882    Description:
1883    
1884    o Made the mdl tool stuff compile and run again.
1885    o I've disabled all the stuff that depends on RTL specifications; they
1886      are all badly broken anyway.
1887    
1888    ----------------------------------------------------------------------
1889    Name: Matthias Blume
1890    Date: 2002/05/17 16:49:00 EDT
1891    Tag: blume-20020517-controls
1892    Description:
1893    
1894    0. John Reppy made several modifications to the SML/NJ library.
1895       In particular, there is a shiny new controls-lib.cm.
1896    
1897    1. Pushed new controls interface through compiler so that everything
1898       compiles again.
1899    
1900    2. Added FormatComb and FORMAT_COMB to the CML version of the
1901       SML/NJ library (so that CML compiles again).
1902    
1903    3. Modified init scripts because XXX_DEFAULT environment variables
1904       are no longer with us.  (Boot-time initialization is now done
1905       using the same environment variables that are also used for
1906       startup-time initialization of controls.)
1907    
1908    ----------------------------------------------------------------------
1909    Name: Lal George
1910    Date: 2002/05/15 09:20:10 EDT
1911    Tag: george-20020515-pseudo-op-decls
1912    Description:
1913    
1914            All pseudo-ops emitted before the first segment declaration
1915    such as TEXT, DATA, and BSS directives are assumed to be global
1916    declarations and are emitted first in the assembly file. This is
1917    useful in a number of situations where one has pseudo-ops that are not
1918    specific to any segment, and also works around the constraint that one
1919    cannot have client pseudo-ops in the TEXT segment.
1920    
1921            Because no segment is associated with these declarations it is
1922    an error to allocate any space or objects before the first segment
1923    directive and an exception will be raised. However, we cannot make
1924    this check for client pseudo-ops.
1925    
1926            These top level declarations are a field in the CFG graph_info.
1927    In theory you can continue to add to this field after the CFG has been
1928    built -- provided you know what you are doing;-)
1929    
1930    ----------------------------------------------------------------------
1931    Name: Matthias Blume
1932    Date: 2002/05/13 16:40:00 EDT
1933    Tag: blume-20020513-pp-etc
1934    Description:
1935    
1936    A few minor bugfixes:
1937    
1938      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1939        (Bogus printouts for redefined bindings still occur.  Compiler
1940        bug should no longer occur now.  We need to redo the prettyprinter
1941        from scratch.)
1942    
1943      - CM pathname printer now also adds escape sequences for ( and )
1944    
1945      - commend and docu fixes for ml-nlffi
1946    
1947    ----------------------------------------------------------------------
1948    Name: Matthias Blume
1949    Date: 2002/05/10 16:40:00 EDT
1950    Tag: blume-20020510-erg-textio
1951    Description:
1952    
1953    Applied the following bugfix provided by Emden Gansner:
1954    
1955        Output is corrupted when outputSubstr is used rather than output.
1956    
1957        The problem occurs when a substring
1958    
1959            ss = (s, dataStart, dataLen)
1960    
1961        where dataStart > 0, fills a stream buffer with avail bytes left.
1962        avail bytes of s, starting at index dataStart, are copied into the
1963        buffer, the buffer is flushed, and then the remaining dataLen-avail
1964        bytes of ss are copied into the beginning of the buffer. Instead of
1965        starting this copy at index dataStart+avail in s, the current code
1966        starts the copy at index avail.
1967    
1968        Fix:
1969        In text-io-fn.sml, change line 695 from
1970             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1971        to
1972             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1973    
1974    ----------------------------------------------------------------------
1975    Name: Matthias Blume
1976    Date: 2002/04/12 13:55:00 EDT
1977    Tag: blume-20020412-assyntax
1978    Description:
1979    
1980    1. Grabbed newer assyntax.h from the XFree86 project.
1981    2. Fiddled with how to compile X86.prim.asm without warnings.
1982    3. (Very) Minor cleanup in CM.
1983    
1984    ----------------------------------------------------------------------
1985    Name: Matthias Blume
1986    Date: 2002/04/01 (no joke!) 17:07:00 EST
1987    Tag: blume-20020401-x86div
1988    Description:
1989    
1990    Added full support for div/mod/rem/quot on the x86, using the machine
1991    instruction's two results (without clumsily recomputing the remainder)
1992    directly where appropriate.
1993    
1994    Some more extensive power-of-two support was added to the x86 instruction
1995    selector (avoiding expensive divs, mods, and muls where they can be
1996    replaced with cheaper shifts and masks).  However, this sort of thing
1997    ought to be done earlier, e.g., within the CPS optimizer so that
1998    all architectures benefit from it.
1999    
2000    The compiler compiles to a fixed point, but changes might be somewhat
2001    fragile nevertheless.  Please, report any strange things that you might
2002    see wrt. div/mod/quot/rem...
2003    
2004    ----------------------------------------------------------------------
2005    Name: Matthias Blume
2006    Date: 2002/03/29 17:22:00
2007    Tag: blume-20020329-div
2008    Description:
2009    
2010    Fixed my broken div/mod logic.  Unfortunately, this means that the
2011    inline code for div/mod now has one more comparison than before.
2012    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2013    The problem was with quotient = 0, because that alone does not tell
2014    us which way the rounding went.  One then has to look at whether
2015    remainder and divisor have the same sign...  :(
2016    
2017    Anyway, I replaced the bootfiles with fresh ones...
2018    
2019    ----------------------------------------------------------------------
2020    Name: Matthias Blume
2021    Date: 2002/03/29 14:10:00 EST
2022    Tag: blume-20020329-inlprims
2023    Description:
2024    
2025    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2026    
2027    Primops have changed. This means that the bin/boot-file formats have
2028    changed as well.
2029    
2030    To make sure that there is no confusion, I made a new version.
2031    
2032    
2033    CHANGES:
2034    
2035    * removed REMT from mltree (remainder should never overflow).
2036    
2037    * added primops to deal with divisions of all flavors to the frontend
2038    
2039    * handled these primops all the way through so they map to their respective
2040      MLRISC support
2041    
2042    * used these primops in the implementation of Int, Int32, Word, Word32
2043    
2044    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2045    
2046    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2047    
2048    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2049      floating point
2050    
2051    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2052      and Real (but Real.abs maps to a separate floating-point-only primop)
2053    
2054    
2055    TODO items:
2056    
2057    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2058      instructions because the machine instruction encoder cannot handle
2059      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2060      REM in the first place, and REM is supported by the x86 machine coder.)
2061    
2062    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2063      selection phase.  (The two can be streamlined because the hardware
2064      delivers both quotient and remainder at the same time anyway.)
2065    
2066    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2067      (Currently the behavior is inconsistent both across architectures and
2068      wrt. the draft Basis spec.)
2069    
2070    * Word8 should eventually be handled natively, too.
2071    
2072    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2073      as if there currently is no execution path that could trigger it in
2074      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2075      hold for things like multiplication and division.)
2076    
2077    ----------------------------------------------------------------------
2078    Name: Matthias Blume
2079    Date: 2002/03/27 16:27:00 EST
2080    Tag: blume-20020327-mlrisc-divisions
2081    Description:
2082    
2083    Added support for all four division operations (ML's div, mod, quot,
2084    and rem) to MLRISC.  In the course of doing so, I also rationalized
2085    the naming (no more annoying switch-around of DIV and QUOT), by
2086    parameterizing the operation by div_rounding_mode (which can be either
2087    DIV_TO_ZERO or DIV_TO_NEGINF).
2088    
2089    The generic MLTreeGen functor takes care of compiling all four
2090    operations down to only round-to-zero div.
2091    
2092    Missing pieces:
2093    
2094      * Doing something smarter than relying on MLTreeGen on architectures
2095        like, e.g., the x86 where hardware division delivers both quotient and
2096        remainder at the same time.  With this, the implementation of the
2097        round-to-neginf operations could be further streamlined.
2098    
2099      * Remove inlining support for div/mod/rem from the frontend and replace it
2100        with primops that get carried through to the backend.  Do this for all
2101        int and word types.
2102    
2103    ----------------------------------------------------------------------
2104    Name: Matthias Blume
2105    Date: 2002/03/25 17:25:00 EST
2106    Tag: blume-20020325-divmod
2107    Description:
2108    
2109    I improved (hopefully without breaking them) the implementation of Int.div,
2110    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2111    advantage of the following observations:
2112    
2113      Let  q = x quot y      r = x rem y
2114           d = x div  y      m = x mod y
2115    
2116    where "quot" is the round-to-zero version of integer division that
2117    hardware usually provides.  Then we have:
2118    
2119         r = x - q * y        where neither the * nor the - will overflow
2120         d = if q >= 0 orelse x = q * y then q else q - 1
2121                              where neither the * nor the - will overflow
2122         m = if q >= 0 orelse r = 0 then r else r + y
2123                              where the + will not overflow
2124    
2125    This results in substantial simplification of the generated code.
2126    The following table shows the number of CFG nodes and edges generated
2127    for
2128            fun f (x, y) = x OPER y
2129            (* with OPER \in div, mod, quot, rem *)
2130    
2131    
2132        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2133        --------------------------------------------------------
2134         div |         24 |         39 |         12 |         16
2135         mod |         41 |         71 |         12 |         16
2136        quot |          8 |         10 |          8 |         10
2137         rem |         10 |         14 |          8 |         10
2138    
2139    
2140    ----------------------------------------------------------------------
2141    Name: Matthias Blume
2142    Date: 2002/03/25 22:06:00 EST
2143    Tag: blume-20020325-cprotobug
2144    Description:
2145    
2146    Fixed a bug in cproto (c prototype decoder).
2147    
2148    ----------------------------------------------------------------------
2149    Name: Matthias Blume
2150    Date: 2002/03/25 16:00:00 EST
2151    Tag: blume-20020325-raw-primops
2152    Description:
2153    
2154    I did some cleanup to Allen's new primop code and
2155    replaced yesterday's bootfiles with new ones.
2156    (But they are stored in the same place.)
2157    
2158    ----------------------------------------------------------------------
2159    Name: Matthias Blume
2160    Date: 2002/03/24 22:40:00 EST
2161    Tag: blume-20020324-bootfiles
2162    Description:
2163    
2164    Made the bootfiles that Allen asked for.
2165    
2166    ----------------------------------------------------------------------
2167    Name: Allen Leung
2168    Date: 2002/03/23 15:50:00 EST
2169    Tag: leunga-20020323-flint-cps-rcc-primops
2170    Description:
2171    
2172      1. Changes to FLINT primops:
2173    
2174        (* make a call to a C-function;
2175         * The primop carries C function prototype information and specifies
2176         * which of its (ML-) arguments are floating point. C prototype
2177         * information is for use by the backend, ML information is for
2178         * use by the CPS converter. *)
2179      | RAW_CCALL of { c_proto: CTypes.c_proto,
2180                       ml_args: ccall_type list,
2181                       ml_res_opt: ccall_type option,
2182                       reentrant : bool
2183                     } option
2184       (* Allocate uninitialized storage on the heap.
2185        * The record is meant to hold short-lived C objects, i.e., they
2186        * are not ML pointers.  With the tag, the representation is
2187        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2188        *)
2189      | RAW_RECORD of {tag:bool,sz:int}
2190      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2191    
2192      2.  These CPS primops are now overloaded:
2193    
2194           rawload of {kind:numkind}
2195           rawstore of {kind:numkind}
2196    
2197          The one argument form is:
2198    
2199             rawload {kind} address
2200    
2201          The two argument form is:
2202    
2203             rawload {kind} [ml object, byte-offset]
2204    
2205      3. RAW_CCALL/RCC now takes two extra arguments:
2206    
2207         a. The first is whether the C call is reentrant, i.e., whether
2208            ML state should be saved and restored.
2209         b. The second argument is a string argument specifying the name of
2210            library and the C function.
2211    
2212         These things are currently not handled in the code generator, yet.
2213    
2214      4. In CProto,
2215    
2216         An encoding type of "bool" means "ml object" and is mapped into
2217         C prototype of PTR.  Note that "bool" is different than "string",
2218         even though "string" is also mapped into PTR, because "bool"
2219         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2220    
2221      5. Pickler/unpicker
2222    
2223         Changed to handle RAW_RECORD and newest RAW_CCALL
2224    
2225      6. MLRiscGen,
2226    
2227         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2228         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2229            in the file CodeGen/cpscompile/cps-c-calls.sml
2230    
2231      7. Added the conditional move operator
2232    
2233             condmove of branch
2234    
2235         to cps.  Generation of this is still buggy so it is currently
2236         disabled.
2237    
2238    ----------------------------------------------------------------------
2239    Name: Lal George
2240    Date: 2002/03/22 14:18:25 EST
2241    Tag: george-20020322-cps-branch-prob
2242    Description:
2243    
2244    Implemented the Ball-Larus branch prediction-heuristics, and
2245    incorporated graphical viewers for control flow graphs.
2246    
2247    Ball-Larus Heuristics:
2248    ---------------------
2249    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2250    
2251    By design it uses the Dempster-Shafer theory for combining
2252    probabilities.  For example, in the function:
2253    
2254        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2255    
2256    the ball-larus heuristics predicts that the n=0 is unlikely
2257    (OH-heuristic), and the 'then' branch is unlikely because of the
2258    RH-heuristic -- giving the 'then' branch an even lower combined
2259    probability using the Dempster-Shafer theory.
2260    
2261    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2262    probability of the 'then' branch because of the loop in the else
2263    branch.
2264    
2265    
2266    Graphical Viewing:
2267    ------------------
2268    I merely plugged in Allen's graphical viewers into the compiler. The
2269    additional code is not much. At the top level, saying:
2270    
2271            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2272    
2273    will display the graphical view of the control flow graph just before
2274    back-patching.  daVinci must be in your path for this to work. If
2275    daVinci is not available, then the default viewer can be changed
2276    using:
2277    
2278            Control.MLRISC.getString "viewer"
2279    
2280    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2281    course, these viewers must be in your path.
2282    
2283    The above will display the compilation unit at the level of clusters,
2284    many of which are small, boring, and un-interesting. Also setting:
2285    
2286            Control.MLRISC.getInt "cfg-graphical-view_size"
2287    
2288    will display clusters that are larger than the value set by the above.
2289    
2290    
2291    ----------------------------------------------------------------------
2292    Name: Matthias Blume
2293    Date: 2002/03/21 22:20:00 EST
2294    Tag: blume-20020321-kmp-bugfix
2295    Description:
2296    
2297    Changed the interface to the KMP routine in PreString and fixed
2298    a minor bug in one place where it was used.
2299    
2300    ----------------------------------------------------------------------
2301    Name: Allen Leung
2302    Date: 2002/03/21 20:30:00 EST
2303    Tag: leunga-20020321-cfg
2304    Description:
2305    
2306      Fixed a potential problem in cfg edge splitting.
2307    
2308    ----------------------------------------------------------------------
2309    Name: Allen Leung
2310    Date: 2002/03/21 17:15:00 EST
2311    Tag: leunga-20020321-x86-fp-cfg
2312    Description:
2313    
2314      1. Recoded the buggy parts of x86-fp.
2315    
2316         a. All the block reordering code has been removed.
2317            We now depend on the block placement phases to do this work.
2318    
2319         b. Critical edge splitting code has been simplified and moved into the
2320            CFG modules, as where they belong.
2321    
2322         Both of these were quite buggy and complex.  The code is now much, much
2323         simpler.
2324    
2325      2. X86 backend.
2326    
2327         a. Added instructions for 64-bit support.  Instruction selection for
2328            64-bit has not been committed, however, since that
2329            requires changes to MLTREE which haven't been approved by
2330            Lal and John.
2331    
2332         b. Added support for FUCOMI and FUCOMIP when generating code for
2333            PentiumPro and above.  We only generate these instructions in
2334            the fast-fp mode.
2335    
2336         c. Added cases for JP and JNP in X86FreqProps.
2337    
2338      3. CFG
2339    
2340         CFG now has a bunch of methods for edge splitting and merging.
2341    
2342      4. Machine description.
2343    
2344         John's simplification of MLTREE_BASIS.fcond broke a few machine
2345         description things:
2346    
2347         rtl-build.{sig,sml} and hppa.mdl fixed.
2348    
2349         NOTE: the machine description stuff in the repository is still broken.
2350               Again, I can't put my fixes in because that involves
2351               changes to MLTREE.
2352    
2353    ----------------------------------------------------------------------
2354    Name: Matthias Blume
2355    Date: 2002/03/20 15:55:00 EST
2356    Tag: blume-20020320-kmp
2357    Description:
2358    
2359    Implemented Knuth-Morris-Pratt string matching in PreString and used
2360    it for String.isSubstring, Substring.isSubstring, and
2361    Substring.position.
2362    
2363    (Might need some stress-testing.  Simple examples worked fine.)
2364    
2365    ----------------------------------------------------------------------
2366    Name: Matthias Blume
2367    Date: 2002/03/19 16:37:00 EST
2368    Tag: blume-20020319-witnesses
2369    Description:
2370    
2371    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2372    
2373    This implements a generic mechanism for changing constness qualifiers
2374    anywhere within big C types without resorting to outright "casts".
2375    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2376    modify the constness at the outermost level.)
2377    The implementation of "convert" is based on the idea of "witness"
2378    values -- values that are not used by the operation but whose types
2379    "testify" to their applicability.  On the implementation side, "convert"
2380    is simply a projection (returning its second curried argument).  With
2381    cross-module inlining, it should not result in any machine code being
2382    generated.
2383    
2384    ----------------------------------------------------------------------
2385    Name: Matthias Blume
2386    Date: 2002/03/15 16:40:00 EST
2387    Tag: blume-20020315-basis
2388    Description:
2389    
2390    Provided (preliminary?) implementations for
2391    
2392      {String,Substring}.{concatWith,isSuffix,isSubstring}
2393    
2394    and
2395    
2396      Substring.full
2397    
2398    Those are in the Basis spec but they were missing in SML/NJ.
2399    
2400    ----------------------------------------------------------------------
2401    Name: Matthias Blume
2402    Date: 2002/03/14 21:30:00 EST
2403    Tag: blume-20020314-controls
2404    Description:
2405    
2406    Controls:
2407    ---------
2408    
2409    1. Factored out the recently-added Controls : CONTROLS stuff and put
2410       it into its own library $/controls-lib.cm.  The source tree for
2411       this is under src/smlnj-lib/Controls.
2412    
2413    2. Changed the names of types and functions in this interface, so they
2414       make a bit more "sense":
2415    
2416          module -> registry
2417          'a registry -> 'a group
2418    
2419    3. The interface now deals in ref cells only.  The getter/setter interface
2420       is (mostly) gone.
2421    
2422    4. Added a function that lets one register an already-existing ref cell.
2423    
2424    5. Made the corresponding modifications to the rest of the code so that
2425       everything compiles again.
2426    
2427    6. Changed the implementation of Controls.MLRISC back to something closer
2428       to the original.  In particular, this module (and therefore MLRISC)
2429       does not depend on Controls.  There now is some link-time code in
2430       int-sys.sml that registers the MLRISC controls with the Controls
2431       module.
2432    
2433    CM:
2434    ---
2435    
2436      * One can now specify the lambda-split aggressiveness in init.cmi.
2437    
2438    ----------------------------------------------------------------------
2439    Name: Allen Leung
2440    Date: 2002/03/13 17:30:00 EST
2441    Tag: leunga-20020313-x86-fp-unary
2442    Description:
2443    
2444    Bug fix for:
2445    
2446    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2447    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2448    > - fun f(x,(y,z)) = Real.~ y;
2449    > [autoloading]
2450    > [autoloading done]
2451    >       fchsl   (%eax), 184(%esp)
2452    > Error: MLRisc bug: X86MCEmitter.emitInstr
2453    >
2454    > uncaught exception Error
2455    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2456    
2457    The problem was that the code generator did not generate any fp registers
2458    in this case, and the ra didn't know that it needed to run the X86FP phase to
2459    translate the pseudo fp instruction.   This only happened with unary fp
2460    operators in certain situations.
2461    
2462    ----------------------------------------------------------------------
2463    Name: Matthias Blume
2464    Date: 2002/03/13 14:00:00 EST
2465    Tag: blume-20020313-overload-etc
2466    Description:
2467    
2468    1. Added _overload as a synonym for overload for backward compatibility.
2469       (Control.overloadKW must be true for either version to be accepted.)
2470    
2471    2. Fixed bug in install script that caused more things to be installed
2472       than what was requested in config/targets.
2473    
2474    3. Made CM aware of the (_)overload construct so that autoloading
2475       works.
2476    
2477    ----------------------------------------------------------------------
2478    Name: Matthias Blume
2479    Date: 2002/03/12 22:03:00 EST
2480    Tag: blume-20020312-url
2481    Description:
2482    
2483    Forgot to update BOOT and srcarchiveurl.
2484    
2485    ----------------------------------------------------------------------
2486    Name: Matthias Blume
2487    Date: 2002/03/12 17:30:00 EST
2488    Tag: blume-20020312-version110392
2489    Description:
2490    
2491    Yet another version number bump (because of small changes to the
2492    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2493    
2494    Changes:
2495    
2496      The new pid generation scheme described a few weeks ago was overly
2497      complicated.  I implemented a new mechanism that is simpler and
2498      provides a bit more "stability":  Once CM has seen a compilation
2499      unit, it keeps its identity constant (as long as you do not delete
2500      those crucial CM/GUID/* files).  This means that when you change
2501      an interface, compile, then go back to the old interface, and
2502      compile again, you arrive at the original pid.
2503    
2504      There now also is a mechanism that instructs CM to use the plain
2505      environment hash as a module's pid (effectively making its GUID
2506      the empty string).  For this, "noguid" must be specified as an
2507      option to the .sml file in question within its .cm file.
2508      This is most useful for code that is being generated by tools such
2509      as ml-nlffigen (because during development programmers tend to
2510      erase the tool's entire output directory tree including CM's cached
2511      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2512      revert to the old, broken behavior of SML/NJ, but in specific cases
2513      where there is no danger of interface confusion, its use is ok
2514      (I think).
2515    
2516      ml-nlffigen by default generates "noguid" annotations.  They can be
2517      turned off by specifying -guid in its command line.
2518    
2519    ----------------------------------------------------------------------
2520    Name: Lal George
2521    Date: 2002/03/12 12 14:42:36 EST
2522    Tag: george-20020312-frequency-computation
2523    Description:
2524    
2525    Integrated jump chaining and static block frequency into the
2526    compiler. More details and numbers later.
2527    
2528    ----------------------------------------------------------------------
2529    Name: Lal George
2530    Date: 2002/03/11 11 22:38:53 EST
2531    Tag: george-20020311-jump-chain-elim
2532    Description:
2533    
2534    Tested the jump chain elimination on all architectures (except the
2535    hppa).  This is on by default right now and is profitable for the
2536    alpha and x86, however, it may not be profitable for the sparc and ppc
2537    when compiling the compiler.
2538    
2539    The gc test will typically jump to a label at the end of the cluster,
2540    where there is another jump to an external cluster containing the actual
2541    code to invoke gc. This is to allow factoring of common gc invocation
2542    sequences. That is to say, we generate:
2543    
2544            f:
2545               testgc
2546               ja   L1      % jump if above to L1
2547    
2548            L1:
2549               jmp L2
2550    
2551    
2552    After jump chain elimination the 'ja L1' instructions is converted to
2553    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2554    up being implemented in their long form (if L2 is far away) using:
2555    
2556            jbe     L3      % jump if below or equal to L3
2557            jmp     L2
2558         L3:
2559            ...
2560    
2561    
2562    For large compilation units L2  may be far away.
2563    
2564    
2565    ----------------------------------------------------------------------
2566    Name: Matthias Blume
2567    Date: 2002/03/11 13:30:00 EST
2568    Tag: blume-20020311-mltreeeval
2569    Description:
2570    
2571    A functor parameter was missing.
2572    
2573    ----------------------------------------------------------------------
2574    Name: Allen Leung
2575    Date: 2002/03/11 10:30:00 EST
2576    Tag: leunga-20020311-runtime-string0
2577    Description:
2578    
2579       The representation of the empty string now points to a
2580    legal null terminated C string instead of unit.  It is now possible
2581    to convert an ML string into C string with InlineT.CharVector.getData.
2582    This compiles into one single machine instruction.
2583    
2584    ----------------------------------------------------------------------
2585    Name: Allen Leung
2586    Date: 2002/03/10 23:55:00 EST
2587    Tag: leunga-20020310-x86-call
2588    Description:
2589    
2590       Added machine generation for CALL instruction (relative displacement mode)
2591    
2592    ----------------------------------------------------------------------
2593    Name: Matthias Blume
2594    Date: 2002/03/08 16:05:00
2595    Tag: blume-20020308-entrypoints
2596    Description:
2597    
2598    Version number bumped to 110.39.1.  NEW BOOTFILES!
2599    
2600    Entrypoints: non-zero offset into a code object where execution should begin.
2601    
2602    - Added the notion of an entrypoint to CodeObj.
2603    - Added reading/writing of entrypoint info to Binfile.
2604    - Made runtime system bootloader aware of entrypoints.
2605    - Use the address of the label of the first function given to mlriscGen
2606      as the entrypoint.  This address is currently always 0, but it will
2607      not be 0 once we turn on block placement.
2608    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2609      with entry points) from mlriscGen.
2610    
2611    ----------------------------------------------------------------------
2612    Name: Allen Leung
2613    Date: 2002/03/07 20:45:00 EST
2614    Tag: leunga-20020307-x86-cmov
2615    Description:
2616    
2617       Bug fixes for CMOVcc on x86.
2618    
2619       1. Added machine code generation for CMOVcc
2620       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2621       3. CMOVcc cannot have an immediate operand as argument.
2622    
2623    ----------------------------------------------------------------------
2624    Name: Matthias Blume
2625    Date: 2002/03/07 16:15:00 EST
2626    Tag: blume-20020307-controls
2627    Description:
2628    
2629    This is a very large but mostly boring patch which makes (almost)
2630    every tuneable compiler knob (i.e., pretty much everything under
2631    Control.* plus a few other things) configurable via both the command
2632    line and environment variables in the style CM did its configuration
2633    until now.
2634    
2635    Try starting sml with '-h' (or, if you are brave, '-H')
2636    
2637    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2638    implements the underlying generic mechanism.
2639    
2640    The interface to some of the existing such facilities has changed somewhat.
2641    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2642    (The getFoo interface is still there for backward-compatibility, but its
2643    use is deprecated.)
2644    
2645    The ml-build script passes -Cxxx=yyy command-line arguments through so
2646    that one can now twiddle the compiler settings when using this "batch"
2647    compiler.
2648    
2649    TODO items:
2650    
2651    We should go through and throw out all controls that are no longer
2652    connected to anything.  Moreover, we should go through and provide
2653    meaningful (and correct!) documentation strings for those controls
2654    that still are connected.
2655    
2656    Currently, multiple calls to Controls.new are accepted (only the first
2657    has any effect).  Eventually we should make sure that every control
2658    is being made (via Controls.new) exactly once.  Future access can then
2659    be done using Controls.acc.
2660    
2661    Finally, it would probably be a good idea to use the getter-setter
2662    interface to controls rather than ref cells.  For the time being, both
2663    styles are provided by the Controls module, but getter-setter pairs are
2664    better if thread-safety is of any concern because they can be wrapped.
2665    
2666    *****************************************
2667    
2668    One bug fix: The function blockPlacement in three of the MLRISC
2669    backpatch files used to be hard-wired to one of two possibilities at
2670    link time (according to the value of the placementFlag).  But (I
2671    think) it should rather sense the flag every time.
2672    
2673    *****************************************
2674    
2675    Other assorted changes (by other people who did not supply a HISTORY entry):
2676    
2677    1. the cross-module inliner now works much better (Monnier)
2678    2. representation of weights, frequencies, and probabilities in MLRISC
2679       changed in preparation of using those for weighted block placement
2680       (Reppy, George)
2681    
2682    ----------------------------------------------------------------------
2683    Name: Lal George
2684    Date: 2002/03/07 14:44:24 EST 2002
2685    Tag: george-20020307-weighted-block-placement
2686    
2687    Tested the weighted block placement optimization on all architectures
2688    (except the hppa) using AMPL to generate the block and edge frequencies.
2689    Changes were required in the machine properties to correctly
2690    categorize trap instructions. There is an MLRISC flag
2691    "weighted-block-placement" that can be used to enable weighted block
2692    placement, but this will be ineffective without block/edge
2693    frequencies (coming soon).
2694    
2695    
2696    ----------------------------------------------------------------------
2697    Name: Lal George
2698    Date: 2002/03/05 17:24:48 EST
2699    Tag: george-20020305-linkage-cluster
2700    
2701    In order to support the block placement optimization, a new cluster
2702    is generated as the very first cluster (called the linkage cluster).
2703    It contains a single jump to the 'real' entry point for the compilation
2704    unit. Block placement has no effect on the linkage cluster itself, but
2705    all the other clusters  have full freedom in the manner in which they
2706    reorder blocks or functions.
2707    
2708    On the x86 the typical linkage code that is generated is:
2709       ----------------------
2710            .align 2
2711       L0:
2712            addl    $L1-L0, 72(%esp)
2713            jmp     L1
2714    
2715    
2716            .align  2
2717       L1:
2718       ----------------------
2719    
2720    72(%esp) is the memory location for the stdlink register. This
2721    must contain the address of the CPS function being called. In the
2722    above example, it contains the address of  L0; before
2723    calling L1 (the real entry point for the compilation unit), it
2724    must contain the address for L1, and hence
2725    
2726            addl $L1-L0, 72(%esp)
2727    
2728    I have tested this on all architectures except the hppa.The increase
2729    in code size is of course negligible
2730    
2731    ----------------------------------------------------------------------
2732    Name: Allen Leung
2733    Date: 2002/03/03 13:20:00 EST
2734    Tag: leunga-20020303-mlrisc-tools
2735    
2736      Added #[ ... ] expressions to mlrisc tools
2737    
2738    ----------------------------------------------------------------------
2739    Name: Matthias Blume
2740    Date: 2002/02/27 12:29:00 EST
2741    Tag: blume-20020227-cdebug
2742    Description:
2743    
2744    - made types in structure C and C_Debug to be equal
2745    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2746    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2747    
2748    ----------------------------------------------------------------------
2749    Name: Matthias Blume
2750    Date: 2002/02/26 12:00:00 EST
2751    Tag: blume-20020226-ffi
2752    Description:
2753    
2754    1. Fixed a minor bug in CM's "noweb" tool:
2755       If numbering is turned off, then truly don't number (i.e., do not
2756       supply the -L option to noweb).  The previous behavior was to supply
2757       -L'' -- which caused noweb to use the "default" line numbering scheme.
2758       Thanks to Chris Richards for pointing this out (and supplying the fix).
2759    
2760    2. Once again, I reworked some aspects of the FFI:
2761    
2762       A. The incomplete/complete type business:
2763    
2764       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2765         gone!
2766       - ML types representing an incomplete type are now *equal* to
2767         ML types representing their corresponding complete types (just like
2768         in C).  This is still safe because ml-nlffigen will not generate
2769         RTTI for incomplete types, nor will it generate functions that
2770         require access to such RTTI.   But when ML code generated from both
2771         incomplete and complete versions of the C type meet, the ML types
2772         are trivially interoperable.
2773    
2774         NOTE:  These changes restore the full generality of the translation
2775         (which was previously lost when I eliminated functorization)!
2776    
2777       B. Enum types:
2778    
2779       - Structure C now has a type constructor "enum" that is similar to
2780         how the "su" constructor works.  However, "enum" is not a phantom
2781         type because each "T enum" has values (and is isomorphic to
2782         MLRep.Signed.int).
2783       - There are generic access operations for enum objects (using
2784         MLRep.Signed.int).
2785       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2786         * The structure contains the definition of type "mlrep" (the ML-side
2787         representation type of the enum).  Normally, mlrep is the same
2788         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2789         then mlrep will be defined as a datatype -- thus facilitating
2790         pattern matching on mlrep values.
2791         ("-ec" will be suppressed if there are duplicate values in an
2792          enumeration.)
2793         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2794         will be generated for each C enum constant xxx.
2795         * Conversion functions m2i and i2m convert between mlrep and
2796         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2797         * Coversion functions c and ml convert between mlrep and "tag enum".
2798         * Access functions (get/set) fetch and store mlrep values.
2799       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2800         enumerations are merged into one single enumeration represented by
2801         structure E_'.
2802    
2803    ----------------------------------------------------------------------
2804    Name: Allen Leung
2805    Date: 2002/02/25 04:45:00 EST
2806    Tag: leunga-20020225-cps-spill
2807    
2808    This is a new implementation of the CPS spill phase.
2809    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2810    In case of problems, replace it with the old file spill.sml
2811    
2812    The current compiler runs into some serious performance problems when
2813    constructing a large record.  This can happen when we try to compile a
2814    structure with many items.  Even a very simple structure like the following
2815    makes the compiler slow down.
2816    
2817        structure Foo = struct
2818           val x_1 = 0w1 : Word32.int
2819           val x_2 = 0w2 : Word32.int
2820           val x_3 = 0w3 : Word32.int
2821           ...
2822           val x_N = 0wN : Word32.int
2823        end
2824    
2825    The following table shows the compile time, from N=1000 to N=4000,
2826    with the old compiler:
2827    
2828    N
2829    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2830           MLRISC ra                               0.06u  0.00s  0.05g
2831              (spills = 0 reloads = 0)
2832           TOTAL                                   0.63u  0.07s  0.21g
2833    
2834    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2835           MLRISC ra                               5.68u  0.59s  3.93g
2836              (spills = 0 reloads = 0)
2837           TOTAL                                   14.71u  0.99s  4.81g
2838    
2839    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2840           MLRISC ra                               5.54u  0.65s  3.91g
2841              (spills = 543 reloads = 1082)
2842           TOTAL                                   65.40u  3.13s  6.00g
2843    
2844    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2845           MLRISC ra                               0.80u  0.10s  0.55g
2846              (spills = 42 reloads = 84)
2847           TOTAL                                   129.42u  5.10s  4.13g
2848    
2849    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2850           MLRISC ra                               2.69u  0.27s  1.38g
2851              (spills = 62 reloads = 124)
2852           TOTAL                                   682.48u  19.61s  13.99g
2853    
2854    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2855           MLRISC ra                               4.96u  0.27s  2.72g
2856              (spills = 85 reloads = 170)
2857           TOTAL                                   2375.26u  57.21s  48.00g
2858    
2859    As you can see the old cps spill module suffers from some serious
2860    performance problem.  But since I cannot decipher the old code fully,
2861    instead of patching the problems up, I'm reimplementing it
2862    with a different algorithm.  The new code is more modular,
2863    smaller when compiled, and substantially faster
2864    (O(n log n) time and O(n) space).  Timing of the new spill module:
2865    
2866    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2867          MLRISC ra                               0.25u  0.02s  0.15g
2868             (spills=1 reloads=3)
2869          TOTAL                                   7.74u  0.34s  1.62g
2870    
2871    Implementation details:
2872    
2873    As far as I can tell, the purpose of the CPS spill module is to make sure the
2874    number of live variables at any program point (the bandwidth)
2875    does not exceed a certain limit, which is determined by the
2876    size of the spill area.
2877    
2878    When the bandwidth is too large, we decrease the register pressure by
2879    packing live variables into spill records.  How we achieve this is
2880    completely different than what we did in the old code.
2881    
2882    First, there is something about the MLRiscGen code generator
2883    that we should be aware of:
2884    
2885    o MLRiscGen performs code motion!
2886    
2887       In particular, it will move floating point computations and
2888       address computations involving only the heap pointer to
2889       their use sites (if there is only a single use).
2890       What this means is that if we have a CPS record construction
2891       statement
2892    
2893           RECORD(k,vl,w,e)
2894    
2895       we should never count the new record address w as live if w
2896       has only one use (which is often the case).
2897    
2898       We should do something similar to floating point, but the transformation
2899       there is much more complex, so I won't deal with that.
2900    
2901    Secondly, there are now two new cps primops at our disposal:
2902    
2903     1. rawrecord of record_kind option
2904        This pure operator allocates some uninitialized storage from the heap.
2905        There are two forms:
2906    
2907         rawrecord NONE [INT n]  allocates a tagless record of length n
2908         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2909                                     and initializes the tag.
2910    
2911     2. rawupdate of cty
2912          rawupdate cty (v,i,x)
2913          Assigns to x to the ith component of record v.
2914          The storelist is not updated.
2915    
2916    We use these new primops for both spilling and increment record construction.
2917    
2918     1. Spilling.
2919    
2920        This is implemented with a linear scan algorithm (but generalized
2921        to trees).  The algorithm will create a single spill record at the
2922        beginning of the cps function and use rawupdate to spill to it,
2923        and SELECT or SELp to reload from it.  So both spills and reloads
2924        are fine-grain operations.  In contrast, in the old algorithm
2925        "spills" have to be bundled together in records.
2926    
2927        Ideally, we should sink the spill record construction to where
2928        it is needed.  We can even split the spill record into multiple ones
2929        at the places where they are needed.  But CPS is not a good
2930        representation for global code motion, so I'll keep it simple and
2931        am not attempting this.
2932    
2933     2. Incremental record construction (aka record splitting).
2934    
2935        Long records with many component values which are simulatenously live
2936        (recall that single use record addresses are not considered to
2937         be live) are constructed with rawrecord and rawupdate.
2938        We allocate space on the heap with rawrecord first, then gradually
2939        fill it in with rawupdate.  This is the technique suggested to me
2940        by Matthias.
2941    
2942        Some restrictions on when this is applicable:
2943        1. It is not a VECTOR record.  The code generator currently does not handle
2944           this case. VECTOR record uses double indirection like arrays.
2945        2. All the record component values are defined in the same "basic block"
2946           as the record constructor.  This is to prevent speculative
2947           record construction.
2948    
2949    ----------------------------------------------------------------------
2950    Name: Allen Leung
2951    Date: 2002/02/22 01:02:00 EST
2952    Tag: leunga-20020222-mlrisc-tools
2953    
2954    Minor bug fixes in the parser and rewriter
2955    
2956    ----------------------------------------------------------------------
2957    Name: Allen Leung
2958    Date: 2002/02/21 20:20:00 EST
2959    Tag: leunga-20020221-peephole
2960    
2961    Regenerated the peephole files.  Some contained typos in the specification
2962    and some didn't compile because of pretty printing bugs in the old version
2963    of 'nowhere'.
2964    
2965    ----------------------------------------------------------------------
2966    Name: Allen Leung
2967    Date: 2002/02/19 20:20:00 EST
2968    Tag: leunga-20020219-mlrisc-tools
2969    Description:
2970    
2971       Minor bug fixes to the mlrisc-tools library:
2972    
2973       1.  Fixed up parsing colon suffixed keywords
2974       2.  Added the ability to shut the error messages up
2975       3.  Reimplemented the pretty printer and fixed up/improved
2976           the pretty printing of handle and -> types.
2977       4.  Fixed up generation of literal symbols in the nowhere tool.
2978       5.  Added some SML keywords to to sml.sty
2979    
2980    ----------------------------------------------------------------------
2981    Name: Matthias Blume
2982    Date: 2002/02/19 16:20:00 EST
2983    Tag: blume-20020219-cmffi
2984    Description:
2985    
2986    A wild mix of changes, some minor, some major:
2987    
2988    * All C FFI-related libraries are now anchored under $c:
2989        $/c.cm      --> $c/c.cm
2990        $/c-int.cm  --> $c/internals/c-int.cm
2991        $/memory.cm --> $c/memory/memory.cm
2992    
2993    * "make" tool (in CM) now treats its argument pathname slightly
2994      differently:
2995        1. If the native expansion is an absolute name, then before invoking
2996           the "make" command on it, CM will apply OS.Path.mkRelative
2997           (with relativeTo = OS.FileSys.getDir()) to it.
2998        2. The argument will be passed through to subsequent phases of CM
2999           processing without "going native".  In particular, if the argument
3000           was an anchored path, then "make" will not lose track of that anchor.
3001    
3002    * Compiler backends now "know" their respective C calling conventions
3003      instead of having to be told about it by ml-nlffigen.  This relieves
3004      ml-nlffigen from one of its burdens.
3005    
3006    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3007    
3008    * Export C_DEBUG and C_Debug from $c/c.cm.
3009    
3010    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3011      subtyping relationship between incomplete pointers and their complete
3012      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3013      with the convention of instantiating 'o with ('t, 'c) obj whenever
3014      the pointer target type is complete.  In the incomplete case, 'o
3015      will be instantiated with some "'c iobj" -- a type obtained by
3016      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3017    
3018      Operations that work on both incomplete and complete pointer types are
3019      typed as taking an 'o ptr while operations that require the target to
3020      be known are typed as taking some ('t, 'c) obj ptr.
3021    
3022      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3023      where void is an eqtype without any values.  This makes it possible
3024      to work on voidptr values using functions meant to operate on light
3025      incomplete pointers.
3026    
3027    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3028      been vastly simplified.
3029    
3030    ----------------------------------------------------------------------
3031    Name: Matthias Blume
3032    Date: 2002/02/19 10:48:00 EST
3033    Tag: blume-20020219-pqfix
3034    Description:
3035    
3036    Applied Chris Okasaki's bug fix for priority queues.
3037    
3038    ----------------------------------------------------------------------
3039    Name: Matthias Blume
3040    Date: 2002/02/15 17:05:00
3041    Tag: Release_110_39
3042    Description:
3043    
3044    Last-minute retagging is becoming a tradition... :-(
3045    
3046    This is the working release 110.39.
3047    
3048    ----------------------------------------------------------------------
3049    Name: Matthias Blume
3050    Date: 2002/02/15 16:00:00 EST
3051    Tag: Release_110_39-orig
3052    Description:
3053    
3054    Working release 110.39.  New bootfiles.
3055    
3056    (Update: There was a small bug in the installer so it wouldn't work
3057    with all shells.  So I retagged. -Matthias)
3058    
3059    ----------------------------------------------------------------------
3060    Name: Matthias Blume
3061    Date: 2002/02/15 14:17:00 EST
3062    Tag: blume-20020215-showbindings
3063    Description:
3064    
3065    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3066    the latter can be useful for exploring what bindings are available at
3067    the interactive prompt.  (The first function returns only the list
3068    of symbols that are really bound, the second prints those but also the
3069    ones that CM's autoloading mechanism knows about.)
3070    
3071    ----------------------------------------------------------------------
3072    Name: Matthias Blume
3073    Date: 2002/02/15 12:08:00 EST
3074    Tag: blume-20020215-iptrs
3075    Description:
3076    
3077    Two improvements to ml-nlffigen:
3078    
3079      1. Write files only if they do not exist or if their current contents
3080         do not coincide with what's being written.  (That is, avoid messing
3081         with the time stamps unless absolutely necessary.)
3082    
3083      2. Implement a "repository" mechanism for generated files related
3084         to "incomplete pointer types".   See the README file for details.
3085    
3086    ----------------------------------------------------------------------
3087    Name: Matthias Blume
3088    Date: 2002/02/14 11:50:00 EST
3089    Tag: blume-20020214-quote
3090    Description:
3091    
3092    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3093    because of the new and improved tag generation scheme.  (Thanks to Allen
3094    Leung for pointing it out.)
3095    
3096    ----------------------------------------------------------------------
3097    Name: Lal George
3098    Date: 2002/02/14 09:55:27 EST 2002
3099    Tag: george-20020214-isabelle-bug
3100    Description:
3101    
3102    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3103    of Isabelle on the x86.
3104    
3105    From Allen:
3106    -----------
3107     I've found the problem:
3108    
3109         in ra-core.sml, I use the counter "blocked" to keep track of the
3110         true number of elements in the freeze queue.  When the counter goes
3111         to zero, I skip examining the queue.  But I've messed up the
3112         bookkeeping in combine():
3113    
3114             else ();
3115             case !ucol of
3116               PSEUDO => (if !cntv > 0 then
3117                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3118                                        ^^^^^^^^^^^^^^^^^^^^^^^
3119                      moveu := mergeMoveList(!movev, !moveu)
3120                     )
3121                  else ();
3122    
3123         combine() is called to coalesce two nodes u and v.
3124         I think I was thinking that if the move counts of u and v are both
3125         greater than zero then after they are coalesced then one node is
3126         removed from the freeze queue.  Apparently I was thinking that
3127         both u and v are of low degree, but that's clearly not necessarily true.
3128    
3129    
3130    02/12/2002:
3131        Here's the patch.  HOL now compiles.
3132    
3133        I don't know how this impact on performance (compile
3134        time or runtime).  This bug caused the RA (especially on the x86)
3135        to go thru the potential spill phase when there are still nodes on the
3136        freeze queue.
3137    
3138    
3139    
3140    
3141    ----------------------------------------------------------------------
3142    Name: Matthias Blume
3143    Date: 2002/02/13 22:40:00 EST
3144    Tag: blume-20020213-fptr-rtti
3145    Description:
3146    
3147    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3148    updates.
3149    
3150    ----------------------------------------------------------------------
3151    Name: Matthias Blume
3152    Date: 2002/02/13 16:41:00 EST
3153    Tag: blume-20020213-cmlpq
3154    Description:
3155    
3156    Added new priority queue export symbols (which have just been added to
3157    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3158    not compile and the installer would choke.)
3159    
3160    ----------------------------------------------------------------------
3161    Name: Matthias Blume
3162    Date: 2002/02/13 16:15:00 EST
3163    Tag: blume-20020213-various
3164    Description:
3165    
3166    1. More tweaks to ml-nlffigen:
3167    
3168       - better internal datastructures (resulting in slight speedup)
3169       - "-match" option requires exact match
3170       - "localized" gensym counters (untagged structs/unions nested within
3171         other structs/unions or within typedefs get a fresh counter; their
3172         tag will be prefixed by a concatenation of their parents' tags)
3173       - bug fixes (related to calculation of transitive closure of types
3174         to be included in the output)
3175    
3176    2. Minor Basis updates:
3177    
3178       - added implementations for List.collate and Option.app
3179    
3180    ----------------------------------------------------------------------
3181    Name: Matthias Blume
3182    Date: 2002/02/11 15:55:00 EST
3183    Tag: blume-20020211-gensym
3184    Description:
3185    
3186    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3187    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3188    names (ML structure names that correspond to unnamed C structs, unions,
3189    and enums), so that separate runs of ml-nlffigen do not clash.
3190    
3191    ----------------------------------------------------------------------
3192    Name: Matthias Blume
3193    Date: 2002/02/11 12:05:00 EST
3194    Tag: blume-20020211-gensml
3195    Description:
3196    
3197    A quick fix for a problem with GenSML (in the pgraph-util library):
3198    Make generation of toplevel "local" optional.  (Strictly speaking,
3199    signature definitions within "local" are not legal SML.)
3200    
3201    Other than that: updates to INSTALL and cm/TODO.
3202    
3203    ----------------------------------------------------------------------
3204    Name: Matthias Blume
3205    Date: 2002/02/08 15:00:00 EST
3206    Tag: blume-20020208-uniquepid
3207    Description:
3208    
3209    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3210    
3211    1. The installer (config/install.sh) has gotten smarter:
3212    
3213         - Configuration options are a bit easier to specify now
3214           (in config/targets).
3215         - Bug in recognizing .tar.bz2 files fixed.
3216         - Installer automatically resolves dependencies between
3217           configuration options (e.g., if you ask for eXene, you will
3218           also get cml -- regardless whether you asked for it or not).
3219         - Installer can run in "quieter mode" by setting the environment
3220           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3221           "completely silent", though.
3222         - Build HashCons library as part of smlnj-lib.
3223    
3224    2. A new scheme for assigning persistent identifiers to compilation
3225       units (and, by extension, to types etc.) has been put into place.
3226       This fixes a long-standing bug where types and even dynamic values
3227       can get internally confused, thereby compromising type safety
3228       (abstraction) and dynamic correctness.  See
3229    
3230         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3231    
3232       for an example of how things could go wrong until now.
3233    
3234       The downside of the new scheme is that pids are not quite as
3235       persistent as they used to be: CM will generate a fresh pid
3236       for every compilation unit that it thinks it sees for the first
3237       time.  That means that if you compile starting from a clean, fresh
3238       source tree at two different times, you end up with different
3239       binaries.
3240    
3241       Cutoff recompilation, however, has not been compromised because
3242       CM keeps pid information in special caches between runs.
3243    
3244    ----------------------------------------------------------------------
3245    Name: Lal George
3246    Date: 2002/02/07 15:34:13 EST 2002
3247    Tag: <none>
3248    Description:
3249    
3250    Compilers that generate assembly code may produce  global labels
3251    whose value is resolved at link time. The various peephole optimization
3252    modules did not take this in account.
3253    
3254    TODO. The Labels.addrOf function should really return an option
3255    type so that clients are forced to deal with this issue, rather
3256    than an exception being raised.
3257    
3258    ----------------------------------------------------------------------
3259    Name: Lal George
3260    Date: 2002/02/06 13:55:02 EST
3261    Tag: george-20020206-ra-breakup
3262    Description:
3263    
3264    1. A bug fix from Allen.
3265    
3266        A typo causes extra fstp %st(0)'s to be generated at compensation
3267        edges, which might cause stack underflow traps at runtime.  This
3268        occurs in fft where there are extraneous fstps right before the 'into'
3269        trap instruction (in this case they are harmless since none of the
3270        integers overflow.)
3271    
3272    2. Pulled out various utility modules that were embedded in the modules
3273       of the register allocator. I need these modules for other purposes, but
3274       they are not complete enough to put into a library (just yet).
3275    ----------------------------------------------------------------------
3276    Name: Matthias Blume
3277    Date: 2002/01/31 16:05:00 EST
3278    Tag: blume-20020131-sparc-ccalls
3279    Description:
3280    
3281    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3282       of extra stack space by mistake.  Fixed.
3283    
3284    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3285    
3286    ----------------------------------------------------------------------
3287    Name: Allen Leung
3288    Date: 2002/01/30
3289    Tag: leunga-20020130-nowhere-bug-fix
3290    Description:
3291    
3292       MLRISC bug fixes:
3293       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3294       2. MachineInt.fromString was negating its value.
3295    
3296    ----------------------------------------------------------------------
3297    Name: Matthias Blume
3298    Date: 2002/01/29
3299    Tag: blume-20020129-INSTALL
3300    Description:
3301    
3302    - Added somewhat detailed installation instructions (file INSTALL).
3303    - Fixed curl-detection bug in config/install.sh.
3304    - It is now possible to select the URL getter using the URLGETTER
3305      environment variable:
3306    
3307          not set / "unknown"      --> automatic detection (script tries wget,
3308                                       curl, and lynx)
3309          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3310                                       how to properly invoke them)
3311          other                    --> use $URLGETTER directly, it must take
3312                                       precisely two command-line arguments
3313                                       (source URL and destination file name)
3314    
3315    ----------------------------------------------------------------------
3316    Name: Matthias Blume
3317    Date: 2002/01/28
3318    Tag: blume-20020128-sparc-ccalls
3319    Description:
3320    
3321    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3322    - Make use of the allocParam argument in sparc-c-calls.
3323    
3324    ----------------------------------------------------------------------
3325    Name: Matthias Blume
3326    Date: 2002/01/28
3327    Tag: blume-20020128-allocParam
3328    Description:
3329    
3330    John Reppy:  Changes c-calls API to accept client-callback for
3331    allocating extra stack space.
3332    me: Corresponding changes to mlriscGen (using a dummy argument that
3333        does not change the current behavior).
3334    
3335    ----------------------------------------------------------------------
3336    Name: Matthias Blume
3337    Date: 2002/01/28 12:00:00
3338    Tag: Release_110_38
3339    Description:
3340    
3341    This time for real!!!
3342    
3343    ----------------------------------------------------------------------
3344    Name: Matthias Blume
3345    Date: 2002/01/28 10:56:00 EST
3346    Tag: blume-20020128-retraction
3347    Description:
3348    
3349    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3350       with blume-Release_110_38-retracted.)
3351    
3352    1. Fixed a problem with incorrect rounding modes in real64.sml.
3353       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3354    
3355    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3356       and enums fixed.  The naming of corresponding ML identifiers should
3357       now be consistent again.
3358    
3359    ----------------------------------------------------------------------
3360    Name: Allen Leung
3361    Date: 2002/01/27
3362    Tag: leunga-20020127-nowhere
3363    Description:
3364    
3365       Added a target called nowhere in the configuration scripts.
3366       Enabling this will build the MLRISC 'nowhere' tool (for translating
3367       programs with where-clauses into legal SML code) during installation.
3368    
3369    ----------------------------------------------------------------------
3370    Name: Matthias Blume
3371    Date: 2002/01/25 21:27:00 EST
3372    Tag: blume-Release_110_38-retracted
3373    Description:
3374    
3375    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3376    
3377    README will be added later.
3378    
3379    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3380    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3381    
3382    ----------------------------------------------------------------------
3383    Name: Matthias Blume
3384    Date: 2002/01/25
3385    Tag: blume-20020125-ffi
3386    Description:
3387    
3388    A large number of tweaks and improvements to ml-nlffi-lib and
3389    ml-nlffigen:
3390    
3391       - ML represenation types have been streamlined
3392       - getter and setter functions work with concrete values, not abstract
3393         ones where possible
3394       - ml-nlffigen command line more flexible (see README file there)
3395       - some bugs have been fixed (hopefully)
3396    
3397    ----------------------------------------------------------------------
3398    Name: Lal George
3399    Date: 2002/01/24
3400    Tag: george-20020124-risc-ra-interface
3401    Description:
3402    
3403       There is a dramatic simplification in the interface to the
3404       register allocator for RISC architectures as a result of making
3405       parallel copy instructions explicit.
3406    
3407    ----------------------------------------------------------------------
3408    Name: Matthias Blume
3409    Date: 2002/01/22
3410    Tag: blume-20020122-x86-ccalls
3411    Description:
3412    
3413    Bug fix for c-calls on x86 (having to do with how char- and
3414    short-arguments are being handled).
3415    
3416    ----------------------------------------------------------------------
3417    Name: Matthias Blume
3418    Date: 2002/01/21
3419    Tag: blume-20020121-ff
3420    Description:
3421    
3422    Another day of fiddling with the FFI...
3423    
3424    1. Bug fix/workaround:  CKIT does not complain about negative array
3425       dimensions, so ml-nlffigen has to guard itself against this possibility.
3426       (Otherwise a negative dimension would send it into an infinite loop.)
3427    
3428    2. Some of the abstract types (light objects, light pointers, most "base"
3429       types) in structure C are now eqtypes.
3430    
3431    3. Added constructors and test functions for NULL function pointers.
3432    
3433    ----------------------------------------------------------------------
3434    Name: Matthias Blume
3435    Date: 2002/01/18
3436    Tag: blume-20020118-ready-for-new-release
3437    Description:
3438    
3439    Made config/srcarchiveurl point to a new place.  (Will provide boot
3440    files shortly.)
3441    
3442    Maybe we christen this to be 110.38?
3443    
3444    ----------------------------------------------------------------------
3445    Name: Matthias Blume
3446    Date: 2002/01/18
3447    Tag: blume-20020118-more-ffifiddle
3448    Description:
3449    
3450    Today's FFI fiddling:
3451    
3452      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3453        C.Set.* that use concrete (MLRep.*) arguments and results instead
3454        of abstract ones.
3455    
3456      - Provided word-style bit operations etc. for "int" representation
3457        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3458        and Long.
3459    
3460    ----------------------------------------------------------------------
3461    Name: Matthias Blume
3462    Date: 2002/01/18
3463    Tag: blume-20020118-use-x86-fp
3464    Description:
3465    
3466    Now that x86-fast-fp seems to be working, I turned it back on again
3467    by default.  (Seems to work fine now, even with the FFI.)
3468    
3469    Other than that, I added some documentation about the FFI to
3470    src/ml-nlffigen/README and updated the FFI test examples in
3471    src/ml-nlffi-lib/Tests/*.
3472    
3473    ----------------------------------------------------------------------
3474    Name: Allen Leung
3475    Date: 2002/01/17
3476    Tag: leunga-20020117-x86-fast-fp-call
3477    Description:
3478    
3479       1. Fixed a problem with handling return fp values when x86's fast fp
3480          mode is turned on.
3481    
3482       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3483          of %f32.
3484    
3485       3. Added a constructor INT32lit to the ast of MLRISC tools.
3486    
3487    ----------------------------------------------------------------------
3488    Name: Matthias Blume
3489    Date: 2002/01/16
3490    Tag: blume-20020116-ffifiddle
3491    Description:
3492    
3493    More fiddling with the FFI interface:
3494    
3495     - Make constness 'c instead of rw wherever possible.  This eliminates
3496       the need for certain explicit coercions.  (However, due to ML's
3497       value polymorphism, there will still be many cases where explicit
3498       coercions are necessary.  Phantom types are not the whole answer
3499       to modeling a subtyping relationship in ML.)
3500    
3501     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3502    
3503     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3504       clumsy workaround for varargs problem.)
3505    
3506    ----------------------------------------------------------------------
3507    Name: Lal George
3508    Date: 2002/01/15
3509    Tag: <none>
3510    Description:
3511    
3512    1. Since COPY instructions are no longer native to the architecture,
3513       a generic functor can be used to implement the expandCopies function.
3514    
3515    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3516       TEXT segment.
3517    
3518    ----------------------------------------------------------------------
3519    Name: Matthias Blume
3520    Date: 2002/01/15
3521    Tag: blume-20020115-ffiupdates
3522    Description:
3523    
3524    1. Fix for bug resulting in single-precision float values being returned
3525       incorrectly from FFI calls.
3526    
3527    2. Small modifications to C FFI API:
3528    
3529        - memory-allocation routines return straight objects (no options)
3530          and raise an exception in out-of-memory situations
3531        - unsafe extensions to cast between function pointers and pointers
3532          from/to ints
3533        - added structure C_Debug as an alternative to structure C where
3534          pointer-dereferencing (|*| and |*!) always check for null-pointers
3535        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3536          but also takes a (possibly empty) list of existing library handles
3537          that the current library depends on
3538    
3539    ----------------------------------------------------------------------
3540    Name: Matthias Blume
3541    Date: 2002/01/10
3542    Tag: blume-20020110-newffigen
3543    Description:
3544    
3545    1. Updates to portable graph code.
3546    
3547    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3548       more scalable now so that even huge interfaces such as the one
3549       for GTK compile in finite time and space. :-)
3550       See src/ml-nlffigen/README for details on what's new.
3551    
3552    ----------------------------------------------------------------------
3553    Name: Lal George
3554    Date: 2001/01/09 14:31:35 EST 2002
3555    Tag: george-20011206-rm-native-copy
3556    Description:
3557    
3558            Removed the native COPY and FCOPY instructions
3559            from all the architectures and replaced it with the
3560            explicit COPY instruction from the previous commit.
3561    
3562            It is now possible to simplify many of the optimizations
3563            modules that manipulate copies. This has not been
3564            done in this change.
3565    
3566    ----------------------------------------------------------------------
3567    Name: Lal George
3568    Date: 2001/12/06 16:50:13 EST 2001
3569    Tag: george-20011206-mlrisc-instruction
3570    Description:
3571    
3572    Changed the representation of instructions from being fully abstract
3573    to being partially concrete. That is to say:
3574    
3575      from
3576            type instruction
3577    
3578      to
3579            type instr                              (* machine instruction *)
3580    
3581            datatype instruction =
3582                LIVE of {regs: C.cellset, spilled: C.cellset}
3583              | KILL of {regs: C.cellset, spilled: C.cellset}
3584              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3585              | ANNOTATION of {i: instruction, a: Annotations.annotation}
3586              | INSTR of instr
3587    
3588    This makes the handling of certain special instructions that appear on
3589    all architectures easier and uniform.
3590    
3591    LIVE and KILL say that a list of registers are live or killed at the
3592    program point where they appear. No spill code is generated when an
3593    element of the 'regs' field is spilled, but the register is moved to
3594    the 'spilled' (which is present, more for debugging than anything else).
3595    
3596    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3597    We used to generate:
3598    
3599            DEFFREG f1
3600            f1 := f2 + f3
3601            trapb
3602    
3603    but now generate:
3604    
3605            f1 := f2 + f3
3606            trapb
3607            LIVE {regs=[f1,f2,f3], spilled=[]}
3608    
3609    Furthermore, the DEFFREG (hack) required that all floating point instruction
3610    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3611    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3612    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3613    this architecture).
3614    
3615    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3616    all the architectures. This will result in further simplification of the
3617    register allocator that must be aware of them for coalescing purposes, and
3618    will also simplify certain aspects of the machine description that provides
3619    callbacks related to parallel copies.