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

SCM Repository

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

Diff of /sml/trunk/NOTES/HISTORY

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

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