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

SCM Repository

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

Diff of /sml/trunk/HISTORY

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

revision 666, Fri Jun 16 08:27:00 2000 UTC revision 1375, Mon Sep 15 22:44:08 2003 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2003/09/15 17:45:00 CDT
18    Tag: blume-20030915-rbase
19    Description:
20    
21    Fixed bug in Real.fromLargeInt.
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume (blume (at) tti - c (dot) org)
25    Date: 2003/09/13 18:11:00 CDT
26    Tag: blume-20030913-libinstall
27    Description:
28    
29    Minor bugfix in config/libinstall (set anchor with path to
30    standalone tool after installing it, otherwise libraries that
31    need ml-lex or ml-yacc won't compile the first time the installer
32    runs).
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume (blume (at) tti - c (dot) org)
36    Date: 2003/09/12 11:45:00 CDT
37    Tag: blume-20030912-various
38    Description:
39    
40     - fixed bug in Real.toLargeInt
41     - fixed bug in Posix.ProcEnv.times
42     - changed inputLine functions to return an option
43     - minor installer improvements / bugfixes
44     - changed default @SMLalloc parameter for x86/celeron to 64k
45    
46    ----------------------------------------------------------------------
47    Name: Matthias Blume (blume (at) tti - c (dot) org)
48    Date: 2003/09/09 22:00:00 CDT
49    Tag: Release_110_43
50    Description:
51    
52        New working release 110.43.  New bootfiles.
53    
54    ----------------------------------------------------------------------
55    Name: Matthias Blume (blume (at) tti - c (dot) org)
56    Date: 2003/09/09 19:20:00 CDT
57    Tag: blume-20030909-installer
58    Description:
59    
60    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
61    Modified config/install.bat to take advantage of it.  Also modified
62    config/install.sh (and called it config/new-install.sh) to take advantage
63    of it on Unix systems.  (The SML code is (supposed to be) platform-
64    independent.)
65    
66    The installer can now install everything under Win32
67    as well as under *nix as long as it compiles.
68    
69    Other changes:
70    
71      - made CML compile again under Win32
72      - made eXene compile under Win32 (by providing a fake structure UnixSock
73        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
74      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
75        same as type int;  under Win32 it isn't
76      - fixed some slice-related problems in the win32-specific parts of CML
77      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
78        (under Win32, the volume name is case-insensitive, and the
79        OS.Path code compares volume names for equality)
80    
81    ----------------------------------------------------------------------
82    Name: Matthias Blume (blume (at) tti - c (dot) org)
83    Date: 2003/09/08 11:55:00 CDT
84    Tag: blume-20030908-fullpath
85    Description:
86    
87    Made Win32 version of OS.FileSys.fullPath return current directory
88    when given an empty string.  This is what the spec says, and incidentally,
89    CM depends on it.  (CM otherwise goes into an infinite loop in certain
90    cases when presented with the name of a non-existing .cm file.)
91    
92    ----------------------------------------------------------------------
93    Name: Matthias Blume (blume (at) tti - c (dot) org)
94    Date: 2003/09/04 16:30:00 CDT
95    Tag: blume-20030905-slices-etc
96    Description:
97    
98      1. Changed interface to vectors and arrays in Basis to match
99         (draft) Basis spec.
100      2. Added signatures and implementations of slices according to
101         Basis spec.
102      3. Edited source code throughout the system to make it compile again
103         under 1. and 2.  (In some cases code had to be added to have it
104         match the new signatures.)
105      4. MLRISC should be backward-compatible: the copies of the originals
106         of files that needed to change under 3. were retained, the .cm files
107         check the compiler version number and use old versions when
108         appropriate.
109      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
110         dirstream -> string option (in accordance with Basis spec).
111      6. When generating code that counts lines, ml-lex used function
112         CharVector.foldli, taking advantage of its old interface.
113         This has been replaced with the corresponding code from
114         CharVectorSlice.  (html-lex must be re-lexed!)
115      7. BitArray in smlnj-lib/Util has been extended/modified to match the
116         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
117      8. Removed temporary additions (fromInternal, toInternal) from the
118         (now obsolete) IntInf in smlnj-lib/Util.
119      9. Cleaned up structure Byte.
120     10. Added localOffset, scan, and fromString to Date (according to spec).
121         Cleaned/corrected implementation of Date.
122         (Still need to check for correctness; implement better canonicalizeDate.)
123     11. Added "scan" to signature IEEE_REAL.
124     12. Some improvements to IntInf [in particular: efficiency-hack for
125         mod and rem when second operand is 2 (for parity checks).]
126     13. Changed representation of type Time.time, using a single IntInf.int
127         value counting microseconds.  This considerably simplified the
128         implementation of structure Time.  We now support negative time
129         values; scan and fromString handle signs.
130     14. Functor PrimIO now takes two additional arguments (VectorSlice and
131         ArraySlice).
132    
133    ----------------------------------------------------------------------
134    Name: Matthias Blume (blume (at) tti - c (dot) org)
135    Date: 2003/08/28 17:00:00 CDT
136    Tag: blume-20030828-intinf
137    Description:
138    
139    This is a major update which comes with a version number bump
140    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
141    and an implementation of IntInf in the Basis.
142    
143    There are a fairly large number of related changes and updates throughout
144    the system:
145    
146    Basis:
147            - Implemented IntInf.
148            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
149            - Added some missing Real64 operations, most notably Real.toLargeInt.
150            - Added FixedInt as a synonym for Int32.
151    
152    compiler:
153            * Added support for a built-in intinf type.
154              - literals
155              - pattern matching
156              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
157              - overloading on literals and operations
158    
159            This required adding a primitive type intinf, some additional
160            primops, and implementations for several non-trivial intinf
161            operations in Core.  (The intinf type is completely abstract
162            to the compiler; all operations get delegated back to the Core.)
163    
164            * Intinf equality is handled by polyequal.  However, the compiler
165            does not print its usual warning in this case (since polyequal
166            is the right thing to do there).
167    
168            * Improved the organization of structure InlineT.
169    
170            * A word about conversion primops:
171            If conversions involving intinf do not cancel out during
172            CPS contract, then the compiler must insert calls to Core functions.
173            Since all core access must be resolved already during the FLINT
174            translate phase, it would be too late a the time of CPS contract
175            to add new Core calls.  For this reason, conversion primops
176            for intinf carry two arguments:  1. the numeric argument that
177            they are supposed to convert, and 2. the Core function that
178            can help with this conversion if necessary.  If CPS contract
179            eliminates a primop, then the associated Core function becomes
180            dead and goes away.  Intinf conversion primops that do not get
181            eliminated by CPS contract get rewritten into calls of their
182            core functions by a separate, new phase.
183    
184    interactive system:
185            - Control.Print.intinfDepth controls max length of intinf constants
186            being printed. (Analogous to Control.Print.stringDepth.)
187            - Cleanup in printutil and pputil: got rid of unused stuff and
188            duplicates; replaced some of the code with code that makes better
189            use of library functionality.
190    
191    CM:
192            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
193            are no longer silent.
194    
195    CKIT:
196            Fixed mismatched uses of Int32 and LargeInt.  I always decided
197            in favor of LargeInt -- which is now the same as IntInf.
198            CKIT-knowledgable people should check whether this is what's
199            intended and otherwise change things back to using Int32 or
200            FixedInt.
201    
202    Throughout the code:
203            Started using IntInf.int literals and built-in operations
204            (e.g., comparison with 0) where this seems appropriate.
205    
206    
207    ----------------------------------------------------------------------
208    Name: Dave MacQueen (dbm@cs.uchicago.edu)
209    Date: 2003/08/13 11:36:00 CDT
210    Tag: dbm-20030813-mcz-merge1
211    Description:
212    
213    Merging changes from the mcz-branch development branch into trunk.
214    These changes involve replacement of the emulated old prettyprinter
215    interface with direct use of the SML/NJ Lib PP library, and fixing
216    of a couple of bugs (895, 1186) relating to error messages.  A new
217    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
218    has been added.
219    
220    ----------------------------------------------------------------------
221    Name: Matthias Blume (blume (at) tti - c (dot) org)
222    Date: 2003/08/11 15:45:00 CDT
223    Tag: blume-20030811-windows
224    Description:
225    
226    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
227    
228        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
229    
230    This patch restores SML/NJ's ability to run under win32.  There are a
231    number of changes, including fixes for several bugs that had gone
232    unnoticed until now:
233    
234      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
235        trying to run the win32 version from within cygwin.)
236    
237      - There are a number of simple .bat scripts that substitute for their
238        corresponding Unix shell-scripts.  (See below.)
239    
240      - The internals of ml-build have been modified slightly.  The main
241        difference is that instead of calling ".link-sml" (or link-sml.bat)
242        using OS.Process.system, the ML process delegates this task back
243        to the script.  Otherwise problems arise in mixed environments such
244        as Cygwin where scripts look and work like Unix scripts, but
245        where OS.Process.system cannot run them.
246    
247      - In CM, the srcpath pickler used native pathname syntax -- which
248        is incorrect in the case of cross-compilation.  The new pickle format
249        is independent of platform-specific naming conventions.
250    
251      - Path configuration files (such as lib/pathconfig) can now choose
252        between native and standard syntax.  Placing a line of the form
253    
254            standard!
255    
256        into the file causes all subsequent paths to be interpreted using
257        CM standard pathname syntax (= Unix conventions); a line
258    
259            native!
260    
261        switches back to native style.  This was needed so that
262        path config files can be written portably, see src/system/pathconfig.
263    
264      - Runtime system:
265    
266          - win32-filesys.c:  get_file_time and set_file_time now
267            access modification time, not creation time.
268    
269          - I/O code made aware of new array representation.
270    
271          - Bug fixes in X86.prim.masm.
272    
273       - src/system/makeml made aware of win32. (For use under cygwin
274            and other Unix-environments for windows.)
275    
276       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
277         which caused BinIO.inputAll to fail consistently.
278    
279    .bat scripts:
280    
281       Windows .bat scripts assume that SMLNJ_HOME is defined.
282    
283       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
284         applications (sml, ml-yacc, ml-lex).
285       - ml-build.bat: analogous to ml-build.
286       - config\install.bat:  Analogous to config/install.sh.  This requires
287         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
288         (nmake etc. must be on the path, and vcvars32 must have been run.)
289         Moreover, sources for ml-lex and ml-yacc need to exist under src,
290         and the bootfile hierarchy must have been unpacked under
291         sml.boot.x86-win32.
292         The script is very primitive and does a poor job at error checking.
293         It only installs the base system, ml-lex, and ml-yacc.  No other
294         libraries are being installed (i.e., you get only those that
295         are part of the compiler.)
296       - link-sml.bat:  analogous to .link-sml, but not currently used
297    
298    Unrelated bug fixes:
299    
300      - ml-nlffigen now exports structures ST_* corresponding to incomplete
301        types.
302      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
303        otherwise.)
304    
305    ----------------------------------------------------------------------
306    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
307    Date: 2003/06/17
308    Tag: macqueen-20030617-bug895
309    Description:
310    
311    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
312    Tag will be used for new development branch (mcz-branch) for
313    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
314    
315    ----------------------------------------------------------------------
316    Name: Matthias Blume (blume (at) tti - c (dot) org)
317    Date: 2003/05/27 16:55:00 CDT
318    Tag: blume-20030527-polyeq
319    Description:
320    
321    Tried to eliminated most cases of polymorphic equality.
322    
323    ----------------------------------------------------------------------
324    Name: Matthias Blume (blume (at) tti - c (dot) org)
325    Date: 2003/05/21 17:45:00 CDT
326    Tag: blume-20030517-complete
327    Description:
328    
329    Two changes:
330    
331      1. Added a flag for controlling whether non-exhaustive bindings will
332         be treated as errors (default is false).
333      2. Cleaned up the *entire* source tree so that CMB.make goes through
334         without a single non-exhaustive match- or bind warning.
335    
336    ----------------------------------------------------------------------
337    Name: Matthias Blume (blume (at) tti - c (dot) org)
338    Date: 2003/05/17 10:20:00 CDT
339    Tag: blume-20030517-absyn
340    Description:
341    
342    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
343    
344       This mainly affects the quality of error messages.  However, some
345       of the code is now more straightforward than before. (Treatment of
346       the above four constructs in translate.sml is much simpler than
347       the "macro-expansion" that was going on before.  Plus, the mach-
348       compiler no longer gets invoked just to be able to compile an
349       if-expression.)
350    
351    2. The ErrorMsg.Error exception is now caught and absorbed by the
352       interactive loop.
353    
354    ----------------------------------------------------------------------
355    Name: Allen Leung
356    Date: 2003/05/16 13:05:00 CDT
357    Tag: leunga-20030516-cygwin-runtime
358    Description:
359    
360        Ported the runtime system to cygwin, which uses the unix
361    x86-unix bin files.  Missing/buggy features:
362    
363       o getnetbyname, getnetbyaddr: these functions seem to be missing in
364         the Cygwin library.
365       o Ctrl-C handling may be flaky.
366       o Windows system calls and Windows I/O are not supported.
367    
368       A new set of binfiles is located at:
369    
370         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
371    
372       This is only needed for bootstrapping the cygwin version of smlnj.
373       Other x86 versions can use the existing binfiles.
374    
375    ----------------------------------------------------------------------
376    Name: Matthias Blume
377    Date: 2003/04/08 15:42:00 CDT
378    Tag: blume-20030408-listpair
379    Description:
380    
381    1. Added a target 'mlrisc' to installer.
382    
383    2. Added missing elements to structure ListPair.
384    
385    ----------------------------------------------------------------------
386    Name: Allen Leung
387    Date: 2003/01/07 10:40:00 EST
388    Tag: leunga-20030107-int-rem
389    Description:
390    
391    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
392    The arguments to the SUBL instruction were swapped.
393    
394    ----------------------------------------------------------------------
395    Name: Matthias Blume
396    Date: 2002/12/12 16:25:00 EST
397    Tag: blume-20021212-risc-ra
398    Description:
399    
400    Fixed a serious bug in the rewrite code for FP spilling/reloading that
401    sent the RA into an infinite loop when floating point registers get
402    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
403    110.37 and 110.38.)
404    There was another set of potential problems related to the handling of
405    MLRISC annotations (but those did not yet cause real problems, apparently).
406    
407    ----------------------------------------------------------------------
408    Name: Matthias Blume
409    Date: 2002/12/06 22:40:00 EST
410    Tag: blume-20021206-cm-fileid
411    Description:
412    
413    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
414    This fixes the problem of CM getting confused by files that suddenly
415    change their identity (e.g., by getting unlinked and recreated by some
416    text editor such as vi).  There might be a better/cheaper/cleaner way
417    of doing this, but for now this will have to do.
418    
419    ----------------------------------------------------------------------
420    Name: Matthias Blume
421    Date: 2002/10/28 09:50:00 EST
422    Tag: blume-20021028-typecheck
423    Description:
424    
425    Exported structure Typecheck from $smlnj/viscomp/core.cm.
426    
427    ----------------------------------------------------------------------
428    Name: Matthias Blume
429    Date: 2002/10/17 09:10:00 EDT
430    Tag: Release_110_42
431    Description:
432    
433    In good old tradition, there has been a slight hiccup so that we have
434    to patch 110.42 after the fact.   The old release tag has been replaced
435    (see below).
436    
437    The change solves a problem with two competing approaches the
438    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
439    each other's way.
440    
441    This change only affects the runtime system code and the installer script.
442    (No new bootfiles.)
443    
444    ----------------------------------------------------------------------
445    Name: Matthias Blume
446    Date: 2002/10/16 12:00:00 EDT
447    Tag: Release_110_42_removed
448    Description:
449    
450    New working release.  New bootfiles.
451    
452    ----------------------------------------------------------------------
453    Name: Matthias Blume
454    Date: 2002/10/10 13:10:00 EDT
455    Tag: blume-20021010-ppc-divs
456    Description:
457    
458    The mltree operator DIVS must be implemented with an overflow check on
459    the PPC because the hardware indicates divide-by-zero using "overflow" as
460    well.
461    
462    ----------------------------------------------------------------------
463    Name: Matthias Blume
464    Date: 2002/07/23 11:45:00
465    Tag: blume-20020723-smlnj-home
466    Description:
467    
468    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
469    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
470    CM_PATHCONFIG is also set), the path configuration file is assumed
471    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
472    the entire tree to some other place and everything will "just work".
473    
474    (Companion commands such as ml-build and ml-makedepend also sense this
475    variable.)
476    
477    ----------------------------------------------------------------------
478    Name: Matthias Blume
479    Date: 2002/07/12 21:19:00 EDT
480    Tag: blume-20020712-liveness
481    Description:
482    
483    Exported two useful "step" functions from liveness module (MLRISC).
484    
485    ----------------------------------------------------------------------
486    Name: Matthias Blume
487    Date: 2002/07/05 16:00 EDT
488    Tag: Release_110_41
489    Description:
490    
491    New working release.  New bootfiles.
492    
493    ----------------------------------------------------------------------
494    Name: Matthias Blume
495    Date: 2002/07/05 10:25:00 EDT
496    Tag: blume-20020705-btimp
497    Description:
498    
499    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
500    other clients can set up backtracing support.
501    
502    ----------------------------------------------------------------------
503    Name: Matthias Blume
504    Date: 2002/06/25 17:23:00 EDT
505    Tag: blume-20020625-fpmax
506    Description:
507    
508    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
509    case.  (The sense of the isNaN test was reversed -- which made min and
510    max always return their first argument.)
511    
512    ----------------------------------------------------------------------
513    Name: Matthias Blume
514    Date: 2002/06/11
515    Tag: blume-20020611-unixpath
516    Description:
517    
518    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
519    
520    ----------------------------------------------------------------------
521    Name: Matthias Blume
522    Date: 2002/06/10 16:35:00 EDT
523    Tag: blume-20020610-ieeereal
524    Description:
525    
526    I back-ported my implementation of IEEEReal.fromString from the
527    idlbasis-devel branch so that we can test it.
528    
529    Another small change is that ppDec tries to give more information
530    than just "<sig>" in the case of functors.  However, this code is
531    broken in some mysterious way if the functor's body's signature
532    has not been declared by ascription but gets inferred from the
533    implementation.  This needs fixing...
534    
535    ----------------------------------------------------------------------
536    Name: Matthias Blume
537    Date: 2002/05/31
538    Tag: blume-20020531-btrace-mode
539    Description:
540    
541    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
542    the wayside when I switched over to using Controls everywhere.)
543    
544    ----------------------------------------------------------------------
545    Name: Lal George
546    Date: 2002/05/23 12:21:40 EDT
547    Tag: george-20020523-visual-labels
548    Description:
549    
550    Labels are now displayed in the graphical output to make
551    the fall-through and target blocks obvious.
552    
553    ----------------------------------------------------------------------
554    Name: Matthias Blume
555    Date: 2002/05/22 11:03:00 EDT
556    Tag: blume-20020522-shrink
557    Description:
558    
559    John tweaked yesterday's fix for 1131 to handle an out-of-memory
560    situation that comes up when allocating huge arrays.
561    
562    ----------------------------------------------------------------------
563    Name: Matthias Blume
564    Date: 2002/05/21 16:00:00 EDT
565    Tag: Release_110_40
566    Description:
567    
568    New working release (110.40).  New bootfiles.
569    
570    [Also: John Reppy fixed GC bug 1131.]
571    
572    ----------------------------------------------------------------------
573    Name: Matthias Blume
574    Date: 2002/05/21 12:35:00 EDT
575    Tag: blume-20020521-cmdoc
576    Description:
577    
578    CM documentation update.
579    
580    ----------------------------------------------------------------------
581    Name: Matthias Blume
582    Date: 2002/05/21 10:55:00 EDT
583    Tag: blume-20020521-misc
584    Description:
585    
586    - John tweaked runtime to be silent on heap export (except when
587      GC messages are on).
588    - I added a few more things (cross-compiling versions of CMB) to
589      config/preloads (as suggestions).
590    
591    ----------------------------------------------------------------------
592    Name: Matthias Blume
593    Date: 2002/05/20 22:25:00 EDT
594    Tag: blume-20020520-controls
595    Description:
596    
597    - Added ControlUtil structure to control-lib.cm.
598    - Use it throughout.
599    - Used Controls facility to define MLRISC controls (as opposed to
600      registering MLRISC control ref cells with Controls after the fact)
601    - Fixed messed-up controls priorities.
602    
603    * Removed again all the stuff from config/preloads that one wouldn't
604      be able to preload at the time the initial heap image is built.
605      (Many libraries, e.g., CML, do not exist yet at this time.  The
606       only libraries that can be preloaded via config/preloads are those
607       that come bundled with the bootfiles.)
608    
609    ----------------------------------------------------------------------
610    Name: Matthias Blume
611    Date: 2002/05/20 10:59:00 EDT
612    Tag: blume-20020520-preloads
613    Description:
614    
615    Added a lot of commented-out suggestions for things to be included
616    in config/preloads.
617    
618    ----------------------------------------------------------------------
619    Name: Allen Leung
620    Date: 2002/05/18 14:20:00 EDT
621    Tag: leunga-20020518-mdl
622    Description:
623    
624    o Made the mdl tool stuff compile and run again.
625    o I've disabled all the stuff that depends on RTL specifications; they
626      are all badly broken anyway.
627    
628    ----------------------------------------------------------------------
629    Name: Matthias Blume
630    Date: 2002/05/17 16:49:00 EDT
631    Tag: blume-20020517-controls
632    Description:
633    
634    0. John Reppy made several modifications to the SML/NJ library.
635       In particular, there is a shiny new controls-lib.cm.
636    
637    1. Pushed new controls interface through compiler so that everything
638       compiles again.
639    
640    2. Added FormatComb and FORMAT_COMB to the CML version of the
641       SML/NJ library (so that CML compiles again).
642    
643    3. Modified init scripts because XXX_DEFAULT environment variables
644       are no longer with us.  (Boot-time initialization is now done
645       using the same environment variables that are also used for
646       startup-time initialization of controls.)
647    
648    ----------------------------------------------------------------------
649    Name: Lal George
650    Date: 2002/05/15 09:20:10 EDT
651    Tag: george-20020515-pseudo-op-decls
652    Description:
653    
654            All pseudo-ops emitted before the first segment declaration
655    such as TEXT, DATA, and BSS directives are assumed to be global
656    declarations and are emitted first in the assembly file. This is
657    useful in a number of situations where one has pseudo-ops that are not
658    specific to any segment, and also works around the constraint that one
659    cannot have client pseudo-ops in the TEXT segment.
660    
661            Because no segment is associated with these declarations it is
662    an error to allocate any space or objects before the first segment
663    directive and an exception will be raised. However, we cannot make
664    this check for client pseudo-ops.
665    
666            These top level declarations are a field in the CFG graph_info.
667    In theory you can continue to add to this field after the CFG has been
668    built -- provided you know what you are doing;-)
669    
670    ----------------------------------------------------------------------
671    Name: Matthias Blume
672    Date: 2002/05/13 16:40:00 EDT
673    Tag: blume-20020513-pp-etc
674    Description:
675    
676    A few minor bugfixes:
677    
678      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
679        (Bogus printouts for redefined bindings still occur.  Compiler
680        bug should no longer occur now.  We need to redo the prettyprinter
681        from scratch.)
682    
683      - CM pathname printer now also adds escape sequences for ( and )
684    
685      - commend and docu fixes for ml-nlffi
686    
687    ----------------------------------------------------------------------
688    Name: Matthias Blume
689    Date: 2002/05/10 16:40:00 EDT
690    Tag: blume-20020510-erg-textio
691    Description:
692    
693    Applied the following bugfix provided by Emden Gansner:
694    
695        Output is corrupted when outputSubstr is used rather than output.
696    
697        The problem occurs when a substring
698    
699            ss = (s, dataStart, dataLen)
700    
701        where dataStart > 0, fills a stream buffer with avail bytes left.
702        avail bytes of s, starting at index dataStart, are copied into the
703        buffer, the buffer is flushed, and then the remaining dataLen-avail
704        bytes of ss are copied into the beginning of the buffer. Instead of
705        starting this copy at index dataStart+avail in s, the current code
706        starts the copy at index avail.
707    
708        Fix:
709        In text-io-fn.sml, change line 695 from
710             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
711        to
712             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
713    
714    ----------------------------------------------------------------------
715    Name: Matthias Blume
716    Date: 2002/04/12 13:55:00 EDT
717    Tag: blume-20020412-assyntax
718    Description:
719    
720    1. Grabbed newer assyntax.h from the XFree86 project.
721    2. Fiddled with how to compile X86.prim.asm without warnings.
722    3. (Very) Minor cleanup in CM.
723    
724    ----------------------------------------------------------------------
725    Name: Matthias Blume
726    Date: 2002/04/01 (no joke!) 17:07:00 EST
727    Tag: blume-20020401-x86div
728    Description:
729    
730    Added full support for div/mod/rem/quot on the x86, using the machine
731    instruction's two results (without clumsily recomputing the remainder)
732    directly where appropriate.
733    
734    Some more extensive power-of-two support was added to the x86 instruction
735    selector (avoiding expensive divs, mods, and muls where they can be
736    replaced with cheaper shifts and masks).  However, this sort of thing
737    ought to be done earlier, e.g., within the CPS optimizer so that
738    all architectures benefit from it.
739    
740    The compiler compiles to a fixed point, but changes might be somewhat
741    fragile nevertheless.  Please, report any strange things that you might
742    see wrt. div/mod/quot/rem...
743    
744    ----------------------------------------------------------------------
745    Name: Matthias Blume
746    Date: 2002/03/29 17:22:00
747    Tag: blume-20020329-div
748    Description:
749    
750    Fixed my broken div/mod logic.  Unfortunately, this means that the
751    inline code for div/mod now has one more comparison than before.
752    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
753    The problem was with quotient = 0, because that alone does not tell
754    us which way the rounding went.  One then has to look at whether
755    remainder and divisor have the same sign...  :(
756    
757    Anyway, I replaced the bootfiles with fresh ones...
758    
759    ----------------------------------------------------------------------
760    Name: Matthias Blume
761    Date: 2002/03/29 14:10:00 EST
762    Tag: blume-20020329-inlprims
763    Description:
764    
765    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
766    
767    Primops have changed. This means that the bin/boot-file formats have
768    changed as well.
769    
770    To make sure that there is no confusion, I made a new version.
771    
772    
773    CHANGES:
774    
775    * removed REMT from mltree (remainder should never overflow).
776    
777    * added primops to deal with divisions of all flavors to the frontend
778    
779    * handled these primops all the way through so they map to their respective
780      MLRISC support
781    
782    * used these primops in the implementation of Int, Int32, Word, Word32
783    
784    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
785    
786    * parameterized INLMIN, INLMAX, and INLABS by a numkind
787    
788    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
789      floating point
790    
791    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
792      and Real (but Real.abs maps to a separate floating-point-only primop)
793    
794    
795    TODO items:
796    
797    * Hacked Alpha32 instruction selection, disabling the selection of REMx
798      instructions because the machine instruction encoder cannot handle
799      them.  (Hppa, PPC, and Sparc instruction selection did not handle
800      REM in the first place, and REM is supported by the x86 machine coder.)
801    
802    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
803      selection phase.  (The two can be streamlined because the hardware
804      delivers both quotient and remainder at the same time anyway.)
805    
806    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
807      (Currently the behavior is inconsistent both across architectures and
808      wrt. the draft Basis spec.)
809    
810    * Word8 should eventually be handled natively, too.
811    
812    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
813      as if there currently is no execution path that could trigger it in
814      SML/NJ.  (The assumptions underlying functions arith and promotable do not
815      hold for things like multiplication and division.)
816    
817    ----------------------------------------------------------------------
818    Name: Matthias Blume
819    Date: 2002/03/27 16:27:00 EST
820    Tag: blume-20020327-mlrisc-divisions
821    Description:
822    
823    Added support for all four division operations (ML's div, mod, quot,
824    and rem) to MLRISC.  In the course of doing so, I also rationalized
825    the naming (no more annoying switch-around of DIV and QUOT), by
826    parameterizing the operation by div_rounding_mode (which can be either
827    DIV_TO_ZERO or DIV_TO_NEGINF).
828    
829    The generic MLTreeGen functor takes care of compiling all four
830    operations down to only round-to-zero div.
831    
832    Missing pieces:
833    
834      * Doing something smarter than relying on MLTreeGen on architectures
835        like, e.g., the x86 where hardware division delivers both quotient and
836        remainder at the same time.  With this, the implementation of the
837        round-to-neginf operations could be further streamlined.
838    
839      * Remove inlining support for div/mod/rem from the frontend and replace it
840        with primops that get carried through to the backend.  Do this for all
841        int and word types.
842    
843    ----------------------------------------------------------------------
844    Name: Matthias Blume
845    Date: 2002/03/25 17:25:00 EST
846    Tag: blume-20020325-divmod
847    Description:
848    
849    I improved (hopefully without breaking them) the implementation of Int.div,
850    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
851    advantage of the following observations:
852    
853      Let  q = x quot y      r = x rem y
854           d = x div  y      m = x mod y
855    
856    where "quot" is the round-to-zero version of integer division that
857    hardware usually provides.  Then we have:
858    
859         r = x - q * y        where neither the * nor the - will overflow
860         d = if q >= 0 orelse x = q * y then q else q - 1
861                              where neither the * nor the - will overflow
862         m = if q >= 0 orelse r = 0 then r else r + y
863                              where the + will not overflow
864    
865    This results in substantial simplification of the generated code.
866    The following table shows the number of CFG nodes and edges generated
867    for
868            fun f (x, y) = x OPER y
869            (* with OPER \in div, mod, quot, rem *)
870    
871    
872        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
873        --------------------------------------------------------
874         div |         24 |         39 |         12 |         16
875         mod |         41 |         71 |         12 |         16
876        quot |          8 |         10 |          8 |         10
877         rem |         10 |         14 |          8 |         10
878    
879    
880    ----------------------------------------------------------------------
881    Name: Matthias Blume
882    Date: 2002/03/25 22:06:00 EST
883    Tag: blume-20020325-cprotobug
884    Description:
885    
886    Fixed a bug in cproto (c prototype decoder).
887    
888    ----------------------------------------------------------------------
889    Name: Matthias Blume
890    Date: 2002/03/25 16:00:00 EST
891    Tag: blume-20020325-raw-primops
892    Description:
893    
894    I did some cleanup to Allen's new primop code and
895    replaced yesterday's bootfiles with new ones.
896    (But they are stored in the same place.)
897    
898    ----------------------------------------------------------------------
899    Name: Matthias Blume
900    Date: 2002/03/24 22:40:00 EST
901    Tag: blume-20020324-bootfiles
902    Description:
903    
904    Made the bootfiles that Allen asked for.
905    
906    ----------------------------------------------------------------------
907    Name: Allen Leung
908    Date: 2002/03/23 15:50:00 EST
909    Tag: leunga-20020323-flint-cps-rcc-primops
910    Description:
911    
912      1. Changes to FLINT primops:
913    
914        (* make a call to a C-function;
915         * The primop carries C function prototype information and specifies
916         * which of its (ML-) arguments are floating point. C prototype
917         * information is for use by the backend, ML information is for
918         * use by the CPS converter. *)
919      | RAW_CCALL of { c_proto: CTypes.c_proto,
920                       ml_args: ccall_type list,
921                       ml_res_opt: ccall_type option,
922                       reentrant : bool
923                     } option
924       (* Allocate uninitialized storage on the heap.
925        * The record is meant to hold short-lived C objects, i.e., they
926        * are not ML pointers.  With the tag, the representation is
927        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
928        *)
929      | RAW_RECORD of {tag:bool,sz:int}
930      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
931    
932      2.  These CPS primops are now overloaded:
933    
934           rawload of {kind:numkind}
935           rawstore of {kind:numkind}
936    
937          The one argument form is:
938    
939             rawload {kind} address
940    
941          The two argument form is:
942    
943             rawload {kind} [ml object, byte-offset]
944    
945      3. RAW_CCALL/RCC now takes two extra arguments:
946    
947         a. The first is whether the C call is reentrant, i.e., whether
948            ML state should be saved and restored.
949         b. The second argument is a string argument specifying the name of
950            library and the C function.
951    
952         These things are currently not handled in the code generator, yet.
953    
954      4. In CProto,
955    
956         An encoding type of "bool" means "ml object" and is mapped into
957         C prototype of PTR.  Note that "bool" is different than "string",
958         even though "string" is also mapped into PTR, because "bool"
959         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
960    
961      5. Pickler/unpicker
962    
963         Changed to handle RAW_RECORD and newest RAW_CCALL
964    
965      6. MLRiscGen,
966    
967         1. Changed to handle the new rawload/rawstore/rawrecord operators.
968         2. Code for handling C Calls has been moved to a new module CPSCCalls,
969            in the file CodeGen/cpscompile/cps-c-calls.sml
970    
971      7. Added the conditional move operator
972    
973             condmove of branch
974    
975         to cps.  Generation of this is still buggy so it is currently
976         disabled.
977    
978    ----------------------------------------------------------------------
979    Name: Lal George
980    Date: 2002/03/22 14:18:25 EST
981    Tag: george-20020322-cps-branch-prob
982    Description:
983    
984    Implemented the Ball-Larus branch prediction-heuristics, and
985    incorporated graphical viewers for control flow graphs.
986    
987    Ball-Larus Heuristics:
988    ---------------------
989    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
990    
991    By design it uses the Dempster-Shafer theory for combining
992    probabilities.  For example, in the function:
993    
994        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
995    
996    the ball-larus heuristics predicts that the n=0 is unlikely
997    (OH-heuristic), and the 'then' branch is unlikely because of the
998    RH-heuristic -- giving the 'then' branch an even lower combined
999    probability using the Dempster-Shafer theory.
1000    
1001    Finally, John Reppy's loop analysis in MLRISC, further lowers the
1002    probability of the 'then' branch because of the loop in the else
1003    branch.
1004    
1005    
1006    Graphical Viewing:
1007    ------------------
1008    I merely plugged in Allen's graphical viewers into the compiler. The
1009    additional code is not much. At the top level, saying:
1010    
1011            Control.MLRISC.getFlag "cfg-graphical-view" := true;
1012    
1013    will display the graphical view of the control flow graph just before
1014    back-patching.  daVinci must be in your path for this to work. If
1015    daVinci is not available, then the default viewer can be changed
1016    using:
1017    
1018            Control.MLRISC.getString "viewer"
1019    
1020    which can be set to "dot" or "vcg" for the corresponding viewers. Of
1021    course, these viewers must be in your path.
1022    
1023    The above will display the compilation unit at the level of clusters,
1024    many of which are small, boring, and un-interesting. Also setting:
1025    
1026            Control.MLRISC.getInt "cfg-graphical-view_size"
1027    
1028    will display clusters that are larger than the value set by the above.
1029    
1030    
1031    ----------------------------------------------------------------------
1032    Name: Matthias Blume
1033    Date: 2002/03/21 22:20:00 EST
1034    Tag: blume-20020321-kmp-bugfix
1035    Description:
1036    
1037    Changed the interface to the KMP routine in PreString and fixed
1038    a minor bug in one place where it was used.
1039    
1040    ----------------------------------------------------------------------
1041    Name: Allen Leung
1042    Date: 2002/03/21 20:30:00 EST
1043    Tag: leunga-20020321-cfg
1044    Description:
1045    
1046      Fixed a potential problem in cfg edge splitting.
1047    
1048    ----------------------------------------------------------------------
1049    Name: Allen Leung
1050    Date: 2002/03/21 17:15:00 EST
1051    Tag: leunga-20020321-x86-fp-cfg
1052    Description:
1053    
1054      1. Recoded the buggy parts of x86-fp.
1055    
1056         a. All the block reordering code has been removed.
1057            We now depend on the block placement phases to do this work.
1058    
1059         b. Critical edge splitting code has been simplified and moved into the
1060            CFG modules, as where they belong.
1061    
1062         Both of these were quite buggy and complex.  The code is now much, much
1063         simpler.
1064    
1065      2. X86 backend.
1066    
1067         a. Added instructions for 64-bit support.  Instruction selection for
1068            64-bit has not been committed, however, since that
1069            requires changes to MLTREE which haven't been approved by
1070            Lal and John.
1071    
1072         b. Added support for FUCOMI and FUCOMIP when generating code for
1073            PentiumPro and above.  We only generate these instructions in
1074            the fast-fp mode.
1075    
1076         c. Added cases for JP and JNP in X86FreqProps.
1077    
1078      3. CFG
1079    
1080         CFG now has a bunch of methods for edge splitting and merging.
1081    
1082      4. Machine description.
1083    
1084         John's simplification of MLTREE_BASIS.fcond broke a few machine
1085         description things:
1086    
1087         rtl-build.{sig,sml} and hppa.mdl fixed.
1088    
1089         NOTE: the machine description stuff in the repository is still broken.
1090               Again, I can't put my fixes in because that involves
1091               changes to MLTREE.
1092    
1093    ----------------------------------------------------------------------
1094    Name: Matthias Blume
1095    Date: 2002/03/20 15:55:00 EST
1096    Tag: blume-20020320-kmp
1097    Description:
1098    
1099    Implemented Knuth-Morris-Pratt string matching in PreString and used
1100    it for String.isSubstring, Substring.isSubstring, and
1101    Substring.position.
1102    
1103    (Might need some stress-testing.  Simple examples worked fine.)
1104    
1105    ----------------------------------------------------------------------
1106    Name: Matthias Blume
1107    Date: 2002/03/19 16:37:00 EST
1108    Tag: blume-20020319-witnesses
1109    Description:
1110    
1111    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
1112    
1113    This implements a generic mechanism for changing constness qualifiers
1114    anywhere within big C types without resorting to outright "casts".
1115    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
1116    modify the constness at the outermost level.)
1117    The implementation of "convert" is based on the idea of "witness"
1118    values -- values that are not used by the operation but whose types
1119    "testify" to their applicability.  On the implementation side, "convert"
1120    is simply a projection (returning its second curried argument).  With
1121    cross-module inlining, it should not result in any machine code being
1122    generated.
1123    
1124    ----------------------------------------------------------------------
1125    Name: Matthias Blume
1126    Date: 2002/03/15 16:40:00 EST
1127    Tag: blume-20020315-basis
1128    Description:
1129    
1130    Provided (preliminary?) implementations for
1131    
1132      {String,Substring}.{concatWith,isSuffix,isSubstring}
1133    
1134    and
1135    
1136      Substring.full
1137    
1138    Those are in the Basis spec but they were missing in SML/NJ.
1139    
1140    ----------------------------------------------------------------------
1141    Name: Matthias Blume
1142    Date: 2002/03/14 21:30:00 EST
1143    Tag: blume-20020314-controls
1144    Description:
1145    
1146    Controls:
1147    ---------
1148    
1149    1. Factored out the recently-added Controls : CONTROLS stuff and put
1150       it into its own library $/controls-lib.cm.  The source tree for
1151       this is under src/smlnj-lib/Controls.
1152    
1153    2. Changed the names of types and functions in this interface, so they
1154       make a bit more "sense":
1155    
1156          module -> registry
1157          'a registry -> 'a group
1158    
1159    3. The interface now deals in ref cells only.  The getter/setter interface
1160       is (mostly) gone.
1161    
1162    4. Added a function that lets one register an already-existing ref cell.
1163    
1164    5. Made the corresponding modifications to the rest of the code so that
1165       everything compiles again.
1166    
1167    6. Changed the implementation of Controls.MLRISC back to something closer
1168       to the original.  In particular, this module (and therefore MLRISC)
1169       does not depend on Controls.  There now is some link-time code in
1170       int-sys.sml that registers the MLRISC controls with the Controls
1171       module.
1172    
1173    CM:
1174    ---
1175    
1176      * One can now specify the lambda-split aggressiveness in init.cmi.
1177    
1178    ----------------------------------------------------------------------
1179    Name: Allen Leung
1180    Date: 2002/03/13 17:30:00 EST
1181    Tag: leunga-20020313-x86-fp-unary
1182    Description:
1183    
1184    Bug fix for:
1185    
1186    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
1187    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
1188    > - fun f(x,(y,z)) = Real.~ y;
1189    > [autoloading]
1190    > [autoloading done]
1191    >       fchsl   (%eax), 184(%esp)
1192    > Error: MLRisc bug: X86MCEmitter.emitInstr
1193    >
1194    > uncaught exception Error
1195    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
1196    
1197    The problem was that the code generator did not generate any fp registers
1198    in this case, and the ra didn't know that it needed to run the X86FP phase to
1199    translate the pseudo fp instruction.   This only happened with unary fp
1200    operators in certain situations.
1201    
1202    ----------------------------------------------------------------------
1203    Name: Matthias Blume
1204    Date: 2002/03/13 14:00:00 EST
1205    Tag: blume-20020313-overload-etc
1206    Description:
1207    
1208    1. Added _overload as a synonym for overload for backward compatibility.
1209       (Control.overloadKW must be true for either version to be accepted.)
1210    
1211    2. Fixed bug in install script that caused more things to be installed
1212       than what was requested in config/targets.
1213    
1214    3. Made CM aware of the (_)overload construct so that autoloading
1215       works.
1216    
1217    ----------------------------------------------------------------------
1218    Name: Matthias Blume
1219    Date: 2002/03/12 22:03:00 EST
1220    Tag: blume-20020312-url
1221    Description:
1222    
1223    Forgot to update BOOT and srcarchiveurl.
1224    
1225    ----------------------------------------------------------------------
1226    Name: Matthias Blume
1227    Date: 2002/03/12 17:30:00 EST
1228    Tag: blume-20020312-version110392
1229    Description:
1230    
1231    Yet another version number bump (because of small changes to the
1232    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
1233    
1234    Changes:
1235    
1236      The new pid generation scheme described a few weeks ago was overly
1237      complicated.  I implemented a new mechanism that is simpler and
1238      provides a bit more "stability":  Once CM has seen a compilation
1239      unit, it keeps its identity constant (as long as you do not delete
1240      those crucial CM/GUID/* files).  This means that when you change
1241      an interface, compile, then go back to the old interface, and
1242      compile again, you arrive at the original pid.
1243    
1244      There now also is a mechanism that instructs CM to use the plain
1245      environment hash as a module's pid (effectively making its GUID
1246      the empty string).  For this, "noguid" must be specified as an
1247      option to the .sml file in question within its .cm file.
1248      This is most useful for code that is being generated by tools such
1249      as ml-nlffigen (because during development programmers tend to
1250      erase the tool's entire output directory tree including CM's cached
1251      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
1252      revert to the old, broken behavior of SML/NJ, but in specific cases
1253      where there is no danger of interface confusion, its use is ok
1254      (I think).
1255    
1256      ml-nlffigen by default generates "noguid" annotations.  They can be
1257      turned off by specifying -guid in its command line.
1258    
1259    ----------------------------------------------------------------------
1260    Name: Lal George
1261    Date: 2002/03/12 12 14:42:36 EST
1262    Tag: george-20020312-frequency-computation
1263    Description:
1264    
1265    Integrated jump chaining and static block frequency into the
1266    compiler. More details and numbers later.
1267    
1268    ----------------------------------------------------------------------
1269    Name: Lal George
1270    Date: 2002/03/11 11 22:38:53 EST
1271    Tag: george-20020311-jump-chain-elim
1272    Description:
1273    
1274    Tested the jump chain elimination on all architectures (except the
1275    hppa).  This is on by default right now and is profitable for the
1276    alpha and x86, however, it may not be profitable for the sparc and ppc
1277    when compiling the compiler.
1278    
1279    The gc test will typically jump to a label at the end of the cluster,
1280    where there is another jump to an external cluster containing the actual
1281    code to invoke gc. This is to allow factoring of common gc invocation
1282    sequences. That is to say, we generate:
1283    
1284            f:
1285               testgc
1286               ja   L1      % jump if above to L1
1287    
1288            L1:
1289               jmp L2
1290    
1291    
1292    After jump chain elimination the 'ja L1' instructions is converted to
1293    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
1294    up being implemented in their long form (if L2 is far away) using:
1295    
1296            jbe     L3      % jump if below or equal to L3
1297            jmp     L2
1298         L3:
1299            ...
1300    
1301    
1302    For large compilation units L2  may be far away.
1303    
1304    
1305    ----------------------------------------------------------------------
1306    Name: Matthias Blume
1307    Date: 2002/03/11 13:30:00 EST
1308    Tag: blume-20020311-mltreeeval
1309    Description:
1310    
1311    A functor parameter was missing.
1312    
1313    ----------------------------------------------------------------------
1314    Name: Allen Leung
1315    Date: 2002/03/11 10:30:00 EST
1316    Tag: leunga-20020311-runtime-string0
1317    Description:
1318    
1319       The representation of the empty string now points to a
1320    legal null terminated C string instead of unit.  It is now possible
1321    to convert an ML string into C string with InlineT.CharVector.getData.
1322    This compiles into one single machine instruction.
1323    
1324    ----------------------------------------------------------------------
1325    Name: Allen Leung
1326    Date: 2002/03/10 23:55:00 EST
1327    Tag: leunga-20020310-x86-call
1328    Description:
1329    
1330       Added machine generation for CALL instruction (relative displacement mode)
1331    
1332    ----------------------------------------------------------------------
1333    Name: Matthias Blume
1334    Date: 2002/03/08 16:05:00
1335    Tag: blume-20020308-entrypoints
1336    Description:
1337    
1338    Version number bumped to 110.39.1.  NEW BOOTFILES!
1339    
1340    Entrypoints: non-zero offset into a code object where execution should begin.
1341    
1342    - Added the notion of an entrypoint to CodeObj.
1343    - Added reading/writing of entrypoint info to Binfile.
1344    - Made runtime system bootloader aware of entrypoints.
1345    - Use the address of the label of the first function given to mlriscGen
1346      as the entrypoint.  This address is currently always 0, but it will
1347      not be 0 once we turn on block placement.
1348    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
1349      with entry points) from mlriscGen.
1350    
1351    ----------------------------------------------------------------------
1352    Name: Allen Leung
1353    Date: 2002/03/07 20:45:00 EST
1354    Tag: leunga-20020307-x86-cmov
1355    Description:
1356    
1357       Bug fixes for CMOVcc on x86.
1358    
1359       1. Added machine code generation for CMOVcc
1360       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
1361       3. CMOVcc cannot have an immediate operand as argument.
1362    
1363    ----------------------------------------------------------------------
1364    Name: Matthias Blume
1365    Date: 2002/03/07 16:15:00 EST
1366    Tag: blume-20020307-controls
1367    Description:
1368    
1369    This is a very large but mostly boring patch which makes (almost)
1370    every tuneable compiler knob (i.e., pretty much everything under
1371    Control.* plus a few other things) configurable via both the command
1372    line and environment variables in the style CM did its configuration
1373    until now.
1374    
1375    Try starting sml with '-h' (or, if you are brave, '-H')
1376    
1377    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
1378    implements the underlying generic mechanism.
1379    
1380    The interface to some of the existing such facilities has changed somewhat.
1381    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
1382    (The getFoo interface is still there for backward-compatibility, but its
1383    use is deprecated.)
1384    
1385    The ml-build script passes -Cxxx=yyy command-line arguments through so
1386    that one can now twiddle the compiler settings when using this "batch"
1387    compiler.
1388    
1389    TODO items:
1390    
1391    We should go through and throw out all controls that are no longer
1392    connected to anything.  Moreover, we should go through and provide
1393    meaningful (and correct!) documentation strings for those controls
1394    that still are connected.
1395    
1396    Currently, multiple calls to Controls.new are accepted (only the first
1397    has any effect).  Eventually we should make sure that every control
1398    is being made (via Controls.new) exactly once.  Future access can then
1399    be done using Controls.acc.
1400    
1401    Finally, it would probably be a good idea to use the getter-setter
1402    interface to controls rather than ref cells.  For the time being, both
1403    styles are provided by the Controls module, but getter-setter pairs are
1404    better if thread-safety is of any concern because they can be wrapped.
1405    
1406    *****************************************
1407    
1408    One bug fix: The function blockPlacement in three of the MLRISC
1409    backpatch files used to be hard-wired to one of two possibilities at
1410    link time (according to the value of the placementFlag).  But (I
1411    think) it should rather sense the flag every time.
1412    
1413    *****************************************
1414    
1415    Other assorted changes (by other people who did not supply a HISTORY entry):
1416    
1417    1. the cross-module inliner now works much better (Monnier)
1418    2. representation of weights, frequencies, and probabilities in MLRISC
1419       changed in preparation of using those for weighted block placement
1420       (Reppy, George)
1421    
1422    ----------------------------------------------------------------------
1423    Name: Lal George
1424    Date: 2002/03/07 14:44:24 EST 2002
1425    Tag: george-20020307-weighted-block-placement
1426    
1427    Tested the weighted block placement optimization on all architectures
1428    (except the hppa) using AMPL to generate the block and edge frequencies.
1429    Changes were required in the machine properties to correctly
1430    categorize trap instructions. There is an MLRISC flag
1431    "weighted-block-placement" that can be used to enable weighted block
1432    placement, but this will be ineffective without block/edge
1433    frequencies (coming soon).
1434    
1435    
1436    ----------------------------------------------------------------------
1437    Name: Lal George
1438    Date: 2002/03/05 17:24:48 EST
1439    Tag: george-20020305-linkage-cluster
1440    
1441    In order to support the block placement optimization, a new cluster
1442    is generated as the very first cluster (called the linkage cluster).
1443    It contains a single jump to the 'real' entry point for the compilation
1444    unit. Block placement has no effect on the linkage cluster itself, but
1445    all the other clusters  have full freedom in the manner in which they
1446    reorder blocks or functions.
1447    
1448    On the x86 the typical linkage code that is generated is:
1449       ----------------------
1450            .align 2
1451       L0:
1452            addl    $L1-L0, 72(%esp)
1453            jmp     L1
1454    
1455    
1456            .align  2
1457       L1:
1458       ----------------------
1459    
1460    72(%esp) is the memory location for the stdlink register. This
1461    must contain the address of the CPS function being called. In the
1462    above example, it contains the address of  L0; before
1463    calling L1 (the real entry point for the compilation unit), it
1464    must contain the address for L1, and hence
1465    
1466            addl $L1-L0, 72(%esp)
1467    
1468    I have tested this on all architectures except the hppa.The increase
1469    in code size is of course negligible
1470    
1471    ----------------------------------------------------------------------
1472    Name: Allen Leung
1473    Date: 2002/03/03 13:20:00 EST
1474    Tag: leunga-20020303-mlrisc-tools
1475    
1476      Added #[ ... ] expressions to mlrisc tools
1477    
1478    ----------------------------------------------------------------------
1479    Name: Matthias Blume
1480    Date: 2002/02/27 12:29:00 EST
1481    Tag: blume-20020227-cdebug
1482    Description:
1483    
1484    - made types in structure C and C_Debug to be equal
1485    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
1486    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
1487    
1488    ----------------------------------------------------------------------
1489    Name: Matthias Blume
1490    Date: 2002/02/26 12:00:00 EST
1491    Tag: blume-20020226-ffi
1492    Description:
1493    
1494    1. Fixed a minor bug in CM's "noweb" tool:
1495       If numbering is turned off, then truly don't number (i.e., do not
1496       supply the -L option to noweb).  The previous behavior was to supply
1497       -L'' -- which caused noweb to use the "default" line numbering scheme.
1498       Thanks to Chris Richards for pointing this out (and supplying the fix).
1499    
1500    2. Once again, I reworked some aspects of the FFI:
1501    
1502       A. The incomplete/complete type business:
1503    
1504       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
1505         gone!
1506       - ML types representing an incomplete type are now *equal* to
1507         ML types representing their corresponding complete types (just like
1508         in C).  This is still safe because ml-nlffigen will not generate
1509         RTTI for incomplete types, nor will it generate functions that
1510         require access to such RTTI.   But when ML code generated from both
1511         incomplete and complete versions of the C type meet, the ML types
1512         are trivially interoperable.
1513    
1514         NOTE:  These changes restore the full generality of the translation
1515         (which was previously lost when I eliminated functorization)!
1516    
1517       B. Enum types:
1518    
1519       - Structure C now has a type constructor "enum" that is similar to
1520         how the "su" constructor works.  However, "enum" is not a phantom
1521         type because each "T enum" has values (and is isomorphic to
1522         MLRep.Signed.int).
1523       - There are generic access operations for enum objects (using
1524         MLRep.Signed.int).
1525       - ml-nlffigen will generate a structure E_foo for each "enum foo".
1526         * The structure contains the definition of type "mlrep" (the ML-side
1527         representation type of the enum).  Normally, mlrep is the same
1528         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
1529         then mlrep will be defined as a datatype -- thus facilitating
1530         pattern matching on mlrep values.
1531         ("-ec" will be suppressed if there are duplicate values in an
1532          enumeration.)
1533         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
1534         will be generated for each C enum constant xxx.
1535         * Conversion functions m2i and i2m convert between mlrep and
1536         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
1537         * Coversion functions c and ml convert between mlrep and "tag enum".
1538         * Access functions (get/set) fetch and store mlrep values.
1539       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
1540         enumerations are merged into one single enumeration represented by
1541         structure E_'.
1542    
1543    ----------------------------------------------------------------------
1544    Name: Allen Leung
1545    Date: 2002/02/25 04:45:00 EST
1546    Tag: leunga-20020225-cps-spill
1547    
1548    This is a new implementation of the CPS spill phase.
1549    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
1550    In case of problems, replace it with the old file spill.sml
1551    
1552    The current compiler runs into some serious performance problems when
1553    constructing a large record.  This can happen when we try to compile a
1554    structure with many items.  Even a very simple structure like the following
1555    makes the compiler slow down.
1556    
1557        structure Foo = struct
1558           val x_1 = 0w1 : Word32.int
1559           val x_2 = 0w2 : Word32.int
1560           val x_3 = 0w3 : Word32.int
1561           ...
1562           val x_N = 0wN : Word32.int
1563        end
1564    
1565    The following table shows the compile time, from N=1000 to N=4000,
1566    with the old compiler:
1567    
1568    N
1569    1000   CPS 100 spill                           0.04u  0.00s  0.00g
1570           MLRISC ra                               0.06u  0.00s  0.05g
1571              (spills = 0 reloads = 0)
1572           TOTAL                                   0.63u  0.07s  0.21g
1573    
1574    1100   CPS 100 spill                           8.25u  0.32s  0.64g
1575           MLRISC ra                               5.68u  0.59s  3.93g
1576              (spills = 0 reloads = 0)
1577           TOTAL                                   14.71u  0.99s  4.81g
1578    
1579    1500   CPS 100 spill                           58.55u  2.34s  1.74g
1580           MLRISC ra                               5.54u  0.65s  3.91g
1581              (spills = 543 reloads = 1082)
1582           TOTAL                                   65.40u  3.13s  6.00g
1583    
1584    2000   CPS 100 spill                           126.69u  4.84s  3.08g
1585           MLRISC ra                               0.80u  0.10s  0.55g
1586              (spills = 42 reloads = 84)
1587           TOTAL                                   129.42u  5.10s  4.13g
1588    
1589    3000   CPS 100 spill                           675.59u  19.03s  11.64g
1590           MLRISC ra                               2.69u  0.27s  1.38g
1591              (spills = 62 reloads = 124)
1592           TOTAL                                   682.48u  19.61s  13.99g
1593    
1594    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
1595           MLRISC ra                               4.96u  0.27s  2.72g
1596              (spills = 85 reloads = 170)
1597           TOTAL                                   2375.26u  57.21s  48.00g
1598    
1599    As you can see the old cps spill module suffers from some serious
1600    performance problem.  But since I cannot decipher the old code fully,
1601    instead of patching the problems up, I'm reimplementing it
1602    with a different algorithm.  The new code is more modular,
1603    smaller when compiled, and substantially faster
1604    (O(n log n) time and O(n) space).  Timing of the new spill module:
1605    
1606    4000  CPS 100 spill                           0.02u  0.00s  0.00g
1607          MLRISC ra                               0.25u  0.02s  0.15g
1608             (spills=1 reloads=3)
1609          TOTAL                                   7.74u  0.34s  1.62g
1610    
1611    Implementation details:
1612    
1613    As far as I can tell, the purpose of the CPS spill module is to make sure the
1614    number of live variables at any program point (the bandwidth)
1615    does not exceed a certain limit, which is determined by the
1616    size of the spill area.
1617    
1618    When the bandwidth is too large, we decrease the register pressure by
1619    packing live variables into spill records.  How we achieve this is
1620    completely different than what we did in the old code.
1621    
1622    First, there is something about the MLRiscGen code generator
1623    that we should be aware of:
1624    
1625    o MLRiscGen performs code motion!
1626    
1627       In particular, it will move floating point computations and
1628       address computations involving only the heap pointer to
1629       their use sites (if there is only a single use).
1630       What this means is that if we have a CPS record construction
1631       statement
1632    
1633           RECORD(k,vl,w,e)
1634    
1635       we should never count the new record address w as live if w
1636       has only one use (which is often the case).
1637    
1638       We should do something similar to floating point, but the transformation
1639       there is much more complex, so I won't deal with that.
1640    
1641    Secondly, there are now two new cps primops at our disposal:
1642    
1643     1. rawrecord of record_kind option
1644        This pure operator allocates some uninitialized storage from the heap.
1645        There are two forms:
1646    
1647         rawrecord NONE [INT n]  allocates a tagless record of length n
1648         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
1649                                     and initializes the tag.
1650    
1651     2. rawupdate of cty
1652          rawupdate cty (v,i,x)
1653          Assigns to x to the ith component of record v.
1654          The storelist is not updated.
1655    
1656    We use these new primops for both spilling and increment record construction.
1657    
1658     1. Spilling.
1659    
1660        This is implemented with a linear scan algorithm (but generalized
1661        to trees).  The algorithm will create a single spill record at the
1662        beginning of the cps function and use rawupdate to spill to it,
1663        and SELECT or SELp to reload from it.  So both spills and reloads
1664        are fine-grain operations.  In contrast, in the old algorithm
1665        "spills" have to be bundled together in records.
1666    
1667        Ideally, we should sink the spill record construction to where
1668        it is needed.  We can even split the spill record into multiple ones
1669        at the places where they are needed.  But CPS is not a good
1670        representation for global code motion, so I'll keep it simple and
1671        am not attempting this.
1672    
1673     2. Incremental record construction (aka record splitting).
1674    
1675        Long records with many component values which are simulatenously live
1676        (recall that single use record addresses are not considered to
1677         be live) are constructed with rawrecord and rawupdate.
1678        We allocate space on the heap with rawrecord first, then gradually
1679        fill it in with rawupdate.  This is the technique suggested to me
1680        by Matthias.
1681    
1682        Some restrictions on when this is applicable:
1683        1. It is not a VECTOR record.  The code generator currently does not handle
1684           this case. VECTOR record uses double indirection like arrays.
1685        2. All the record component values are defined in the same "basic block"
1686           as the record constructor.  This is to prevent speculative
1687           record construction.
1688    
1689    ----------------------------------------------------------------------
1690    Name: Allen Leung
1691    Date: 2002/02/22 01:02:00 EST
1692    Tag: leunga-20020222-mlrisc-tools
1693    
1694    Minor bug fixes in the parser and rewriter
1695    
1696    ----------------------------------------------------------------------
1697    Name: Allen Leung
1698    Date: 2002/02/21 20:20:00 EST
1699    Tag: leunga-20020221-peephole
1700    
1701    Regenerated the peephole files.  Some contained typos in the specification
1702    and some didn't compile because of pretty printing bugs in the old version
1703    of 'nowhere'.
1704    
1705    ----------------------------------------------------------------------
1706    Name: Allen Leung
1707    Date: 2002/02/19 20:20:00 EST
1708    Tag: leunga-20020219-mlrisc-tools
1709    Description:
1710    
1711       Minor bug fixes to the mlrisc-tools library:
1712    
1713       1.  Fixed up parsing colon suffixed keywords
1714       2.  Added the ability to shut the error messages up
1715       3.  Reimplemented the pretty printer and fixed up/improved
1716           the pretty printing of handle and -> types.
1717       4.  Fixed up generation of literal symbols in the nowhere tool.
1718       5.  Added some SML keywords to to sml.sty
1719    
1720    ----------------------------------------------------------------------
1721    Name: Matthias Blume
1722    Date: 2002/02/19 16:20:00 EST
1723    Tag: blume-20020219-cmffi
1724    Description:
1725    
1726    A wild mix of changes, some minor, some major:
1727    
1728    * All C FFI-related libraries are now anchored under $c:
1729        $/c.cm      --> $c/c.cm
1730        $/c-int.cm  --> $c/internals/c-int.cm
1731        $/memory.cm --> $c/memory/memory.cm
1732    
1733    * "make" tool (in CM) now treats its argument pathname slightly
1734      differently:
1735        1. If the native expansion is an absolute name, then before invoking
1736           the "make" command on it, CM will apply OS.Path.mkRelative
1737           (with relativeTo = OS.FileSys.getDir()) to it.
1738        2. The argument will be passed through to subsequent phases of CM
1739           processing without "going native".  In particular, if the argument
1740           was an anchored path, then "make" will not lose track of that anchor.
1741    
1742    * Compiler backends now "know" their respective C calling conventions
1743      instead of having to be told about it by ml-nlffigen.  This relieves
1744      ml-nlffigen from one of its burdens.
1745    
1746    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
1747    
1748    * Export C_DEBUG and C_Debug from $c/c.cm.
1749    
1750    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
1751      subtyping relationship between incomplete pointers and their complete
1752      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
1753      with the convention of instantiating 'o with ('t, 'c) obj whenever
1754      the pointer target type is complete.  In the incomplete case, 'o
1755      will be instantiated with some "'c iobj" -- a type obtained by
1756      using one of the functors PointerToIncompleteType or PointerToCompleteType.
1757    
1758      Operations that work on both incomplete and complete pointer types are
1759      typed as taking an 'o ptr while operations that require the target to
1760      be known are typed as taking some ('t, 'c) obj ptr.
1761    
1762      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
1763      where void is an eqtype without any values.  This makes it possible
1764      to work on voidptr values using functions meant to operate on light
1765      incomplete pointers.
1766    
1767    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
1768      been vastly simplified.
1769    
1770    ----------------------------------------------------------------------
1771    Name: Matthias Blume
1772    Date: 2002/02/19 10:48:00 EST
1773    Tag: blume-20020219-pqfix
1774    Description:
1775    
1776    Applied Chris Okasaki's bug fix for priority queues.
1777    
1778    ----------------------------------------------------------------------
1779    Name: Matthias Blume
1780    Date: 2002/02/15 17:05:00
1781    Tag: Release_110_39
1782    Description:
1783    
1784    Last-minute retagging is becoming a tradition... :-(
1785    
1786    This is the working release 110.39.
1787    
1788    ----------------------------------------------------------------------
1789    Name: Matthias Blume
1790    Date: 2002/02/15 16:00:00 EST
1791    Tag: Release_110_39-orig
1792    Description:
1793    
1794    Working release 110.39.  New bootfiles.
1795    
1796    (Update: There was a small bug in the installer so it wouldn't work
1797    with all shells.  So I retagged. -Matthias)
1798    
1799    ----------------------------------------------------------------------
1800    Name: Matthias Blume
1801    Date: 2002/02/15 14:17:00 EST
1802    Tag: blume-20020215-showbindings
1803    Description:
1804    
1805    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
1806    the latter can be useful for exploring what bindings are available at
1807    the interactive prompt.  (The first function returns only the list
1808    of symbols that are really bound, the second prints those but also the
1809    ones that CM's autoloading mechanism knows about.)
1810    
1811    ----------------------------------------------------------------------
1812    Name: Matthias Blume
1813    Date: 2002/02/15 12:08:00 EST
1814    Tag: blume-20020215-iptrs
1815    Description:
1816    
1817    Two improvements to ml-nlffigen:
1818    
1819      1. Write files only if they do not exist or if their current contents
1820         do not coincide with what's being written.  (That is, avoid messing
1821         with the time stamps unless absolutely necessary.)
1822    
1823      2. Implement a "repository" mechanism for generated files related
1824         to "incomplete pointer types".   See the README file for details.
1825    
1826    ----------------------------------------------------------------------
1827    Name: Matthias Blume
1828    Date: 2002/02/14 11:50:00 EST
1829    Tag: blume-20020214-quote
1830    Description:
1831    
1832    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
1833    because of the new and improved tag generation scheme.  (Thanks to Allen
1834    Leung for pointing it out.)
1835    
1836    ----------------------------------------------------------------------
1837    Name: Lal George
1838    Date: 2002/02/14 09:55:27 EST 2002
1839    Tag: george-20020214-isabelle-bug
1840    Description:
1841    
1842    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
1843    of Isabelle on the x86.
1844    
1845    From Allen:
1846    -----------
1847     I've found the problem:
1848    
1849         in ra-core.sml, I use the counter "blocked" to keep track of the
1850         true number of elements in the freeze queue.  When the counter goes
1851         to zero, I skip examining the queue.  But I've messed up the
1852         bookkeeping in combine():
1853    
1854             else ();
1855             case !ucol of
1856               PSEUDO => (if !cntv > 0 then
1857                     (if !cntu > 0 then blocked := !blocked - 1 else ();
1858                                        ^^^^^^^^^^^^^^^^^^^^^^^
1859                      moveu := mergeMoveList(!movev, !moveu)
1860                     )
1861                  else ();
1862    
1863         combine() is called to coalesce two nodes u and v.
1864         I think I was thinking that if the move counts of u and v are both
1865         greater than zero then after they are coalesced then one node is
1866         removed from the freeze queue.  Apparently I was thinking that
1867         both u and v are of low degree, but that's clearly not necessarily true.
1868    
1869    
1870    02/12/2002:
1871        Here's the patch.  HOL now compiles.
1872    
1873        I don't know how this impact on performance (compile
1874        time or runtime).  This bug caused the RA (especially on the x86)
1875        to go thru the potential spill phase when there are still nodes on the
1876        freeze queue.
1877    
1878    
1879    
1880    
1881    ----------------------------------------------------------------------
1882    Name: Matthias Blume
1883    Date: 2002/02/13 22:40:00 EST
1884    Tag: blume-20020213-fptr-rtti
1885    Description:
1886    
1887    Fixed a bug in ml-nlffigen that was introduced with one of the previous
1888    updates.
1889    
1890    ----------------------------------------------------------------------
1891    Name: Matthias Blume
1892    Date: 2002/02/13 16:41:00 EST
1893    Tag: blume-20020213-cmlpq
1894    Description:
1895    
1896    Added new priority queue export symbols (which have just been added to
1897    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
1898    not compile and the installer would choke.)
1899    
1900    ----------------------------------------------------------------------
1901    Name: Matthias Blume
1902    Date: 2002/02/13 16:15:00 EST
1903    Tag: blume-20020213-various
1904    Description:
1905    
1906    1. More tweaks to ml-nlffigen:
1907    
1908       - better internal datastructures (resulting in slight speedup)
1909       - "-match" option requires exact match
1910       - "localized" gensym counters (untagged structs/unions nested within
1911         other structs/unions or within typedefs get a fresh counter; their
1912         tag will be prefixed by a concatenation of their parents' tags)
1913       - bug fixes (related to calculation of transitive closure of types
1914         to be included in the output)
1915    
1916    2. Minor Basis updates:
1917    
1918       - added implementations for List.collate and Option.app
1919    
1920    ----------------------------------------------------------------------
1921    Name: Matthias Blume
1922    Date: 2002/02/11 15:55:00 EST
1923    Tag: blume-20020211-gensym
1924    Description:
1925    
1926    Added a "-gensym" option to command line of ml-nlffigen.  This can be
1927    used to specify a "stem" -- a string that is inserted in all "gensym'd"
1928    names (ML structure names that correspond to unnamed C structs, unions,
1929    and enums), so that separate runs of ml-nlffigen do not clash.
1930    
1931    ----------------------------------------------------------------------
1932    Name: Matthias Blume
1933    Date: 2002/02/11 12:05:00 EST
1934    Tag: blume-20020211-gensml
1935    Description:
1936    
1937    A quick fix for a problem with GenSML (in the pgraph-util library):
1938    Make generation of toplevel "local" optional.  (Strictly speaking,
1939    signature definitions within "local" are not legal SML.)
1940    
1941    Other than that: updates to INSTALL and cm/TODO.
1942    
1943    ----------------------------------------------------------------------
1944    Name: Matthias Blume
1945    Date: 2002/02/08 15:00:00 EST
1946    Tag: blume-20020208-uniquepid
1947    Description:
1948    
1949    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
1950    
1951    1. The installer (config/install.sh) has gotten smarter:
1952    
1953         - Configuration options are a bit easier to specify now
1954           (in config/targets).
1955         - Bug in recognizing .tar.bz2 files fixed.
1956         - Installer automatically resolves dependencies between
1957           configuration options (e.g., if you ask for eXene, you will
1958           also get cml -- regardless whether you asked for it or not).
1959         - Installer can run in "quieter mode" by setting the environment
1960           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
1961           "completely silent", though.
1962         - Build HashCons library as part of smlnj-lib.
1963    
1964    2. A new scheme for assigning persistent identifiers to compilation
1965       units (and, by extension, to types etc.) has been put into place.
1966       This fixes a long-standing bug where types and even dynamic values
1967       can get internally confused, thereby compromising type safety
1968       (abstraction) and dynamic correctness.  See
1969    
1970         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
1971    
1972       for an example of how things could go wrong until now.
1973    
1974       The downside of the new scheme is that pids are not quite as
1975       persistent as they used to be: CM will generate a fresh pid
1976       for every compilation unit that it thinks it sees for the first
1977       time.  That means that if you compile starting from a clean, fresh
1978       source tree at two different times, you end up with different
1979       binaries.
1980    
1981       Cutoff recompilation, however, has not been compromised because
1982       CM keeps pid information in special caches between runs.
1983    
1984    ----------------------------------------------------------------------
1985    Name: Lal George
1986    Date: 2002/02/07 15:34:13 EST 2002
1987    Tag: <none>
1988    Description:
1989    
1990    Compilers that generate assembly code may produce  global labels
1991    whose value is resolved at link time. The various peephole optimization
1992    modules did not take this in account.
1993    
1994    TODO. The Labels.addrOf function should really return an option
1995    type so that clients are forced to deal with this issue, rather
1996    than an exception being raised.
1997    
1998    ----------------------------------------------------------------------
1999    Name: Lal George
2000    Date: 2002/02/06 13:55:02 EST
2001    Tag: george-20020206-ra-breakup
2002    Description:
2003    
2004    1. A bug fix from Allen.
2005    
2006        A typo causes extra fstp %st(0)'s to be generated at compensation
2007        edges, which might cause stack underflow traps at runtime.  This
2008        occurs in fft where there are extraneous fstps right before the 'into'
2009        trap instruction (in this case they are harmless since none of the
2010        integers overflow.)
2011    
2012    2. Pulled out various utility modules that were embedded in the modules
2013       of the register allocator. I need these modules for other purposes, but
2014       they are not complete enough to put into a library (just yet).
2015    ----------------------------------------------------------------------
2016    Name: Matthias Blume
2017    Date: 2002/01/31 16:05:00 EST
2018    Tag: blume-20020131-sparc-ccalls
2019    Description:
2020    
2021    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
2022       of extra stack space by mistake.  Fixed.
2023    
2024    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
2025    
2026    ----------------------------------------------------------------------
2027    Name: Allen Leung
2028    Date: 2002/01/30
2029    Tag: leunga-20020130-nowhere-bug-fix
2030    Description:
2031    
2032       MLRISC bug fixes:
2033       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
2034       2. MachineInt.fromString was negating its value.
2035    
2036    ----------------------------------------------------------------------
2037    Name: Matthias Blume
2038    Date: 2002/01/29
2039    Tag: blume-20020129-INSTALL
2040    Description:
2041    
2042    - Added somewhat detailed installation instructions (file INSTALL).
2043    - Fixed curl-detection bug in config/install.sh.
2044    - It is now possible to select the URL getter using the URLGETTER
2045      environment variable:
2046    
2047          not set / "unknown"      --> automatic detection (script tries wget,
2048                                       curl, and lynx)
2049          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
2050                                       how to properly invoke them)
2051          other                    --> use $URLGETTER directly, it must take
2052                                       precisely two command-line arguments
2053                                       (source URL and destination file name)
2054    
2055    ----------------------------------------------------------------------
2056    Name: Matthias Blume
2057    Date: 2002/01/28
2058    Tag: blume-20020128-sparc-ccalls
2059    Description:
2060    
2061    - Fixed problem with calculation of "used" registers in sparc-c-calls.
2062    - Make use of the allocParam argument in sparc-c-calls.
2063    
2064    ----------------------------------------------------------------------
2065    Name: Matthias Blume
2066    Date: 2002/01/28
2067    Tag: blume-20020128-allocParam
2068    Description:
2069    
2070    John Reppy:  Changes c-calls API to accept client-callback for
2071    allocating extra stack space.
2072    me: Corresponding changes to mlriscGen (using a dummy argument that
2073        does not change the current behavior).
2074    
2075    ----------------------------------------------------------------------
2076    Name: Matthias Blume
2077    Date: 2002/01/28 12:00:00
2078    Tag: Release_110_38
2079    Description:
2080    
2081    This time for real!!!
2082    
2083    ----------------------------------------------------------------------
2084    Name: Matthias Blume
2085    Date: 2002/01/28 10:56:00 EST
2086    Tag: blume-20020128-retraction
2087    Description:
2088    
2089    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
2090       with blume-Release_110_38-retracted.)
2091    
2092    1. Fixed a problem with incorrect rounding modes in real64.sml.
2093       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
2094    
2095    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
2096       and enums fixed.  The naming of corresponding ML identifiers should
2097       now be consistent again.
2098    
2099    ----------------------------------------------------------------------
2100    Name: Allen Leung
2101    Date: 2002/01/27
2102    Tag: leunga-20020127-nowhere
2103    Description:
2104    
2105       Added a target called nowhere in the configuration scripts.
2106       Enabling this will build the MLRISC 'nowhere' tool (for translating
2107       programs with where-clauses into legal SML code) during installation.
2108    
2109    ----------------------------------------------------------------------
2110    Name: Matthias Blume
2111    Date: 2002/01/25 21:27:00 EST
2112    Tag: blume-Release_110_38-retracted
2113    Description:
2114    
2115    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
2116    
2117    README will be added later.
2118    
2119    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
2120    (Release_110_38) removed.  Reason: Last-minute bug fixes.
2121    
2122    ----------------------------------------------------------------------
2123    Name: Matthias Blume
2124    Date: 2002/01/25
2125    Tag: blume-20020125-ffi
2126    Description:
2127    
2128    A large number of tweaks and improvements to ml-nlffi-lib and
2129    ml-nlffigen:
2130    
2131       - ML represenation types have been streamlined
2132       - getter and setter functions work with concrete values, not abstract
2133         ones where possible
2134       - ml-nlffigen command line more flexible (see README file there)
2135       - some bugs have been fixed (hopefully)
2136    
2137    ----------------------------------------------------------------------
2138    Name: Lal George
2139    Date: 2002/01/24
2140    Tag: george-20020124-risc-ra-interface
2141    Description:
2142    
2143       There is a dramatic simplification in the interface to the
2144       register allocator for RISC architectures as a result of making
2145       parallel copy instructions explicit.
2146    
2147    ----------------------------------------------------------------------
2148    Name: Matthias Blume
2149    Date: 2002/01/22
2150    Tag: blume-20020122-x86-ccalls
2151    Description:
2152    
2153    Bug fix for c-calls on x86 (having to do with how char- and
2154    short-arguments are being handled).
2155    
2156    ----------------------------------------------------------------------
2157    Name: Matthias Blume
2158    Date: 2002/01/21
2159    Tag: blume-20020121-ff
2160    Description:
2161    
2162    Another day of fiddling with the FFI...
2163    
2164    1. Bug fix/workaround:  CKIT does not complain about negative array
2165       dimensions, so ml-nlffigen has to guard itself against this possibility.
2166       (Otherwise a negative dimension would send it into an infinite loop.)
2167    
2168    2. Some of the abstract types (light objects, light pointers, most "base"
2169       types) in structure C are now eqtypes.
2170    
2171    3. Added constructors and test functions for NULL function pointers.
2172    
2173    ----------------------------------------------------------------------
2174    Name: Matthias Blume
2175    Date: 2002/01/18
2176    Tag: blume-20020118-ready-for-new-release
2177    Description:
2178    
2179    Made config/srcarchiveurl point to a new place.  (Will provide boot
2180    files shortly.)
2181    
2182    Maybe we christen this to be 110.38?
2183    
2184    ----------------------------------------------------------------------
2185    Name: Matthias Blume
2186    Date: 2002/01/18
2187    Tag: blume-20020118-more-ffifiddle
2188    Description:
2189    
2190    Today's FFI fiddling:
2191    
2192      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
2193        C.Set.* that use concrete (MLRep.*) arguments and results instead
2194        of abstract ones.
2195    
2196      - Provided word-style bit operations etc. for "int" representation
2197        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
2198        and Long.
2199    
2200    ----------------------------------------------------------------------
2201    Name: Matthias Blume
2202    Date: 2002/01/18
2203    Tag: blume-20020118-use-x86-fp
2204    Description:
2205    
2206    Now that x86-fast-fp seems to be working, I turned it back on again
2207    by default.  (Seems to work fine now, even with the FFI.)
2208    
2209    Other than that, I added some documentation about the FFI to
2210    src/ml-nlffigen/README and updated the FFI test examples in
2211    src/ml-nlffi-lib/Tests/*.
2212    
2213    ----------------------------------------------------------------------
2214    Name: Allen Leung
2215    Date: 2002/01/17
2216    Tag: leunga-20020117-x86-fast-fp-call
2217    Description:
2218    
2219       1. Fixed a problem with handling return fp values when x86's fast fp
2220          mode is turned on.
2221    
2222       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
2223          of %f32.
2224    
2225       3. Added a constructor INT32lit to the ast of MLRISC tools.
2226    
2227    ----------------------------------------------------------------------
2228    Name: Matthias Blume
2229    Date: 2002/01/16
2230    Tag: blume-20020116-ffifiddle
2231    Description:
2232    
2233    More fiddling with the FFI interface:
2234    
2235     - Make constness 'c instead of rw wherever possible.  This eliminates
2236       the need for certain explicit coercions.  (However, due to ML's
2237       value polymorphism, there will still be many cases where explicit
2238       coercions are necessary.  Phantom types are not the whole answer
2239       to modeling a subtyping relationship in ML.)
2240    
2241     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
2242    
2243     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
2244       clumsy workaround for varargs problem.)
2245    
2246    ----------------------------------------------------------------------
2247    Name: Lal George
2248    Date: 2002/01/15
2249    Tag: <none>
2250    Description:
2251    
2252    1. Since COPY instructions are no longer native to the architecture,
2253       a generic functor can be used to implement the expandCopies function.
2254    
2255    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
2256       TEXT segment.
2257    
2258    ----------------------------------------------------------------------
2259    Name: Matthias Blume
2260    Date: 2002/01/15
2261    Tag: blume-20020115-ffiupdates
2262    Description:
2263    
2264    1. Fix for bug resulting in single-precision float values being returned
2265       incorrectly from FFI calls.
2266    
2267    2. Small modifications to C FFI API:
2268    
2269        - memory-allocation routines return straight objects (no options)
2270          and raise an exception in out-of-memory situations
2271        - unsafe extensions to cast between function pointers and pointers
2272          from/to ints
2273        - added structure C_Debug as an alternative to structure C where
2274          pointer-dereferencing (|*| and |*!) always check for null-pointers
2275        - added open_lib' to DynLinkage;  open_lib' works like open_lib
2276          but also takes a (possibly empty) list of existing library handles
2277          that the current library depends on
2278    
2279    ----------------------------------------------------------------------
2280    Name: Matthias Blume
2281    Date: 2002/01/10
2282    Tag: blume-20020110-newffigen
2283    Description:
2284    
2285    1. Updates to portable graph code.
2286    
2287    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
2288       more scalable now so that even huge interfaces such as the one
2289       for GTK compile in finite time and space. :-)
2290       See src/ml-nlffigen/README for details on what's new.
2291    
2292    ----------------------------------------------------------------------
2293    Name: Lal George
2294    Date: 2001/01/09 14:31:35 EST 2002
2295    Tag: george-20011206-rm-native-copy
2296    Description:
2297    
2298            Removed the native COPY and FCOPY instructions
2299            from all the architectures and replaced it with the
2300            explicit COPY instruction from the previous commit.
2301    
2302            It is now possible to simplify many of the optimizations
2303            modules that manipulate copies. This has not been
2304            done in this change.
2305    
2306    ----------------------------------------------------------------------
2307    Name: Lal George
2308    Date: 2001/12/06 16:50:13 EST 2001
2309    Tag: george-20011206-mlrisc-instruction
2310    Description:
2311    
2312    Changed the representation of instructions from being fully abstract
2313    to being partially concrete. That is to say:
2314    
2315      from
2316            type instruction
2317    
2318      to
2319            type instr                              (* machine instruction *)
2320    
2321            datatype instruction =
2322                LIVE of {regs: C.cellset, spilled: C.cellset}
2323              | KILL of {regs: C.cellset, spilled: C.cellset}
2324              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
2325              | ANNOTATION of {i: instruction, a: Annotations.annotation}
2326              | INSTR of instr
2327    
2328    This makes the handling of certain special instructions that appear on
2329    all architectures easier and uniform.
2330    
2331    LIVE and KILL say that a list of registers are live or killed at the
2332    program point where they appear. No spill code is generated when an
2333    element of the 'regs' field is spilled, but the register is moved to
2334    the 'spilled' (which is present, more for debugging than anything else).
2335    
2336    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
2337    We used to generate:
2338    
2339            DEFFREG f1
2340            f1 := f2 + f3
2341            trapb
2342    
2343    but now generate:
2344    
2345            f1 := f2 + f3
2346            trapb
2347            LIVE {regs=[f1,f2,f3], spilled=[]}
2348    
2349    Furthermore, the DEFFREG (hack) required that all floating point instruction
2350    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
2351    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
2352    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
2353    this architecture).
2354    
2355    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
2356    all the architectures. This will result in further simplification of the
2357    register allocator that must be aware of them for coalescing purposes, and
2358    will also simplify certain aspects of the machine description that provides
2359    callbacks related to parallel copies.
2360    
2361    ANNOTATION should be obvious, and now INSTR represents the honest to God
2362    machine instruction set!
2363    
2364    The <arch>/instructions/<arch>Instr.sml files define certain utility
2365    functions for making porting easier -- essentially converting upper case
2366    to lower case. All machine instructions (of type instr) are in upper case,
2367    and the lower case form generates an MLRISC instruction. For example on
2368    the alpha we have:
2369    
2370      datatype instr =
2371         LDA of {r:cell, b:cell, d:operand}
2372       | ...
2373    
2374      val lda : {r:cell, b:cell, d:operand} -> instruction
2375        ...
2376    
2377    where lda is just (INSTR o LDA), etc.
2378    
2379    ----------------------------------------------------------------------
2380    Name: Matthias Blume
2381    Date: 2001/11/22 21:40:00 EST
2382    Tag: Release_110_37
2383    Description:
2384    
2385    Release 110.37.  This time for real.
2386    
2387    ----------------------------------------------------------------------
2388    Name: Matthias Blume
2389    Date: 2001/11/21 16:35:00 EST
2390    Tag: blume-20011121-foot-in-mouth
2391    Description:
2392    
2393    Removed the "Release_110_37" tag because of a serious bug.
2394    This will be re-tagged once the bug is fixed.
2395    
2396    ----------------------------------------------------------------------
2397    Name: Matthias Blume
2398    Date: 2001/11/21 16:14:00 EST
2399    Tag: blume-20011121-forgottenfile
2400    Description:
2401    
2402    Forgot to add a file.  (Just a .tex-file -- part of
2403    the CM manual source.)
2404    
2405    ----------------------------------------------------------------------
2406    Name: Matthias Blume
2407    Date: 2001/11/21 16:10:00 EST
2408    Tag: blume-20011121-invalid_110_37
2409    Description:
2410    
2411    Note: I removed the original tag "Release_110_37" from this commit
2412          because we found a serious bug in all non-x86 backends.
2413          - Matthias
2414    
2415    1. Modifications to the SML/NJ code generator and to the runtime system
2416       so that code object name strings are directly inserted into code
2417       objects at code generation time.  The only business the runtime system
2418       has with this is now to read the name strings on occasions.
2419       (The encoding of the name string has also changed somewhat.)
2420    
2421    2. CM now implements a simple "set calculus" for specifying export lists.
2422       In particular, it is now possible to refer to the export lists of
2423       other libraries/groups/sources and form unions as well as differences.
2424       See the latest CM manual for details.
2425    
2426    3. An separate notion of "proxy" libraries has again be eliminated from
2427       CM's model.  (Proxy libraries are now simply a special case of using
2428       the export list calculus.)
2429    
2430    4. Some of the existing libraries now take advantage of the new set
2431       calculus.
2432       (Notice that not all libraries have been converted because some
2433       of the existing .cm-files are supposed to be backward compatible
2434       with 110.0.x.)
2435    
2436    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
2437       "General.exnMessage"!  The former relies on a certain hook to be
2438       initialized, and that often does not happen in the stand-alone case.)
2439    
2440    ----------------------------------------------------------------------
2441    Name: Lal George
2442    Date: 2001/11/21  13:56:18 EST
2443    Tag: george-2001121-pseudo-ops
2444    Description:
2445    
2446      Implemented a complete redesign of MLRISC pseudo-ops. Now there
2447      ought to never be any question of incompatabilities with
2448      pseudo-op syntax expected by host assemblers.
2449    
2450      For now, only modules supporting GAS syntax are implemented
2451      but more should follow, such as MASM, and vendor assembler
2452      syntax, e.g. IBM as, Sun as, etc.
2453    
2454    ----------------------------------------------------------------------
2455    Name: Matthias Blume
2456    Date: 2001/11/14 11:52:00 EST
2457    Tag: blume-20011114-srcname
2458    Description:
2459    
2460    1. Routed the name of the current source file to mlriscgen where it
2461       should be directly emitted into the code object.  (This last part
2462       is yet to be done.)
2463    
2464    2. Some cleanup of the pgraph code to make it match the proposal that
2465       I put out the other day.  (The proposal notwithstanding, things are
2466       still in flux here.)
2467    
2468    ----------------------------------------------------------------------
2469    Name: Lal George
2470    Date: 2001/11/14 09:44:04 EST
2471    Tag:
2472    Description:
2473    
2474      Fix for a backpatching bug reported by Allen.
2475    
2476      Because the boundary between short and long span-dependent
2477      instructions is +/- 128, there are an astounding number of
2478      span-dependent instructions whose size is over estimated.
2479    
2480      Allen came up with the idea of letting the size of span
2481      dependent instructions be non-monotonic, for a maxIter
2482      number of times, after which the size must be monotonically
2483      increasing.
2484    
2485      This table shows the number of span-dependent instructions
2486      whose size was over-estimated as a function of maxIter, for the
2487      file Parse/parse/ml.grm.sml:
2488    
2489         maxIter            # of instructions:
2490            10                      687
2491            20                      438
2492            30                      198
2493            40                        0
2494    
2495      In compiling the compiler, there is no significant difference in
2496      compilation speed between maxIter=10 and maxIter=40. Actually,
2497      my measurements showed that maxIter=40 was a tad faster than
2498      maxIter=10! Also 96% of the  files in the compiler reach a fix
2499      point within 13 iterations, so fixing maxIter at 40, while high,
2500      is okay.
2501    
2502    ----------------------------------------------------------------------
2503    Name: Matthias Blume
2504    Date: 2001/10/31 15:25:00 EST
2505    Tag: blume-20011031-pgraph
2506    Description:
2507    
2508    CKIT:
2509    * Changed the "Function" constructor of type Ast.ctype to carry optional
2510      argument identifiers.
2511    * Changed the return type of TypeUtil.getFunction accordingly.
2512    * Type equality ignores the argument names.
2513    * TypeUtil.composite tries to preserve argument names but gives up quickly
2514      if there is a mismatch.
2515    
2516    installation script:
2517    * attempts to use "curl" if available (unless "wget" is available as well)
2518    
2519    CM:
2520    * has an experimental implementation of "portable graphs" which I will
2521      soon propose as an implementation-independent library format
2522    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
2523    
2524    NLFFI-LIB:
2525    * some cleanup (all cosmetic)
2526    
2527    NLFFIGEN:
2528    * temporarily disabled the mechanism that suppresses ML output for
2529      C definitions whose identifiers start with an underscore character
2530    * generate val bindings for enum constants
2531    * user can request that only one style (light or heavy) is being used;
2532      default is to use both (command-line arguments: -heavy and -light)
2533    * fixed bug in handling of function types involving incomplete pointers
2534    * generate ML entry points that take record arguments (i.e., using
2535      named arguments) for C functions that have a prototype with named
2536      arguments
2537      (see changes to CKIT)
2538    
2539    ----------------------------------------------------------------------
2540    Name: Allen Leung
2541    Date: 2001/10/27 20:34:00 EDT
2542    Tag: leunga-20011027-x86-fast-fp-call
2543    Description:
2544    
2545       Fixed the bug described in blume-20010920-slowfp.
2546    
2547       The fix involves
2548          1. generating FCOPYs in FSTP in ia32-svid
2549          2. marking a CALL with the appropriate annotation
2550    
2551    ----------------------------------------------------------------------
2552    Name: Matthias Blume
2553    Date: 2001/10/16 11:32:00 EDT
2554    Tag: blume-20011016-netbsd
2555    Description:
2556    
2557    Underscore patch from Chris Richards (fixing problem with compiling
2558    runtime system under recent NetBSD).
2559    
2560    ----------------------------------------------------------------------
2561    Name: Allen Leung
2562    Date: 2001/10/12 17:18:32 EDT 2001
2563    Tag: leung-20011012-x86-printflowgraph
2564    Description:
2565    
2566    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
2567    
2568    ----------------------------------------------------------------------
2569    Name: Lal George
2570    Date: 2001/10/11 23:51:34 EDT
2571    Tag: george-20011011-too-many-instrs
2572    Description:
2573    
2574    The representation of a program point never expected to see more
2575    than 65536 instructions in a basic block!
2576    
2577    ----------------------------------------------------------------------
2578    Name: Lal George
2579    Date: 2001/10/09 09:41:37 EDT
2580    Tag: george-20011008-mlrisc-labels
2581    Description:
2582    
2583    Changed the machine description files to support printing of
2584    local and global labels in assembly code, based on host assembler
2585    conventions.
2586    
2587    ----------------------------------------------------------------------
2588    Name: Matthias Blume
2589    Date: 2001/09/25 15:25:00 EDT
2590    Tag: blume-20010925-exninfo
2591    Description:
2592    
2593    I provided a non-hook implementation of exnName (at the toplevel) and
2594    made the "dummy" implementation of exnMessage (at the toplevel) more
2595    useful: if nothing gets "hooked in", then at least you are going to
2596    see the exception name and a message indicating why you don't see more.
2597    
2598    [For the time being, programs that need exnMessage and want to use
2599    ml-build should either use General.exnMessage (strongly recommended) or
2600    refer to structure General at some other point so that CM sees a
2601    static dependency.]
2602    
2603    [Similar remarks go for "print" and "use":  If you want to use their
2604    functionality in stand-alone programs generated by ml-build, then use
2605    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
2606    
2607    ----------------------------------------------------------------------
2608    Name: Matthias Blume
2609    Date: 2001/09/20 17:28:00 EDT
2610    Tag: blume-20010920-slowfp
2611    Description:
2612    
2613    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
2614    
2615    ----------------------------------------------------------------------
2616    Name: Matthias Blume
2617    Date: 2001/09/20 17:20:00 EDT
2618    Tag: blume-20010920-canonicalpaths
2619    Description:
2620    
2621    0. Updated the BOOT file (something that I forgot to do earlier).
2622    
2623    1. Small internal change to CM so that it avoids "/../" in filenames
2624       as much as possible (but only where it is safe).
2625    
2626    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
2627       that arguments that contain delimiters are passed through correctly.
2628       This change also means that all "special" arguments of the form
2629       @SMLxxx... must come first.
2630    
2631    3. Changed install script to put relative anchor names for tool commands
2632       into pathconfig.
2633    
2634    ----------------------------------------------------------------------
2635    Name: Matthias Blume
2636    Date: 2001/09/18 15:35:00 EDT
2637    Tag: blume-20010918-readme11036
2638    Description:
2639    
2640    Added README files.
2641    
2642    ----------------------------------------------------------------------
2643    Name: Matthias Blume
2644    Date: 2001/09/18 11:45:00 EDT
2645    Tag: Release_110_36 (retag)
2646    Description:
2647    
2648    Fixed mistake in config/preloads. Retagged as 110.36.
2649    
2650    ----------------------------------------------------------------------
2651    Name: Matthias Blume
2652    Date: 2001/09/18 09:40:00 EDT
2653    Tag: Release_110_36_orig (tag changed)
2654    Description:
2655    
2656    New version (110.36).  New bootfiles.
2657    
2658    ----------------------------------------------------------------------
2659    Name: Matthias Blume
2660    Date: 2001/09/14 16:15:00 EDT
2661    Tag: blume-20010914-x86fastfp
2662    Description:
2663    
2664    John committed some changes that Allen made, in particular a (hopefully)
2665    correctly working version of the x86-fp module.
2666    
2667    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
2668    flag to "true".  Everything seems to compile to a fixpoint ok, and
2669    "mandelbrot" speeds up by about 15%.
2670    
2671    ----------------------------------------------------------------------
2672    Name: Matthias Blume
2673    Date: 2001/09/13 11:20:00 EDT
2674    Tag: blume-20010913-minimal
2675    Description:
2676    
2677    1. Stefan Monnier's patch to fix a miscompilation problem that
2678       was brought to light by John Reppy's work on Moby.
2679    
2680    2. Implemented a minimal "structure Compiler" that contains just
2681       "version" and "architecture".  The minimal version will be
2682       available when the full version is not.  This is for backward-
2683       compatibility with code that wants to test Compiler.version.
2684    
2685    ----------------------------------------------------------------------
2686    Name: Matthias Blume
2687    Date: 2001/08/28 14:03:00 EDT
2688    Tag: blume-20010828-ml-lex
2689    Description:
2690    
2691    Fix for bug 1581, received from Neophytos Michael.
2692    
2693    ----------------------------------------------------------------------
2694    Name: Matthias Blume
2695    Date: 2001/08/27 11:20:00 EDT
2696    Tag: blume-20010827-readme11035
2697    Description:
2698    
2699    Fleshed out the README file for 110.35.
2700    
2701    ----------------------------------------------------------------------
2702    Name: Matthias Blume
2703    Date: 2001/08/24 17:10:00 EDT
2704    Tag: Release_110_35
2705    Description:
2706    
2707    New version number (110.35).  New bootfiles.
2708    
2709    ----------------------------------------------------------------------
2710    Name: Lal George
2711    Date: 2001/08/24 13:47:18 EDT 2001
2712    Tag: george-20010824-MLRISC-graphs
2713    Description:
2714    
2715     removed clusters from MLRISC completely and replaced with graphs.
2716    
2717    ----------------------------------------------------------------------
2718    Name: Matthias Blume
2719    Date: 2001/08/23 17:50:00 EDT
2720    Tag: blume-20010823-toplevel
2721    Description:
2722    
2723    - some reorganization of the code that implements various kinds of
2724      environments in the compiler (static, dynamic, symbolic, combined)
2725    - re-implemented the EnvRef module so that evalStream works properly
2726      (if the stream contains references to "use", "CM.make", etc.)
2727    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
2728    
2729    ----------------------------------------------------------------------
2730    Name: Matthias Blume
2731    Date: 2001/08/20 15:50 EDT
2732    Tag: blume20010820-slipup
2733    Description:
2734    
2735    I forgot to commit a few files.  Here they are...
2736    
2737    ----------------------------------------------------------------------
2738    Name: Matthias Blume
2739    Date: 2001/08/20 15:35:00 EDT
2740    Tag: blume-20010820-debugprof
2741    Description:
2742    
2743    !!!! NEW BOOTFILES !!!!
2744    
2745    This is another round of reorganizing the compiler sources.  This
2746    time the main goal was to factor out all the "instrumentation"
2747    passes (for profiling and backtracing) into their own library.
2748    The difficulty was to do it in such a way that it does not depend
2749    on elaborate.cm but only on elabdata.cm.
2750    
2751    Therefore there have been further changes to both elaborate.cm and
2752    elabdata.cm -- more "generic" things have been moved from the former
2753    to the latter.  As a result, I was forced to split the assignment
2754    of numbers indicating "primtyc"s into two portions: SML-generic and
2755    SML/NJ-specific.  Since it would have been awkward to maintain,
2756    I bit the bullet and actually _changed_ the mapping between these
2757    numbers and primtycs.  The bottom line of this is that you need
2758    a new set of bin- and bootfiles.
2759    
2760    I have built new bootfiles for all architectures, so doing a fresh
2761    checkout and config/install.sh should be all you need.
2762    
2763    The newly created library's name is
2764    
2765        $smlnj/viscomp/debugprof.cm
2766    
2767    and its sources live under
2768    
2769        src/compiler/DebugProf
2770    
2771    ----------------------------------------------------------------------
2772    Name: Matthias Blume
2773    Date: 2001/08/15 17:15:00 EDT
2774    Tag: blume-20010815-compreorg
2775    Description:
2776    
2777    This is a first cut at reorganizing the CM libraries that make up the
2778    core of the compiler.  The idea is to separate out pieces that could
2779    be used independently by tools, e.g., the parser, the typechecker, etc.
2780    
2781    The current status is a step in this direction, but it is not quite
2782    satisfactory yet.  Expect more changes in the future.
2783    
2784    Here is the current (new) organization...
2785    
2786        What used to be $smlnj/viscomp/core.cm is now divided into
2787        six CM libraries:
2788    
2789             $smlnj/viscomp/basics.cm
2790                           /parser.cm
2791                           /elabdata.cm
2792                           /elaborate.cm
2793                           /execute.cm
2794                           /core.cm
2795    
2796        The CM files for these libraries live under src/system/smlnj/viscomp.
2797        All these libraries are proxy libraries that contain precisely
2798        one CM library component.  Here are the locations of the components
2799        (all within the src/compiler tree):
2800    
2801             Basics/basics.cm
2802             Parse/parser.cm
2803             ElabData/elabdata.cm
2804             Elaborator/elaborate.cm
2805             Execution/execute.cm
2806             core.cm
2807    
2808         [This organization is the same that has been used already
2809         for a while for the architecture-specific parts of the visible
2810         compiler and for the old version of core.cm.]
2811    
2812         As you will notice, many source files have been moved from their
2813         respective original locations to a new home in one of the above
2814         subtrees.
2815    
2816         The division of labor between the new libraries is the following:
2817    
2818             basics.cm:
2819                - Simple, basic definitions that pertain to many (or all) of
2820                  the other libraries.
2821             parser.cm:
2822                - The SML parser, producing output of type Ast.dec.
2823                - The type family for Ast is also defined and exported here.
2824             elabdata.cm:
2825                - The datatypes that describe input and output of the elaborator.
2826                  This includes types, absyn, and static environments.
2827             elaborator.cm:
2828                - The SML/NJ type checker and elaborator.
2829                  This maps an Ast.dec (with a given static environment) to
2830                  an Absyn.dec (with a new static environment).
2831                - This libraries implements certain modules that used to be
2832                  structures as functors (to remove dependencies on FLINT).
2833             execute.cm:
2834                - Everything having to do with executing binary code objects.
2835                - Dynamic environments.
2836             core.cm:
2837                - SML/NJ-specific instantiations of the elaborator and MLRISC.
2838                - Top-level modules.
2839                - FLINT (this should eventually become its own library)
2840    
2841    Notes:
2842    
2843    I am not 100% happy with the way I separated the elaborator (and its
2844    data structures) from FLINT.  Two instances of the same problem:
2845    
2846        1. Data structures contain certain fields that carry FLINT-specific
2847           information.  I hacked around this using exn and the property list
2848           module from smlnj-lib.  But the fact that there are middle-end
2849           specific fields around at all is a bit annoying.
2850    
2851        2. The elaborator calculates certain FLINT-related information.  I tried
2852           to make this as abstract as I could using functorization, but, again,
2853           the fact that the elaborator has to perform calculations on behalf
2854           of the middle-end at all is not nice.
2855    
2856        3. Having to used exn and property lists is unfortunate because it
2857           weakens type checking.  The other alternative (parameterizing
2858           nearly *everything*) is not appealing, though.
2859    
2860    I removed the "rebinding =" warning hack because due to the new organization
2861    it was awkward to maintain it.  As a result, the compiler now issues some of
2862    these warnings when compiling init.cmi during bootstrap compilation. On
2863    the plus side, you also get a warning when you do, for example:
2864       val op = = Int32.+
2865    which was not the case up to now.
2866    
2867    I placed "assign" and "deref" into the _Core structure so that the
2868    code that deals with the "lazy" keyword can find them there.  This
2869    removes the need for having access to the primitive environment
2870    during elaboration.
2871    
2872    ----------------------------------------------------------------------
2873    Name: Matthias Blume
2874    Date: 2001/08/13
2875    Tag: blume-20010813-closures
2876    Description:
2877    
2878    This fix was sent to us by Zhong Shao.  It is supposed to improve the
2879    performance of certain loops by avoiding needless closure allocation.
2880    
2881    ----------------------------------------------------------------------
2882    Name: Lal George
2883    Date: 2001/07/31 10:03:23 EDT 2001
2884    Tag: george-20010731-x86-fmalloc
2885    Description: Fixed bug in x86 calls
2886    
2887        There was a bug where call instructions would mysteriously
2888        vanish. The call instruction had to be one that returned
2889        a floating point value.
2890    
2891    ----------------------------------------------------------------------
2892    Name: Lal George
2893    Date: 2001/07/19 16:36:29 EDT 2001
2894    Tag: george-20010719-simple-cells
2895    Description:
2896    
2897    I have dramatically simplified the interface for CELLS in MLRISC.
2898    
2899    In summary, the cells interface is broken up into three parts:
2900    
2901      1. CellsBasis : CELLS_BASIS
2902    
2903            CellsBasis is a top level structure and common for all
2904            architectures.  it contains the definitions of basic datatypes
2905            and utility  functions over these types.
2906    
2907      2. functor Cells() : CELLS
2908    
2909            Cells generates an interface for CELLS that incorporates the
2910            specific resources on the target architecture, such as the
2911            presence of special register classes, their number and size,
2912            and various useful substructures.
2913    
2914      3. <ARCH>CELLS
2915    
2916            e.g. SparcCells: SPARCCELLS
2917    
2918            <ARCH>CELLS usually contains additional bindings for special
2919            registers  on the architecture, such as:
2920    
2921                    val r0 : cell           (* register zero *)
2922                    val y : cell            (* Y register *)
2923                    val psr : cell          (* processor status register *)
2924                    ...
2925    
2926            The structure returned by applying the Cells functor is opened
2927            in this interface.
2928    
2929    The main implication of all this is that the datatypes for cells is
2930    split between CellsBasis and CELLS -- a fairly simple change for user
2931    code.
2932    
2933    In the old scheme the CELLS interface had a definitional binding of
2934    the form:
2935    
2936            signature CELLS = sig
2937    
2938               structure CellsBasis = CellsBasis
2939    
2940               ...
2941    
2942            end
2943    
2944    With all the sharing constraints that goes on in MLRISC, this old
2945    design  quickly leads to errors such as:
2946    
2947            "structure definition spec inside of sharing ... "
2948    
2949    
2950    and appears to require an unacceptable amount of sharing and where
2951    constraint hackery.
2952    
2953    I think this error message (the interaction of definitional specs and
2954    sharing) requires more explanation on our web page.
2955    
2956    ----------------------------------------------------------------------
2957    Name: Matthias Blume
2958    Date: 2001/07/19 15:00:00 EDT
2959    Tag: blume-20010719-libreorg
2960    Description:
2961    
2962    This update puts together a fairly extensive but straightforward change
2963    to the way the libraries that implement the interactive system are
2964    organized:
2965    
2966       The biggest change is the elimination of structure Compiler.  As a
2967       replacement for this structure, there is now a CM library
2968       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
2969       that exports all the substructures of the original structure Compiler
2970       directly.  So instead of saying Compiler.Foo.bar one now simply
2971       says Foo.bar.  (The CM libraries actually export a collection of
2972       structures that is richer than the collection of substructures of
2973       structure Compiler.)
2974    
2975       To make the transition smooth, there is a separate library called
2976       $smlnj/compiler/compiler.cm which puts together and exports the
2977       original structure Compiler (or at least something very close to it).
2978    
2979       There are five members of the original structure Compiler
2980       that are not exported directly but which instead became members
2981       of a new structure Backend (described by signature BACKEND).  These are:
2982       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
2983       Interact (: INTERACT), structure Machine (: MACHINE), and val
2984       architecture (: string).
2985    
2986       Structure Compiler.Version has become structure CompilerVersion.
2987    
2988       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
2989       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
2990       or x86, respectively.
2991       Each of these exports the same frontend structures that
2992       $smlnj/compiler.cm exports.  But they do not have a structure Backend
2993       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
2994       Hppa, PPC, Sparc, or X86, respectively.
2995    
2996       Library $smlnj/compiler/all.cm exports the union of the exports of
2997       $smlnj/compiler/<arch>.cm
2998    
2999       There are no structures <Arch>Compiler anymore, use
3000       $smlnj/compiler/<arch>.cm instead.
3001    
3002       Library host-compiler-0.cm is gone.  Instead, the internal library
3003       that instantiates CM is now called cm0.cm.  Selection of the host
3004       compiler (backend) is no longer done here but. (Responsibility for it
3005       now lies with $smlnj/compiler/current.cm.  This seems to be more
3006       logical.)
3007    
3008       Many individual files have been moved or renamed.  Some files have
3009       been split into multiple files, and some "dead" files have been deleted.
3010    
3011    Aside from these changes to library organization, there are also changes
3012    to the way the code itself is organized:
3013    
3014       Structure Binfile has been re-implemented in such a way that it no
3015       longer needs any knowledge of the compiler.  It exclusively deals
3016       with the details of binfile layout.  It no longer invokes the
3017       compiler (for the purpose of creating new prospective binfile
3018       content), and it no longer has any knowledge of how to interpret
3019       pickles.
3020    
3021       Structure Compile (: COMPILE) has been stripped down to the bare
3022       essentials of compilation.  It no longer deals with linking/execution.
3023       The interface has been cleaned up considerably.
3024    
3025       Utility routines for dealing with linking and execution have been
3026       moved into their own substructures.
3027    
3028       (The ultimate goal of these changes is to provide a light-weight
3029       binfile loader/linker (at least for, e.g., stable libraries) that
3030       does not require CM or the compiler to be present.)
3031    
3032    CM documentation has been updated to reflect the changes to library
3033    organization.
3034    
3035    ----------------------------------------------------------------------
3036    Name: Matthias Blume
3037    Date: 2001/07/10 17:30:00 EDT
3038    Tag: Release_110_34
3039    Description:
3040    
3041    Minor tweak to 110.34 (re-tagged):
3042    
3043      - README.html file added to CVS repository
3044      - runtime compiles properly under FreeBSD 3.X and 4.X
3045    
3046    ----------------------------------------------------------------------
3047    Name: Matthias Blume
3048    Date: 2001/07/10 17:30:00 EDT
3049    Tag: Release_110_34
3050    Description:
3051    
3052    New version number (110.34). New bootfiles.
3053    
3054    ----------------------------------------------------------------------
3055    Name: Matthias Blume
3056    Date: 2001/07/09 16:00:00 EDT
3057    Tag: blume-20010709-more-varargs
3058    Description:
3059    
3060    I changed the handling of varargs in ml-nlffigen again:
3061    The ellipsis ... will now simply be ignored (with an accompanying warning).
3062    
3063    The immediate effect is that you can actually call a varargs function
3064    from ML -- but you can't actually supply any arguments beyond the ones
3065    specified explicitly.  (For example, you can call printf with its format
3066    string, but you cannot pass additional arguments.)
3067    
3068    This behavior is only marginally more useful than the one before, but
3069    it has the advantage that a function or, more importantly, a function
3070    type never gets dropped on the floor, thus avoiding follow-up problems with
3071    other types that refer to the offending one.
3072    
3073    ----------------------------------------------------------------------
3074    Name: Matthias Blume
3075    Date: 2001/07/09 11:25:00 EDT
3076    Tag: blume-20010709-varargs
3077    Description:
3078    
3079    1. ckit-lib.cm now exports structure Error
3080    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
3081       with a warning accompanied by a source location.  Moreover, it
3082       merely skips the offending function or type and proceeds with the
3083       rest of its work.u  As a result, one can safely feed C code containing
3084       "..." to ml-nlffigen.
3085    3. There are some internal improvements to CM, providing slightly
3086       more general string substitutions in the tools subsystem.
3087    
3088    ----------------------------------------------------------------------
3089    Name: Matthias Blume
3090    Date: 2001/06/27 15:10:00 EDT
3091    Tag: blume-20010627-concur
3092    Description:
3093    
3094    Fixed a small bug in CM's handling of parallel compilation.
3095    (You could observe the bug by Control-C-interrupting an ordinary
3096    CMB.make or CM.stabilize and then attaching some compile servers.
3097    The result was that all of a sudden the previously interrupted
3098    compilation would continue on its own.  This was because of
3099    an over-optimization: CM did not bother to clean out certain queues
3100    when no servers were attached "anyway", resulting in the contents
3101    of these queues to grab control when new servers did get attached.)
3102    
3103    There is also another minor update to the CM manual.
3104    
3105    ----------------------------------------------------------------------
3106    Name: Matthias Blume
3107    Date: 2001/06/26 16:15:00 EDT
3108    Tag: blume-20010626-cmdoc
3109    Description:
3110    
3111    Minor typo fixed in CM manual (syntax diagram for libraries).
3112    
3113    ----------------------------------------------------------------------
3114    Name: Matthias Blume
3115    Date: 2001/06/25 22:55:00 EDT
3116    Tag: blume-20010625-x86pc
3117    Description:
3118    
3119    Fixed a nasty bug in the X86 assembly code that caused signal
3120    handlers to fail (crash) randomly.
3121    
3122    ----------------------------------------------------------------------
3123    Name: Matthias Blume
3124    Date: 2001/06/25 12:05:00 EDT
3125    Tag: blume-20010625-nlffigen
3126    Description:
3127    
3128    This update fixes a number of minor bugs in ml-nlffigen as reported by
3129    Nick Carter <nbc@andrew.cmu.edu>.
3130    
3131      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
3132      2. Default names for generated files are now derived from the name of
3133         the C file *without its directory*.  In particular, this causes generated
3134         files to be placed locally even if the C file is in some system directory.
3135      3. Default names for generated signatures and structures are also derived
3136         from the C file name without its directory.  This avoids silly things
3137         like "structure GL/GL".
3138         (Other silly names are still possible because ml-nlffigen does not do
3139          a thorough check of whether generated names are legal ML identifiers.
3140          When in doubt, use command line arguments to force particular names.)
3141    
3142    ----------------------------------------------------------------------
3143    Name: Matthias Blume
3144    Date: 2001/06/21 12:25:00 EDT
3145    Tag: blume-20010621-eXene
3146    Description:
3147    
3148    eXene now compiles and (sort of) works again.
3149    
3150    The library name (for version > 110.33) is $/eXene.cm.
3151    
3152    I also added an new example in src/eXene/examples/nbody.  See the
3153    README file there for details.
3154    
3155    ----------------------------------------------------------------------
3156    Name: Matthias Blume
3157    Date: 2001/06/20 16:40:00 EDT
3158    Tag: blume-20010620-cml
3159    Description:
3160    
3161    CML now compiles and works again.
3162    
3163    Libraries (for version > 110.33):
3164    
3165      $cml/cml.cm            Main CML library.
3166      $cml/basis.cm          CML's version of $/basis.cm.
3167      $cml/cml-internal.cm   Internal helper library.
3168      $cml/core-cml.cm       Internal helper library.
3169      $cml-lib/trace-cml.cm  Tracing facility.
3170      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
3171    
3172    The installer (config/install.sh) has been taught how to properly
3173    install this stuff.
3174    
3175    ----------------------------------------------------------------------
3176    Name: Matthias Blume
3177    Date: 2001/06/19 17:55:00 EDT
3178    Tag: blume-20010619-instantiate
3179    Description:
3180    
3181    This un-breaks the fix for bug 1432.
3182    (The bug was originally fixed in 110.9 but I broke it again some
3183    time after that.)
3184    
3185    ----------------------------------------------------------------------
3186    Name: Matthias Blume
3187    Date: 2001/06/19 17:25:00 EDT
3188    Tag: blume-20010619-signals
3189    Description:
3190    
3191    This should (hopefully) fix the long-standing signal handling bug.
3192    (The runtime system was constructing a continuation record with an
3193    incorrect descriptor which would cause the GC to drop data on the floor...)
3194    
3195    ----------------------------------------------------------------------
3196    Name: Matthias Blume
3197    Date: 2001/06/15 15:05:00 EDT
3198    Tag: blume-20010615-moresparc
3199    Description:
3200    
3201    Here is a short late-hour update related to Sparc c-calls:
3202    
3203     -- made handling of double-word arguments a bit smarter
3204    
3205     -- instruction selection phase tries to collapse certain clumsily
3206        constructed ML-Trees; typical example:
3207    
3208            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
3209    
3210        This currently has no further impact on SML/NJ since mlriscGen does
3211        not seem to generate such patterns in the first place, and c-calls
3212        (which did generate them in the beginning) has meanwhile been fixed
3213        so as to avoid them as well.
3214    
3215    ----------------------------------------------------------------------
3216    Name: Matthias Blume
3217    Date: 2001/06/15 15:05:00 EDT
3218    Tag: blume-20010615-sparc
3219    Description:
3220    
3221    The purpose of this update is to provide an implementation of NLFFI
3222    on Sparc machines.
3223    
3224    Here are the changes in detail:
3225    
3226       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
3227       the Sparc implementation of the c-calls API.
3228       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
3229       for accessing the ML frame.  Thus, we have a real frame pointer and
3230       can freely modify %sp without need for an omit-frame-ptr phase.
3231       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
3232       this case.
3233       * ml-nlffigen has been taught to produce code for different architectures
3234       and calling conventions.
3235       * In a way similar to what was done in the x86 case, the Sparc
3236       backend uses its own specific extension to mltree.  (For example,
3237       it needs to be able to generate UNIMP instructions which are part
3238       of the calling convention.)
3239       * ml-nlffi-lib was reorganized to make it more modular (in particular,
3240       to make it easier to plug in new machine- and os-dependent parts).
3241    
3242    There are some other fairly unrelated bug fixes and cleanups as well:
3243    
3244       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
3245       that they properly share their libraries with existing SML/NJ libraries.
3246       * I fixed a minor cosmetic bug in CM, supressing certain spurious
3247       follow-up error messages.
3248       * Updates to CM/CMB documentation.
3249    
3250    TODO items:
3251    
3252       * MLRISC should use a different register as its asmTemp on the Sparc.
3253         (The current %o2 is a really bad choice because it is part of the
3254         calling conventions, so things might interfere in unexpected ways.)
3255    
3256    ----------------------------------------------------------------------
3257    Name: Matthias Blume
3258    Date: 2001/06/07
3259    Tag: blume-20010607-calls
3260    Description:
3261    
3262    A number of internal changes related to C calls and calling conventions:
3263    
3264    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
3265       number of bytes popped implicitly (by the callee).  In most cases
3266       this field is 0 but on x86/win32 it is some non-zero value.  This
3267       is information provided for the benefit of the "omit-frameptr" pass.
3268    2. The CALL instruction on the x86 carries a similar "pops" field.
3269       The instruction selection phase copies its value from the ML-Tree
3270       CALL statement.
3271    3. On all other architectures, the instruction selection phase checks
3272       whether "pops=0" and complains if not.
3273    4. The c-calls implementation for x86 now accepts two calling conventions:
3274       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
3275       up after the call and pops is set to 0.  For "stdcall", the caller
3276       does nothing, leaving the cleanup to the callee; pops is set to
3277       the number of bytes that were pushed onto the stack.
3278    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
3279       distinguish between "ccall" and "stdcall".
3280    6. The UNIMP instruction has been added to the supported Sparc instruction
3281       set. (This is needed for implementing the official C calling convention
3282       on this architecture.)
3283    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
3284       work with the latest CM.
3285    
3286    ----------------------------------------------------------------------
3287    Name: Matthias Blume
3288    Date: 2001/06/05 15:10:00 EDT
3289    Tag: blume-20010605-cm-index
3290    Description:
3291    
3292    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
3293    
3294    1. CM can now generate "index files".  These are human-readable files
3295       that list on a per-.cm-file basis each toplevel symbol defined or
3296       imported.  The location of the index file for
3297       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
3298       To enable index-file generation, set CM.Control.generate_index to true
3299       or export an environment-symbol: export CM_GENERATE_INDEX=true.
3300    
3301       The CM manual has been updated accordingly.
3302    
3303    2. I made some slight modifications to the c-calls API in MLRISC.
3304    
3305         a) There is now a callback to support saving/restoring of
3306            dedicated but caller-save registers around the actual call
3307            instruction.
3308         b) One can optionally specify a comment-annotation for the
3309            call instruction.
3310    
3311    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
3312       (For example, the comment annotation shows the C prototype of
3313        the function being called.)
3314    
3315    ----------------------------------------------------------------------
3316    Name: Matthias Blume
3317    Date: 2001/06/01 13:30:00 EDT
3318    Tag: blume-20010601-nlffi-cleanup
3319    Description:
3320    
3321    This is mostly a cleanup of MLFFI stuff:
3322    
3323       - some signature files have been put into a more exposed place
3324       - the ugly 'f type parameter is gone (simplifies types tremendously!)
3325       - ml-nlffigen changed accordingly
3326       - tutorial updated
3327    
3328    Other changes:
3329    
3330       - author's affiliation in CM manual(s) updated
3331       - some more recognized keywords added to Allen's sml.sty
3332    
3333    ----------------------------------------------------------------------
3334    Name: Matthias Blume
3335    Date: 2001/05/25 15:30:00 EDT
3336    Tag: blume-20010525-iptr
3337    Description:
3338    
3339      - put the official 110.33-README (as it appears on the ftp server) under
3340        CVS
3341      - fixed a small bug related to incomplete pointer types in
3342        ml-nlffigen
3343      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
3344        (it does not need the 'f type parameter)
3345    
3346    ----------------------------------------------------------------------
3347    Name: Matthias Blume
3348    Date: 2001/05/23 14:30:00 EDT
3349    Tag: Release_110_33
3350    Description:
3351    
3352    New version number (110.33).  New bootfiles.
3353    
3354    ----------------------------------------------------------------------
3355    Name: Matthias Blume
3356    Date: 2001/05/22 18:06:00 EDT
3357    Tag: blume-20010522-targets
3358    Description:
3359    
3360    Made install.sh use file config/targets.customized if it exists, falling
3361    back to config/targets if it doesn't.  This way one can have a customized
3362    version of the targets file without touching the "real thing", thus
3363    eliminating the constant fear of accidentally checking something bogus
3364    back into the CVS repository...  (File config/targets.customized must
3365    not be added to the repository!)
3366    
3367    ----------------------------------------------------------------------
3368    Name: Matthias Blume
3369    Date: 2001/05/22 16:30:00 EDT
3370    Tag: blume-20010522-minitut
3371    Description:
3372    
3373    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
3374       struct returns.
3375    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
3376       incomplete, preliminary documentation for NLFFI.
3377    
3378    ----------------------------------------------------------------------
3379    Name: Matthias Blume
3380    Date: 2001/05/14 11:30:00 EDT
3381    Tag: blume-20010514-script
3382    Description:
3383    
3384    Some bugs in install script fixed.
3385    
3386    In addition to that I also made a slight change to the NLFFI API:
3387    Functors generated by ml-nlffigen now take the dynamic library as a
3388    straight functor argument, not as a suspended one.  (The original
3389    functor code used to force the suspension right away anyway, so there
3390    was nothing gained by this complication of the interface.)
3391    
3392    ----------------------------------------------------------------------
3393    Name: Matthias Blume
3394    Date: 2001/05/11 14:35:00 EDT
3395    Tag: blume-20010511-ml-nlffi
3396    Description:
3397    
3398    I finally took the plunge and added my new FFI code to the main
3399    repository.  For x86-linux it is now ready for prime-time.
3400    
3401    There are two new subdirectories of "src":
3402    
3403      - ml-nlffi-lib:
3404           The utility library for programs using the FFI interface.
3405           Here is the implementation of $/c.cm and its associated low-level
3406           partners $/c-int.cm and $/memory.cm.
3407      - ml-nlffigen:
3408           A stand-alone program for generating ML glue code from C source
3409           code.
3410    
3411    Building ml-nlffigen requires $/ckit-lib.cm.
3412    
3413    The config/install.sh script has been updates to do the Right Thing
3414    (hopefully).
3415    
3416    Notice that the source tree for the C-Kit will not be put under "src"
3417    but directly under the installation root directory.  (This is the
3418    structure that currently exists on the CVS server when you check out
3419    module "sml".)  Fortunately, config/install.sh knows about this oddity.
3420    
3421    Bugs: No documentation yet.
3422    
3423    ----------------------------------------------------------------------
3424    Name: Matthias Blume
3425    Date: 2001/05/09 16:35:00 EDT
3426    Tag: blume-20010509-cpscontract
3427    Description:
3428    
3429    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
3430    wrapper/unwrapper elimination did not decrement usage counts and some
3431    dead variables got overlooked by the dead-up logic.)
3432    
3433    ----------------------------------------------------------------------
3434    Name: Lal George
3435    Date: 2001/05/08  17:26:09 EDT
3436    Tag: george-20010508-omit-frameptr
3437    Description:
3438    
3439    Changes to implement the omit-frame-pointer optimization to support
3440    raw C calls. For now, there is only support on the Intel x86, but
3441    other architectures will follow as more experience is gained with this.
3442    
3443    
3444    ----------------------------------------------------------------------
3445    Name: Matthias Blume
3446    Date: 2001/05/07 14:40:00 EDT
3447    Tag: blume-20010507-proxies
3448    Description:
3449    
3450    I made into "proxy libraries" all libraries that qualify for such a
3451    change.  (A qualifying library is a library that has another library or
3452    groups as its sole member and repeats that member's export list
3453    verbatim.  A proxy library avoids this repetition by omitting its export
3454    list, effectively inheriting the list that its (only) member exports.
3455    See the CM manual for more explanation.)
3456    The main effect is that explicit export lists for these libraries
3457    do not have to be kepts in sync, making maintenance a bit easier.
3458    
3459    I also added copyright notices to many .cm-files.
3460    
3461    Last but not least, I made a new set of bootfiles.
3462    
3463    ----------------------------------------------------------------------
3464    Name: Matthias Blume
3465    Date: 2001/05/04 17:00:00 EDT
3466    Tag: blume-20010504-cm-lsplit
3467    Description:
3468    
3469    0. John merged pending changes to $/smlnj-lib.cm
3470    
3471    1. Allen's previous change accidentally backed out of one of Lal's
3472       earlier changes.  I undid this mistake (re-introducing Lal's change).
3473    
3474    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
3475       within the compiler where applicable.  There is some code simplification
3476       because of that.
3477    
3478    3. The "split" phase (in FLINT) is now part of the default list of phases.
3479       Compiler.Control.LambdaSplitting.* can be used to globally control the
3480       lambda-splitting (cross-module-inlining) engine.  In addition to that,
3481       it can now also be controlled on a per-source basis: CM has been taught
3482       a new tool parameter applicable to ML source files.
3483    
3484       - To turn lambda-splitting off completely:
3485            local open Compiler.Control.LambdaSplitting in
3486                val _ = set Off
3487            end
3488       - To make "no lambda-splitting" the global default (but allow per-source
3489         overriding); this is the initial setting:
3490            local open Compiler.Control.LambdaSplitting in
3491                val _ = set (Default NONE)
3492            end
3493       - To make "lambda-splitting with aggressiveness a" the global default
3494         (and allow per-source overriding):
3495            local open Compiler.Control.LambdaSplitting in
3496                val _ = set (Default (SOME a))
3497            end
3498    
3499       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
3500         write (in the respective .cm-file):
3501            a.sml (lambdasplitting:off)
3502       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
3503            a.sml (lambdasplitting:on)
3504       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
3505         <a> is a decimal non-negative integer):
3506            a.sml (lambdasplitting:<a>)
3507       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
3508            a.sml (lambdasplitting:infinity)
3509       - To use the global default for a.sml:
3510            a.sml (lambdasplitting:default)
3511         or simply
3512            a.sml
3513    
3514    ----------------------------------------------------------------------
3515    Name: Allen Leung
3516    Date: 2001/05/04 01:57:00 EDT
3517    Tag: leunga-20010504-sync
3518    Description:
3519    
3520      MLRISC features.
3521    
3522      1. Fix to CMPXCHG instructions.
3523      2. Changed RA interface to allow annotations in callbacks.
3524      3. Added a new method to the stream interface to allow annotations updates.
3525    
3526    ----------------------------------------------------------------------
3527    Name: Matthias Blume
3528    Date: 2001/05/01 11:45:00 EDT
3529    Tag: blume-20010501-pcedittmp
3530    Description:
3531    
3532    Changed install.sh to use the current working directory instead of
3533    /usr/tmp for a temporary file (pcedittmp).  The previous choice
3534    of /usr/tmp caused trouble with MacOS X because of file premission
3535    problems.
3536    
3537    ----------------------------------------------------------------------
3538    Name: Matthias Blume
3539    Date: 2001/04/20 11:10:00 EDT
3540    Tag: blume-20010420-inMLflag
3541    Description:
3542    
3543     - added vp_limitPtrMask to vproc-state.h
3544       (for use by the raw-C-calls mechanism to implement proper interrupt
3545        handling)
3546     - made the ML compiler aware of various data-structure offsets so it
3547       can generate code for accessing the vp_inML flag and vp_limitPtrMask
3548     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
3549       raw C-calls
3550    
3551    ----------------------------------------------------------------------
3552    Name: Lal George
3553    Date: 2001/04/20 09:15:28 EDT
3554    Tag: george-20010420-macosX
3555    Description:
3556    
3557     - Changes to port to Mac OS X; Darwin.
3558    
3559     - In the process I found that sqrt was broken on the PPC, because the
3560       fsqrt instruction is not implemented.
3561    
3562    ----------------------------------------------------------------------
3563    Name: Matthias Blume
3564    Date: 2001/04/18 12:45:00 EDT
3565    Tag: blume-20010418-ccalls
3566    Description:
3567    
3568     - fixed two off-by-4 errors in the x86-specific c-calls implementation
3569       (this bug prevented structure arguments containing pointers from being
3570        passed correctly)
3571     - changed the raw-C-call code in mlriscGen.sml in such a way that
3572       structure arguments are represented as a pointer to the beginning
3573       of the structure (instead of having a series of synthesized arguments,
3574       one for each structure member)
3575    
3576     - made makeml script's verbosity level configurable via environment
3577       variable (MAKEML_VERBOSITY)
3578    
3579     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
3580       in rawmem-x86.sml; we are now using the real thing
3581    
3582    ----------------------------------------------------------------------
3583    Name: Matthias Blume
3584    Date: 2001/03/22 16:25:00 EST
3585    Tag: blume-20010322-bootfiles
3586    Description:
3587    
3588    Created a new set of bootfiles (for your automatic installation convenience).
3589    
3590    ----------------------------------------------------------------------
3591    Name: Matthias Blume
3592    Date: 2001/03/22 15:10:00 EST
3593    Tag: blume-20010322-rawmem-parcm
3594    Description:
3595    
3596    1. All "raw memory access" primitives for the new FFI are implemented now
3597       (at least on the x86).
3598    2. Some further cleanup of CM's parallel make mechanism.
3599    
3600    ----------------------------------------------------------------------
3601    Name: Matthias Blume
3602    Date: 2001/03/19 17:53:00 EST
3603    Tag: blume-20010319-parallel
3604    Description:
3605    
3606    Parallel make (using compile servers) now works again.
3607    
3608    To this end, CM.stabilize and CMB.make have been modified to work in
3609    two passes when compile servers are attached:
3610       1. Compile everything, do not perform stabilization; this pass
3611          uses compile servers
3612       2. Stabilize everything; this pass does not use compile servers
3613    If there are no compile servers, the two passes are combined into one
3614    (as before).  Splitting the passes increases the inherent parallelism
3615    in the dependency graph because the entire graph including all
3616    libraries is available at the same time.  This, in turn, improves
3617    server utilization.  The downside is that the master process will
3618    have to do some extra work after compilation is done (because for
3619    technical reasons it must re-read all the binfiles during stabilization).
3620    
3621    ----------------------------------------------------------------------
3622    Name: Matthias Blume
3623    Date: 2001/03/16 12:22:00 EST
3624    Tag: blume-20010316-bootfiles
3625    Description:
3626    
3627    Created a new set of bootfiles (for your automatic installation convenience).
3628    
3629    ----------------------------------------------------------------------
3630    Name: Matthias Blume
3631    Date: 2001/03/16 11:00:00 EST
3632    Tag: blume-20010316-MLTREE-fixup
3633    Description:
3634    
3635    This is a minor fixup for an (untagged) earlier commit by Allen.
3636    (A file was missing).
3637    
3638    ----------------------------------------------------------------------
3639    Name: Allen Leung
3640    Date: Mon Mar  5 18:54:57 EST 2001
3641    Tag: leunga-20010305-cut-support
3642    
3643    1. New support for alternative control-flow in MLTREE.
3644       Currently we support
3645    
3646          FLOW_TO(CALL ...., [k1,...,kn])
3647    
3648       This is needed for 'cuts to' in C-- and try/handle-like constructs
3649       in Moby
3650    
3651       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
3652    
3653    2. Register Allocator
3654    
3655       Changes in interface [from Fermin, John]
3656    
3657    3. Alpha 8-bit SLL support [Fermin]
3658    
3659    4. All architectures
3660    
3661       A new module (ClusterExpandCopies) for expanding parallel copies.
3662    
3663    ----------------------------------------------------------------------
3664    Name: Allen Leung
3665    Date: 2001/02/27 23:07:00 EST
3666    Tag: leunga-20010227-minor-stuff
3667    
3668    1. Alpha bug fix for CMOVNE
3669    2. Handle mltree COND(..,FCMP ...,...)
3670    3. Bug fix in simplifier
3671    
3672    ----------------------------------------------------------------------
3673    Name: Matthias Blume
3674    Date: 2001/01/30 17:50:00 EST
3675    Tag: blume-20010130-sync
3676    Description:
3677    
3678    This is just a minor update to sync my devel branch with the main brach.
3679    The only visible change is the addition of some README files.
3680    
3681    ----------------------------------------------------------------------
3682    Name: Matthias Blume
3683    Date: 2001/01/12 23:30:00 JST
3684    Tag: blume-20010112-bootfiles
3685    Description:
3686    
3687    Made a new set of bootfiles that goes with the current state of the
3688    repository.
3689    
3690    ----------------------------------------------------------------------
3691    Name: Matthias Blume
3692    Date: 2001/01/12 21:20:00 JST
3693    Tag: blume-20010112-sync
3694    Description:
3695    
3696    I am just flushing out some minor changes that had accumulated in
3697    my private branch in order to sync with the main tree.  (This is
3698    mainly because I had CVS trouble when trying to merge _into_ my
3699    private branch.)
3700    
3701    Most people should be completely unaffected by this.
3702    
3703    ----------------------------------------------------------------------
3704    Name: Allen Leung
3705    Date: Thu Jan 11 21:03:00 EST 2001
3706    Tag: leunga-20010111-labexp=mltree
3707    Description:
3708    
3709    1.  Removed the type LabelExp and replace it by MLTree.
3710    2.  Rewritten mltree-simplify with the pattern matcher tool.
3711    3.  There were some bugs in alpha code generator which would break
3712        64-bit code generation.
3713    4.  Redo the tools to generate code with the
3714    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
3715        are now generated by perl scripts.
3716    
3717    ----------------------------------------------------------------------
3718    Name: Matthias Blume
3719    Date: 2001/01/10 21:55:00 JST
3720    Tag: blume-20010110-rcc
3721    Description:
3722    
3723    The RCC stuff now seems to work (but only on the x86).
3724    This required hacking of the c-calls interface (and -implementation) in
3725    MLRISC.
3726    
3727    Normal compiler users should be unaffected.
3728    
3729    ----------------------------------------------------------------------
3730    Name: Matthias Blume
3731    Date: 2001/01/09 01:20:00 JST
3732    Tag: blume-20010109-rcc
3733    Description:
3734    
3735    This is a fairly big patch, flushing out a large number of pending
3736    changes that I made to my development copy over the last couple of days.
3737    
3738    Of practical relevance at this moment is a workaround for a pickling
3739    bug that Allen ran into the other day.  The cause of the bug itself is
3740    still unknown and it might be hard to fix it properly, but the
3741    workaround has some merits of its own (namely somewhat reducing pickling
3742    overhead for certain libraries).  Therefore, I think this solution should
3743    be satisfactory at this time.
3744    
3745    The rest of the changes (i.e., the vast majority) has to do with my
3746    ongoing efforts of providing direct support for C function calls from
3747    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
3748    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
3749    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
3750    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
3751    handle RCC.
3752    
3753    The last part (the changes to mlriscGen.sml) are still known to be
3754    wrong on the x86 and not implemented on all other architectures.  But
3755    the infrastructure is in place. I had to change a few functor
3756    signatures in the backend to be able to route the CCalls interface
3757    from MLRISC there, and I had to specialize the mltree type (on the
3758    x86) to include the necessary extensions. (The extensions themselves
3759    were already there and redy to go in MLRISC/x86).
3760    
3761    Everything should be very happy as soon as someone helps me with
3762    mlriscGen.sml...
3763    
3764    In any case, nothing of this should matter to anyone as long as the
3765    new primop is not being used (which is going to be the case unless you
3766    find it where I hid it :). The rest of the compiler is completely
3767    unaffected.
3768    
3769    ----------------------------------------------------------------------
3770    Name: Matthias Blume
3771    Date: 2001/01/05 00:30:00 JST
3772    Tag: blume-20010105-primops
3773    Description:
3774    
3775    Added some experimental support for work that I am doing right now.
3776    These changes mostly concern added primops, but there is also a new
3777    experimental C library in the runtime system (but currently not enabled
3778    anywhere except on Linux/X86).
3779    
3780    In the course of adding primops (and playing with them), I discovered that
3781    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
3782    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
3783    To recover, I made type information in INL_PRIM mandatory and changed
3784    prim.sml as well as built-in.sml accordingly.  The InLine structure now
3785    has complete, correct type information (i.e., no bottom types).
3786    
3787    Since all these changes mean that we need new binfiles, I also bumped the
3788    version number to 110.32.1.
3789    
3790    ----------------------------------------------------------------------
3791    Name: Matthias Blume
3792    Date: 2000/12/30 22:10:00 JST
3793    Tag: blume-20001230-various
3794    Description:
3795    
3796    Added proxy libraries for MLRISC and let MLRISC libraries refer
3797    to each other using path anchors.  (See CM manual for explanation.)
3798    
3799    Updated CM documentation.
3800    
3801    Fixed some bugs in CM.
3802    
3803    Implemented "proxy" libraries (= syntactic sugar for CM).
3804    
3805    Added "-quiet" option to makeml and changed runtime system accordingly.
3806    
3807    Added cleanup handler for exportML to reset timers and compiler stats.
3808    
3809    ----------------------------------------------------------------------
3810    Name: Lal George
3811    Date: 2000/12/22 22:22:58 EST 2000
3812    Tag: Release_110_32
3813    Description:
3814    
3815            Infinite precision used throughout MLRISC.
3816            see MLRISC/mltree/machine-int.sig
3817    
3818    ----------------------------------------------------------------------
3819    Name: Matthias Blume
3820    Date: 2000/12/22 23:16:00 JST
3821    Tag: blume-20001222-warn
3822    Description:
3823    
3824    Corrected wording and formatting of some CM warning message which I
3825    broke in my previous patch.
3826    
3827    ----------------------------------------------------------------------
3828    Name: Matthias Blume
3829    Date: 2000/12/22 21:20:00 JST
3830    Tag: blume-20001222-anchorenv
3831    Description:
3832    
3833    Fixed CM's handling of anchor environments in connection with CMB.make.
3834    
3835    ----------------------------------------------------------------------
3836    Name: Matthias Blume
3837    Date: 2000/12/22 13:15:00 JST
3838    Tag: blume-20001222-cleanup
3839    Description: