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 1062, Wed Feb 13 21:15:14 2002 UTC revision 1390, Wed Sep 24 20:34:06 2003 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2003/09/24 15:35:00 CDT
18    Tag: blume-20030924-synsock
19    Description:
20    
21    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
22    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
23    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
24    implementations of non-blocking operations from CML's Socket
25    structure.
26    
27    ----------------------------------------------------------------------
28    Name: Matthias Blume (blume (at) tti - c (dot) org)
29    Date: 2003/09/24 15:10:05 CDT
30    Tag: blume-20030924-sockets
31    Description:
32    
33    1. Fixed SOCKET API and implementation to match Basis spec.
34       This required changing the internal representation of sockets to one
35       that remembers (for each socket file descriptor) whether it is currently
36       blocking or non-blocking.  This state is maintained lazily (i.e., a system
37       call is made only if the state actually needs to change).
38    
39    2. OS-specific details of sockets were moved into separate files, thus
40       making it possible to unify the bulk of the socket implementations
41       between Unix and Win32.
42    
43    3. CML's socket API changed accordingly.
44       (Note that we need to remove non-blocking functions from this API
45       since they are redundant in the case of CML!)
46    
47    4. CML's socket implementation now makes use of non-blocking functions
48       provided by Basis, thus removing all OS-dependent code from this part
49       of CML.
50    
51    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
52    
53    ----------------------------------------------------------------------
54    Name: Matthias Blume (blume (at) tti - c (dot) org)
55    Date: 2003/09/22 12:10:00 CDT
56    Tag: blume-20030922-110_43_2
57    Description:
58    
59    Made a new interim version and bootfiles for developer's bootstrapping
60    convenience.
61    
62       110.43.2 -- NEW BOOTFILES
63    
64    ----------------------------------------------------------------------
65    Name: Matthias Blume (blume (at) tti - c (dot) org)
66    Date: 2003/09/19 15:55:00 CDT
67    Tag: blume-20030919-cmdir
68    Description:
69    
70    1. new-install.sh -> install.sh
71    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
72    3. tweaked installer so that another name instead of .cm can be chosen
73       at install time (by setting the CM_DIR_ARC environment variable
74       during installation); once installation is complete, the name is
75       fixed
76    
77    ----------------------------------------------------------------------
78    Name: Matthias Blume (blume (at) tti - c (dot) org)
79    Date: 2003/09/18 16:00:00 CDT
80    Tag: blume-20030918-110_43_1
81    Description:
82    
83    Made a new interim version and bootfiles for developer's bootstrapping
84    convenience.
85    
86       110.43.1 -- NEW BOOTFILES
87    
88    ----------------------------------------------------------------------
89    Name: Matthias Blume (blume (at) tti - c (dot) org)
90    Date: 2003/09/18 15:20:00 CDT
91    Tag: blume-20030918-misc
92    Description:
93    
94    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
95       this seems to be controversial at the moment) and used those in
96       Posix.ProcEnv.times.
97    
98    2. Added Time.{from,to}Nanoseconds to Time.
99    
100    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
101       For example, fromLargeInt never needs to look at more than 3 "big
102       digits" to get its 53 bits of precision.
103    
104    ----------------------------------------------------------------------
105    Name: Matthias Blume (blume (at) tti - c (dot) org)
106    Date: 2003/09/17 16:30:00 CDT
107    Tag: blume-20030917-real32-slices
108    Description:
109    
110    Added an entry to the primitive environment
111    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
112    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
113    
114    Removed some of the "magic" constants in real64.sml and replaced them
115    with code that generates these values from their corresponding
116    integer counterparts.
117    
118    Made all(?) the slice-related changes to the Basis and made everything
119    compile again...
120    
121    ----------------------------------------------------------------------
122    Name: Matthias Blume (blume (at) tti - c (dot) org)
123    Date: 2003/09/15 17:45:00 CDT
124    Tag: blume-20030915-rbase
125    Description:
126    
127    Fixed bug in Real.fromLargeInt.
128    
129    ----------------------------------------------------------------------
130    Name: Matthias Blume (blume (at) tti - c (dot) org)
131    Date: 2003/09/13 18:11:00 CDT
132    Tag: blume-20030913-libinstall
133    Description:
134    
135    Minor bugfix in config/libinstall (set anchor with path to
136    standalone tool after installing it, otherwise libraries that
137    need ml-lex or ml-yacc won't compile the first time the installer
138    runs).
139    
140    ----------------------------------------------------------------------
141    Name: Matthias Blume (blume (at) tti - c (dot) org)
142    Date: 2003/09/12 11:45:00 CDT
143    Tag: blume-20030912-various
144    Description:
145    
146     - fixed bug in Real.toLargeInt
147     - fixed bug in Posix.ProcEnv.times
148     - changed inputLine functions to return an option
149     - minor installer improvements / bugfixes
150     - changed default @SMLalloc parameter for x86/celeron to 64k
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume (blume (at) tti - c (dot) org)
154    Date: 2003/09/09 22:00:00 CDT
155    Tag: Release_110_43
156    Description:
157    
158        New working release 110.43.  New bootfiles.
159    
160    ----------------------------------------------------------------------
161    Name: Matthias Blume (blume (at) tti - c (dot) org)
162    Date: 2003/09/09 19:20:00 CDT
163    Tag: blume-20030909-installer
164    Description:
165    
166    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
167    Modified config/install.bat to take advantage of it.  Also modified
168    config/install.sh (and called it config/new-install.sh) to take advantage
169    of it on Unix systems.  (The SML code is (supposed to be) platform-
170    independent.)
171    
172    The installer can now install everything under Win32
173    as well as under *nix as long as it compiles.
174    
175    Other changes:
176    
177      - made CML compile again under Win32
178      - made eXene compile under Win32 (by providing a fake structure UnixSock
179        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
180      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
181        same as type int;  under Win32 it isn't
182      - fixed some slice-related problems in the win32-specific parts of CML
183      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
184        (under Win32, the volume name is case-insensitive, and the
185        OS.Path code compares volume names for equality)
186    
187    ----------------------------------------------------------------------
188    Name: Matthias Blume (blume (at) tti - c (dot) org)
189    Date: 2003/09/08 11:55:00 CDT
190    Tag: blume-20030908-fullpath
191    Description:
192    
193    Made Win32 version of OS.FileSys.fullPath return current directory
194    when given an empty string.  This is what the spec says, and incidentally,
195    CM depends on it.  (CM otherwise goes into an infinite loop in certain
196    cases when presented with the name of a non-existing .cm file.)
197    
198    ----------------------------------------------------------------------
199    Name: Matthias Blume (blume (at) tti - c (dot) org)
200    Date: 2003/09/04 16:30:00 CDT
201    Tag: blume-20030905-slices-etc
202    Description:
203    
204      1. Changed interface to vectors and arrays in Basis to match
205         (draft) Basis spec.
206      2. Added signatures and implementations of slices according to
207         Basis spec.
208      3. Edited source code throughout the system to make it compile again
209         under 1. and 2.  (In some cases code had to be added to have it
210         match the new signatures.)
211      4. MLRISC should be backward-compatible: the copies of the originals
212         of files that needed to change under 3. were retained, the .cm files
213         check the compiler version number and use old versions when
214         appropriate.
215      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
216         dirstream -> string option (in accordance with Basis spec).
217      6. When generating code that counts lines, ml-lex used function
218         CharVector.foldli, taking advantage of its old interface.
219         This has been replaced with the corresponding code from
220         CharVectorSlice.  (html-lex must be re-lexed!)
221      7. BitArray in smlnj-lib/Util has been extended/modified to match the
222         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
223      8. Removed temporary additions (fromInternal, toInternal) from the
224         (now obsolete) IntInf in smlnj-lib/Util.
225      9. Cleaned up structure Byte.
226     10. Added localOffset, scan, and fromString to Date (according to spec).
227         Cleaned/corrected implementation of Date.
228         (Still need to check for correctness; implement better canonicalizeDate.)
229     11. Added "scan" to signature IEEE_REAL.
230     12. Some improvements to IntInf [in particular: efficiency-hack for
231         mod and rem when second operand is 2 (for parity checks).]
232     13. Changed representation of type Time.time, using a single IntInf.int
233         value counting microseconds.  This considerably simplified the
234         implementation of structure Time.  We now support negative time
235         values; scan and fromString handle signs.
236     14. Functor PrimIO now takes two additional arguments (VectorSlice and
237         ArraySlice).
238    
239    ----------------------------------------------------------------------
240    Name: Matthias Blume (blume (at) tti - c (dot) org)
241    Date: 2003/08/28 17:00:00 CDT
242    Tag: blume-20030828-intinf
243    Description:
244    
245    This is a major update which comes with a version number bump
246    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
247    and an implementation of IntInf in the Basis.
248    
249    There are a fairly large number of related changes and updates throughout
250    the system:
251    
252    Basis:
253            - Implemented IntInf.
254            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
255            - Added some missing Real64 operations, most notably Real.toLargeInt.
256            - Added FixedInt as a synonym for Int32.
257    
258    compiler:
259            * Added support for a built-in intinf type.
260              - literals
261              - pattern matching
262              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
263              - overloading on literals and operations
264    
265            This required adding a primitive type intinf, some additional
266            primops, and implementations for several non-trivial intinf
267            operations in Core.  (The intinf type is completely abstract
268            to the compiler; all operations get delegated back to the Core.)
269    
270            * Intinf equality is handled by polyequal.  However, the compiler
271            does not print its usual warning in this case (since polyequal
272            is the right thing to do there).
273    
274            * Improved the organization of structure InlineT.
275    
276            * A word about conversion primops:
277            If conversions involving intinf do not cancel out during
278            CPS contract, then the compiler must insert calls to Core functions.
279            Since all core access must be resolved already during the FLINT
280            translate phase, it would be too late a the time of CPS contract
281            to add new Core calls.  For this reason, conversion primops
282            for intinf carry two arguments:  1. the numeric argument that
283            they are supposed to convert, and 2. the Core function that
284            can help with this conversion if necessary.  If CPS contract
285            eliminates a primop, then the associated Core function becomes
286            dead and goes away.  Intinf conversion primops that do not get
287            eliminated by CPS contract get rewritten into calls of their
288            core functions by a separate, new phase.
289    
290    interactive system:
291            - Control.Print.intinfDepth controls max length of intinf constants
292            being printed. (Analogous to Control.Print.stringDepth.)
293            - Cleanup in printutil and pputil: got rid of unused stuff and
294            duplicates; replaced some of the code with code that makes better
295            use of library functionality.
296    
297    CM:
298            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
299            are no longer silent.
300    
301    CKIT:
302            Fixed mismatched uses of Int32 and LargeInt.  I always decided
303            in favor of LargeInt -- which is now the same as IntInf.
304            CKIT-knowledgable people should check whether this is what's
305            intended and otherwise change things back to using Int32 or
306            FixedInt.
307    
308    Throughout the code:
309            Started using IntInf.int literals and built-in operations
310            (e.g., comparison with 0) where this seems appropriate.
311    
312    
313    ----------------------------------------------------------------------
314    Name: Dave MacQueen (dbm@cs.uchicago.edu)
315    Date: 2003/08/13 11:36:00 CDT
316    Tag: dbm-20030813-mcz-merge1
317    Description:
318    
319    Merging changes from the mcz-branch development branch into trunk.
320    These changes involve replacement of the emulated old prettyprinter
321    interface with direct use of the SML/NJ Lib PP library, and fixing
322    of a couple of bugs (895, 1186) relating to error messages.  A new
323    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
324    has been added.
325    
326    ----------------------------------------------------------------------
327    Name: Matthias Blume (blume (at) tti - c (dot) org)
328    Date: 2003/08/11 15:45:00 CDT
329    Tag: blume-20030811-windows
330    Description:
331    
332    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
333    
334        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
335    
336    This patch restores SML/NJ's ability to run under win32.  There are a
337    number of changes, including fixes for several bugs that had gone
338    unnoticed until now:
339    
340      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
341        trying to run the win32 version from within cygwin.)
342    
343      - There are a number of simple .bat scripts that substitute for their
344        corresponding Unix shell-scripts.  (See below.)
345    
346      - The internals of ml-build have been modified slightly.  The main
347        difference is that instead of calling ".link-sml" (or link-sml.bat)
348        using OS.Process.system, the ML process delegates this task back
349        to the script.  Otherwise problems arise in mixed environments such
350        as Cygwin where scripts look and work like Unix scripts, but
351        where OS.Process.system cannot run them.
352    
353      - In CM, the srcpath pickler used native pathname syntax -- which
354        is incorrect in the case of cross-compilation.  The new pickle format
355        is independent of platform-specific naming conventions.
356    
357      - Path configuration files (such as lib/pathconfig) can now choose
358        between native and standard syntax.  Placing a line of the form
359    
360            standard!
361    
362        into the file causes all subsequent paths to be interpreted using
363        CM standard pathname syntax (= Unix conventions); a line
364    
365            native!
366    
367        switches back to native style.  This was needed so that
368        path config files can be written portably, see src/system/pathconfig.
369    
370      - Runtime system:
371    
372          - win32-filesys.c:  get_file_time and set_file_time now
373            access modification time, not creation time.
374    
375          - I/O code made aware of new array representation.
376    
377          - Bug fixes in X86.prim.masm.
378    
379       - src/system/makeml made aware of win32. (For use under cygwin
380            and other Unix-environments for windows.)
381    
382       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
383         which caused BinIO.inputAll to fail consistently.
384    
385    .bat scripts:
386    
387       Windows .bat scripts assume that SMLNJ_HOME is defined.
388    
389       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
390         applications (sml, ml-yacc, ml-lex).
391       - ml-build.bat: analogous to ml-build.
392       - config\install.bat:  Analogous to config/install.sh.  This requires
393         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
394         (nmake etc. must be on the path, and vcvars32 must have been run.)
395         Moreover, sources for ml-lex and ml-yacc need to exist under src,
396         and the bootfile hierarchy must have been unpacked under
397         sml.boot.x86-win32.
398         The script is very primitive and does a poor job at error checking.
399         It only installs the base system, ml-lex, and ml-yacc.  No other
400         libraries are being installed (i.e., you get only those that
401         are part of the compiler.)
402       - link-sml.bat:  analogous to .link-sml, but not currently used
403    
404    Unrelated bug fixes:
405    
406      - ml-nlffigen now exports structures ST_* corresponding to incomplete
407        types.
408      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
409        otherwise.)
410    
411    ----------------------------------------------------------------------
412    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
413    Date: 2003/06/17
414    Tag: macqueen-20030617-bug895
415    Description:
416    
417    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
418    Tag will be used for new development branch (mcz-branch) for
419    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
420    
421    ----------------------------------------------------------------------
422    Name: Matthias Blume (blume (at) tti - c (dot) org)
423    Date: 2003/05/27 16:55:00 CDT
424    Tag: blume-20030527-polyeq
425    Description:
426    
427    Tried to eliminated most cases of polymorphic equality.
428    
429    ----------------------------------------------------------------------
430    Name: Matthias Blume (blume (at) tti - c (dot) org)
431    Date: 2003/05/21 17:45:00 CDT
432    Tag: blume-20030517-complete
433    Description:
434    
435    Two changes:
436    
437      1. Added a flag for controlling whether non-exhaustive bindings will
438         be treated as errors (default is false).
439      2. Cleaned up the *entire* source tree so that CMB.make goes through
440         without a single non-exhaustive match- or bind warning.
441    
442    ----------------------------------------------------------------------
443    Name: Matthias Blume (blume (at) tti - c (dot) org)
444    Date: 2003/05/17 10:20:00 CDT
445    Tag: blume-20030517-absyn
446    Description:
447    
448    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
449    
450       This mainly affects the quality of error messages.  However, some
451       of the code is now more straightforward than before. (Treatment of
452       the above four constructs in translate.sml is much simpler than
453       the "macro-expansion" that was going on before.  Plus, the mach-
454       compiler no longer gets invoked just to be able to compile an
455       if-expression.)
456    
457    2. The ErrorMsg.Error exception is now caught and absorbed by the
458       interactive loop.
459    
460    ----------------------------------------------------------------------
461    Name: Allen Leung
462    Date: 2003/05/16 13:05:00 CDT
463    Tag: leunga-20030516-cygwin-runtime
464    Description:
465    
466        Ported the runtime system to cygwin, which uses the unix
467    x86-unix bin files.  Missing/buggy features:
468    
469       o getnetbyname, getnetbyaddr: these functions seem to be missing in
470         the Cygwin library.
471       o Ctrl-C handling may be flaky.
472       o Windows system calls and Windows I/O are not supported.
473    
474       A new set of binfiles is located at:
475    
476         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
477    
478       This is only needed for bootstrapping the cygwin version of smlnj.
479       Other x86 versions can use the existing binfiles.
480    
481    ----------------------------------------------------------------------
482    Name: Matthias Blume
483    Date: 2003/04/08 15:42:00 CDT
484    Tag: blume-20030408-listpair
485    Description:
486    
487    1. Added a target 'mlrisc' to installer.
488    
489    2. Added missing elements to structure ListPair.
490    
491    ----------------------------------------------------------------------
492    Name: Allen Leung
493    Date: 2003/01/07 10:40:00 EST
494    Tag: leunga-20030107-int-rem
495    Description:
496    
497    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
498    The arguments to the SUBL instruction were swapped.
499    
500    ----------------------------------------------------------------------
501    Name: Matthias Blume
502    Date: 2002/12/12 16:25:00 EST
503    Tag: blume-20021212-risc-ra
504    Description:
505    
506    Fixed a serious bug in the rewrite code for FP spilling/reloading that
507    sent the RA into an infinite loop when floating point registers get
508    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
509    110.37 and 110.38.)
510    There was another set of potential problems related to the handling of
511    MLRISC annotations (but those did not yet cause real problems, apparently).
512    
513    ----------------------------------------------------------------------
514    Name: Matthias Blume
515    Date: 2002/12/06 22:40:00 EST
516    Tag: blume-20021206-cm-fileid
517    Description:
518    
519    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
520    This fixes the problem of CM getting confused by files that suddenly
521    change their identity (e.g., by getting unlinked and recreated by some
522    text editor such as vi).  There might be a better/cheaper/cleaner way
523    of doing this, but for now this will have to do.
524    
525    ----------------------------------------------------------------------
526    Name: Matthias Blume
527    Date: 2002/10/28 09:50:00 EST
528    Tag: blume-20021028-typecheck
529    Description:
530    
531    Exported structure Typecheck from $smlnj/viscomp/core.cm.
532    
533    ----------------------------------------------------------------------
534    Name: Matthias Blume
535    Date: 2002/10/17 09:10:00 EDT
536    Tag: Release_110_42
537    Description:
538    
539    In good old tradition, there has been a slight hiccup so that we have
540    to patch 110.42 after the fact.   The old release tag has been replaced
541    (see below).
542    
543    The change solves a problem with two competing approaches the
544    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
545    each other's way.
546    
547    This change only affects the runtime system code and the installer script.
548    (No new bootfiles.)
549    
550    ----------------------------------------------------------------------
551    Name: Matthias Blume
552    Date: 2002/10/16 12:00:00 EDT
553    Tag: Release_110_42_removed
554    Description:
555    
556    New working release.  New bootfiles.
557    
558    ----------------------------------------------------------------------
559    Name: Matthias Blume
560    Date: 2002/10/10 13:10:00 EDT
561    Tag: blume-20021010-ppc-divs
562    Description:
563    
564    The mltree operator DIVS must be implemented with an overflow check on
565    the PPC because the hardware indicates divide-by-zero using "overflow" as
566    well.
567    
568    ----------------------------------------------------------------------
569    Name: Matthias Blume
570    Date: 2002/07/23 11:45:00
571    Tag: blume-20020723-smlnj-home
572    Description:
573    
574    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
575    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
576    CM_PATHCONFIG is also set), the path configuration file is assumed
577    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
578    the entire tree to some other place and everything will "just work".
579    
580    (Companion commands such as ml-build and ml-makedepend also sense this
581    variable.)
582    
583    ----------------------------------------------------------------------
584    Name: Matthias Blume
585    Date: 2002/07/12 21:19:00 EDT
586    Tag: blume-20020712-liveness
587    Description:
588    
589    Exported two useful "step" functions from liveness module (MLRISC).
590    
591    ----------------------------------------------------------------------
592    Name: Matthias Blume
593    Date: 2002/07/05 16:00 EDT
594    Tag: Release_110_41
595    Description:
596    
597    New working release.  New bootfiles.
598    
599    ----------------------------------------------------------------------
600    Name: Matthias Blume
601    Date: 2002/07/05 10:25:00 EDT
602    Tag: blume-20020705-btimp
603    Description:
604    
605    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
606    other clients can set up backtracing support.
607    
608    ----------------------------------------------------------------------
609    Name: Matthias Blume
610    Date: 2002/06/25 17:23:00 EDT
611    Tag: blume-20020625-fpmax
612    Description:
613    
614    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
615    case.  (The sense of the isNaN test was reversed -- which made min and
616    max always return their first argument.)
617    
618    ----------------------------------------------------------------------
619    Name: Matthias Blume
620    Date: 2002/06/11
621    Tag: blume-20020611-unixpath
622    Description:
623    
624    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
625    
626    ----------------------------------------------------------------------
627    Name: Matthias Blume
628    Date: 2002/06/10 16:35:00 EDT
629    Tag: blume-20020610-ieeereal
630    Description:
631    
632    I back-ported my implementation of IEEEReal.fromString from the
633    idlbasis-devel branch so that we can test it.
634    
635    Another small change is that ppDec tries to give more information
636    than just "<sig>" in the case of functors.  However, this code is
637    broken in some mysterious way if the functor's body's signature
638    has not been declared by ascription but gets inferred from the
639    implementation.  This needs fixing...
640    
641    ----------------------------------------------------------------------
642    Name: Matthias Blume
643    Date: 2002/05/31
644    Tag: blume-20020531-btrace-mode
645    Description:
646    
647    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
648    the wayside when I switched over to using Controls everywhere.)
649    
650    ----------------------------------------------------------------------
651    Name: Lal George
652    Date: 2002/05/23 12:21:40 EDT
653    Tag: george-20020523-visual-labels
654    Description:
655    
656    Labels are now displayed in the graphical output to make
657    the fall-through and target blocks obvious.
658    
659    ----------------------------------------------------------------------
660    Name: Matthias Blume
661    Date: 2002/05/22 11:03:00 EDT
662    Tag: blume-20020522-shrink
663    Description:
664    
665    John tweaked yesterday's fix for 1131 to handle an out-of-memory
666    situation that comes up when allocating huge arrays.
667    
668    ----------------------------------------------------------------------
669    Name: Matthias Blume
670    Date: 2002/05/21 16:00:00 EDT
671    Tag: Release_110_40
672    Description:
673    
674    New working release (110.40).  New bootfiles.
675    
676    [Also: John Reppy fixed GC bug 1131.]
677    
678    ----------------------------------------------------------------------
679    Name: Matthias Blume
680    Date: 2002/05/21 12:35:00 EDT
681    Tag: blume-20020521-cmdoc
682    Description:
683    
684    CM documentation update.
685    
686    ----------------------------------------------------------------------
687    Name: Matthias Blume
688    Date: 2002/05/21 10:55:00 EDT
689    Tag: blume-20020521-misc
690    Description:
691    
692    - John tweaked runtime to be silent on heap export (except when
693      GC messages are on).
694    - I added a few more things (cross-compiling versions of CMB) to
695      config/preloads (as suggestions).
696    
697    ----------------------------------------------------------------------
698    Name: Matthias Blume
699    Date: 2002/05/20 22:25:00 EDT
700    Tag: blume-20020520-controls
701    Description:
702    
703    - Added ControlUtil structure to control-lib.cm.
704    - Use it throughout.
705    - Used Controls facility to define MLRISC controls (as opposed to
706      registering MLRISC control ref cells with Controls after the fact)
707    - Fixed messed-up controls priorities.
708    
709    * Removed again all the stuff from config/preloads that one wouldn't
710      be able to preload at the time the initial heap image is built.
711      (Many libraries, e.g., CML, do not exist yet at this time.  The
712       only libraries that can be preloaded via config/preloads are those
713       that come bundled with the bootfiles.)
714    
715    ----------------------------------------------------------------------
716    Name: Matthias Blume
717    Date: 2002/05/20 10:59:00 EDT
718    Tag: blume-20020520-preloads
719    Description:
720    
721    Added a lot of commented-out suggestions for things to be included
722    in config/preloads.
723    
724    ----------------------------------------------------------------------
725    Name: Allen Leung
726    Date: 2002/05/18 14:20:00 EDT
727    Tag: leunga-20020518-mdl
728    Description:
729    
730    o Made the mdl tool stuff compile and run again.
731    o I've disabled all the stuff that depends on RTL specifications; they
732      are all badly broken anyway.
733    
734    ----------------------------------------------------------------------
735    Name: Matthias Blume
736    Date: 2002/05/17 16:49:00 EDT
737    Tag: blume-20020517-controls
738    Description:
739    
740    0. John Reppy made several modifications to the SML/NJ library.
741       In particular, there is a shiny new controls-lib.cm.
742    
743    1. Pushed new controls interface through compiler so that everything
744       compiles again.
745    
746    2. Added FormatComb and FORMAT_COMB to the CML version of the
747       SML/NJ library (so that CML compiles again).
748    
749    3. Modified init scripts because XXX_DEFAULT environment variables
750       are no longer with us.  (Boot-time initialization is now done
751       using the same environment variables that are also used for
752       startup-time initialization of controls.)
753    
754    ----------------------------------------------------------------------
755    Name: Lal George
756    Date: 2002/05/15 09:20:10 EDT
757    Tag: george-20020515-pseudo-op-decls
758    Description:
759    
760            All pseudo-ops emitted before the first segment declaration
761    such as TEXT, DATA, and BSS directives are assumed to be global
762    declarations and are emitted first in the assembly file. This is
763    useful in a number of situations where one has pseudo-ops that are not
764    specific to any segment, and also works around the constraint that one
765    cannot have client pseudo-ops in the TEXT segment.
766    
767            Because no segment is associated with these declarations it is
768    an error to allocate any space or objects before the first segment
769    directive and an exception will be raised. However, we cannot make
770    this check for client pseudo-ops.
771    
772            These top level declarations are a field in the CFG graph_info.
773    In theory you can continue to add to this field after the CFG has been
774    built -- provided you know what you are doing;-)
775    
776    ----------------------------------------------------------------------
777    Name: Matthias Blume
778    Date: 2002/05/13 16:40:00 EDT
779    Tag: blume-20020513-pp-etc
780    Description:
781    
782    A few minor bugfixes:
783    
784      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
785        (Bogus printouts for redefined bindings still occur.  Compiler
786        bug should no longer occur now.  We need to redo the prettyprinter
787        from scratch.)
788    
789      - CM pathname printer now also adds escape sequences for ( and )
790    
791      - commend and docu fixes for ml-nlffi
792    
793    ----------------------------------------------------------------------
794    Name: Matthias Blume
795    Date: 2002/05/10 16:40:00 EDT
796    Tag: blume-20020510-erg-textio
797    Description:
798    
799    Applied the following bugfix provided by Emden Gansner:
800    
801        Output is corrupted when outputSubstr is used rather than output.
802    
803        The problem occurs when a substring
804    
805            ss = (s, dataStart, dataLen)
806    
807        where dataStart > 0, fills a stream buffer with avail bytes left.
808        avail bytes of s, starting at index dataStart, are copied into the
809        buffer, the buffer is flushed, and then the remaining dataLen-avail
810        bytes of ss are copied into the beginning of the buffer. Instead of
811        starting this copy at index dataStart+avail in s, the current code
812        starts the copy at index avail.
813    
814        Fix:
815        In text-io-fn.sml, change line 695 from
816             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
817        to
818             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
819    
820    ----------------------------------------------------------------------
821    Name: Matthias Blume
822    Date: 2002/04/12 13:55:00 EDT
823    Tag: blume-20020412-assyntax
824    Description:
825    
826    1. Grabbed newer assyntax.h from the XFree86 project.
827    2. Fiddled with how to compile X86.prim.asm without warnings.
828    3. (Very) Minor cleanup in CM.
829    
830    ----------------------------------------------------------------------
831    Name: Matthias Blume
832    Date: 2002/04/01 (no joke!) 17:07:00 EST
833    Tag: blume-20020401-x86div
834    Description:
835    
836    Added full support for div/mod/rem/quot on the x86, using the machine
837    instruction's two results (without clumsily recomputing the remainder)
838    directly where appropriate.
839    
840    Some more extensive power-of-two support was added to the x86 instruction
841    selector (avoiding expensive divs, mods, and muls where they can be
842    replaced with cheaper shifts and masks).  However, this sort of thing
843    ought to be done earlier, e.g., within the CPS optimizer so that
844    all architectures benefit from it.
845    
846    The compiler compiles to a fixed point, but changes might be somewhat
847    fragile nevertheless.  Please, report any strange things that you might
848    see wrt. div/mod/quot/rem...
849    
850    ----------------------------------------------------------------------
851    Name: Matthias Blume
852    Date: 2002/03/29 17:22:00
853    Tag: blume-20020329-div
854    Description:
855    
856    Fixed my broken div/mod logic.  Unfortunately, this means that the
857    inline code for div/mod now has one more comparison than before.
858    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
859    The problem was with quotient = 0, because that alone does not tell
860    us which way the rounding went.  One then has to look at whether
861    remainder and divisor have the same sign...  :(
862    
863    Anyway, I replaced the bootfiles with fresh ones...
864    
865    ----------------------------------------------------------------------
866    Name: Matthias Blume
867    Date: 2002/03/29 14:10:00 EST
868    Tag: blume-20020329-inlprims
869    Description:
870    
871    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
872    
873    Primops have changed. This means that the bin/boot-file formats have
874    changed as well.
875    
876    To make sure that there is no confusion, I made a new version.
877    
878    
879    CHANGES:
880    
881    * removed REMT from mltree (remainder should never overflow).
882    
883    * added primops to deal with divisions of all flavors to the frontend
884    
885    * handled these primops all the way through so they map to their respective
886      MLRISC support
887    
888    * used these primops in the implementation of Int, Int32, Word, Word32
889    
890    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
891    
892    * parameterized INLMIN, INLMAX, and INLABS by a numkind
893    
894    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
895      floating point
896    
897    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
898      and Real (but Real.abs maps to a separate floating-point-only primop)
899    
900    
901    TODO items:
902    
903    * Hacked Alpha32 instruction selection, disabling the selection of REMx
904      instructions because the machine instruction encoder cannot handle
905      them.  (Hppa, PPC, and Sparc instruction selection did not handle
906      REM in the first place, and REM is supported by the x86 machine coder.)
907    
908    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
909      selection phase.  (The two can be streamlined because the hardware
910      delivers both quotient and remainder at the same time anyway.)
911    
912    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
913      (Currently the behavior is inconsistent both across architectures and
914      wrt. the draft Basis spec.)
915    
916    * Word8 should eventually be handled natively, too.
917    
918    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
919      as if there currently is no execution path that could trigger it in
920      SML/NJ.  (The assumptions underlying functions arith and promotable do not
921      hold for things like multiplication and division.)
922    
923    ----------------------------------------------------------------------
924    Name: Matthias Blume
925    Date: 2002/03/27 16:27:00 EST
926    Tag: blume-20020327-mlrisc-divisions
927    Description:
928    
929    Added support for all four division operations (ML's div, mod, quot,
930    and rem) to MLRISC.  In the course of doing so, I also rationalized
931    the naming (no more annoying switch-around of DIV and QUOT), by
932    parameterizing the operation by div_rounding_mode (which can be either
933    DIV_TO_ZERO or DIV_TO_NEGINF).
934    
935    The generic MLTreeGen functor takes care of compiling all four
936    operations down to only round-to-zero div.
937    
938    Missing pieces:
939    
940      * Doing something smarter than relying on MLTreeGen on architectures
941        like, e.g., the x86 where hardware division delivers both quotient and
942        remainder at the same time.  With this, the implementation of the
943        round-to-neginf operations could be further streamlined.
944    
945      * Remove inlining support for div/mod/rem from the frontend and replace it
946        with primops that get carried through to the backend.  Do this for all
947        int and word types.
948    
949    ----------------------------------------------------------------------
950    Name: Matthias Blume
951    Date: 2002/03/25 17:25:00 EST
952    Tag: blume-20020325-divmod
953    Description:
954    
955    I improved (hopefully without breaking them) the implementation of Int.div,
956    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
957    advantage of the following observations:
958    
959      Let  q = x quot y      r = x rem y
960           d = x div  y      m = x mod y
961    
962    where "quot" is the round-to-zero version of integer division that
963    hardware usually provides.  Then we have:
964    
965         r = x - q * y        where neither the * nor the - will overflow
966         d = if q >= 0 orelse x = q * y then q else q - 1
967                              where neither the * nor the - will overflow
968         m = if q >= 0 orelse r = 0 then r else r + y
969                              where the + will not overflow
970    
971    This results in substantial simplification of the generated code.
972    The following table shows the number of CFG nodes and edges generated
973    for
974            fun f (x, y) = x OPER y
975            (* with OPER \in div, mod, quot, rem *)
976    
977    
978        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
979        --------------------------------------------------------
980         div |         24 |         39 |         12 |         16
981         mod |         41 |         71 |         12 |         16
982        quot |          8 |         10 |          8 |         10
983         rem |         10 |         14 |          8 |         10
984    
985    
986    ----------------------------------------------------------------------
987    Name: Matthias Blume
988    Date: 2002/03/25 22:06:00 EST
989    Tag: blume-20020325-cprotobug
990    Description:
991    
992    Fixed a bug in cproto (c prototype decoder).
993    
994    ----------------------------------------------------------------------
995    Name: Matthias Blume
996    Date: 2002/03/25 16:00:00 EST
997    Tag: blume-20020325-raw-primops
998    Description:
999    
1000    I did some cleanup to Allen's new primop code and
1001    replaced yesterday's bootfiles with new ones.
1002    (But they are stored in the same place.)
1003    
1004    ----------------------------------------------------------------------
1005    Name: Matthias Blume
1006    Date: 2002/03/24 22:40:00 EST
1007    Tag: blume-20020324-bootfiles
1008    Description:
1009    
1010    Made the bootfiles that Allen asked for.
1011    
1012    ----------------------------------------------------------------------
1013    Name: Allen Leung
1014    Date: 2002/03/23 15:50:00 EST
1015    Tag: leunga-20020323-flint-cps-rcc-primops
1016    Description:
1017    
1018      1. Changes to FLINT primops:
1019    
1020        (* make a call to a C-function;
1021         * The primop carries C function prototype information and specifies
1022         * which of its (ML-) arguments are floating point. C prototype
1023         * information is for use by the backend, ML information is for
1024         * use by the CPS converter. *)
1025      | RAW_CCALL of { c_proto: CTypes.c_proto,
1026                       ml_args: ccall_type list,
1027                       ml_res_opt: ccall_type option,
1028                       reentrant : bool
1029                     } option
1030       (* Allocate uninitialized storage on the heap.
1031        * The record is meant to hold short-lived C objects, i.e., they
1032        * are not ML pointers.  With the tag, the representation is
1033        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1034        *)
1035      | RAW_RECORD of {tag:bool,sz:int}
1036      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1037    
1038      2.  These CPS primops are now overloaded:
1039    
1040           rawload of {kind:numkind}
1041           rawstore of {kind:numkind}
1042    
1043          The one argument form is:
1044    
1045             rawload {kind} address
1046    
1047          The two argument form is:
1048    
1049             rawload {kind} [ml object, byte-offset]
1050    
1051      3. RAW_CCALL/RCC now takes two extra arguments:
1052    
1053         a. The first is whether the C call is reentrant, i.e., whether
1054            ML state should be saved and restored.
1055         b. The second argument is a string argument specifying the name of
1056            library and the C function.
1057    
1058         These things are currently not handled in the code generator, yet.
1059    
1060      4. In CProto,
1061    
1062         An encoding type of "bool" means "ml object" and is mapped into
1063         C prototype of PTR.  Note that "bool" is different than "string",
1064         even though "string" is also mapped into PTR, because "bool"
1065         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
1066    
1067      5. Pickler/unpicker
1068    
1069         Changed to handle RAW_RECORD and newest RAW_CCALL
1070    
1071      6. MLRiscGen,
1072    
1073         1. Changed to handle the new rawload/rawstore/rawrecord operators.
1074         2. Code for handling C Calls has been moved to a new module CPSCCalls,
1075            in the file CodeGen/cpscompile/cps-c-calls.sml
1076    
1077      7. Added the conditional move operator
1078    
1079             condmove of branch
1080    
1081         to cps.  Generation of this is still buggy so it is currently
1082         disabled.
1083    
1084    ----------------------------------------------------------------------
1085    Name: Lal George
1086    Date: 2002/03/22 14:18:25 EST
1087    Tag: george-20020322-cps-branch-prob
1088    Description:
1089    
1090    Implemented the Ball-Larus branch prediction-heuristics, and
1091    incorporated graphical viewers for control flow graphs.
1092    
1093    Ball-Larus Heuristics:
1094    ---------------------
1095    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
1096    
1097    By design it uses the Dempster-Shafer theory for combining
1098    probabilities.  For example, in the function:
1099    
1100        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
1101    
1102    the ball-larus heuristics predicts that the n=0 is unlikely
1103    (OH-heuristic), and the 'then' branch is unlikely because of the
1104    RH-heuristic -- giving the 'then' branch an even lower combined
1105    probability using the Dempster-Shafer theory.
1106    
1107    Finally, John Reppy's loop analysis in MLRISC, further lowers the
1108    probability of the 'then' branch because of the loop in the else
1109    branch.
1110    
1111    
1112    Graphical Viewing:
1113    ------------------
1114    I merely plugged in Allen's graphical viewers into the compiler. The
1115    additional code is not much. At the top level, saying:
1116    
1117            Control.MLRISC.getFlag "cfg-graphical-view" := true;
1118    
1119    will display the graphical view of the control flow graph just before
1120    back-patching.  daVinci must be in your path for this to work. If
1121    daVinci is not available, then the default viewer can be changed
1122    using:
1123    
1124            Control.MLRISC.getString "viewer"
1125    
1126    which can be set to "dot" or "vcg" for the corresponding viewers. Of
1127    course, these viewers must be in your path.
1128    
1129    The above will display the compilation unit at the level of clusters,
1130    many of which are small, boring, and un-interesting. Also setting:
1131    
1132            Control.MLRISC.getInt "cfg-graphical-view_size"
1133    
1134    will display clusters that are larger than the value set by the above.
1135    
1136    
1137    ----------------------------------------------------------------------
1138    Name: Matthias Blume
1139    Date: 2002/03/21 22:20:00 EST
1140    Tag: blume-20020321-kmp-bugfix
1141    Description:
1142    
1143    Changed the interface to the KMP routine in PreString and fixed
1144    a minor bug in one place where it was used.
1145    
1146    ----------------------------------------------------------------------
1147    Name: Allen Leung
1148    Date: 2002/03/21 20:30:00 EST
1149    Tag: leunga-20020321-cfg
1150    Description:
1151    
1152      Fixed a potential problem in cfg edge splitting.
1153    
1154    ----------------------------------------------------------------------
1155    Name: Allen Leung
1156    Date: 2002/03/21 17:15:00 EST
1157    Tag: leunga-20020321-x86-fp-cfg
1158    Description:
1159    
1160      1. Recoded the buggy parts of x86-fp.
1161    
1162         a. All the block reordering code has been removed.
1163            We now depend on the block placement phases to do this work.
1164    
1165         b. Critical edge splitting code has been simplified and moved into the
1166            CFG modules, as where they belong.
1167    
1168         Both of these were quite buggy and complex.  The code is now much, much
1169         simpler.
1170    
1171      2. X86 backend.
1172    
1173         a. Added instructions for 64-bit support.  Instruction selection for
1174            64-bit has not been committed, however, since that
1175            requires changes to MLTREE which haven't been approved by
1176            Lal and John.
1177    
1178         b. Added support for FUCOMI and FUCOMIP when generating code for
1179            PentiumPro and above.  We only generate these instructions in
1180            the fast-fp mode.
1181    
1182         c. Added cases for JP and JNP in X86FreqProps.
1183    
1184      3. CFG
1185    
1186         CFG now has a bunch of methods for edge splitting and merging.
1187    
1188      4. Machine description.
1189    
1190         John's simplification of MLTREE_BASIS.fcond broke a few machine
1191         description things:
1192    
1193         rtl-build.{sig,sml} and hppa.mdl fixed.
1194    
1195         NOTE: the machine description stuff in the repository is still broken.
1196               Again, I can't put my fixes in because that involves
1197               changes to MLTREE.
1198    
1199    ----------------------------------------------------------------------
1200    Name: Matthias Blume
1201    Date: 2002/03/20 15:55:00 EST
1202    Tag: blume-20020320-kmp
1203    Description:
1204    
1205    Implemented Knuth-Morris-Pratt string matching in PreString and used
1206    it for String.isSubstring, Substring.isSubstring, and
1207    Substring.position.
1208    
1209    (Might need some stress-testing.  Simple examples worked fine.)
1210    
1211    ----------------------------------------------------------------------
1212    Name: Matthias Blume
1213    Date: 2002/03/19 16:37:00 EST
1214    Tag: blume-20020319-witnesses
1215    Description:
1216    
1217    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
1218    
1219    This implements a generic mechanism for changing constness qualifiers
1220    anywhere within big C types without resorting to outright "casts".
1221    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
1222    modify the constness at the outermost level.)
1223    The implementation of "convert" is based on the idea of "witness"
1224    values -- values that are not used by the operation but whose types
1225    "testify" to their applicability.  On the implementation side, "convert"
1226    is simply a projection (returning its second curried argument).  With
1227    cross-module inlining, it should not result in any machine code being
1228    generated.
1229    
1230    ----------------------------------------------------------------------
1231    Name: Matthias Blume
1232    Date: 2002/03/15 16:40:00 EST
1233    Tag: blume-20020315-basis
1234    Description:
1235    
1236    Provided (preliminary?) implementations for
1237    
1238      {String,Substring}.{concatWith,isSuffix,isSubstring}
1239    
1240    and
1241    
1242      Substring.full
1243    
1244    Those are in the Basis spec but they were missing in SML/NJ.
1245    
1246    ----------------------------------------------------------------------
1247    Name: Matthias Blume
1248    Date: 2002/03/14 21:30:00 EST
1249    Tag: blume-20020314-controls
1250    Description:
1251    
1252    Controls:
1253    ---------
1254    
1255    1. Factored out the recently-added Controls : CONTROLS stuff and put
1256       it into its own library $/controls-lib.cm.  The source tree for
1257       this is under src/smlnj-lib/Controls.
1258    
1259    2. Changed the names of types and functions in this interface, so they
1260       make a bit more "sense":
1261    
1262          module -> registry
1263          'a registry -> 'a group
1264    
1265    3. The interface now deals in ref cells only.  The getter/setter interface
1266       is (mostly) gone.
1267    
1268    4. Added a function that lets one register an already-existing ref cell.
1269    
1270    5. Made the corresponding modifications to the rest of the code so that
1271       everything compiles again.
1272    
1273    6. Changed the implementation of Controls.MLRISC back to something closer
1274       to the original.  In particular, this module (and therefore MLRISC)
1275       does not depend on Controls.  There now is some link-time code in
1276       int-sys.sml that registers the MLRISC controls with the Controls
1277       module.
1278    
1279    CM:
1280    ---
1281    
1282      * One can now specify the lambda-split aggressiveness in init.cmi.
1283    
1284    ----------------------------------------------------------------------
1285    Name: Allen Leung
1286    Date: 2002/03/13 17:30:00 EST
1287    Tag: leunga-20020313-x86-fp-unary
1288    Description:
1289    
1290    Bug fix for:
1291    
1292    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
1293    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
1294    > - fun f(x,(y,z)) = Real.~ y;
1295    > [autoloading]
1296    > [autoloading done]
1297    >       fchsl   (%eax), 184(%esp)
1298    > Error: MLRisc bug: X86MCEmitter.emitInstr
1299    >
1300    > uncaught exception Error
1301    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
1302    
1303    The problem was that the code generator did not generate any fp registers
1304    in this case, and the ra didn't know that it needed to run the X86FP phase to
1305    translate the pseudo fp instruction.   This only happened with unary fp
1306    operators in certain situations.
1307    
1308    ----------------------------------------------------------------------
1309    Name: Matthias Blume
1310    Date: 2002/03/13 14:00:00 EST
1311    Tag: blume-20020313-overload-etc
1312    Description:
1313    
1314    1. Added _overload as a synonym for overload for backward compatibility.
1315       (Control.overloadKW must be true for either version to be accepted.)
1316    
1317    2. Fixed bug in install script that caused more things to be installed
1318       than what was requested in config/targets.
1319    
1320    3. Made CM aware of the (_)overload construct so that autoloading
1321       works.
1322    
1323    ----------------------------------------------------------------------
1324    Name: Matthias Blume
1325    Date: 2002/03/12 22:03:00 EST
1326    Tag: blume-20020312-url
1327    Description:
1328    
1329    Forgot to update BOOT and srcarchiveurl.
1330    
1331    ----------------------------------------------------------------------
1332    Name: Matthias Blume
1333    Date: 2002/03/12 17:30:00 EST
1334    Tag: blume-20020312-version110392
1335    Description:
1336    
1337    Yet another version number bump (because of small changes to the
1338    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
1339    
1340    Changes:
1341    
1342      The new pid generation scheme described a few weeks ago was overly
1343      complicated.  I implemented a new mechanism that is simpler and
1344      provides a bit more "stability":  Once CM has seen a compilation
1345      unit, it keeps its identity constant (as long as you do not delete
1346      those crucial CM/GUID/* files).  This means that when you change
1347      an interface, compile, then go back to the old interface, and
1348      compile again, you arrive at the original pid.
1349    
1350      There now also is a mechanism that instructs CM to use the plain
1351      environment hash as a module's pid (effectively making its GUID
1352      the empty string).  For this, "noguid" must be specified as an
1353      option to the .sml file in question within its .cm file.
1354      This is most useful for code that is being generated by tools such
1355      as ml-nlffigen (because during development programmers tend to
1356      erase the tool's entire output directory tree including CM's cached
1357      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
1358      revert to the old, broken behavior of SML/NJ, but in specific cases
1359      where there is no danger of interface confusion, its use is ok
1360      (I think).
1361    
1362      ml-nlffigen by default generates "noguid" annotations.  They can be
1363      turned off by specifying -guid in its command line.
1364    
1365    ----------------------------------------------------------------------
1366    Name: Lal George
1367    Date: 2002/03/12 12 14:42:36 EST
1368    Tag: george-20020312-frequency-computation
1369    Description:
1370    
1371    Integrated jump chaining and static block frequency into the
1372    compiler. More details and numbers later.
1373    
1374    ----------------------------------------------------------------------
1375    Name: Lal George
1376    Date: 2002/03/11 11 22:38:53 EST
1377    Tag: george-20020311-jump-chain-elim
1378    Description:
1379    
1380    Tested the jump chain elimination on all architectures (except the
1381    hppa).  This is on by default right now and is profitable for the
1382    alpha and x86, however, it may not be profitable for the sparc and ppc
1383    when compiling the compiler.
1384    
1385    The gc test will typically jump to a label at the end of the cluster,
1386    where there is another jump to an external cluster containing the actual
1387    code to invoke gc. This is to allow factoring of common gc invocation
1388    sequences. That is to say, we generate:
1389    
1390            f:
1391               testgc
1392               ja   L1      % jump if above to L1
1393    
1394            L1:
1395               jmp L2
1396    
1397    
1398    After jump chain elimination the 'ja L1' instructions is converted to
1399    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
1400    up being implemented in their long form (if L2 is far away) using:
1401    
1402            jbe     L3      % jump if below or equal to L3
1403            jmp     L2
1404         L3:
1405            ...
1406    
1407    
1408    For large compilation units L2  may be far away.
1409    
1410    
1411    ----------------------------------------------------------------------
1412    Name: Matthias Blume
1413    Date: 2002/03/11 13:30:00 EST
1414    Tag: blume-20020311-mltreeeval
1415    Description:
1416    
1417    A functor parameter was missing.
1418    
1419    ----------------------------------------------------------------------
1420    Name: Allen Leung
1421    Date: 2002/03/11 10:30:00 EST
1422    Tag: leunga-20020311-runtime-string0
1423    Description:
1424    
1425       The representation of the empty string now points to a
1426    legal null terminated C string instead of unit.  It is now possible
1427    to convert an ML string into C string with InlineT.CharVector.getData.
1428    This compiles into one single machine instruction.
1429    
1430    ----------------------------------------------------------------------
1431    Name: Allen Leung
1432    Date: 2002/03/10 23:55:00 EST
1433    Tag: leunga-20020310-x86-call
1434    Description:
1435    
1436       Added machine generation for CALL instruction (relative displacement mode)
1437    
1438    ----------------------------------------------------------------------
1439    Name: Matthias Blume
1440    Date: 2002/03/08 16:05:00
1441    Tag: blume-20020308-entrypoints
1442    Description:
1443    
1444    Version number bumped to 110.39.1.  NEW BOOTFILES!
1445    
1446    Entrypoints: non-zero offset into a code object where execution should begin.
1447    
1448    - Added the notion of an entrypoint to CodeObj.
1449    - Added reading/writing of entrypoint info to Binfile.
1450    - Made runtime system bootloader aware of entrypoints.
1451    - Use the address of the label of the first function given to mlriscGen
1452      as the entrypoint.  This address is currently always 0, but it will
1453      not be 0 once we turn on block placement.
1454    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
1455      with entry points) from mlriscGen.
1456    
1457    ----------------------------------------------------------------------
1458    Name: Allen Leung
1459    Date: 2002/03/07 20:45:00 EST
1460    Tag: leunga-20020307-x86-cmov
1461    Description:
1462    
1463       Bug fixes for CMOVcc on x86.
1464    
1465       1. Added machine code generation for CMOVcc
1466       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
1467       3. CMOVcc cannot have an immediate operand as argument.
1468    
1469    ----------------------------------------------------------------------
1470    Name: Matthias Blume
1471    Date: 2002/03/07 16:15:00 EST
1472    Tag: blume-20020307-controls
1473    Description:
1474    
1475    This is a very large but mostly boring patch which makes (almost)
1476    every tuneable compiler knob (i.e., pretty much everything under
1477    Control.* plus a few other things) configurable via both the command
1478    line and environment variables in the style CM did its configuration
1479    until now.
1480    
1481    Try starting sml with '-h' (or, if you are brave, '-H')
1482    
1483    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
1484    implements the underlying generic mechanism.
1485    
1486    The interface to some of the existing such facilities has changed somewhat.
1487    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
1488    (The getFoo interface is still there for backward-compatibility, but its
1489    use is deprecated.)
1490    
1491    The ml-build script passes -Cxxx=yyy command-line arguments through so
1492    that one can now twiddle the compiler settings when using this "batch"
1493    compiler.
1494    
1495    TODO items:
1496    
1497    We should go through and throw out all controls that are no longer
1498    connected to anything.  Moreover, we should go through and provide
1499    meaningful (and correct!) documentation strings for those controls
1500    that still are connected.
1501    
1502    Currently, multiple calls to Controls.new are accepted (only the first
1503    has any effect).  Eventually we should make sure that every control
1504    is being made (via Controls.new) exactly once.  Future access can then
1505    be done using Controls.acc.
1506    
1507    Finally, it would probably be a good idea to use the getter-setter
1508    interface to controls rather than ref cells.  For the time being, both
1509    styles are provided by the Controls module, but getter-setter pairs are
1510    better if thread-safety is of any concern because they can be wrapped.
1511    
1512    *****************************************
1513    
1514    One bug fix: The function blockPlacement in three of the MLRISC
1515    backpatch files used to be hard-wired to one of two possibilities at
1516    link time (according to the value of the placementFlag).  But (I
1517    think) it should rather sense the flag every time.
1518    
1519    *****************************************
1520    
1521    Other assorted changes (by other people who did not supply a HISTORY entry):
1522    
1523    1. the cross-module inliner now works much better (Monnier)
1524    2. representation of weights, frequencies, and probabilities in MLRISC
1525       changed in preparation of using those for weighted block placement
1526       (Reppy, George)
1527    
1528    ----------------------------------------------------------------------
1529    Name: Lal George
1530    Date: 2002/03/07 14:44:24 EST 2002
1531    Tag: george-20020307-weighted-block-placement
1532    
1533    Tested the weighted block placement optimization on all architectures
1534    (except the hppa) using AMPL to generate the block and edge frequencies.
1535    Changes were required in the machine properties to correctly
1536    categorize trap instructions. There is an MLRISC flag
1537    "weighted-block-placement" that can be used to enable weighted block
1538    placement, but this will be ineffective without block/edge
1539    frequencies (coming soon).
1540    
1541    
1542    ----------------------------------------------------------------------
1543    Name: Lal George
1544    Date: 2002/03/05 17:24:48 EST
1545    Tag: george-20020305-linkage-cluster
1546    
1547    In order to support the block placement optimization, a new cluster
1548    is generated as the very first cluster (called the linkage cluster).
1549    It contains a single jump to the 'real' entry point for the compilation
1550    unit. Block placement has no effect on the linkage cluster itself, but
1551    all the other clusters  have full freedom in the manner in which they
1552    reorder blocks or functions.
1553    
1554    On the x86 the typical linkage code that is generated is:
1555       ----------------------
1556            .align 2
1557       L0:
1558            addl    $L1-L0, 72(%esp)
1559            jmp     L1
1560    
1561    
1562            .align  2
1563       L1:
1564       ----------------------
1565    
1566    72(%esp) is the memory location for the stdlink register. This
1567    must contain the address of the CPS function being called. In the
1568    above example, it contains the address of  L0; before
1569    calling L1 (the real entry point for the compilation unit), it
1570    must contain the address for L1, and hence
1571    
1572            addl $L1-L0, 72(%esp)
1573    
1574    I have tested this on all architectures except the hppa.The increase
1575    in code size is of course negligible
1576    
1577    ----------------------------------------------------------------------
1578    Name: Allen Leung
1579    Date: 2002/03/03 13:20:00 EST
1580    Tag: leunga-20020303-mlrisc-tools
1581    
1582      Added #[ ... ] expressions to mlrisc tools
1583    
1584    ----------------------------------------------------------------------
1585    Name: Matthias Blume
1586    Date: 2002/02/27 12:29:00 EST
1587    Tag: blume-20020227-cdebug
1588    Description:
1589    
1590    - made types in structure C and C_Debug to be equal
1591    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
1592    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
1593    
1594    ----------------------------------------------------------------------
1595    Name: Matthias Blume
1596    Date: 2002/02/26 12:00:00 EST
1597    Tag: blume-20020226-ffi
1598    Description:
1599    
1600    1. Fixed a minor bug in CM's "noweb" tool:
1601       If numbering is turned off, then truly don't number (i.e., do not
1602       supply the -L option to noweb).  The previous behavior was to supply
1603       -L'' -- which caused noweb to use the "default" line numbering scheme.
1604       Thanks to Chris Richards for pointing this out (and supplying the fix).
1605    
1606    2. Once again, I reworked some aspects of the FFI:
1607    
1608       A. The incomplete/complete type business:
1609    
1610       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
1611         gone!
1612       - ML types representing an incomplete type are now *equal* to
1613         ML types representing their corresponding complete types (just like
1614         in C).  This is still safe because ml-nlffigen will not generate
1615         RTTI for incomplete types, nor will it generate functions that
1616         require access to such RTTI.   But when ML code generated from both
1617         incomplete and complete versions of the C type meet, the ML types
1618         are trivially interoperable.
1619    
1620         NOTE:  These changes restore the full generality of the translation
1621         (which was previously lost when I eliminated functorization)!
1622    
1623       B. Enum types:
1624    
1625       - Structure C now has a type constructor "enum" that is similar to
1626         how the "su" constructor works.  However, "enum" is not a phantom
1627         type because each "T enum" has values (and is isomorphic to
1628         MLRep.Signed.int).
1629       - There are generic access operations for enum objects (using
1630         MLRep.Signed.int).
1631       - ml-nlffigen will generate a structure E_foo for each "enum foo".
1632         * The structure contains the definition of type "mlrep" (the ML-side
1633         representation type of the enum).  Normally, mlrep is the same
1634         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
1635         then mlrep will be defined as a datatype -- thus facilitating
1636         pattern matching on mlrep values.
1637         ("-ec" will be suppressed if there are duplicate values in an
1638          enumeration.)
1639         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
1640         will be generated for each C enum constant xxx.
1641         * Conversion functions m2i and i2m convert between mlrep and
1642         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
1643         * Coversion functions c and ml convert between mlrep and "tag enum".
1644         * Access functions (get/set) fetch and store mlrep values.
1645       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
1646         enumerations are merged into one single enumeration represented by
1647         structure E_'.
1648    
1649    ----------------------------------------------------------------------
1650    Name: Allen Leung
1651    Date: 2002/02/25 04:45:00 EST
1652    Tag: leunga-20020225-cps-spill
1653    
1654    This is a new implementation of the CPS spill phase.
1655    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
1656    In case of problems, replace it with the old file spill.sml
1657    
1658    The current compiler runs into some serious performance problems when
1659    constructing a large record.  This can happen when we try to compile a
1660    structure with many items.  Even a very simple structure like the following
1661    makes the compiler slow down.
1662    
1663        structure Foo = struct
1664           val x_1 = 0w1 : Word32.int
1665           val x_2 = 0w2 : Word32.int
1666           val x_3 = 0w3 : Word32.int
1667           ...
1668           val x_N = 0wN : Word32.int
1669        end
1670    
1671    The following table shows the compile time, from N=1000 to N=4000,
1672    with the old compiler:
1673    
1674    N
1675    1000   CPS 100 spill                           0.04u  0.00s  0.00g
1676           MLRISC ra                               0.06u  0.00s  0.05g
1677              (spills = 0 reloads = 0)
1678           TOTAL                                   0.63u  0.07s  0.21g
1679    
1680    1100   CPS 100 spill                           8.25u  0.32s  0.64g
1681           MLRISC ra                               5.68u  0.59s  3.93g
1682              (spills = 0 reloads = 0)
1683           TOTAL                                   14.71u  0.99s  4.81g
1684    
1685    1500   CPS 100 spill                           58.55u  2.34s  1.74g
1686           MLRISC ra                               5.54u  0.65s  3.91g
1687              (spills = 543 reloads = 1082)
1688           TOTAL                                   65.40u  3.13s  6.00g
1689    
1690    2000   CPS 100 spill                           126.69u  4.84s  3.08g
1691           MLRISC ra                               0.80u  0.10s  0.55g
1692              (spills = 42 reloads = 84)
1693           TOTAL                                   129.42u  5.10s  4.13g
1694    
1695    3000   CPS 100 spill                           675.59u  19.03s  11.64g
1696           MLRISC ra                               2.69u  0.27s  1.38g
1697              (spills = 62 reloads = 124)
1698           TOTAL                                   682.48u  19.61s  13.99g
1699    
1700    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
1701           MLRISC ra                               4.96u  0.27s  2.72g
1702              (spills = 85 reloads = 170)
1703           TOTAL                                   2375.26u  57.21s  48.00g
1704    
1705    As you can see the old cps spill module suffers from some serious
1706    performance problem.  But since I cannot decipher the old code fully,
1707    instead of patching the problems up, I'm reimplementing it
1708    with a different algorithm.  The new code is more modular,
1709    smaller when compiled, and substantially faster
1710    (O(n log n) time and O(n) space).  Timing of the new spill module:
1711    
1712    4000  CPS 100 spill                           0.02u  0.00s  0.00g
1713          MLRISC ra                               0.25u  0.02s  0.15g
1714             (spills=1 reloads=3)
1715          TOTAL                                   7.74u  0.34s  1.62g
1716    
1717    Implementation details:
1718    
1719    As far as I can tell, the purpose of the CPS spill module is to make sure the
1720    number of live variables at any program point (the bandwidth)
1721    does not exceed a certain limit, which is determined by the
1722    size of the spill area.
1723    
1724    When the bandwidth is too large, we decrease the register pressure by
1725    packing live variables into spill records.  How we achieve this is
1726    completely different than what we did in the old code.
1727    
1728    First, there is something about the MLRiscGen code generator
1729    that we should be aware of:
1730    
1731    o MLRiscGen performs code motion!
1732    
1733       In particular, it will move floating point computations and
1734       address computations involving only the heap pointer to
1735       their use sites (if there is only a single use).
1736       What this means is that if we have a CPS record construction
1737       statement
1738    
1739           RECORD(k,vl,w,e)
1740    
1741       we should never count the new record address w as live if w
1742       has only one use (which is often the case).
1743    
1744       We should do something similar to floating point, but the transformation
1745       there is much more complex, so I won't deal with that.
1746    
1747    Secondly, there are now two new cps primops at our disposal:
1748    
1749     1. rawrecord of record_kind option
1750        This pure operator allocates some uninitialized storage from the heap.
1751        There are two forms:
1752    
1753         rawrecord NONE [INT n]  allocates a tagless record of length n
1754         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
1755                                     and initializes the tag.
1756    
1757     2. rawupdate of cty
1758          rawupdate cty (v,i,x)
1759          Assigns to x to the ith component of record v.
1760          The storelist is not updated.
1761    
1762    We use these new primops for both spilling and increment record construction.
1763    
1764     1. Spilling.
1765    
1766        This is implemented with a linear scan algorithm (but generalized
1767        to trees).  The algorithm will create a single spill record at the
1768        beginning of the cps function and use rawupdate to spill to it,
1769        and SELECT or SELp to reload from it.  So both spills and reloads
1770        are fine-grain operations.  In contrast, in the old algorithm
1771        "spills" have to be bundled together in records.
1772    
1773        Ideally, we should sink the spill record construction to where
1774        it is needed.  We can even split the spill record into multiple ones
1775        at the places where they are needed.  But CPS is not a good
1776        representation for global code motion, so I'll keep it simple and
1777        am not attempting this.
1778    
1779     2. Incremental record construction (aka record splitting).
1780    
1781        Long records with many component values which are simulatenously live
1782        (recall that single use record addresses are not considered to
1783         be live) are constructed with rawrecord and rawupdate.
1784        We allocate space on the heap with rawrecord first, then gradually
1785        fill it in with rawupdate.  This is the technique suggested to me
1786        by Matthias.
1787    
1788        Some restrictions on when this is applicable:
1789        1. It is not a VECTOR record.  The code generator currently does not handle
1790           this case. VECTOR record uses double indirection like arrays.
1791        2. All the record component values are defined in the same "basic block"
1792           as the record constructor.  This is to prevent speculative
1793           record construction.
1794    
1795    ----------------------------------------------------------------------
1796    Name: Allen Leung
1797    Date: 2002/02/22 01:02:00 EST
1798    Tag: leunga-20020222-mlrisc-tools
1799    
1800    Minor bug fixes in the parser and rewriter
1801    
1802    ----------------------------------------------------------------------
1803    Name: Allen Leung
1804    Date: 2002/02/21 20:20:00 EST
1805    Tag: leunga-20020221-peephole
1806    
1807    Regenerated the peephole files.  Some contained typos in the specification
1808    and some didn't compile because of pretty printing bugs in the old version
1809    of 'nowhere'.
1810    
1811    ----------------------------------------------------------------------
1812    Name: Allen Leung
1813    Date: 2002/02/19 20:20:00 EST
1814    Tag: leunga-20020219-mlrisc-tools
1815    Description:
1816    
1817       Minor bug fixes to the mlrisc-tools library:
1818    
1819       1.  Fixed up parsing colon suffixed keywords
1820       2.  Added the ability to shut the error messages up
1821       3.  Reimplemented the pretty printer and fixed up/improved
1822           the pretty printing of handle and -> types.
1823       4.  Fixed up generation of literal symbols in the nowhere tool.
1824       5.  Added some SML keywords to to sml.sty
1825    
1826    ----------------------------------------------------------------------
1827    Name: Matthias Blume
1828    Date: 2002/02/19 16:20:00 EST
1829    Tag: blume-20020219-cmffi
1830    Description:
1831    
1832    A wild mix of changes, some minor, some major:
1833    
1834    * All C FFI-related libraries are now anchored under $c:
1835        $/c.cm      --> $c/c.cm
1836        $/c-int.cm  --> $c/internals/c-int.cm
1837        $/memory.cm --> $c/memory/memory.cm
1838    
1839    * "make" tool (in CM) now treats its argument pathname slightly
1840      differently:
1841        1. If the native expansion is an absolute name, then before invoking
1842           the "make" command on it, CM will apply OS.Path.mkRelative
1843           (with relativeTo = OS.FileSys.getDir()) to it.
1844        2. The argument will be passed through to subsequent phases of CM
1845           processing without "going native".  In particular, if the argument
1846           was an anchored path, then "make" will not lose track of that anchor.
1847    
1848    * Compiler backends now "know" their respective C calling conventions
1849      instead of having to be told about it by ml-nlffigen.  This relieves
1850      ml-nlffigen from one of its burdens.
1851    
1852    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
1853    
1854    * Export C_DEBUG and C_Debug from $c/c.cm.
1855    
1856    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
1857      subtyping relationship between incomplete pointers and their complete
1858      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
1859      with the convention of instantiating 'o with ('t, 'c) obj whenever
1860      the pointer target type is complete.  In the incomplete case, 'o
1861      will be instantiated with some "'c iobj" -- a type obtained by
1862      using one of the functors PointerToIncompleteType or PointerToCompleteType.
1863    
1864      Operations that work on both incomplete and complete pointer types are
1865      typed as taking an 'o ptr while operations that require the target to
1866      be known are typed as taking some ('t, 'c) obj ptr.
1867    
1868      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
1869      where void is an eqtype without any values.  This makes it possible
1870      to work on voidptr values using functions meant to operate on light
1871      incomplete pointers.
1872    
1873    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
1874      been vastly simplified.
1875    
1876    ----------------------------------------------------------------------
1877    Name: Matthias Blume
1878    Date: 2002/02/19 10:48:00 EST
1879    Tag: blume-20020219-pqfix
1880    Description:
1881    
1882    Applied Chris Okasaki's bug fix for priority queues.
1883    
1884    ----------------------------------------------------------------------
1885    Name: Matthias Blume
1886    Date: 2002/02/15 17:05:00
1887    Tag: Release_110_39
1888    Description:
1889    
1890    Last-minute retagging is becoming a tradition... :-(
1891    
1892    This is the working release 110.39.
1893    
1894    ----------------------------------------------------------------------
1895    Name: Matthias Blume
1896    Date: 2002/02/15 16:00:00 EST
1897    Tag: Release_110_39-orig
1898    Description:
1899    
1900    Working release 110.39.  New bootfiles.
1901    
1902    (Update: There was a small bug in the installer so it wouldn't work
1903    with all shells.  So I retagged. -Matthias)
1904    
1905    ----------------------------------------------------------------------
1906    Name: Matthias Blume
1907    Date: 2002/02/15 14:17:00 EST
1908    Tag: blume-20020215-showbindings
1909    Description:
1910    
1911    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
1912    the latter can be useful for exploring what bindings are available at
1913    the interactive prompt.  (The first function returns only the list
1914    of symbols that are really bound, the second prints those but also the
1915    ones that CM's autoloading mechanism knows about.)
1916    
1917    ----------------------------------------------------------------------
1918    Name: Matthias Blume
1919    Date: 2002/02/15 12:08:00 EST
1920    Tag: blume-20020215-iptrs
1921    Description:
1922    
1923    Two improvements to ml-nlffigen:
1924    
1925      1. Write files only if they do not exist or if their current contents
1926         do not coincide with what's being written.  (That is, avoid messing
1927         with the time stamps unless absolutely necessary.)
1928    
1929      2. Implement a "repository" mechanism for generated files related
1930         to "incomplete pointer types".   See the README file for details.
1931    
1932    ----------------------------------------------------------------------
1933    Name: Matthias Blume
1934    Date: 2002/02/14 11:50:00 EST
1935    Tag: blume-20020214-quote
1936    Description:
1937    
1938    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
1939    because of the new and improved tag generation scheme.  (Thanks to Allen
1940    Leung for pointing it out.)
1941    
1942    ----------------------------------------------------------------------
1943    Name: Lal George
1944    Date: 2002/02/14 09:55:27 EST 2002
1945    Tag: george-20020214-isabelle-bug
1946    Description:
1947    
1948    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
1949    of Isabelle on the x86.
1950    
1951    From Allen:
1952    -----------
1953     I've found the problem:
1954    
1955         in ra-core.sml, I use the counter "blocked" to keep track of the
1956         true number of elements in the freeze queue.  When the counter goes
1957         to zero, I skip examining the queue.  But I've messed up the
1958         bookkeeping in combine():
1959    
1960             else ();
1961             case !ucol of
1962               PSEUDO => (if !cntv > 0 then
1963                     (if !cntu > 0 then blocked := !blocked - 1 else ();
1964                                        ^^^^^^^^^^^^^^^^^^^^^^^
1965                      moveu := mergeMoveList(!movev, !moveu)
1966                     )
1967                  else ();
1968    
1969         combine() is called to coalesce two nodes u and v.
1970         I think I was thinking that if the move counts of u and v are both
1971         greater than zero then after they are coalesced then one node is
1972         removed from the freeze queue.  Apparently I was thinking that
1973         both u and v are of low degree, but that's clearly not necessarily true.
1974    
1975    
1976    02/12/2002:
1977        Here's the patch.  HOL now compiles.
1978    
1979        I don't know how this impact on performance (compile
1980        time or runtime).  This bug caused the RA (especially on the x86)
1981        to go thru the potential spill phase when there are still nodes on the
1982        freeze queue.
1983    
1984    
1985    
1986    
1987    ----------------------------------------------------------------------
1988    Name: Matthias Blume
1989    Date: 2002/02/13 22:40:00 EST
1990    Tag: blume-20020213-fptr-rtti
1991    Description:
1992    
1993    Fixed a bug in ml-nlffigen that was introduced with one of the previous
1994    updates.
1995    
1996    ----------------------------------------------------------------------
1997    Name: Matthias Blume
1998    Date: 2002/02/13 16:41:00 EST
1999    Tag: blume-20020213-cmlpq
2000    Description:
2001    
2002    Added new priority queue export symbols (which have just been added to
2003    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
2004    not compile and the installer would choke.)
2005    
2006    ----------------------------------------------------------------------
2007  Name: Matthias Blume  Name: Matthias Blume
2008  Date: 2002/02/13 16:15:00 EST  Date: 2002/02/13 16:15:00 EST
2009  Tag: blume-20020213-various  Tag: blume-20020213-various

Legend:
Removed from v.1062  
changed lines
  Added in v.1390

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