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 847, Mon Jun 25 16:11:22 2001 UTC revision 1409, Mon Nov 3 22:00:35 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: 2003/11/03 16:00:00 CST
18    Tag: blume-20031103-installdir
19    Description:
20    
21    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
22    Richards for pointing out the problem and providing the solution.)
23    
24    ----------------------------------------------------------------------
25    Name: Matthias Blume (blume (at) tti - c (dot) org)
26    Date: 2003/10/01 17:05:00 CDT
27    Tag: blume-20031001-lal-mlrisc
28    Description:
29    
30    MLRISC bug fix from Lal.
31    
32    ----------------------------------------------------------------------
33    Name: Matthias Blume (blume (at) tti - c (dot) org)
34    Date: 2003/09/30 16:10:00 CDT
35    Tag: blume-20030930-primio-bat
36    Description:
37    
38    1. Added openVector, nullRd, and nullWr to PRIM_IO.
39    2. Improved .bat files (for Win32 port) to make things work under Win95.
40       (thanks to Aaron S. Hawley for this one)
41    
42    ----------------------------------------------------------------------
43    Name: Matthias Blume (blume (at) tti - c (dot) org)
44    Date: 2003/09/26 16:05:00 CDT
45    Tag: blume-20030926-wrappriv
46    Description:
47    
48    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
49    
50    ----------------------------------------------------------------------
51    Name: Matthias Blume (blume (at) tti - c (dot) org)
52    Date: 2003/09/26 15:00:00 CDT
53    Tag: blume-20030926-110_43_3
54    Description:
55    
56    - additional cleanup
57    - version number bump, NEW BOOTFILES
58    
59    ----------------------------------------------------------------------
60    Name: Matthias Blume (blume (at) tti - c (dot) org)
61    Date: 2003/09/26 12:00:00 CDT
62    Tag: blume-20030926-ppautoload
63    Description:
64    
65    I modified the read-eval-print loop so that the autoloader gets
66    invoked whenever the prettyprinter tries to look up a symbol that
67    is not currently defined in the toplevel environment but which
68    appears in CM's autoload registry.  As a result, we see far fewer of
69    those ?.Foo.Bar.xxx names in the prettyprinter's output.
70    
71    In addition to this I tried to clean up some pieces of the Basis
72    implementation (e.g., Socket, Word8Array) in order to prevent other
73    instances of these ?.Foo.Bar.xxx names from being printed.
74    
75    The mechanism that picks names for types still needs some work, though.
76    (Right now it seems that if there is a type A.t which is defined to
77    be B.u, but B is unavailable at toplevel, then A.t gets printed as
78    "?.B.u" although the perhaps more sensible solution would be to use
79    "A.t" in this case.  In other words, the prettyprinter should follow
80    a chain of DEFtycs not farther than there are corresponding toplevel
81    names in the current environment.)
82    
83    ----------------------------------------------------------------------
84    Name: Matthias Blume (blume (at) tti - c (dot) org)
85    Date: 2003/09/24 16:31:00 CDT
86    Tag: blume-20030924-installer
87    Description:
88    
89    Another installer tweak:  All the ML code for the installer is now
90    compiled during CMB.make and put into a little library called
91    $smlnj/installer.cm.  The installation then simply invokes
92    
93       sml -m $smlnj/installer.cm
94    
95    and everything happens automagically.
96    
97    Win32:  ML code senses value of environment variable SMLNJ_HOME.
98    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
99            and BINDIR.
100    
101    The new scheme guarantees that the ML code responsible for the installation
102    is in sync with the APIs of the main system.  Also, the installer is
103    somewhat faster because the installer script is precompiled.
104    
105    ----------------------------------------------------------------------
106    Name: Matthias Blume (blume (at) tti - c (dot) org)
107    Date: 2003/09/24 15:35:00 CDT
108    Tag: blume-20030924-synsock
109    Description:
110    
111    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
112    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
113    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
114    implementations of non-blocking operations from CML's Socket
115    structure.
116    
117    ----------------------------------------------------------------------
118    Name: Matthias Blume (blume (at) tti - c (dot) org)
119    Date: 2003/09/24 15:10:05 CDT
120    Tag: blume-20030924-sockets
121    Description:
122    
123    1. Fixed SOCKET API and implementation to match Basis spec.
124       This required changing the internal representation of sockets to one
125       that remembers (for each socket file descriptor) whether it is currently
126       blocking or non-blocking.  This state is maintained lazily (i.e., a system
127       call is made only if the state actually needs to change).
128    
129    2. OS-specific details of sockets were moved into separate files, thus
130       making it possible to unify the bulk of the socket implementations
131       between Unix and Win32.
132    
133    3. CML's socket API changed accordingly.
134       (Note that we need to remove non-blocking functions from this API
135       since they are redundant in the case of CML!)
136    
137    4. CML's socket implementation now makes use of non-blocking functions
138       provided by Basis, thus removing all OS-dependent code from this part
139       of CML.
140    
141    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
142    
143    ----------------------------------------------------------------------
144    Name: Matthias Blume (blume (at) tti - c (dot) org)
145    Date: 2003/09/22 12:10:00 CDT
146    Tag: blume-20030922-110_43_2
147    Description:
148    
149    Made a new interim version and bootfiles for developer's bootstrapping
150    convenience.
151    
152       110.43.2 -- NEW BOOTFILES
153    
154    ----------------------------------------------------------------------
155    Name: Matthias Blume (blume (at) tti - c (dot) org)
156    Date: 2003/09/19 15:55:00 CDT
157    Tag: blume-20030919-cmdir
158    Description:
159    
160    1. new-install.sh -> install.sh
161    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
162    3. tweaked installer so that another name instead of .cm can be chosen
163       at install time (by setting the CM_DIR_ARC environment variable
164       during installation); once installation is complete, the name is
165       fixed
166    
167    ----------------------------------------------------------------------
168    Name: Matthias Blume (blume (at) tti - c (dot) org)
169    Date: 2003/09/18 16:00:00 CDT
170    Tag: blume-20030918-110_43_1
171    Description:
172    
173    Made a new interim version and bootfiles for developer's bootstrapping
174    convenience.
175    
176       110.43.1 -- NEW BOOTFILES
177    
178    ----------------------------------------------------------------------
179    Name: Matthias Blume (blume (at) tti - c (dot) org)
180    Date: 2003/09/18 15:20:00 CDT
181    Tag: blume-20030918-misc
182    Description:
183    
184    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
185       this seems to be controversial at the moment) and used those in
186       Posix.ProcEnv.times.
187    
188    2. Added Time.{from,to}Nanoseconds to Time.
189    
190    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
191       For example, fromLargeInt never needs to look at more than 3 "big
192       digits" to get its 53 bits of precision.
193    
194    ----------------------------------------------------------------------
195    Name: Matthias Blume (blume (at) tti - c (dot) org)
196    Date: 2003/09/17 16:30:00 CDT
197    Tag: blume-20030917-real32-slices
198    Description:
199    
200    Added an entry to the primitive environment
201    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
202    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
203    
204    Removed some of the "magic" constants in real64.sml and replaced them
205    with code that generates these values from their corresponding
206    integer counterparts.
207    
208    Made all(?) the slice-related changes to the Basis and made everything
209    compile again...
210    
211    ----------------------------------------------------------------------
212    Name: Matthias Blume (blume (at) tti - c (dot) org)
213    Date: 2003/09/15 17:45:00 CDT
214    Tag: blume-20030915-rbase
215    Description:
216    
217    Fixed bug in Real.fromLargeInt.
218    
219    ----------------------------------------------------------------------
220    Name: Matthias Blume (blume (at) tti - c (dot) org)
221    Date: 2003/09/13 18:11:00 CDT
222    Tag: blume-20030913-libinstall
223    Description:
224    
225    Minor bugfix in config/libinstall (set anchor with path to
226    standalone tool after installing it, otherwise libraries that
227    need ml-lex or ml-yacc won't compile the first time the installer
228    runs).
229    
230    ----------------------------------------------------------------------
231    Name: Matthias Blume (blume (at) tti - c (dot) org)
232    Date: 2003/09/12 11:45:00 CDT
233    Tag: blume-20030912-various
234    Description:
235    
236     - fixed bug in Real.toLargeInt
237     - fixed bug in Posix.ProcEnv.times
238     - changed inputLine functions to return an option
239     - minor installer improvements / bugfixes
240     - changed default @SMLalloc parameter for x86/celeron to 64k
241    
242    ----------------------------------------------------------------------
243    Name: Matthias Blume (blume (at) tti - c (dot) org)
244    Date: 2003/09/09 22:00:00 CDT
245    Tag: Release_110_43
246    Description:
247    
248        New working release 110.43.  New bootfiles.
249    
250    ----------------------------------------------------------------------
251    Name: Matthias Blume (blume (at) tti - c (dot) org)
252    Date: 2003/09/09 19:20:00 CDT
253    Tag: blume-20030909-installer
254    Description:
255    
256    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
257    Modified config/install.bat to take advantage of it.  Also modified
258    config/install.sh (and called it config/new-install.sh) to take advantage
259    of it on Unix systems.  (The SML code is (supposed to be) platform-
260    independent.)
261    
262    The installer can now install everything under Win32
263    as well as under *nix as long as it compiles.
264    
265    Other changes:
266    
267      - made CML compile again under Win32
268      - made eXene compile under Win32 (by providing a fake structure UnixSock
269        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
270      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
271        same as type int;  under Win32 it isn't
272      - fixed some slice-related problems in the win32-specific parts of CML
273      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
274        (under Win32, the volume name is case-insensitive, and the
275        OS.Path code compares volume names for equality)
276    
277    ----------------------------------------------------------------------
278    Name: Matthias Blume (blume (at) tti - c (dot) org)
279    Date: 2003/09/08 11:55:00 CDT
280    Tag: blume-20030908-fullpath
281    Description:
282    
283    Made Win32 version of OS.FileSys.fullPath return current directory
284    when given an empty string.  This is what the spec says, and incidentally,
285    CM depends on it.  (CM otherwise goes into an infinite loop in certain
286    cases when presented with the name of a non-existing .cm file.)
287    
288    ----------------------------------------------------------------------
289    Name: Matthias Blume (blume (at) tti - c (dot) org)
290    Date: 2003/09/04 16:30:00 CDT
291    Tag: blume-20030905-slices-etc
292    Description:
293    
294      1. Changed interface to vectors and arrays in Basis to match
295         (draft) Basis spec.
296      2. Added signatures and implementations of slices according to
297         Basis spec.
298      3. Edited source code throughout the system to make it compile again
299         under 1. and 2.  (In some cases code had to be added to have it
300         match the new signatures.)
301      4. MLRISC should be backward-compatible: the copies of the originals
302         of files that needed to change under 3. were retained, the .cm files
303         check the compiler version number and use old versions when
304         appropriate.
305      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
306         dirstream -> string option (in accordance with Basis spec).
307      6. When generating code that counts lines, ml-lex used function
308         CharVector.foldli, taking advantage of its old interface.
309         This has been replaced with the corresponding code from
310         CharVectorSlice.  (html-lex must be re-lexed!)
311      7. BitArray in smlnj-lib/Util has been extended/modified to match the
312         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
313      8. Removed temporary additions (fromInternal, toInternal) from the
314         (now obsolete) IntInf in smlnj-lib/Util.
315      9. Cleaned up structure Byte.
316     10. Added localOffset, scan, and fromString to Date (according to spec).
317         Cleaned/corrected implementation of Date.
318         (Still need to check for correctness; implement better canonicalizeDate.)
319     11. Added "scan" to signature IEEE_REAL.
320     12. Some improvements to IntInf [in particular: efficiency-hack for
321         mod and rem when second operand is 2 (for parity checks).]
322     13. Changed representation of type Time.time, using a single IntInf.int
323         value counting microseconds.  This considerably simplified the
324         implementation of structure Time.  We now support negative time
325         values; scan and fromString handle signs.
326     14. Functor PrimIO now takes two additional arguments (VectorSlice and
327         ArraySlice).
328    
329    ----------------------------------------------------------------------
330    Name: Matthias Blume (blume (at) tti - c (dot) org)
331    Date: 2003/08/28 17:00:00 CDT
332    Tag: blume-20030828-intinf
333    Description:
334    
335    This is a major update which comes with a version number bump
336    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
337    and an implementation of IntInf in the Basis.
338    
339    There are a fairly large number of related changes and updates throughout
340    the system:
341    
342    Basis:
343            - Implemented IntInf.
344            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
345            - Added some missing Real64 operations, most notably Real.toLargeInt.
346            - Added FixedInt as a synonym for Int32.
347    
348    compiler:
349            * Added support for a built-in intinf type.
350              - literals
351              - pattern matching
352              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
353              - overloading on literals and operations
354    
355            This required adding a primitive type intinf, some additional
356            primops, and implementations for several non-trivial intinf
357            operations in Core.  (The intinf type is completely abstract
358            to the compiler; all operations get delegated back to the Core.)
359    
360            * Intinf equality is handled by polyequal.  However, the compiler
361            does not print its usual warning in this case (since polyequal
362            is the right thing to do there).
363    
364            * Improved the organization of structure InlineT.
365    
366            * A word about conversion primops:
367            If conversions involving intinf do not cancel out during
368            CPS contract, then the compiler must insert calls to Core functions.
369            Since all core access must be resolved already during the FLINT
370            translate phase, it would be too late a the time of CPS contract
371            to add new Core calls.  For this reason, conversion primops
372            for intinf carry two arguments:  1. the numeric argument that
373            they are supposed to convert, and 2. the Core function that
374            can help with this conversion if necessary.  If CPS contract
375            eliminates a primop, then the associated Core function becomes
376            dead and goes away.  Intinf conversion primops that do not get
377            eliminated by CPS contract get rewritten into calls of their
378            core functions by a separate, new phase.
379    
380    interactive system:
381            - Control.Print.intinfDepth controls max length of intinf constants
382            being printed. (Analogous to Control.Print.stringDepth.)
383            - Cleanup in printutil and pputil: got rid of unused stuff and
384            duplicates; replaced some of the code with code that makes better
385            use of library functionality.
386    
387    CM:
388            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
389            are no longer silent.
390    
391    CKIT:
392            Fixed mismatched uses of Int32 and LargeInt.  I always decided
393            in favor of LargeInt -- which is now the same as IntInf.
394            CKIT-knowledgable people should check whether this is what's
395            intended and otherwise change things back to using Int32 or
396            FixedInt.
397    
398    Throughout the code:
399            Started using IntInf.int literals and built-in operations
400            (e.g., comparison with 0) where this seems appropriate.
401    
402    
403    ----------------------------------------------------------------------
404    Name: Dave MacQueen (dbm@cs.uchicago.edu)
405    Date: 2003/08/13 11:36:00 CDT
406    Tag: dbm-20030813-mcz-merge1
407    Description:
408    
409    Merging changes from the mcz-branch development branch into trunk.
410    These changes involve replacement of the emulated old prettyprinter
411    interface with direct use of the SML/NJ Lib PP library, and fixing
412    of a couple of bugs (895, 1186) relating to error messages.  A new
413    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
414    has been added.
415    
416    ----------------------------------------------------------------------
417    Name: Matthias Blume (blume (at) tti - c (dot) org)
418    Date: 2003/08/11 15:45:00 CDT
419    Tag: blume-20030811-windows
420    Description:
421    
422    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
423    
424        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
425    
426    This patch restores SML/NJ's ability to run under win32.  There are a
427    number of changes, including fixes for several bugs that had gone
428    unnoticed until now:
429    
430      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
431        trying to run the win32 version from within cygwin.)
432    
433      - There are a number of simple .bat scripts that substitute for their
434        corresponding Unix shell-scripts.  (See below.)
435    
436      - The internals of ml-build have been modified slightly.  The main
437        difference is that instead of calling ".link-sml" (or link-sml.bat)
438        using OS.Process.system, the ML process delegates this task back
439        to the script.  Otherwise problems arise in mixed environments such
440        as Cygwin where scripts look and work like Unix scripts, but
441        where OS.Process.system cannot run them.
442    
443      - In CM, the srcpath pickler used native pathname syntax -- which
444        is incorrect in the case of cross-compilation.  The new pickle format
445        is independent of platform-specific naming conventions.
446    
447      - Path configuration files (such as lib/pathconfig) can now choose
448        between native and standard syntax.  Placing a line of the form
449    
450            standard!
451    
452        into the file causes all subsequent paths to be interpreted using
453        CM standard pathname syntax (= Unix conventions); a line
454    
455            native!
456    
457        switches back to native style.  This was needed so that
458        path config files can be written portably, see src/system/pathconfig.
459    
460      - Runtime system:
461    
462          - win32-filesys.c:  get_file_time and set_file_time now
463            access modification time, not creation time.
464    
465          - I/O code made aware of new array representation.
466    
467          - Bug fixes in X86.prim.masm.
468    
469       - src/system/makeml made aware of win32. (For use under cygwin
470            and other Unix-environments for windows.)
471    
472       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
473         which caused BinIO.inputAll to fail consistently.
474    
475    .bat scripts:
476    
477       Windows .bat scripts assume that SMLNJ_HOME is defined.
478    
479       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
480         applications (sml, ml-yacc, ml-lex).
481       - ml-build.bat: analogous to ml-build.
482       - config\install.bat:  Analogous to config/install.sh.  This requires
483         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
484         (nmake etc. must be on the path, and vcvars32 must have been run.)
485         Moreover, sources for ml-lex and ml-yacc need to exist under src,
486         and the bootfile hierarchy must have been unpacked under
487         sml.boot.x86-win32.
488         The script is very primitive and does a poor job at error checking.
489         It only installs the base system, ml-lex, and ml-yacc.  No other
490         libraries are being installed (i.e., you get only those that
491         are part of the compiler.)
492       - link-sml.bat:  analogous to .link-sml, but not currently used
493    
494    Unrelated bug fixes:
495    
496      - ml-nlffigen now exports structures ST_* corresponding to incomplete
497        types.
498      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
499        otherwise.)
500    
501    ----------------------------------------------------------------------
502    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
503    Date: 2003/06/17
504    Tag: macqueen-20030617-bug895
505    Description:
506    
507    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
508    Tag will be used for new development branch (mcz-branch) for
509    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
510    
511    ----------------------------------------------------------------------
512    Name: Matthias Blume (blume (at) tti - c (dot) org)
513    Date: 2003/05/27 16:55:00 CDT
514    Tag: blume-20030527-polyeq
515    Description:
516    
517    Tried to eliminated most cases of polymorphic equality.
518    
519    ----------------------------------------------------------------------
520    Name: Matthias Blume (blume (at) tti - c (dot) org)
521    Date: 2003/05/21 17:45:00 CDT
522    Tag: blume-20030517-complete
523    Description:
524    
525    Two changes:
526    
527      1. Added a flag for controlling whether non-exhaustive bindings will
528         be treated as errors (default is false).
529      2. Cleaned up the *entire* source tree so that CMB.make goes through
530         without a single non-exhaustive match- or bind warning.
531    
532    ----------------------------------------------------------------------
533    Name: Matthias Blume (blume (at) tti - c (dot) org)
534    Date: 2003/05/17 10:20:00 CDT
535    Tag: blume-20030517-absyn
536    Description:
537    
538    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
539    
540       This mainly affects the quality of error messages.  However, some
541       of the code is now more straightforward than before. (Treatment of
542       the above four constructs in translate.sml is much simpler than
543       the "macro-expansion" that was going on before.  Plus, the mach-
544       compiler no longer gets invoked just to be able to compile an
545       if-expression.)
546    
547    2. The ErrorMsg.Error exception is now caught and absorbed by the
548       interactive loop.
549    
550    ----------------------------------------------------------------------
551    Name: Allen Leung
552    Date: 2003/05/16 13:05:00 CDT
553    Tag: leunga-20030516-cygwin-runtime
554    Description:
555    
556        Ported the runtime system to cygwin, which uses the unix
557    x86-unix bin files.  Missing/buggy features:
558    
559       o getnetbyname, getnetbyaddr: these functions seem to be missing in
560         the Cygwin library.
561       o Ctrl-C handling may be flaky.
562       o Windows system calls and Windows I/O are not supported.
563    
564       A new set of binfiles is located at:
565    
566         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
567    
568       This is only needed for bootstrapping the cygwin version of smlnj.
569       Other x86 versions can use the existing binfiles.
570    
571    ----------------------------------------------------------------------
572    Name: Matthias Blume
573    Date: 2003/04/08 15:42:00 CDT
574    Tag: blume-20030408-listpair
575    Description:
576    
577    1. Added a target 'mlrisc' to installer.
578    
579    2. Added missing elements to structure ListPair.
580    
581    ----------------------------------------------------------------------
582    Name: Allen Leung
583    Date: 2003/01/07 10:40:00 EST
584    Tag: leunga-20030107-int-rem
585    Description:
586    
587    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
588    The arguments to the SUBL instruction were swapped.
589    
590    ----------------------------------------------------------------------
591    Name: Matthias Blume
592    Date: 2002/12/12 16:25:00 EST
593    Tag: blume-20021212-risc-ra
594    Description:
595    
596    Fixed a serious bug in the rewrite code for FP spilling/reloading that
597    sent the RA into an infinite loop when floating point registers get
598    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
599    110.37 and 110.38.)
600    There was another set of potential problems related to the handling of
601    MLRISC annotations (but those did not yet cause real problems, apparently).
602    
603    ----------------------------------------------------------------------
604    Name: Matthias Blume
605    Date: 2002/12/06 22:40:00 EST
606    Tag: blume-20021206-cm-fileid
607    Description:
608    
609    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
610    This fixes the problem of CM getting confused by files that suddenly
611    change their identity (e.g., by getting unlinked and recreated by some
612    text editor such as vi).  There might be a better/cheaper/cleaner way
613    of doing this, but for now this will have to do.
614    
615    ----------------------------------------------------------------------
616    Name: Matthias Blume
617    Date: 2002/10/28 09:50:00 EST
618    Tag: blume-20021028-typecheck
619    Description:
620    
621    Exported structure Typecheck from $smlnj/viscomp/core.cm.
622    
623    ----------------------------------------------------------------------
624    Name: Matthias Blume
625    Date: 2002/10/17 09:10:00 EDT
626    Tag: Release_110_42
627    Description:
628    
629    In good old tradition, there has been a slight hiccup so that we have
630    to patch 110.42 after the fact.   The old release tag has been replaced
631    (see below).
632    
633    The change solves a problem with two competing approaches the
634    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
635    each other's way.
636    
637    This change only affects the runtime system code and the installer script.
638    (No new bootfiles.)
639    
640    ----------------------------------------------------------------------
641    Name: Matthias Blume
642    Date: 2002/10/16 12:00:00 EDT
643    Tag: Release_110_42_removed
644    Description:
645    
646    New working release.  New bootfiles.
647    
648    ----------------------------------------------------------------------
649    Name: Matthias Blume
650    Date: 2002/10/10 13:10:00 EDT
651    Tag: blume-20021010-ppc-divs
652    Description:
653    
654    The mltree operator DIVS must be implemented with an overflow check on
655    the PPC because the hardware indicates divide-by-zero using "overflow" as
656    well.
657    
658    ----------------------------------------------------------------------
659    Name: Matthias Blume
660    Date: 2002/07/23 11:45:00
661    Tag: blume-20020723-smlnj-home
662    Description:
663    
664    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
665    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
666    CM_PATHCONFIG is also set), the path configuration file is assumed
667    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
668    the entire tree to some other place and everything will "just work".
669    
670    (Companion commands such as ml-build and ml-makedepend also sense this
671    variable.)
672    
673    ----------------------------------------------------------------------
674    Name: Matthias Blume
675    Date: 2002/07/12 21:19:00 EDT
676    Tag: blume-20020712-liveness
677    Description:
678    
679    Exported two useful "step" functions from liveness module (MLRISC).
680    
681    ----------------------------------------------------------------------
682    Name: Matthias Blume
683    Date: 2002/07/05 16:00 EDT
684    Tag: Release_110_41
685    Description:
686    
687    New working release.  New bootfiles.
688    
689    ----------------------------------------------------------------------
690    Name: Matthias Blume
691    Date: 2002/07/05 10:25:00 EDT
692    Tag: blume-20020705-btimp
693    Description:
694    
695    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
696    other clients can set up backtracing support.
697    
698    ----------------------------------------------------------------------
699    Name: Matthias Blume
700    Date: 2002/06/25 17:23:00 EDT
701    Tag: blume-20020625-fpmax
702    Description:
703    
704    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
705    case.  (The sense of the isNaN test was reversed -- which made min and
706    max always return their first argument.)
707    
708    ----------------------------------------------------------------------
709    Name: Matthias Blume
710    Date: 2002/06/11
711    Tag: blume-20020611-unixpath
712    Description:
713    
714    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
715    
716    ----------------------------------------------------------------------
717    Name: Matthias Blume
718    Date: 2002/06/10 16:35:00 EDT
719    Tag: blume-20020610-ieeereal
720    Description:
721    
722    I back-ported my implementation of IEEEReal.fromString from the
723    idlbasis-devel branch so that we can test it.
724    
725    Another small change is that ppDec tries to give more information
726    than just "<sig>" in the case of functors.  However, this code is
727    broken in some mysterious way if the functor's body's signature
728    has not been declared by ascription but gets inferred from the
729    implementation.  This needs fixing...
730    
731    ----------------------------------------------------------------------
732    Name: Matthias Blume
733    Date: 2002/05/31
734    Tag: blume-20020531-btrace-mode
735    Description:
736    
737    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
738    the wayside when I switched over to using Controls everywhere.)
739    
740    ----------------------------------------------------------------------
741    Name: Lal George
742    Date: 2002/05/23 12:21:40 EDT
743    Tag: george-20020523-visual-labels
744    Description:
745    
746    Labels are now displayed in the graphical output to make
747    the fall-through and target blocks obvious.
748    
749    ----------------------------------------------------------------------
750    Name: Matthias Blume
751    Date: 2002/05/22 11:03:00 EDT
752    Tag: blume-20020522-shrink
753    Description:
754    
755    John tweaked yesterday's fix for 1131 to handle an out-of-memory
756    situation that comes up when allocating huge arrays.
757    
758    ----------------------------------------------------------------------
759    Name: Matthias Blume
760    Date: 2002/05/21 16:00:00 EDT
761    Tag: Release_110_40
762    Description:
763    
764    New working release (110.40).  New bootfiles.
765    
766    [Also: John Reppy fixed GC bug 1131.]
767    
768    ----------------------------------------------------------------------
769    Name: Matthias Blume
770    Date: 2002/05/21 12:35:00 EDT
771    Tag: blume-20020521-cmdoc
772    Description:
773    
774    CM documentation update.
775    
776    ----------------------------------------------------------------------
777    Name: Matthias Blume
778    Date: 2002/05/21 10:55:00 EDT
779    Tag: blume-20020521-misc
780    Description:
781    
782    - John tweaked runtime to be silent on heap export (except when
783      GC messages are on).
784    - I added a few more things (cross-compiling versions of CMB) to
785      config/preloads (as suggestions).
786    
787    ----------------------------------------------------------------------
788    Name: Matthias Blume
789    Date: 2002/05/20 22:25:00 EDT
790    Tag: blume-20020520-controls
791    Description:
792    
793    - Added ControlUtil structure to control-lib.cm.
794    - Use it throughout.
795    - Used Controls facility to define MLRISC controls (as opposed to
796      registering MLRISC control ref cells with Controls after the fact)
797    - Fixed messed-up controls priorities.
798    
799    * Removed again all the stuff from config/preloads that one wouldn't
800      be able to preload at the time the initial heap image is built.
801      (Many libraries, e.g., CML, do not exist yet at this time.  The
802       only libraries that can be preloaded via config/preloads are those
803       that come bundled with the bootfiles.)
804    
805    ----------------------------------------------------------------------
806    Name: Matthias Blume
807    Date: 2002/05/20 10:59:00 EDT
808    Tag: blume-20020520-preloads
809    Description:
810    
811    Added a lot of commented-out suggestions for things to be included
812    in config/preloads.
813    
814    ----------------------------------------------------------------------
815    Name: Allen Leung
816    Date: 2002/05/18 14:20:00 EDT
817    Tag: leunga-20020518-mdl
818    Description:
819    
820    o Made the mdl tool stuff compile and run again.
821    o I've disabled all the stuff that depends on RTL specifications; they
822      are all badly broken anyway.
823    
824    ----------------------------------------------------------------------
825    Name: Matthias Blume
826    Date: 2002/05/17 16:49:00 EDT
827    Tag: blume-20020517-controls
828    Description:
829    
830    0. John Reppy made several modifications to the SML/NJ library.
831       In particular, there is a shiny new controls-lib.cm.
832    
833    1. Pushed new controls interface through compiler so that everything
834       compiles again.
835    
836    2. Added FormatComb and FORMAT_COMB to the CML version of the
837       SML/NJ library (so that CML compiles again).
838    
839    3. Modified init scripts because XXX_DEFAULT environment variables
840       are no longer with us.  (Boot-time initialization is now done
841       using the same environment variables that are also used for
842       startup-time initialization of controls.)
843    
844    ----------------------------------------------------------------------
845    Name: Lal George
846    Date: 2002/05/15 09:20:10 EDT
847    Tag: george-20020515-pseudo-op-decls
848    Description:
849    
850            All pseudo-ops emitted before the first segment declaration
851    such as TEXT, DATA, and BSS directives are assumed to be global
852    declarations and are emitted first in the assembly file. This is
853    useful in a number of situations where one has pseudo-ops that are not
854    specific to any segment, and also works around the constraint that one
855    cannot have client pseudo-ops in the TEXT segment.
856    
857            Because no segment is associated with these declarations it is
858    an error to allocate any space or objects before the first segment
859    directive and an exception will be raised. However, we cannot make
860    this check for client pseudo-ops.
861    
862            These top level declarations are a field in the CFG graph_info.
863    In theory you can continue to add to this field after the CFG has been
864    built -- provided you know what you are doing;-)
865    
866    ----------------------------------------------------------------------
867    Name: Matthias Blume
868    Date: 2002/05/13 16:40:00 EDT
869    Tag: blume-20020513-pp-etc
870    Description:
871    
872    A few minor bugfixes:
873    
874      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
875        (Bogus printouts for redefined bindings still occur.  Compiler
876        bug should no longer occur now.  We need to redo the prettyprinter
877        from scratch.)
878    
879      - CM pathname printer now also adds escape sequences for ( and )
880    
881      - commend and docu fixes for ml-nlffi
882    
883    ----------------------------------------------------------------------
884    Name: Matthias Blume
885    Date: 2002/05/10 16:40:00 EDT
886    Tag: blume-20020510-erg-textio
887    Description:
888    
889    Applied the following bugfix provided by Emden Gansner:
890    
891        Output is corrupted when outputSubstr is used rather than output.
892    
893        The problem occurs when a substring
894    
895            ss = (s, dataStart, dataLen)
896    
897        where dataStart > 0, fills a stream buffer with avail bytes left.
898        avail bytes of s, starting at index dataStart, are copied into the
899        buffer, the buffer is flushed, and then the remaining dataLen-avail
900        bytes of ss are copied into the beginning of the buffer. Instead of
901        starting this copy at index dataStart+avail in s, the current code
902        starts the copy at index avail.
903    
904        Fix:
905        In text-io-fn.sml, change line 695 from
906             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
907        to
908             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
909    
910    ----------------------------------------------------------------------
911    Name: Matthias Blume
912    Date: 2002/04/12 13:55:00 EDT
913    Tag: blume-20020412-assyntax
914    Description:
915    
916    1. Grabbed newer assyntax.h from the XFree86 project.
917    2. Fiddled with how to compile X86.prim.asm without warnings.
918    3. (Very) Minor cleanup in CM.
919    
920    ----------------------------------------------------------------------
921    Name: Matthias Blume
922    Date: 2002/04/01 (no joke!) 17:07:00 EST
923    Tag: blume-20020401-x86div
924    Description:
925    
926    Added full support for div/mod/rem/quot on the x86, using the machine
927    instruction's two results (without clumsily recomputing the remainder)
928    directly where appropriate.
929    
930    Some more extensive power-of-two support was added to the x86 instruction
931    selector (avoiding expensive divs, mods, and muls where they can be
932    replaced with cheaper shifts and masks).  However, this sort of thing
933    ought to be done earlier, e.g., within the CPS optimizer so that
934    all architectures benefit from it.
935    
936    The compiler compiles to a fixed point, but changes might be somewhat
937    fragile nevertheless.  Please, report any strange things that you might
938    see wrt. div/mod/quot/rem...
939    
940    ----------------------------------------------------------------------
941    Name: Matthias Blume
942    Date: 2002/03/29 17:22:00
943    Tag: blume-20020329-div
944    Description:
945    
946    Fixed my broken div/mod logic.  Unfortunately, this means that the
947    inline code for div/mod now has one more comparison than before.
948    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
949    The problem was with quotient = 0, because that alone does not tell
950    us which way the rounding went.  One then has to look at whether
951    remainder and divisor have the same sign...  :(
952    
953    Anyway, I replaced the bootfiles with fresh ones...
954    
955    ----------------------------------------------------------------------
956    Name: Matthias Blume
957    Date: 2002/03/29 14:10:00 EST
958    Tag: blume-20020329-inlprims
959    Description:
960    
961    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
962    
963    Primops have changed. This means that the bin/boot-file formats have
964    changed as well.
965    
966    To make sure that there is no confusion, I made a new version.
967    
968    
969    CHANGES:
970    
971    * removed REMT from mltree (remainder should never overflow).
972    
973    * added primops to deal with divisions of all flavors to the frontend
974    
975    * handled these primops all the way through so they map to their respective
976      MLRISC support
977    
978    * used these primops in the implementation of Int, Int32, Word, Word32
979    
980    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
981    
982    * parameterized INLMIN, INLMAX, and INLABS by a numkind
983    
984    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
985      floating point
986    
987    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
988      and Real (but Real.abs maps to a separate floating-point-only primop)
989    
990    
991    TODO items:
992    
993    * Hacked Alpha32 instruction selection, disabling the selection of REMx
994      instructions because the machine instruction encoder cannot handle
995      them.  (Hppa, PPC, and Sparc instruction selection did not handle
996      REM in the first place, and REM is supported by the x86 machine coder.)
997    
998    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
999      selection phase.  (The two can be streamlined because the hardware
1000      delivers both quotient and remainder at the same time anyway.)
1001    
1002    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
1003      (Currently the behavior is inconsistent both across architectures and
1004      wrt. the draft Basis spec.)
1005    
1006    * Word8 should eventually be handled natively, too.
1007    
1008    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
1009      as if there currently is no execution path that could trigger it in
1010      SML/NJ.  (The assumptions underlying functions arith and promotable do not
1011      hold for things like multiplication and division.)
1012    
1013    ----------------------------------------------------------------------
1014    Name: Matthias Blume
1015    Date: 2002/03/27 16:27:00 EST
1016    Tag: blume-20020327-mlrisc-divisions
1017    Description:
1018    
1019    Added support for all four division operations (ML's div, mod, quot,
1020    and rem) to MLRISC.  In the course of doing so, I also rationalized
1021    the naming (no more annoying switch-around of DIV and QUOT), by
1022    parameterizing the operation by div_rounding_mode (which can be either
1023    DIV_TO_ZERO or DIV_TO_NEGINF).
1024    
1025    The generic MLTreeGen functor takes care of compiling all four
1026    operations down to only round-to-zero div.
1027    
1028    Missing pieces:
1029    
1030      * Doing something smarter than relying on MLTreeGen on architectures
1031        like, e.g., the x86 where hardware division delivers both quotient and
1032        remainder at the same time.  With this, the implementation of the
1033        round-to-neginf operations could be further streamlined.
1034    
1035      * Remove inlining support for div/mod/rem from the frontend and replace it
1036        with primops that get carried through to the backend.  Do this for all
1037        int and word types.
1038    
1039    ----------------------------------------------------------------------
1040    Name: Matthias Blume
1041    Date: 2002/03/25 17:25:00 EST
1042    Tag: blume-20020325-divmod
1043    Description:
1044    
1045    I improved (hopefully without breaking them) the implementation of Int.div,
1046    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
1047    advantage of the following observations:
1048    
1049      Let  q = x quot y      r = x rem y
1050           d = x div  y      m = x mod y
1051    
1052    where "quot" is the round-to-zero version of integer division that
1053    hardware usually provides.  Then we have:
1054    
1055         r = x - q * y        where neither the * nor the - will overflow
1056         d = if q >= 0 orelse x = q * y then q else q - 1
1057                              where neither the * nor the - will overflow
1058         m = if q >= 0 orelse r = 0 then r else r + y
1059                              where the + will not overflow
1060    
1061    This results in substantial simplification of the generated code.
1062    The following table shows the number of CFG nodes and edges generated
1063    for
1064            fun f (x, y) = x OPER y
1065            (* with OPER \in div, mod, quot, rem *)
1066    
1067    
1068        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
1069        --------------------------------------------------------
1070         div |         24 |         39 |         12 |         16
1071         mod |         41 |         71 |         12 |         16
1072        quot |          8 |         10 |          8 |         10
1073         rem |         10 |         14 |          8 |         10
1074    
1075    
1076    ----------------------------------------------------------------------
1077    Name: Matthias Blume
1078    Date: 2002/03/25 22:06:00 EST
1079    Tag: blume-20020325-cprotobug
1080    Description:
1081    
1082    Fixed a bug in cproto (c prototype decoder).
1083    
1084    ----------------------------------------------------------------------
1085    Name: Matthias Blume
1086    Date: 2002/03/25 16:00:00 EST
1087    Tag: blume-20020325-raw-primops
1088    Description:
1089    
1090    I did some cleanup to Allen's new primop code and
1091    replaced yesterday's bootfiles with new ones.
1092    (But they are stored in the same place.)
1093    
1094    ----------------------------------------------------------------------
1095    Name: Matthias Blume
1096    Date: 2002/03/24 22:40:00 EST
1097    Tag: blume-20020324-bootfiles
1098    Description:
1099    
1100    Made the bootfiles that Allen asked for.
1101    
1102    ----------------------------------------------------------------------
1103    Name: Allen Leung
1104    Date: 2002/03/23 15:50:00 EST
1105    Tag: leunga-20020323-flint-cps-rcc-primops
1106    Description:
1107    
1108      1. Changes to FLINT primops:
1109    
1110        (* make a call to a C-function;
1111         * The primop carries C function prototype information and specifies
1112         * which of its (ML-) arguments are floating point. C prototype
1113         * information is for use by the backend, ML information is for
1114         * use by the CPS converter. *)
1115      | RAW_CCALL of { c_proto: CTypes.c_proto,
1116                       ml_args: ccall_type list,
1117                       ml_res_opt: ccall_type option,
1118                       reentrant : bool
1119                     } option
1120       (* Allocate uninitialized storage on the heap.
1121        * The record is meant to hold short-lived C objects, i.e., they
1122        * are not ML pointers.  With the tag, the representation is
1123        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1124        *)
1125      | RAW_RECORD of {tag:bool,sz:int}
1126      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1127    
1128      2.  These CPS primops are now overloaded:
1129    
1130           rawload of {kind:numkind}
1131           rawstore of {kind:numkind}
1132    
1133          The one argument form is:
1134    
1135             rawload {kind} address
1136    
1137          The two argument form is:
1138    
1139             rawload {kind} [ml object, byte-offset]
1140    
1141      3. RAW_CCALL/RCC now takes two extra arguments:
1142    
1143         a. The first is whether the C call is reentrant, i.e., whether
1144            ML state should be saved and restored.
1145         b. The second argument is a string argument specifying the name of
1146            library and the C function.
1147    
1148         These things are currently not handled in the code generator, yet.
1149    
1150      4. In CProto,
1151    
1152         An encoding type of "bool" means "ml object" and is mapped into
1153         C prototype of PTR.  Note that "bool" is different than "string",
1154         even though "string" is also mapped into PTR, because "bool"
1155         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
1156    
1157      5. Pickler/unpicker
1158    
1159         Changed to handle RAW_RECORD and newest RAW_CCALL
1160    
1161      6. MLRiscGen,
1162    
1163         1. Changed to handle the new rawload/rawstore/rawrecord operators.
1164         2. Code for handling C Calls has been moved to a new module CPSCCalls,
1165            in the file CodeGen/cpscompile/cps-c-calls.sml
1166    
1167      7. Added the conditional move operator
1168    
1169             condmove of branch
1170    
1171         to cps.  Generation of this is still buggy so it is currently
1172         disabled.
1173    
1174    ----------------------------------------------------------------------
1175    Name: Lal George
1176    Date: 2002/03/22 14:18:25 EST
1177    Tag: george-20020322-cps-branch-prob
1178    Description:
1179    
1180    Implemented the Ball-Larus branch prediction-heuristics, and
1181    incorporated graphical viewers for control flow graphs.
1182    
1183    Ball-Larus Heuristics:
1184    ---------------------
1185    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
1186    
1187    By design it uses the Dempster-Shafer theory for combining
1188    probabilities.  For example, in the function:
1189    
1190        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
1191    
1192    the ball-larus heuristics predicts that the n=0 is unlikely
1193    (OH-heuristic), and the 'then' branch is unlikely because of the
1194    RH-heuristic -- giving the 'then' branch an even lower combined
1195    probability using the Dempster-Shafer theory.
1196    
1197    Finally, John Reppy's loop analysis in MLRISC, further lowers the
1198    probability of the 'then' branch because of the loop in the else
1199    branch.
1200    
1201    
1202    Graphical Viewing:
1203    ------------------
1204    I merely plugged in Allen's graphical viewers into the compiler. The
1205    additional code is not much. At the top level, saying:
1206    
1207            Control.MLRISC.getFlag "cfg-graphical-view" := true;
1208    
1209    will display the graphical view of the control flow graph just before
1210    back-patching.  daVinci must be in your path for this to work. If
1211    daVinci is not available, then the default viewer can be changed
1212    using:
1213    
1214            Control.MLRISC.getString "viewer"
1215    
1216    which can be set to "dot" or "vcg" for the corresponding viewers. Of
1217    course, these viewers must be in your path.
1218    
1219    The above will display the compilation unit at the level of clusters,
1220    many of which are small, boring, and un-interesting. Also setting:
1221    
1222            Control.MLRISC.getInt "cfg-graphical-view_size"
1223    
1224    will display clusters that are larger than the value set by the above.
1225    
1226    
1227    ----------------------------------------------------------------------
1228    Name: Matthias Blume
1229    Date: 2002/03/21 22:20:00 EST
1230    Tag: blume-20020321-kmp-bugfix
1231    Description:
1232    
1233    Changed the interface to the KMP routine in PreString and fixed
1234    a minor bug in one place where it was used.
1235    
1236    ----------------------------------------------------------------------
1237    Name: Allen Leung
1238    Date: 2002/03/21 20:30:00 EST
1239    Tag: leunga-20020321-cfg
1240    Description:
1241    
1242      Fixed a potential problem in cfg edge splitting.
1243    
1244    ----------------------------------------------------------------------
1245    Name: Allen Leung
1246    Date: 2002/03/21 17:15:00 EST
1247    Tag: leunga-20020321-x86-fp-cfg
1248    Description:
1249    
1250      1. Recoded the buggy parts of x86-fp.
1251    
1252         a. All the block reordering code has been removed.
1253            We now depend on the block placement phases to do this work.
1254    
1255         b. Critical edge splitting code has been simplified and moved into the
1256            CFG modules, as where they belong.
1257    
1258         Both of these were quite buggy and complex.  The code is now much, much
1259         simpler.
1260    
1261      2. X86 backend.
1262    
1263         a. Added instructions for 64-bit support.  Instruction selection for
1264            64-bit has not been committed, however, since that
1265            requires changes to MLTREE which haven't been approved by
1266            Lal and John.
1267    
1268         b. Added support for FUCOMI and FUCOMIP when generating code for
1269            PentiumPro and above.  We only generate these instructions in
1270            the fast-fp mode.
1271    
1272         c. Added cases for JP and JNP in X86FreqProps.
1273    
1274      3. CFG
1275    
1276         CFG now has a bunch of methods for edge splitting and merging.
1277    
1278      4. Machine description.
1279    
1280         John's simplification of MLTREE_BASIS.fcond broke a few machine
1281         description things:
1282    
1283         rtl-build.{sig,sml} and hppa.mdl fixed.
1284    
1285         NOTE: the machine description stuff in the repository is still broken.
1286               Again, I can't put my fixes in because that involves
1287               changes to MLTREE.
1288    
1289    ----------------------------------------------------------------------
1290    Name: Matthias Blume
1291    Date: 2002/03/20 15:55:00 EST
1292    Tag: blume-20020320-kmp
1293    Description:
1294    
1295    Implemented Knuth-Morris-Pratt string matching in PreString and used
1296    it for String.isSubstring, Substring.isSubstring, and
1297    Substring.position.
1298    
1299    (Might need some stress-testing.  Simple examples worked fine.)
1300    
1301    ----------------------------------------------------------------------
1302    Name: Matthias Blume
1303    Date: 2002/03/19 16:37:00 EST
1304    Tag: blume-20020319-witnesses
1305    Description:
1306    
1307    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
1308    
1309    This implements a generic mechanism for changing constness qualifiers
1310    anywhere within big C types without resorting to outright "casts".
1311    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
1312    modify the constness at the outermost level.)
1313    The implementation of "convert" is based on the idea of "witness"
1314    values -- values that are not used by the operation but whose types
1315    "testify" to their applicability.  On the implementation side, "convert"
1316    is simply a projection (returning its second curried argument).  With
1317    cross-module inlining, it should not result in any machine code being
1318    generated.
1319    
1320    ----------------------------------------------------------------------
1321    Name: Matthias Blume
1322    Date: 2002/03/15 16:40:00 EST
1323    Tag: blume-20020315-basis
1324    Description:
1325    
1326    Provided (preliminary?) implementations for
1327    
1328      {String,Substring}.{concatWith,isSuffix,isSubstring}
1329    
1330    and
1331    
1332      Substring.full
1333    
1334    Those are in the Basis spec but they were missing in SML/NJ.
1335    
1336    ----------------------------------------------------------------------
1337    Name: Matthias Blume
1338    Date: 2002/03/14 21:30:00 EST
1339    Tag: blume-20020314-controls
1340    Description:
1341    
1342    Controls:
1343    ---------
1344    
1345    1. Factored out the recently-added Controls : CONTROLS stuff and put
1346       it into its own library $/controls-lib.cm.  The source tree for
1347       this is under src/smlnj-lib/Controls.
1348    
1349    2. Changed the names of types and functions in this interface, so they
1350       make a bit more "sense":
1351    
1352          module -> registry
1353          'a registry -> 'a group
1354    
1355    3. The interface now deals in ref cells only.  The getter/setter interface
1356       is (mostly) gone.
1357    
1358    4. Added a function that lets one register an already-existing ref cell.
1359    
1360    5. Made the corresponding modifications to the rest of the code so that
1361       everything compiles again.
1362    
1363    6. Changed the implementation of Controls.MLRISC back to something closer
1364       to the original.  In particular, this module (and therefore MLRISC)
1365       does not depend on Controls.  There now is some link-time code in
1366       int-sys.sml that registers the MLRISC controls with the Controls
1367       module.
1368    
1369    CM:
1370    ---
1371    
1372      * One can now specify the lambda-split aggressiveness in init.cmi.
1373    
1374    ----------------------------------------------------------------------
1375    Name: Allen Leung
1376    Date: 2002/03/13 17:30:00 EST
1377    Tag: leunga-20020313-x86-fp-unary
1378    Description:
1379    
1380    Bug fix for:
1381    
1382    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
1383    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
1384    > - fun f(x,(y,z)) = Real.~ y;
1385    > [autoloading]
1386    > [autoloading done]
1387    >       fchsl   (%eax), 184(%esp)
1388    > Error: MLRisc bug: X86MCEmitter.emitInstr
1389    >
1390    > uncaught exception Error
1391    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
1392    
1393    The problem was that the code generator did not generate any fp registers
1394    in this case, and the ra didn't know that it needed to run the X86FP phase to
1395    translate the pseudo fp instruction.   This only happened with unary fp
1396    operators in certain situations.
1397    
1398    ----------------------------------------------------------------------
1399    Name: Matthias Blume
1400    Date: 2002/03/13 14:00:00 EST
1401    Tag: blume-20020313-overload-etc
1402    Description:
1403    
1404    1. Added _overload as a synonym for overload for backward compatibility.
1405       (Control.overloadKW must be true for either version to be accepted.)
1406    
1407    2. Fixed bug in install script that caused more things to be installed
1408       than what was requested in config/targets.
1409    
1410    3. Made CM aware of the (_)overload construct so that autoloading
1411       works.
1412    
1413    ----------------------------------------------------------------------
1414    Name: Matthias Blume
1415    Date: 2002/03/12 22:03:00 EST
1416    Tag: blume-20020312-url
1417    Description:
1418    
1419    Forgot to update BOOT and srcarchiveurl.
1420    
1421    ----------------------------------------------------------------------
1422    Name: Matthias Blume
1423    Date: 2002/03/12 17:30:00 EST
1424    Tag: blume-20020312-version110392
1425    Description:
1426    
1427    Yet another version number bump (because of small changes to the
1428    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
1429    
1430    Changes:
1431    
1432      The new pid generation scheme described a few weeks ago was overly
1433      complicated.  I implemented a new mechanism that is simpler and
1434      provides a bit more "stability":  Once CM has seen a compilation
1435      unit, it keeps its identity constant (as long as you do not delete
1436      those crucial CM/GUID/* files).  This means that when you change
1437      an interface, compile, then go back to the old interface, and
1438      compile again, you arrive at the original pid.
1439    
1440      There now also is a mechanism that instructs CM to use the plain
1441      environment hash as a module's pid (effectively making its GUID
1442      the empty string).  For this, "noguid" must be specified as an
1443      option to the .sml file in question within its .cm file.
1444      This is most useful for code that is being generated by tools such
1445      as ml-nlffigen (because during development programmers tend to
1446      erase the tool's entire output directory tree including CM's cached
1447      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
1448      revert to the old, broken behavior of SML/NJ, but in specific cases
1449      where there is no danger of interface confusion, its use is ok
1450      (I think).
1451    
1452      ml-nlffigen by default generates "noguid" annotations.  They can be
1453      turned off by specifying -guid in its command line.
1454    
1455    ----------------------------------------------------------------------
1456    Name: Lal George
1457    Date: 2002/03/12 12 14:42:36 EST
1458    Tag: george-20020312-frequency-computation
1459    Description:
1460    
1461    Integrated jump chaining and static block frequency into the
1462    compiler. More details and numbers later.
1463    
1464    ----------------------------------------------------------------------
1465    Name: Lal George
1466    Date: 2002/03/11 11 22:38:53 EST
1467    Tag: george-20020311-jump-chain-elim
1468    Description:
1469    
1470    Tested the jump chain elimination on all architectures (except the
1471    hppa).  This is on by default right now and is profitable for the
1472    alpha and x86, however, it may not be profitable for the sparc and ppc
1473    when compiling the compiler.
1474    
1475    The gc test will typically jump to a label at the end of the cluster,
1476    where there is another jump to an external cluster containing the actual
1477    code to invoke gc. This is to allow factoring of common gc invocation
1478    sequences. That is to say, we generate:
1479    
1480            f:
1481               testgc
1482               ja   L1      % jump if above to L1
1483    
1484            L1:
1485               jmp L2
1486    
1487    
1488    After jump chain elimination the 'ja L1' instructions is converted to
1489    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
1490    up being implemented in their long form (if L2 is far away) using:
1491    
1492            jbe     L3      % jump if below or equal to L3
1493            jmp     L2
1494         L3:
1495            ...
1496    
1497    
1498    For large compilation units L2  may be far away.
1499    
1500    
1501    ----------------------------------------------------------------------
1502    Name: Matthias Blume
1503    Date: 2002/03/11 13:30:00 EST
1504    Tag: blume-20020311-mltreeeval
1505    Description:
1506    
1507    A functor parameter was missing.
1508    
1509    ----------------------------------------------------------------------
1510    Name: Allen Leung
1511    Date: 2002/03/11 10:30:00 EST
1512    Tag: leunga-20020311-runtime-string0
1513    Description:
1514    
1515       The representation of the empty string now points to a
1516    legal null terminated C string instead of unit.  It is now possible
1517    to convert an ML string into C string with InlineT.CharVector.getData.
1518    This compiles into one single machine instruction.
1519    
1520    ----------------------------------------------------------------------
1521    Name: Allen Leung
1522    Date: 2002/03/10 23:55:00 EST
1523    Tag: leunga-20020310-x86-call
1524    Description:
1525    
1526       Added machine generation for CALL instruction (relative displacement mode)
1527    
1528    ----------------------------------------------------------------------
1529    Name: Matthias Blume
1530    Date: 2002/03/08 16:05:00
1531    Tag: blume-20020308-entrypoints
1532    Description:
1533    
1534    Version number bumped to 110.39.1.  NEW BOOTFILES!
1535    
1536    Entrypoints: non-zero offset into a code object where execution should begin.
1537    
1538    - Added the notion of an entrypoint to CodeObj.
1539    - Added reading/writing of entrypoint info to Binfile.
1540    - Made runtime system bootloader aware of entrypoints.
1541    - Use the address of the label of the first function given to mlriscGen
1542      as the entrypoint.  This address is currently always 0, but it will
1543      not be 0 once we turn on block placement.
1544    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
1545      with entry points) from mlriscGen.
1546    
1547    ----------------------------------------------------------------------
1548    Name: Allen Leung
1549    Date: 2002/03/07 20:45:00 EST
1550    Tag: leunga-20020307-x86-cmov
1551    Description:
1552    
1553       Bug fixes for CMOVcc on x86.
1554    
1555       1. Added machine code generation for CMOVcc
1556       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
1557       3. CMOVcc cannot have an immediate operand as argument.
1558    
1559    ----------------------------------------------------------------------
1560    Name: Matthias Blume
1561    Date: 2002/03/07 16:15:00 EST
1562    Tag: blume-20020307-controls
1563    Description:
1564    
1565    This is a very large but mostly boring patch which makes (almost)
1566    every tuneable compiler knob (i.e., pretty much everything under
1567    Control.* plus a few other things) configurable via both the command
1568    line and environment variables in the style CM did its configuration
1569    until now.
1570    
1571    Try starting sml with '-h' (or, if you are brave, '-H')
1572    
1573    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
1574    implements the underlying generic mechanism.
1575    
1576    The interface to some of the existing such facilities has changed somewhat.
1577    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
1578    (The getFoo interface is still there for backward-compatibility, but its
1579    use is deprecated.)
1580    
1581    The ml-build script passes -Cxxx=yyy command-line arguments through so
1582    that one can now twiddle the compiler settings when using this "batch"
1583    compiler.
1584    
1585    TODO items:
1586    
1587    We should go through and throw out all controls that are no longer
1588    connected to anything.  Moreover, we should go through and provide
1589    meaningful (and correct!) documentation strings for those controls
1590    that still are connected.
1591    
1592    Currently, multiple calls to Controls.new are accepted (only the first
1593    has any effect).  Eventually we should make sure that every control
1594    is being made (via Controls.new) exactly once.  Future access can then
1595    be done using Controls.acc.
1596    
1597    Finally, it would probably be a good idea to use the getter-setter
1598    interface to controls rather than ref cells.  For the time being, both
1599    styles are provided by the Controls module, but getter-setter pairs are
1600    better if thread-safety is of any concern because they can be wrapped.
1601    
1602    *****************************************
1603    
1604    One bug fix: The function blockPlacement in three of the MLRISC
1605    backpatch files used to be hard-wired to one of two possibilities at
1606    link time (according to the value of the placementFlag).  But (I
1607    think) it should rather sense the flag every time.
1608    
1609    *****************************************
1610    
1611    Other assorted changes (by other people who did not supply a HISTORY entry):
1612    
1613    1. the cross-module inliner now works much better (Monnier)
1614    2. representation of weights, frequencies, and probabilities in MLRISC
1615       changed in preparation of using those for weighted block placement
1616       (Reppy, George)
1617    
1618    ----------------------------------------------------------------------
1619    Name: Lal George
1620    Date: 2002/03/07 14:44:24 EST 2002
1621    Tag: george-20020307-weighted-block-placement
1622    
1623    Tested the weighted block placement optimization on all architectures
1624    (except the hppa) using AMPL to generate the block and edge frequencies.
1625    Changes were required in the machine properties to correctly
1626    categorize trap instructions. There is an MLRISC flag
1627    "weighted-block-placement" that can be used to enable weighted block
1628    placement, but this will be ineffective without block/edge
1629    frequencies (coming soon).
1630    
1631    
1632    ----------------------------------------------------------------------
1633    Name: Lal George
1634    Date: 2002/03/05 17:24:48 EST
1635    Tag: george-20020305-linkage-cluster
1636    
1637    In order to support the block placement optimization, a new cluster
1638    is generated as the very first cluster (called the linkage cluster).
1639    It contains a single jump to the 'real' entry point for the compilation
1640    unit. Block placement has no effect on the linkage cluster itself, but
1641    all the other clusters  have full freedom in the manner in which they
1642    reorder blocks or functions.
1643    
1644    On the x86 the typical linkage code that is generated is:
1645       ----------------------
1646            .align 2
1647       L0:
1648            addl    $L1-L0, 72(%esp)
1649            jmp     L1
1650    
1651    
1652            .align  2
1653       L1:
1654       ----------------------
1655    
1656    72(%esp) is the memory location for the stdlink register. This
1657    must contain the address of the CPS function being called. In the
1658    above example, it contains the address of  L0; before
1659    calling L1 (the real entry point for the compilation unit), it
1660    must contain the address for L1, and hence
1661    
1662            addl $L1-L0, 72(%esp)
1663    
1664    I have tested this on all architectures except the hppa.The increase
1665    in code size is of course negligible
1666    
1667    ----------------------------------------------------------------------
1668    Name: Allen Leung
1669    Date: 2002/03/03 13:20:00 EST
1670    Tag: leunga-20020303-mlrisc-tools
1671    
1672      Added #[ ... ] expressions to mlrisc tools
1673    
1674    ----------------------------------------------------------------------
1675    Name: Matthias Blume
1676    Date: 2002/02/27 12:29:00 EST
1677    Tag: blume-20020227-cdebug
1678    Description:
1679    
1680    - made types in structure C and C_Debug to be equal
1681    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
1682    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
1683    
1684    ----------------------------------------------------------------------
1685    Name: Matthias Blume
1686    Date: 2002/02/26 12:00:00 EST
1687    Tag: blume-20020226-ffi
1688    Description:
1689    
1690    1. Fixed a minor bug in CM's "noweb" tool:
1691       If numbering is turned off, then truly don't number (i.e., do not
1692       supply the -L option to noweb).  The previous behavior was to supply
1693       -L'' -- which caused noweb to use the "default" line numbering scheme.
1694       Thanks to Chris Richards for pointing this out (and supplying the fix).
1695    
1696    2. Once again, I reworked some aspects of the FFI:
1697    
1698       A. The incomplete/complete type business:
1699    
1700       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
1701         gone!
1702       - ML types representing an incomplete type are now *equal* to
1703         ML types representing their corresponding complete types (just like
1704         in C).  This is still safe because ml-nlffigen will not generate
1705         RTTI for incomplete types, nor will it generate functions that
1706         require access to such RTTI.   But when ML code generated from both
1707         incomplete and complete versions of the C type meet, the ML types
1708         are trivially interoperable.
1709    
1710         NOTE:  These changes restore the full generality of the translation
1711         (which was previously lost when I eliminated functorization)!
1712    
1713       B. Enum types:
1714    
1715       - Structure C now has a type constructor "enum" that is similar to
1716         how the "su" constructor works.  However, "enum" is not a phantom
1717         type because each "T enum" has values (and is isomorphic to
1718         MLRep.Signed.int).
1719       - There are generic access operations for enum objects (using
1720         MLRep.Signed.int).
1721       - ml-nlffigen will generate a structure E_foo for each "enum foo".
1722         * The structure contains the definition of type "mlrep" (the ML-side
1723         representation type of the enum).  Normally, mlrep is the same
1724         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
1725         then mlrep will be defined as a datatype -- thus facilitating
1726         pattern matching on mlrep values.
1727         ("-ec" will be suppressed if there are duplicate values in an
1728          enumeration.)
1729         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
1730         will be generated for each C enum constant xxx.
1731         * Conversion functions m2i and i2m convert between mlrep and
1732         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
1733         * Coversion functions c and ml convert between mlrep and "tag enum".
1734         * Access functions (get/set) fetch and store mlrep values.
1735       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
1736         enumerations are merged into one single enumeration represented by
1737         structure E_'.
1738    
1739    ----------------------------------------------------------------------
1740    Name: Allen Leung
1741    Date: 2002/02/25 04:45:00 EST
1742    Tag: leunga-20020225-cps-spill
1743    
1744    This is a new implementation of the CPS spill phase.
1745    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
1746    In case of problems, replace it with the old file spill.sml
1747    
1748    The current compiler runs into some serious performance problems when
1749    constructing a large record.  This can happen when we try to compile a
1750    structure with many items.  Even a very simple structure like the following
1751    makes the compiler slow down.
1752    
1753        structure Foo = struct
1754           val x_1 = 0w1 : Word32.int
1755           val x_2 = 0w2 : Word32.int
1756           val x_3 = 0w3 : Word32.int
1757           ...
1758           val x_N = 0wN : Word32.int
1759        end
1760    
1761    The following table shows the compile time, from N=1000 to N=4000,
1762    with the old compiler:
1763    
1764    N
1765    1000   CPS 100 spill                           0.04u  0.00s  0.00g
1766           MLRISC ra                               0.06u  0.00s  0.05g
1767              (spills = 0 reloads = 0)
1768           TOTAL                                   0.63u  0.07s  0.21g
1769    
1770    1100   CPS 100 spill                           8.25u  0.32s  0.64g
1771           MLRISC ra                               5.68u  0.59s  3.93g
1772              (spills = 0 reloads = 0)
1773           TOTAL                                   14.71u  0.99s  4.81g
1774    
1775    1500   CPS 100 spill                           58.55u  2.34s  1.74g
1776           MLRISC ra                               5.54u  0.65s  3.91g
1777              (spills = 543 reloads = 1082)
1778           TOTAL                                   65.40u  3.13s  6.00g
1779    
1780    2000   CPS 100 spill                           126.69u  4.84s  3.08g
1781           MLRISC ra                               0.80u  0.10s  0.55g
1782              (spills = 42 reloads = 84)
1783           TOTAL                                   129.42u  5.10s  4.13g
1784    
1785    3000   CPS 100 spill                           675.59u  19.03s  11.64g
1786           MLRISC ra                               2.69u  0.27s  1.38g
1787              (spills = 62 reloads = 124)
1788           TOTAL                                   682.48u  19.61s  13.99g
1789    
1790    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
1791           MLRISC ra                               4.96u  0.27s  2.72g
1792              (spills = 85 reloads = 170)
1793           TOTAL                                   2375.26u  57.21s  48.00g
1794    
1795    As you can see the old cps spill module suffers from some serious
1796    performance problem.  But since I cannot decipher the old code fully,
1797    instead of patching the problems up, I'm reimplementing it
1798    with a different algorithm.  The new code is more modular,
1799    smaller when compiled, and substantially faster
1800    (O(n log n) time and O(n) space).  Timing of the new spill module:
1801    
1802    4000  CPS 100 spill                           0.02u  0.00s  0.00g
1803          MLRISC ra                               0.25u  0.02s  0.15g
1804             (spills=1 reloads=3)
1805          TOTAL                                   7.74u  0.34s  1.62g
1806    
1807    Implementation details:
1808    
1809    As far as I can tell, the purpose of the CPS spill module is to make sure the
1810    number of live variables at any program point (the bandwidth)
1811    does not exceed a certain limit, which is determined by the
1812    size of the spill area.
1813    
1814    When the bandwidth is too large, we decrease the register pressure by
1815    packing live variables into spill records.  How we achieve this is
1816    completely different than what we did in the old code.
1817    
1818    First, there is something about the MLRiscGen code generator
1819    that we should be aware of:
1820    
1821    o MLRiscGen performs code motion!
1822    
1823       In particular, it will move floating point computations and
1824       address computations involving only the heap pointer to
1825       their use sites (if there is only a single use).
1826       What this means is that if we have a CPS record construction
1827       statement
1828    
1829           RECORD(k,vl,w,e)
1830    
1831       we should never count the new record address w as live if w
1832       has only one use (which is often the case).
1833    
1834       We should do something similar to floating point, but the transformation
1835       there is much more complex, so I won't deal with that.
1836    
1837    Secondly, there are now two new cps primops at our disposal:
1838    
1839     1. rawrecord of record_kind option
1840        This pure operator allocates some uninitialized storage from the heap.
1841        There are two forms:
1842    
1843         rawrecord NONE [INT n]  allocates a tagless record of length n
1844         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
1845                                     and initializes the tag.
1846    
1847     2. rawupdate of cty
1848          rawupdate cty (v,i,x)
1849          Assigns to x to the ith component of record v.
1850          The storelist is not updated.
1851    
1852    We use these new primops for both spilling and increment record construction.
1853    
1854     1. Spilling.
1855    
1856        This is implemented with a linear scan algorithm (but generalized
1857        to trees).  The algorithm will create a single spill record at the
1858        beginning of the cps function and use rawupdate to spill to it,
1859        and SELECT or SELp to reload from it.  So both spills and reloads
1860        are fine-grain operations.  In contrast, in the old algorithm
1861        "spills" have to be bundled together in records.
1862    
1863        Ideally, we should sink the spill record construction to where
1864        it is needed.  We can even split the spill record into multiple ones
1865        at the places where they are needed.  But CPS is not a good
1866        representation for global code motion, so I'll keep it simple and
1867        am not attempting this.
1868    
1869     2. Incremental record construction (aka record splitting).
1870    
1871        Long records with many component values which are simulatenously live
1872        (recall that single use record addresses are not considered to
1873         be live) are constructed with rawrecord and rawupdate.
1874        We allocate space on the heap with rawrecord first, then gradually
1875        fill it in with rawupdate.  This is the technique suggested to me
1876        by Matthias.
1877    
1878        Some restrictions on when this is applicable:
1879        1. It is not a VECTOR record.  The code generator currently does not handle
1880           this case. VECTOR record uses double indirection like arrays.
1881        2. All the record component values are defined in the same "basic block"
1882           as the record constructor.  This is to prevent speculative
1883           record construction.
1884    
1885    ----------------------------------------------------------------------
1886    Name: Allen Leung
1887    Date: 2002/02/22 01:02:00 EST
1888    Tag: leunga-20020222-mlrisc-tools
1889    
1890    Minor bug fixes in the parser and rewriter
1891    
1892    ----------------------------------------------------------------------
1893    Name: Allen Leung
1894    Date: 2002/02/21 20:20:00 EST
1895    Tag: leunga-20020221-peephole
1896    
1897    Regenerated the peephole files.  Some contained typos in the specification
1898    and some didn't compile because of pretty printing bugs in the old version
1899    of 'nowhere'.
1900    
1901    ----------------------------------------------------------------------
1902    Name: Allen Leung
1903    Date: 2002/02/19 20:20:00 EST
1904    Tag: leunga-20020219-mlrisc-tools
1905    Description:
1906    
1907       Minor bug fixes to the mlrisc-tools library:
1908    
1909       1.  Fixed up parsing colon suffixed keywords
1910       2.  Added the ability to shut the error messages up
1911       3.  Reimplemented the pretty printer and fixed up/improved
1912           the pretty printing of handle and -> types.
1913       4.  Fixed up generation of literal symbols in the nowhere tool.
1914       5.  Added some SML keywords to to sml.sty
1915    
1916    ----------------------------------------------------------------------
1917    Name: Matthias Blume
1918    Date: 2002/02/19 16:20:00 EST
1919    Tag: blume-20020219-cmffi
1920    Description:
1921    
1922    A wild mix of changes, some minor, some major:
1923    
1924    * All C FFI-related libraries are now anchored under $c:
1925        $/c.cm      --> $c/c.cm
1926        $/c-int.cm  --> $c/internals/c-int.cm
1927        $/memory.cm --> $c/memory/memory.cm
1928    
1929    * "make" tool (in CM) now treats its argument pathname slightly
1930      differently:
1931        1. If the native expansion is an absolute name, then before invoking
1932           the "make" command on it, CM will apply OS.Path.mkRelative
1933           (with relativeTo = OS.FileSys.getDir()) to it.
1934        2. The argument will be passed through to subsequent phases of CM
1935           processing without "going native".  In particular, if the argument
1936           was an anchored path, then "make" will not lose track of that anchor.
1937    
1938    * Compiler backends now "know" their respective C calling conventions
1939      instead of having to be told about it by ml-nlffigen.  This relieves
1940      ml-nlffigen from one of its burdens.
1941    
1942    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
1943    
1944    * Export C_DEBUG and C_Debug from $c/c.cm.
1945    
1946    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
1947      subtyping relationship between incomplete pointers and their complete
1948      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
1949      with the convention of instantiating 'o with ('t, 'c) obj whenever
1950      the pointer target type is complete.  In the incomplete case, 'o
1951      will be instantiated with some "'c iobj" -- a type obtained by
1952      using one of the functors PointerToIncompleteType or PointerToCompleteType.
1953    
1954      Operations that work on both incomplete and complete pointer types are
1955      typed as taking an 'o ptr while operations that require the target to
1956      be known are typed as taking some ('t, 'c) obj ptr.
1957    
1958      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
1959      where void is an eqtype without any values.  This makes it possible
1960      to work on voidptr values using functions meant to operate on light
1961      incomplete pointers.
1962    
1963    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
1964      been vastly simplified.
1965    
1966    ----------------------------------------------------------------------
1967    Name: Matthias Blume
1968    Date: 2002/02/19 10:48:00 EST
1969    Tag: blume-20020219-pqfix
1970    Description:
1971    
1972    Applied Chris Okasaki's bug fix for priority queues.
1973    
1974    ----------------------------------------------------------------------
1975    Name: Matthias Blume
1976    Date: 2002/02/15 17:05:00
1977    Tag: Release_110_39
1978    Description:
1979    
1980    Last-minute retagging is becoming a tradition... :-(
1981    
1982    This is the working release 110.39.
1983    
1984    ----------------------------------------------------------------------
1985    Name: Matthias Blume
1986    Date: 2002/02/15 16:00:00 EST
1987    Tag: Release_110_39-orig
1988    Description:
1989    
1990    Working release 110.39.  New bootfiles.
1991    
1992    (Update: There was a small bug in the installer so it wouldn't work
1993    with all shells.  So I retagged. -Matthias)
1994    
1995    ----------------------------------------------------------------------
1996    Name: Matthias Blume
1997    Date: 2002/02/15 14:17:00 EST
1998    Tag: blume-20020215-showbindings
1999    Description:
2000    
2001    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
2002    the latter can be useful for exploring what bindings are available at
2003    the interactive prompt.  (The first function returns only the list
2004    of symbols that are really bound, the second prints those but also the
2005    ones that CM's autoloading mechanism knows about.)
2006    
2007    ----------------------------------------------------------------------
2008    Name: Matthias Blume
2009    Date: 2002/02/15 12:08:00 EST
2010    Tag: blume-20020215-iptrs
2011    Description:
2012    
2013    Two improvements to ml-nlffigen:
2014    
2015      1. Write files only if they do not exist or if their current contents
2016         do not coincide with what's being written.  (That is, avoid messing
2017         with the time stamps unless absolutely necessary.)
2018    
2019      2. Implement a "repository" mechanism for generated files related
2020         to "incomplete pointer types".   See the README file for details.
2021    
2022    ----------------------------------------------------------------------
2023    Name: Matthias Blume
2024    Date: 2002/02/14 11:50:00 EST
2025    Tag: blume-20020214-quote
2026    Description:
2027    
2028    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
2029    because of the new and improved tag generation scheme.  (Thanks to Allen
2030    Leung for pointing it out.)
2031    
2032    ----------------------------------------------------------------------
2033    Name: Lal George
2034    Date: 2002/02/14 09:55:27 EST 2002
2035    Tag: george-20020214-isabelle-bug
2036    Description:
2037    
2038    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
2039    of Isabelle on the x86.
2040    
2041    From Allen:
2042    -----------
2043     I've found the problem:
2044    
2045         in ra-core.sml, I use the counter "blocked" to keep track of the
2046         true number of elements in the freeze queue.  When the counter goes
2047         to zero, I skip examining the queue.  But I've messed up the
2048         bookkeeping in combine():
2049    
2050             else ();
2051             case !ucol of
2052               PSEUDO => (if !cntv > 0 then
2053                     (if !cntu > 0 then blocked := !blocked - 1 else ();
2054                                        ^^^^^^^^^^^^^^^^^^^^^^^
2055                      moveu := mergeMoveList(!movev, !moveu)
2056                     )
2057                  else ();
2058    
2059         combine() is called to coalesce two nodes u and v.
2060         I think I was thinking that if the move counts of u and v are both
2061         greater than zero then after they are coalesced then one node is
2062         removed from the freeze queue.  Apparently I was thinking that
2063         both u and v are of low degree, but that's clearly not necessarily true.
2064    
2065    
2066    02/12/2002:
2067        Here's the patch.  HOL now compiles.
2068    
2069        I don't know how this impact on performance (compile
2070        time or runtime).  This bug caused the RA (especially on the x86)
2071        to go thru the potential spill phase when there are still nodes on the
2072        freeze queue.
2073    
2074    
2075    
2076    
2077    ----------------------------------------------------------------------
2078    Name: Matthias Blume
2079    Date: 2002/02/13 22:40:00 EST
2080    Tag: blume-20020213-fptr-rtti
2081    Description:
2082    
2083    Fixed a bug in ml-nlffigen that was introduced with one of the previous
2084    updates.
2085    
2086    ----------------------------------------------------------------------
2087    Name: Matthias Blume
2088    Date: 2002/02/13 16:41:00 EST
2089    Tag: blume-20020213-cmlpq
2090    Description:
2091    
2092    Added new priority queue export symbols (which have just been added to
2093    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
2094    not compile and the installer would choke.)
2095    
2096    ----------------------------------------------------------------------
2097    Name: Matthias Blume
2098    Date: 2002/02/13 16:15:00 EST
2099    Tag: blume-20020213-various
2100    Description:
2101    
2102    1. More tweaks to ml-nlffigen:
2103    
2104       - better internal datastructures (resulting in slight speedup)
2105       - "-match" option requires exact match
2106       - "localized" gensym counters (untagged structs/unions nested within
2107         other structs/unions or within typedefs get a fresh counter; their
2108         tag will be prefixed by a concatenation of their parents' tags)
2109       - bug fixes (related to calculation of transitive closure of types
2110         to be included in the output)
2111    
2112    2. Minor Basis updates:
2113    
2114       - added implementations for List.collate and Option.app
2115    
2116    ----------------------------------------------------------------------
2117    Name: Matthias Blume
2118    Date: 2002/02/11 15:55:00 EST
2119    Tag: blume-20020211-gensym
2120    Description:
2121    
2122    Added a "-gensym" option to command line of ml-nlffigen.  This can be
2123    used to specify a "stem" -- a string that is inserted in all "gensym'd"
2124    names (ML structure names that correspond to unnamed C structs, unions,
2125    and enums), so that separate runs of ml-nlffigen do not clash.
2126    
2127    ----------------------------------------------------------------------
2128    Name: Matthias Blume
2129    Date: 2002/02/11 12:05:00 EST
2130    Tag: blume-20020211-gensml
2131    Description:
2132    
2133    A quick fix for a problem with GenSML (in the pgraph-util library):
2134    Make generation of toplevel "local" optional.  (Strictly speaking,
2135    signature definitions within "local" are not legal SML.)
2136    
2137    Other than that: updates to INSTALL and cm/TODO.
2138    
2139    ----------------------------------------------------------------------
2140    Name: Matthias Blume
2141    Date: 2002/02/08 15:00:00 EST
2142    Tag: blume-20020208-uniquepid
2143    Description:
2144    
2145    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
2146    
2147    1. The installer (config/install.sh) has gotten smarter:
2148    
2149         - Configuration options are a bit easier to specify now
2150           (in config/targets).
2151         - Bug in recognizing .tar.bz2 files fixed.
2152         - Installer automatically resolves dependencies between
2153           configuration options (e.g., if you ask for eXene, you will
2154           also get cml -- regardless whether you asked for it or not).
2155         - Installer can run in "quieter mode" by setting the environment
2156           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
2157           "completely silent", though.
2158         - Build HashCons library as part of smlnj-lib.
2159    
2160    2. A new scheme for assigning persistent identifiers to compilation
2161       units (and, by extension, to types etc.) has been put into place.
2162       This fixes a long-standing bug where types and even dynamic values
2163       can get internally confused, thereby compromising type safety
2164       (abstraction) and dynamic correctness.  See
2165    
2166         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
2167    
2168       for an example of how things could go wrong until now.
2169    
2170       The downside of the new scheme is that pids are not quite as
2171       persistent as they used to be: CM will generate a fresh pid
2172       for every compilation unit that it thinks it sees for the first
2173       time.  That means that if you compile starting from a clean, fresh
2174       source tree at two different times, you end up with different
2175       binaries.
2176    
2177       Cutoff recompilation, however, has not been compromised because
2178       CM keeps pid information in special caches between runs.
2179    
2180    ----------------------------------------------------------------------
2181    Name: Lal George
2182    Date: 2002/02/07 15:34:13 EST 2002
2183    Tag: <none>
2184    Description:
2185    
2186    Compilers that generate assembly code may produce  global labels
2187    whose value is resolved at link time. The various peephole optimization
2188    modules did not take this in account.
2189    
2190    TODO. The Labels.addrOf function should really return an option
2191    type so that clients are forced to deal with this issue, rather
2192    than an exception being raised.
2193    
2194    ----------------------------------------------------------------------
2195    Name: Lal George
2196    Date: 2002/02/06 13:55:02 EST
2197    Tag: george-20020206-ra-breakup
2198    Description:
2199    
2200    1. A bug fix from Allen.
2201    
2202        A typo causes extra fstp %st(0)'s to be generated at compensation
2203        edges, which might cause stack underflow traps at runtime.  This
2204        occurs in fft where there are extraneous fstps right before the 'into'
2205        trap instruction (in this case they are harmless since none of the
2206        integers overflow.)
2207    
2208    2. Pulled out various utility modules that were embedded in the modules
2209       of the register allocator. I need these modules for other purposes, but
2210       they are not complete enough to put into a library (just yet).
2211    ----------------------------------------------------------------------
2212    Name: Matthias Blume
2213    Date: 2002/01/31 16:05:00 EST
2214    Tag: blume-20020131-sparc-ccalls
2215    Description:
2216    
2217    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
2218       of extra stack space by mistake.  Fixed.
2219    
2220    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
2221    
2222    ----------------------------------------------------------------------
2223    Name: Allen Leung
2224    Date: 2002/01/30
2225    Tag: leunga-20020130-nowhere-bug-fix
2226    Description:
2227    
2228       MLRISC bug fixes:
2229       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
2230       2. MachineInt.fromString was negating its value.
2231    
2232    ----------------------------------------------------------------------
2233    Name: Matthias Blume
2234    Date: 2002/01/29
2235    Tag: blume-20020129-INSTALL
2236    Description:
2237    
2238    - Added somewhat detailed installation instructions (file INSTALL).
2239    - Fixed curl-detection bug in config/install.sh.
2240    - It is now possible to select the URL getter using the URLGETTER
2241      environment variable:
2242    
2243          not set / "unknown"      --> automatic detection (script tries wget,
2244                                       curl, and lynx)
2245          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
2246                                       how to properly invoke them)
2247          other                    --> use $URLGETTER directly, it must take
2248                                       precisely two command-line arguments
2249                                       (source URL and destination file name)
2250    
2251    ----------------------------------------------------------------------
2252    Name: Matthias Blume
2253    Date: 2002/01/28
2254    Tag: blume-20020128-sparc-ccalls
2255    Description:
2256    
2257    - Fixed problem with calculation of "used" registers in sparc-c-calls.
2258    - Make use of the allocParam argument in sparc-c-calls.
2259    
2260    ----------------------------------------------------------------------
2261    Name: Matthias Blume
2262    Date: 2002/01/28
2263    Tag: blume-20020128-allocParam
2264    Description:
2265    
2266    John Reppy:  Changes c-calls API to accept client-callback for
2267    allocating extra stack space.
2268    me: Corresponding changes to mlriscGen (using a dummy argument that
2269        does not change the current behavior).
2270    
2271    ----------------------------------------------------------------------
2272    Name: Matthias Blume
2273    Date: 2002/01/28 12:00:00
2274    Tag: Release_110_38
2275    Description:
2276    
2277    This time for real!!!
2278    
2279    ----------------------------------------------------------------------
2280    Name: Matthias Blume
2281    Date: 2002/01/28 10:56:00 EST
2282    Tag: blume-20020128-retraction
2283    Description:
2284    
2285    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
2286       with blume-Release_110_38-retracted.)
2287    
2288    1. Fixed a problem with incorrect rounding modes in real64.sml.
2289       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
2290    
2291    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
2292       and enums fixed.  The naming of corresponding ML identifiers should
2293       now be consistent again.
2294    
2295    ----------------------------------------------------------------------
2296    Name: Allen Leung
2297    Date: 2002/01/27
2298    Tag: leunga-20020127-nowhere
2299    Description:
2300    
2301       Added a target called nowhere in the configuration scripts.
2302       Enabling this will build the MLRISC 'nowhere' tool (for translating
2303       programs with where-clauses into legal SML code) during installation.
2304    
2305    ----------------------------------------------------------------------
2306    Name: Matthias Blume
2307    Date: 2002/01/25 21:27:00 EST
2308    Tag: blume-Release_110_38-retracted
2309    Description:
2310    
2311    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
2312    
2313    README will be added later.
2314    
2315    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
2316    (Release_110_38) removed.  Reason: Last-minute bug fixes.
2317    
2318    ----------------------------------------------------------------------
2319    Name: Matthias Blume
2320    Date: 2002/01/25
2321    Tag: blume-20020125-ffi
2322    Description:
2323    
2324    A large number of tweaks and improvements to ml-nlffi-lib and
2325    ml-nlffigen:
2326    
2327       - ML represenation types have been streamlined
2328       - getter and setter functions work with concrete values, not abstract
2329         ones where possible
2330       - ml-nlffigen command line more flexible (see README file there)
2331       - some bugs have been fixed (hopefully)
2332    
2333    ----------------------------------------------------------------------
2334    Name: Lal George
2335    Date: 2002/01/24
2336    Tag: george-20020124-risc-ra-interface
2337    Description:
2338    
2339       There is a dramatic simplification in the interface to the
2340       register allocator for RISC architectures as a result of making
2341       parallel copy instructions explicit.
2342    
2343    ----------------------------------------------------------------------
2344    Name: Matthias Blume
2345    Date: 2002/01/22
2346    Tag: blume-20020122-x86-ccalls
2347    Description:
2348    
2349    Bug fix for c-calls on x86 (having to do with how char- and
2350    short-arguments are being handled).
2351    
2352    ----------------------------------------------------------------------
2353    Name: Matthias Blume
2354    Date: 2002/01/21
2355    Tag: blume-20020121-ff
2356    Description:
2357    
2358    Another day of fiddling with the FFI...
2359    
2360    1. Bug fix/workaround:  CKIT does not complain about negative array
2361       dimensions, so ml-nlffigen has to guard itself against this possibility.
2362       (Otherwise a negative dimension would send it into an infinite loop.)
2363    
2364    2. Some of the abstract types (light objects, light pointers, most "base"
2365       types) in structure C are now eqtypes.
2366    
2367    3. Added constructors and test functions for NULL function pointers.
2368    
2369    ----------------------------------------------------------------------
2370    Name: Matthias Blume
2371    Date: 2002/01/18
2372    Tag: blume-20020118-ready-for-new-release
2373    Description:
2374    
2375    Made config/srcarchiveurl point to a new place.  (Will provide boot
2376    files shortly.)
2377    
2378    Maybe we christen this to be 110.38?
2379    
2380    ----------------------------------------------------------------------
2381    Name: Matthias Blume
2382    Date: 2002/01/18
2383    Tag: blume-20020118-more-ffifiddle
2384    Description:
2385    
2386    Today's FFI fiddling:
2387    
2388      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
2389        C.Set.* that use concrete (MLRep.*) arguments and results instead
2390        of abstract ones.
2391    
2392      - Provided word-style bit operations etc. for "int" representation
2393        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
2394        and Long.
2395    
2396    ----------------------------------------------------------------------
2397    Name: Matthias Blume
2398    Date: 2002/01/18
2399    Tag: blume-20020118-use-x86-fp
2400    Description:
2401    
2402    Now that x86-fast-fp seems to be working, I turned it back on again
2403    by default.  (Seems to work fine now, even with the FFI.)
2404    
2405    Other than that, I added some documentation about the FFI to
2406    src/ml-nlffigen/README and updated the FFI test examples in
2407    src/ml-nlffi-lib/Tests/*.
2408    
2409    ----------------------------------------------------------------------
2410    Name: Allen Leung
2411    Date: 2002/01/17
2412    Tag: leunga-20020117-x86-fast-fp-call
2413    Description:
2414    
2415       1. Fixed a problem with handling return fp values when x86's fast fp
2416          mode is turned on.
2417    
2418       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
2419          of %f32.
2420    
2421       3. Added a constructor INT32lit to the ast of MLRISC tools.
2422    
2423    ----------------------------------------------------------------------
2424    Name: Matthias Blume
2425    Date: 2002/01/16
2426    Tag: blume-20020116-ffifiddle
2427    Description:
2428    
2429    More fiddling with the FFI interface:
2430    
2431     - Make constness 'c instead of rw wherever possible.  This eliminates
2432       the need for certain explicit coercions.  (However, due to ML's
2433       value polymorphism, there will still be many cases where explicit
2434       coercions are necessary.  Phantom types are not the whole answer
2435       to modeling a subtyping relationship in ML.)
2436    
2437     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
2438    
2439     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
2440       clumsy workaround for varargs problem.)
2441    
2442    ----------------------------------------------------------------------
2443    Name: Lal George
2444    Date: 2002/01/15
2445    Tag: <none>
2446    Description:
2447    
2448    1. Since COPY instructions are no longer native to the architecture,
2449       a generic functor can be used to implement the expandCopies function.
2450    
2451    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
2452       TEXT segment.
2453    
2454    ----------------------------------------------------------------------
2455    Name: Matthias Blume
2456    Date: 2002/01/15
2457    Tag: blume-20020115-ffiupdates
2458    Description:
2459    
2460    1. Fix for bug resulting in single-precision float values being returned
2461       incorrectly from FFI calls.
2462    
2463    2. Small modifications to C FFI API:
2464    
2465        - memory-allocation routines return straight objects (no options)
2466          and raise an exception in out-of-memory situations
2467        - unsafe extensions to cast between function pointers and pointers
2468          from/to ints
2469        - added structure C_Debug as an alternative to structure C where
2470          pointer-dereferencing (|*| and |*!) always check for null-pointers
2471        - added open_lib' to DynLinkage;  open_lib' works like open_lib
2472          but also takes a (possibly empty) list of existing library handles
2473          that the current library depends on
2474    
2475    ----------------------------------------------------------------------
2476    Name: Matthias Blume
2477    Date: 2002/01/10
2478    Tag: blume-20020110-newffigen
2479    Description:
2480    
2481    1. Updates to portable graph code.
2482    
2483    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
2484       more scalable now so that even huge interfaces such as the one
2485       for GTK compile in finite time and space. :-)
2486       See src/ml-nlffigen/README for details on what's new.
2487    
2488    ----------------------------------------------------------------------
2489    Name: Lal George
2490    Date: 2001/01/09 14:31:35 EST 2002
2491    Tag: george-20011206-rm-native-copy
2492    Description:
2493    
2494            Removed the native COPY and FCOPY instructions
2495            from all the architectures and replaced it with the
2496            explicit COPY instruction from the previous commit.
2497    
2498            It is now possible to simplify many of the optimizations
2499            modules that manipulate copies. This has not been
2500            done in this change.
2501    
2502    ----------------------------------------------------------------------
2503    Name: Lal George
2504    Date: 2001/12/06 16:50:13 EST 2001
2505    Tag: george-20011206-mlrisc-instruction
2506    Description:
2507    
2508    Changed the representation of instructions from being fully abstract
2509    to being partially concrete. That is to say:
2510    
2511      from
2512            type instruction
2513    
2514      to
2515            type instr                              (* machine instruction *)
2516    
2517            datatype instruction =
2518                LIVE of {regs: C.cellset, spilled: C.cellset}
2519              | KILL of {regs: C.cellset, spilled: C.cellset}
2520              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
2521              | ANNOTATION of {i: instruction, a: Annotations.annotation}
2522              | INSTR of instr
2523    
2524    This makes the handling of certain special instructions that appear on
2525    all architectures easier and uniform.
2526    
2527    LIVE and KILL say that a list of registers are live or killed at the
2528    program point where they appear. No spill code is generated when an
2529    element of the 'regs' field is spilled, but the register is moved to
2530    the 'spilled' (which is present, more for debugging than anything else).
2531    
2532    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
2533    We used to generate:
2534    
2535            DEFFREG f1
2536            f1 := f2 + f3
2537            trapb
2538    
2539    but now generate:
2540    
2541            f1 := f2 + f3
2542            trapb
2543            LIVE {regs=[f1,f2,f3], spilled=[]}
2544    
2545    Furthermore, the DEFFREG (hack) required that all floating point instruction
2546    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
2547    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
2548    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
2549    this architecture).
2550    
2551    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
2552    all the architectures. This will result in further simplification of the
2553    register allocator that must be aware of them for coalescing purposes, and
2554    will also simplify certain aspects of the machine description that provides
2555    callbacks related to parallel copies.
2556    
2557    ANNOTATION should be obvious, and now INSTR represents the honest to God
2558    machine instruction set!
2559    
2560    The <arch>/instructions/<arch>Instr.sml files define certain utility
2561    functions for making porting easier -- essentially converting upper case
2562    to lower case. All machine instructions (of type instr) are in upper case,
2563    and the lower case form generates an MLRISC instruction. For example on
2564    the alpha we have:
2565    
2566      datatype instr =
2567         LDA of {r:cell, b:cell, d:operand}
2568       | ...
2569    
2570      val lda : {r:cell, b:cell, d:operand} -> instruction
2571        ...
2572    
2573    where lda is just (INSTR o LDA), etc.
2574    
2575    ----------------------------------------------------------------------
2576    Name: Matthias Blume
2577    Date: 2001/11/22 21:40:00 EST
2578    Tag: Release_110_37
2579    Description:
2580    
2581    Release 110.37.  This time for real.
2582    
2583    ----------------------------------------------------------------------
2584    Name: Matthias Blume
2585    Date: 2001/11/21 16:35:00 EST
2586    Tag: blume-20011121-foot-in-mouth
2587    Description:
2588    
2589    Removed the "Release_110_37" tag because of a serious bug.
2590    This will be re-tagged once the bug is fixed.
2591    
2592    ----------------------------------------------------------------------
2593    Name: Matthias Blume
2594    Date: 2001/11/21 16:14:00 EST
2595    Tag: blume-20011121-forgottenfile
2596    Description:
2597    
2598    Forgot to add a file.  (Just a .tex-file -- part of
2599    the CM manual source.)
2600    
2601    ----------------------------------------------------------------------
2602    Name: Matthias Blume
2603    Date: 2001/11/21 16:10:00 EST
2604    Tag: blume-20011121-invalid_110_37
2605    Description:
2606    
2607    Note: I removed the original tag "Release_110_37" from this commit
2608          because we found a serious bug in all non-x86 backends.
2609          - Matthias
2610    
2611    1. Modifications to the SML/NJ code generator and to the runtime system
2612       so that code object name strings are directly inserted into code
2613       objects at code generation time.  The only business the runtime system
2614       has with this is now to read the name strings on occasions.
2615       (The encoding of the name string has also changed somewhat.)
2616    
2617    2. CM now implements a simple "set calculus" for specifying export lists.
2618       In particular, it is now possible to refer to the export lists of
2619       other libraries/groups/sources and form unions as well as differences.
2620       See the latest CM manual for details.
2621    
2622    3. An separate notion of "proxy" libraries has again be eliminated from
2623       CM's model.  (Proxy libraries are now simply a special case of using
2624       the export list calculus.)
2625    
2626    4. Some of the existing libraries now take advantage of the new set
2627       calculus.
2628       (Notice that not all libraries have been converted because some
2629       of the existing .cm-files are supposed to be backward compatible
2630       with 110.0.x.)
2631    
2632    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
2633       "General.exnMessage"!  The former relies on a certain hook to be
2634       initialized, and that often does not happen in the stand-alone case.)
2635    
2636    ----------------------------------------------------------------------
2637    Name: Lal George
2638    Date: 2001/11/21  13:56:18 EST
2639    Tag: george-2001121-pseudo-ops
2640    Description:
2641    
2642      Implemented a complete redesign of MLRISC pseudo-ops. Now there
2643      ought to never be any question of incompatabilities with
2644      pseudo-op syntax expected by host assemblers.
2645    
2646      For now, only modules supporting GAS syntax are implemented
2647      but more should follow, such as MASM, and vendor assembler
2648      syntax, e.g. IBM as, Sun as, etc.
2649    
2650    ----------------------------------------------------------------------
2651    Name: Matthias Blume
2652    Date: 2001/11/14 11:52:00 EST
2653    Tag: blume-20011114-srcname
2654    Description:
2655    
2656    1. Routed the name of the current source file to mlriscgen where it
2657       should be directly emitted into the code object.  (This last part
2658       is yet to be done.)
2659    
2660    2. Some cleanup of the pgraph code to make it match the proposal that
2661       I put out the other day.  (The proposal notwithstanding, things are
2662       still in flux here.)
2663    
2664    ----------------------------------------------------------------------
2665    Name: Lal George
2666    Date: 2001/11/14 09:44:04 EST
2667    Tag:
2668    Description:
2669    
2670      Fix for a backpatching bug reported by Allen.
2671    
2672      Because the boundary between short and long span-dependent
2673      instructions is +/- 128, there are an astounding number of
2674      span-dependent instructions whose size is over estimated.
2675    
2676      Allen came up with the idea of letting the size of span
2677      dependent instructions be non-monotonic, for a maxIter
2678      number of times, after which the size must be monotonically
2679      increasing.
2680    
2681      This table shows the number of span-dependent instructions
2682      whose size was over-estimated as a function of maxIter, for the
2683      file Parse/parse/ml.grm.sml:
2684    
2685         maxIter            # of instructions:
2686            10                      687
2687            20                      438
2688            30                      198
2689            40                        0
2690    
2691      In compiling the compiler, there is no significant difference in
2692      compilation speed between maxIter=10 and maxIter=40. Actually,
2693      my measurements showed that maxIter=40 was a tad faster than
2694      maxIter=10! Also 96% of the  files in the compiler reach a fix
2695      point within 13 iterations, so fixing maxIter at 40, while high,
2696      is okay.
2697    
2698    ----------------------------------------------------------------------
2699    Name: Matthias Blume
2700    Date: 2001/10/31 15:25:00 EST
2701    Tag: blume-20011031-pgraph
2702    Description:
2703    
2704    CKIT:
2705    * Changed the "Function" constructor of type Ast.ctype to carry optional
2706      argument identifiers.
2707    * Changed the return type of TypeUtil.getFunction accordingly.
2708    * Type equality ignores the argument names.
2709    * TypeUtil.composite tries to preserve argument names but gives up quickly
2710      if there is a mismatch.
2711    
2712    installation script:
2713    * attempts to use "curl" if available (unless "wget" is available as well)
2714    
2715    CM:
2716    * has an experimental implementation of "portable graphs" which I will
2717      soon propose as an implementation-independent library format
2718    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
2719    
2720    NLFFI-LIB:
2721    * some cleanup (all cosmetic)
2722    
2723    NLFFIGEN:
2724    * temporarily disabled the mechanism that suppresses ML output for
2725      C definitions whose identifiers start with an underscore character
2726    * generate val bindings for enum constants
2727    * user can request that only one style (light or heavy) is being used;
2728      default is to use both (command-line arguments: -heavy and -light)
2729    * fixed bug in handling of function types involving incomplete pointers
2730    * generate ML entry points that take record arguments (i.e., using
2731      named arguments) for C functions that have a prototype with named
2732      arguments
2733      (see changes to CKIT)
2734    
2735    ----------------------------------------------------------------------
2736    Name: Allen Leung
2737    Date: 2001/10/27 20:34:00 EDT
2738    Tag: leunga-20011027-x86-fast-fp-call
2739    Description:
2740    
2741       Fixed the bug described in blume-20010920-slowfp.
2742    
2743       The fix involves
2744          1. generating FCOPYs in FSTP in ia32-svid
2745          2. marking a CALL with the appropriate annotation
2746    
2747    ----------------------------------------------------------------------
2748    Name: Matthias Blume
2749    Date: 2001/10/16 11:32:00 EDT
2750    Tag: blume-20011016-netbsd
2751    Description:
2752    
2753    Underscore patch from Chris Richards (fixing problem with compiling
2754    runtime system under recent NetBSD).
2755    
2756    ----------------------------------------------------------------------
2757    Name: Allen Leung
2758    Date: 2001/10/12 17:18:32 EDT 2001
2759    Tag: leung-20011012-x86-printflowgraph
2760    Description:
2761    
2762    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
2763    
2764    ----------------------------------------------------------------------
2765    Name: Lal George
2766    Date: 2001/10/11 23:51:34 EDT
2767    Tag: george-20011011-too-many-instrs
2768    Description:
2769    
2770    The representation of a program point never expected to see more
2771    than 65536 instructions in a basic block!
2772    
2773    ----------------------------------------------------------------------
2774    Name: Lal George
2775    Date: 2001/10/09 09:41:37 EDT
2776    Tag: george-20011008-mlrisc-labels
2777    Description:
2778    
2779    Changed the machine description files to support printing of
2780    local and global labels in assembly code, based on host assembler
2781    conventions.
2782    
2783    ----------------------------------------------------------------------
2784    Name: Matthias Blume
2785    Date: 2001/09/25 15:25:00 EDT
2786    Tag: blume-20010925-exninfo
2787    Description:
2788    
2789    I provided a non-hook implementation of exnName (at the toplevel) and
2790    made the "dummy" implementation of exnMessage (at the toplevel) more
2791    useful: if nothing gets "hooked in", then at least you are going to
2792    see the exception name and a message indicating why you don't see more.
2793    
2794    [For the time being, programs that need exnMessage and want to use
2795    ml-build should either use General.exnMessage (strongly recommended) or
2796    refer to structure General at some other point so that CM sees a
2797    static dependency.]
2798    
2799    [Similar remarks go for "print" and "use":  If you want to use their
2800    functionality in stand-alone programs generated by ml-build, then use
2801    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
2802    
2803    ----------------------------------------------------------------------
2804    Name: Matthias Blume
2805    Date: 2001/09/20 17:28:00 EDT
2806    Tag: blume-20010920-slowfp
2807    Description:
2808    
2809    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
2810    
2811    ----------------------------------------------------------------------
2812    Name: Matthias Blume
2813    Date: 2001/09/20 17:20:00 EDT
2814    Tag: blume-20010920-canonicalpaths
2815    Description:
2816    
2817    0. Updated the BOOT file (something that I forgot to do earlier).
2818    
2819    1. Small internal change to CM so that it avoids "/../" in filenames
2820       as much as possible (but only where it is safe).
2821    
2822    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
2823       that arguments that contain delimiters are passed through correctly.
2824       This change also means that all "special" arguments of the form
2825       @SMLxxx... must come first.
2826    
2827    3. Changed install script to put relative anchor names for tool commands
2828       into pathconfig.
2829    
2830    ----------------------------------------------------------------------
2831    Name: Matthias Blume
2832    Date: 2001/09/18 15:35:00 EDT
2833    Tag: blume-20010918-readme11036
2834    Description:
2835    
2836    Added README files.
2837    
2838    ----------------------------------------------------------------------
2839    Name: Matthias Blume
2840    Date: 2001/09/18 11:45:00 EDT
2841    Tag: Release_110_36 (retag)
2842    Description:
2843    
2844    Fixed mistake in config/preloads. Retagged as 110.36.
2845    
2846    ----------------------------------------------------------------------
2847    Name: Matthias Blume
2848    Date: 2001/09/18 09:40:00 EDT
2849    Tag: Release_110_36_orig (tag changed)
2850    Description:
2851    
2852    New version (110.36).  New bootfiles.
2853    
2854    ----------------------------------------------------------------------
2855    Name: Matthias Blume
2856    Date: 2001/09/14 16:15:00 EDT
2857    Tag: blume-20010914-x86fastfp
2858    Description:
2859    
2860    John committed some changes that Allen made, in particular a (hopefully)
2861    correctly working version of the x86-fp module.
2862    
2863    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
2864    flag to "true".  Everything seems to compile to a fixpoint ok, and
2865    "mandelbrot" speeds up by about 15%.
2866    
2867    ----------------------------------------------------------------------
2868    Name: Matthias Blume
2869    Date: 2001/09/13 11:20:00 EDT
2870    Tag: blume-20010913-minimal
2871    Description:
2872    
2873    1. Stefan Monnier's patch to fix a miscompilation problem that
2874       was brought to light by John Reppy's work on Moby.
2875    
2876    2. Implemented a minimal "structure Compiler" that contains just
2877       "version" and "architecture".  The minimal version will be
2878       available when the full version is not.  This is for backward-
2879       compatibility with code that wants to test Compiler.version.
2880    
2881    ----------------------------------------------------------------------
2882    Name: Matthias Blume
2883    Date: 2001/08/28 14:03:00 EDT
2884    Tag: blume-20010828-ml-lex
2885    Description:
2886    
2887    Fix for bug 1581, received from Neophytos Michael.
2888    
2889    ----------------------------------------------------------------------
2890    Name: Matthias Blume
2891    Date: 2001/08/27 11:20:00 EDT
2892    Tag: blume-20010827-readme11035
2893    Description:
2894    
2895    Fleshed out the README file for 110.35.
2896    
2897    ----------------------------------------------------------------------
2898    Name: Matthias Blume
2899    Date: 2001/08/24 17:10:00 EDT
2900    Tag: Release_110_35
2901    Description:
2902    
2903    New version number (110.35).  New bootfiles.
2904    
2905    ----------------------------------------------------------------------
2906    Name: Lal George
2907    Date: 2001/08/24 13:47:18 EDT 2001
2908    Tag: george-20010824-MLRISC-graphs
2909    Description:
2910    
2911     removed clusters from MLRISC completely and replaced with graphs.
2912    
2913    ----------------------------------------------------------------------
2914    Name: Matthias Blume
2915    Date: 2001/08/23 17:50:00 EDT
2916    Tag: blume-20010823-toplevel
2917    Description:
2918    
2919    - some reorganization of the code that implements various kinds of
2920      environments in the compiler (static, dynamic, symbolic, combined)
2921    - re-implemented the EnvRef module so that evalStream works properly
2922      (if the stream contains references to "use", "CM.make", etc.)
2923    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
2924    
2925    ----------------------------------------------------------------------
2926    Name: Matthias Blume
2927    Date: 2001/08/20 15:50 EDT
2928    Tag: blume20010820-slipup
2929    Description:
2930    
2931    I forgot to commit a few files.  Here they are...
2932    
2933    ----------------------------------------------------------------------
2934    Name: Matthias Blume
2935    Date: 2001/08/20 15:35:00 EDT
2936    Tag: blume-20010820-debugprof
2937    Description:
2938    
2939    !!!! NEW BOOTFILES !!!!
2940    
2941    This is another round of reorganizing the compiler sources.  This
2942    time the main goal was to factor out all the "instrumentation"
2943    passes (for profiling and backtracing) into their own library.
2944    The difficulty was to do it in such a way that it does not depend
2945    on elaborate.cm but only on elabdata.cm.
2946    
2947    Therefore there have been further changes to both elaborate.cm and
2948    elabdata.cm -- more "generic" things have been moved from the former
2949    to the latter.  As a result, I was forced to split the assignment
2950    of numbers indicating "primtyc"s into two portions: SML-generic and
2951    SML/NJ-specific.  Since it would have been awkward to maintain,
2952    I bit the bullet and actually _changed_ the mapping between these
2953    numbers and primtycs.  The bottom line of this is that you need
2954    a new set of bin- and bootfiles.
2955    
2956    I have built new bootfiles for all architectures, so doing a fresh
2957    checkout and config/install.sh should be all you need.
2958    
2959    The newly created library's name is
2960    
2961        $smlnj/viscomp/debugprof.cm
2962    
2963    and its sources live under
2964    
2965        src/compiler/DebugProf
2966    
2967    ----------------------------------------------------------------------
2968    Name: Matthias Blume
2969    Date: 2001/08/15 17:15:00 EDT
2970    Tag: blume-20010815-compreorg
2971    Description:
2972    
2973    This is a first cut at reorganizing the CM libraries that make up the
2974    core of the compiler.  The idea is to separate out pieces that could
2975    be used independently by tools, e.g., the parser, the typechecker, etc.
2976    
2977    The current status is a step in this direction, but it is not quite
2978    satisfactory yet.  Expect more changes in the future.
2979    
2980    Here is the current (new) organization...
2981    
2982        What used to be $smlnj/viscomp/core.cm is now divided into
2983        six CM libraries:
2984    
2985             $smlnj/viscomp/basics.cm
2986                           /parser.cm
2987                           /elabdata.cm
2988                           /elaborate.cm
2989                           /execute.cm
2990                           /core.cm
2991    
2992        The CM files for these libraries live under src/system/smlnj/viscomp.
2993        All these libraries are proxy libraries that contain precisely
2994        one CM library component.  Here are the locations of the components
2995        (all within the src/compiler tree):
2996    
2997             Basics/basics.cm
2998             Parse/parser.cm
2999             ElabData/elabdata.cm
3000             Elaborator/elaborate.cm
3001             Execution/execute.cm
3002             core.cm
3003    
3004         [This organization is the same that has been used already
3005         for a while for the architecture-specific parts of the visible
3006         compiler and for the old version of core.cm.]
3007    
3008         As you will notice, many source files have been moved from their
3009         respective original locations to a new home in one of the above
3010         subtrees.
3011    
3012         The division of labor between the new libraries is the following:
3013    
3014             basics.cm:
3015                - Simple, basic definitions that pertain to many (or all) of
3016                  the other libraries.
3017             parser.cm:
3018                - The SML parser, producing output of type Ast.dec.
3019                - The type family for Ast is also defined and exported here.
3020             elabdata.cm:
3021                - The datatypes that describe input and output of the elaborator.
3022                  This includes types, absyn, and static environments.
3023             elaborator.cm:
3024                - The SML/NJ type checker and elaborator.
3025                  This maps an Ast.dec (with a given static environment) to
3026                  an Absyn.dec (with a new static environment).
3027                - This libraries implements certain modules that used to be
3028                  structures as functors (to remove dependencies on FLINT).
3029             execute.cm:
3030                - Everything having to do with executing binary code objects.
3031                - Dynamic environments.
3032             core.cm:
3033                - SML/NJ-specific instantiations of the elaborator and MLRISC.
3034                - Top-level modules.
3035                - FLINT (this should eventually become its own library)
3036    
3037    Notes:
3038    
3039    I am not 100% happy with the way I separated the elaborator (and its
3040    data structures) from FLINT.  Two instances of the same problem:
3041    
3042        1. Data structures contain certain fields that carry FLINT-specific
3043           information.  I hacked around this using exn and the property list
3044           module from smlnj-lib.  But the fact that there are middle-end
3045           specific fields around at all is a bit annoying.
3046    
3047        2. The elaborator calculates certain FLINT-related information.  I tried
3048           to make this as abstract as I could using functorization, but, again,
3049           the fact that the elaborator has to perform calculations on behalf
3050           of the middle-end at all is not nice.
3051    
3052        3. Having to used exn and property lists is unfortunate because it
3053           weakens type checking.  The other alternative (parameterizing
3054           nearly *everything*) is not appealing, though.
3055    
3056    I removed the "rebinding =" warning hack because due to the new organization
3057    it was awkward to maintain it.  As a result, the compiler now issues some of
3058    these warnings when compiling init.cmi during bootstrap compilation. On
3059    the plus side, you also get a warning when you do, for example:
3060       val op = = Int32.+
3061    which was not the case up to now.
3062    
3063    I placed "assign" and "deref" into the _Core structure so that the
3064    code that deals with the "lazy" keyword can find them there.  This
3065    removes the need for having access to the primitive environment
3066    during elaboration.
3067    
3068    ----------------------------------------------------------------------
3069    Name: Matthias Blume
3070    Date: 2001/08/13
3071    Tag: blume-20010813-closures
3072    Description:
3073    
3074    This fix was sent to us by Zhong Shao.  It is supposed to improve the
3075    performance of certain loops by avoiding needless closure allocation.
3076    
3077    ----------------------------------------------------------------------
3078    Name: Lal George
3079    Date: 2001/07/31 10:03:23 EDT 2001
3080    Tag: george-20010731-x86-fmalloc
3081    Description: Fixed bug in x86 calls
3082    
3083        There was a bug where call instructions would mysteriously
3084        vanish. The call instruction had to be one that returned
3085        a floating point value.
3086    
3087    ----------------------------------------------------------------------
3088    Name: Lal George
3089    Date: 2001/07/19 16:36:29 EDT 2001
3090    Tag: george-20010719-simple-cells
3091    Description:
3092    
3093    I have dramatically simplified the interface for CELLS in MLRISC.
3094    
3095    In summary, the cells interface is broken up into three parts:
3096    
3097      1. CellsBasis : CELLS_BASIS
3098    
3099            CellsBasis is a top level structure and common for all
3100            architectures.  it contains the definitions of basic datatypes
3101            and utility  functions over these types.
3102    
3103      2. functor Cells() : CELLS
3104    
3105            Cells generates an interface for CELLS that incorporates the
3106            specific resources on the target architecture, such as the
3107            presence of special register classes, their number and size,
3108            and various useful substructures.
3109    
3110      3. <ARCH>CELLS
3111    
3112            e.g. SparcCells: SPARCCELLS
3113    
3114            <ARCH>CELLS usually contains additional bindings for special
3115            registers  on the architecture, such as:
3116    
3117                    val r0 : cell           (* register zero *)
3118                    val y : cell            (* Y register *)
3119                    val psr : cell          (* processor status register *)
3120                    ...
3121    
3122            The structure returned by applying the Cells functor is opened
3123            in this interface.
3124    
3125    The main implication of all this is that the datatypes for cells is
3126    split between CellsBasis and CELLS -- a fairly simple change for user
3127    code.
3128    
3129    In the old scheme the CELLS interface had a definitional binding of
3130    the form:
3131    
3132            signature CELLS = sig
3133    
3134               structure CellsBasis = CellsBasis
3135    
3136               ...
3137    
3138            end
3139    
3140    With all the sharing constraints that goes on in MLRISC, this old
3141    design  quickly leads to errors such as:
3142    
3143            "structure definition spec inside of sharing ... "
3144    
3145    
3146    and appears to require an unacceptable amount of sharing and where
3147    constraint hackery.
3148    
3149    I think this error message (the interaction of definitional specs and
3150    sharing) requires more explanation on our web page.
3151    
3152    ----------------------------------------------------------------------
3153    Name: Matthias Blume
3154    Date: 2001/07/19 15:00:00 EDT
3155    Tag: blume-20010719-libreorg
3156    Description:
3157    
3158    This update puts together a fairly extensive but straightforward change
3159    to the way the libraries that implement the interactive system are
3160    organized:
3161    
3162       The biggest change is the elimination of structure Compiler.  As a
3163       replacement for this structure, there is now a CM library
3164       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
3165       that exports all the substructures of the original structure Compiler
3166       directly.  So instead of saying Compiler.Foo.bar one now simply
3167       says Foo.bar.  (The CM libraries actually export a collection of
3168       structures that is richer than the collection of substructures of
3169       structure Compiler.)
3170    
3171       To make the transition smooth, there is a separate library called
3172       $smlnj/compiler/compiler.cm which puts together and exports the
3173       original structure Compiler (or at least something very close to it).
3174    
3175       There are five members of the original structure Compiler
3176       that are not exported directly but which instead became members
3177       of a new structure Backend (described by signature BACKEND).  These are:
3178       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
3179       Interact (: INTERACT), structure Machine (: MACHINE), and val
3180       architecture (: string).
3181    
3182       Structure Compiler.Version has become structure CompilerVersion.
3183    
3184       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
3185       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
3186       or x86, respectively.
3187       Each of these exports the same frontend structures that
3188       $smlnj/compiler.cm exports.  But they do not have a structure Backend
3189       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
3190       Hppa, PPC, Sparc, or X86, respectively.
3191    
3192       Library $smlnj/compiler/all.cm exports the union of the exports of
3193       $smlnj/compiler/<arch>.cm
3194    
3195       There are no structures <Arch>Compiler anymore, use
3196       $smlnj/compiler/<arch>.cm instead.
3197    
3198       Library host-compiler-0.cm is gone.  Instead, the internal library
3199       that instantiates CM is now called cm0.cm.  Selection of the host
3200       compiler (backend) is no longer done here but. (Responsibility for it
3201       now lies with $smlnj/compiler/current.cm.  This seems to be more
3202       logical.)
3203    
3204       Many individual files have been moved or renamed.  Some files have
3205       been split into multiple files, and some "dead" files have been deleted.
3206    
3207    Aside from these changes to library organization, there are also changes
3208    to the way the code itself is organized:
3209    
3210       Structure Binfile has been re-implemented in such a way that it no
3211       longer needs any knowledge of the compiler.  It exclusively deals
3212       with the details of binfile layout.  It no longer invokes the
3213       compiler (for the purpose of creating new prospective binfile
3214       content), and it no longer has any knowledge of how to interpret
3215       pickles.
3216    
3217       Structure Compile (: COMPILE) has been stripped down to the bare
3218       essentials of compilation.  It no longer deals with linking/execution.
3219       The interface has been cleaned up considerably.
3220    
3221       Utility routines for dealing with linking and execution have been
3222       moved into their own substructures.
3223    
3224       (The ultimate goal of these changes is to provide a light-weight
3225       binfile loader/linker (at least for, e.g., stable libraries) that
3226       does not require CM or the compiler to be present.)
3227    
3228    CM documentation has been updated to reflect the changes to library
3229    organization.
3230    
3231    ----------------------------------------------------------------------
3232    Name: Matthias Blume
3233    Date: 2001/07/10 17:30:00 EDT
3234    Tag: Release_110_34
3235    Description:
3236    
3237    Minor tweak to 110.34 (re-tagged):
3238    
3239      - README.html file added to CVS repository
3240      - runtime compiles properly under FreeBSD 3.X and 4.X
3241    
3242    ----------------------------------------------------------------------
3243    Name: Matthias Blume
3244    Date: 2001/07/10 17:30:00 EDT
3245    Tag: Release_110_34
3246    Description:
3247    
3248    New version number (110.34). New bootfiles.
3249    
3250    ----------------------------------------------------------------------
3251    Name: Matthias Blume
3252    Date: 2001/07/09 16:00:00 EDT
3253    Tag: blume-20010709-more-varargs
3254    Description:
3255    
3256    I changed the handling of varargs in ml-nlffigen again:
3257    The ellipsis ... will now simply be ignored (with an accompanying warning).
3258    
3259    The immediate effect is that you can actually call a varargs function
3260    from ML -- but you can't actually supply any arguments beyond the ones
3261    specified explicitly.  (For example, you can call printf with its format
3262    string, but you cannot pass additional arguments.)
3263    
3264    This behavior is only marginally more useful than the one before, but
3265    it has the advantage that a function or, more importantly, a function
3266    type never gets dropped on the floor, thus avoiding follow-up problems with
3267    other types that refer to the offending one.
3268    
3269    ----------------------------------------------------------------------
3270    Name: Matthias Blume
3271    Date: 2001/07/09 11:25:00 EDT
3272    Tag: blume-20010709-varargs
3273    Description:
3274    
3275    1. ckit-lib.cm now exports structure Error
3276    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
3277       with a warning accompanied by a source location.  Moreover, it
3278       merely skips the offending function or type and proceeds with the
3279       rest of its work.u  As a result, one can safely feed C code containing
3280       "..." to ml-nlffigen.
3281    3. There are some internal improvements to CM, providing slightly
3282       more general string substitutions in the tools subsystem.
3283    
3284    ----------------------------------------------------------------------
3285    Name: Matthias Blume
3286    Date: 2001/06/27 15:10:00 EDT
3287    Tag: blume-20010627-concur
3288    Description:
3289    
3290    Fixed a small bug in CM's handling of parallel compilation.
3291    (You could observe the bug by Control-C-interrupting an ordinary
3292    CMB.make or CM.stabilize and then attaching some compile servers.
3293    The result was that all of a sudden the previously interrupted
3294    compilation would continue on its own.  This was because of
3295    an over-optimization: CM did not bother to clean out certain queues
3296    when no servers were attached "anyway", resulting in the contents
3297    of these queues to grab control when new servers did get attached.)
3298    
3299    There is also another minor update to the CM manual.
3300    
3301    ----------------------------------------------------------------------
3302    Name: Matthias Blume
3303    Date: 2001/06/26 16:15:00 EDT
3304    Tag: blume-20010626-cmdoc
3305    Description:
3306    
3307    Minor typo fixed in CM manual (syntax diagram for libraries).
3308    
3309    ----------------------------------------------------------------------
3310    Name: Matthias Blume
3311    Date: 2001/06/25 22:55:00 EDT
3312    Tag: blume-20010625-x86pc
3313    Description:
3314    
3315    Fixed a nasty bug in the X86 assembly code that caused signal
3316    handlers to fail (crash) randomly.
3317    
3318    ----------------------------------------------------------------------
3319  Name: Matthias Blume  Name: Matthias Blume
3320  Date: 2001/06/25 12:05:00 EDT  Date: 2001/06/25 12:05:00 EDT
3321  Tag: blume-20010625-nlffigen  Tag: blume-20010625-nlffigen
# Line 2502  Line 5805 
5805       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
5806       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
5807       Please, don't be too harsh on me for messing with this code a bit more       Please, don't be too harsh on me for messing with this code a bit more
5808       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
5809       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
5810       Dave's hands. :)       Dave's hands. :)
5811    

Legend:
Removed from v.847  
changed lines
  Added in v.1409

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