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 1016, Tue Jan 15 23:10:06 2002 UTC revision 1490, Tue May 11 19:32:39 2004 UTC
# Line 11  Line 11 
11  Date: yyyy/mm/dd  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/05/11 14:35:00 CDT
18    Tag: blume-20040511-win32sock
19    Description:
20    
21    Added previously missing support for many socket-related functions
22    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
23    for the voluminous patch!
24    
25    (I have not tested this patch under win32 yet.)
26    
27    Here is David's e-mail:
28    
29    Hi,
30    
31    Attached to this email you find a diff against sml/nj 110.45
32    that will enable socket support under Windows.
33    
34    To apply the patch (using unix or cygwin)
35    1) gunzip runtime.diff.gz
36    2) "cd" into "src/runtime" in the source tree of a fresh
37       110.45 installation.
38    3) patch -p 1 < [your/path/to]runtime.diff
39    
40    The code compiles fine but has NOT yet been extensively tested.
41    I only ran a few tests for basic socket client functionality
42    (which worked fine).  Especially the functions that use ioctl
43    are not tested at all and might not work (see below).
44    
45    I implemented this since we want to move to a newer version of sml/nj
46    but need socket support in order to use it.  This is the first time I
47    even had a look at the sml/nj source,  so please review my changes
48    before making this part of the distribution!  Here are a few issues
49    that I think might be better for someone to solve who is more
50    familiar with the sml/nj source (and socket programming):
51    
52    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
53      exception since I could not figure out what the windows equivalent
54      of these functions is
55    
56    - In sockets-osdep.h there are a some #include statements that are
57      only used in a few files that include sockets-osdep.h
58    
59    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
60      does not process its return value since I don't know how to
61      report an error upwards.
62    
63    - It would probably be good to have a call to WSACleanup() when
64      the library is unloaded (if there is such a possibility).
65      Otherwise I think Windows will take care of this automatically
66      when the process finishes.
67    
68    - I used ioctlsocket() as a replacement for ioctl() but I have
69      no idea if that is actually the proper replacement on Windows.
70    
71    - All these issues are marked in the code by "FIXME" comments.
72    
73    We use sml/nj extensively in our products and are quite happy
74    with it.  I hope this contribution will help you.
75    
76    Keep up the good work!
77    
78    David
79    
80    ----------------------------------------------------------------------
81    Name: Matthias Blume (blume (at) tti - c (dot) org)
82    Date: 2004/05/11 14:20:00 CDT
83    Tag: blume-20040511-installml
84    Description:
85    
86    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
87    the patch.)
88    
89    ----------------------------------------------------------------------
90    Name: Matthias Blume (blume (at) tti - c (dot) org)
91    Date: 2004/05/11 14:05:00 CDT
92    Tag: blume-20040511-nlffi-netbsd
93    Description:
94    
95    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
96    the patch.)
97    
98    ----------------------------------------------------------------------
99    Name: Matthias Blume (blume (at) tti - c (dot) org)
100    Date: 2004/05/11 12:05:00 CDT
101    Tag: blume-20040511-exports
102    Description:
103    
104    As per request by Adam Chlipala <adam@hcoop.net>, extended various
105    export lists in compiler-related .cm-files.
106    
107    ----------------------------------------------------------------------
108    Name: Matthias Blume (blume (at) tti - c (dot) org)
109    Date: 2004/05/11 11:35:00 CDT
110    Tag: blume-20040511-allsource
111    Description:
112    
113    The installer now honors the "src-smlnj" target again, although its meaning
114    has changed from "all sources required for the compiler" to "all sources
115    the installer knows about".  In other words, if you enable "src-smlnj"
116    in the "targets" file, then the installer will pull in sources for
117    everything.  (Notice that this refers to source code only.  Compiled
118    code is still only installed for modules that were requested explicitly
119    or which are required for other modules that were requested explicitly.)
120    
121    ----------------------------------------------------------------------
122    Name: Matthias Blume (blume (at) tti - c (dot) org)
123    Date: 2004/04/23 17:40:00 CDT
124    Tag: blume-20040423-ieee-scan
125    Description:
126    
127    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
128    in the exponent calculation we get INF or ZERO (depending on the mantissa
129    and the sign of the exponent).
130    
131    ----------------------------------------------------------------------
132    Name: Matthias Blume (blume (at) tti - c (dot) org)
133    Date: 2004/04/23 10:40:00 CDT
134    Tag: blume-20040423-ml-build
135    Description:
136    
137    The ml-build script now terminates with a non-0 status when something
138    goes wrong.
139    
140    ----------------------------------------------------------------------
141    Name: Matthias Blume (blume (at) tti - c (dot) org)
142    Date: 2004/04/22 16:35:00 CDT
143    Tag: blume-20040422-Option
144    Description:
145    
146    Made exception Option to be the same as exception Option.Option
147    (as it should be).
148    
149    ----------------------------------------------------------------------
150    Name: Allen Leung (leunga (at) reservoir (dot) com)
151    Date: 2004/03/19 14:40:00 EST
152    Tag: leunga-20040319-cygwin-nlffi
153    Description:
154    
155    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
156    of SML/NJ.  The problem is that
157    
158       lib = dlopen(NULL, ...)
159       f   = dlsym(lib, "malloc");
160    
161    does not work on Windows unless we explicitly export symbols
162    such as 'malloc' during linking.  We fixed this by explicitly
163    exporting the required symbols with the magic gcc incantation:
164    
165        -Wl,--export-all cygwin.def
166    
167    where cygwin.def is a file containing all the symbols that
168    we wish to export.
169    
170    I suspect this is a Windows problem and we'll have to
171    do the same (somehow with windows compilers) when
172    we build the native win32 version with the system
173    calls LoadLibrary/GetProcAddress.
174    
175    ----------------------------------------------------------------------
176    Name: Matthias Blume (blume (at) tti - c (dot) org)
177    Date: 2004/03/04 16:35:00 CST
178    Tag: blume-20040304-intinf-fmt
179    Description:
180    
181    Fixed problem with IntInf.fmt (sign would show up on the right instead
182    of on the left for BIN, OCT, and HEX).
183    
184    ----------------------------------------------------------------------
185    Name: Matthias Blume (blume (at) tti - c (dot) org)
186    Date: 2004/03/04 11:25:00 CST
187    Tag: blume-20040304-symlinks
188    Description:
189    
190    Fixed problem with installer script (unix only) where bin/ml-yacc and
191    friends pointed (via symlinks) to absolute locations instead of just
192    .run-sml.  This was reported by Vesa A Norrman.
193    
194    ----------------------------------------------------------------------
195    Name: Matthias Blume (blume (at) tti - c (dot) org)
196    Date: 2004/02/13 14:50:00 CST
197    Tag: Release_110_45
198    Description:
199    
200    New working version (110.45). New bootfiles.
201    
202    ----------------------------------------------------------------------
203    Name: Matthias Blume (blume (at) tti - c (dot) org)
204    Date: 2004/01/26 15:15:15 CST
205    Tag: blume-20040126-toplevel
206    Description:
207    
208    Improved handling of exceptions at the interactive toplevel.
209    
210    ----------------------------------------------------------------------
211    Name: Matthias Blume (blume (at) tti - c (dot) org)
212    Date: 2004/01/26 11:25:00
213    Tag: blume-20040126-app
214    Description:
215    
216    Type of top-level "app" corrected.
217    Added code for setting vp_limitPtrMask to Win32-specific runtime.
218    
219    ----------------------------------------------------------------------
220    Name: Matthias Blume (blume (at) tti - c (dot) org)
221    Date: 2003/11/18 17:10 CST
222    Tag: blume-20031118-basis-fiddle
223    Description:
224    
225    - changed Timer interface to what might become the spec
226    - POSIX_FLAGS -> BIT_FLAGS according to spec
227    - some other minor discrepancies wrt. spec eliminated
228    
229    ----------------------------------------------------------------------
230    Name: Matthias Blume (blume (at) tti - c (dot) org)
231    Date: 2003/11/06 12:00:00 CST
232    Tag: Release_110_44
233    Description:
234    
235    New working version (110.44). New bootfiles.
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume (blume (at) tti - c (dot) org)
239    Date: 2003/11/04 11:50:00 CST
240    Tag: blume-20031104-move-libraries
241    Description:
242    
243    Eliminated the "dont_move_libraries" directive in config/targets.
244    (The mechanism was broken and could not be fixed easily.  Moreover,
245    there does not seem to be any reason not to move all libraries into
246    lib during installation.  I originally implemented this directive as a
247    backward-compatibility feature when I first introduced the new CM.
248    Now that things have been stable for a long time and going back to the
249    old CM is not an option, there is no reason to keep it around.)
250    
251    ----------------------------------------------------------------------
252    Name: Matthias Blume (blume (at) tti - c (dot) org)
253    Date: 2003/11/03 16:00:00 CST
254    Tag: blume-20031103-installdir
255    Description:
256    
257    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
258    Richards for pointing out the problem and providing the solution.)
259    
260    ----------------------------------------------------------------------
261    Name: Matthias Blume (blume (at) tti - c (dot) org)
262    Date: 2003/10/01 17:05:00 CDT
263    Tag: blume-20031001-lal-mlrisc
264    Description:
265    
266    MLRISC bug fix from Lal.
267    
268    ----------------------------------------------------------------------
269    Name: Matthias Blume (blume (at) tti - c (dot) org)
270    Date: 2003/09/30 16:10:00 CDT
271    Tag: blume-20030930-primio-bat
272    Description:
273    
274    1. Added openVector, nullRd, and nullWr to PRIM_IO.
275    2. Improved .bat files (for Win32 port) to make things work under Win95.
276       (thanks to Aaron S. Hawley for this one)
277    
278    ----------------------------------------------------------------------
279    Name: Matthias Blume (blume (at) tti - c (dot) org)
280    Date: 2003/09/26 16:05:00 CDT
281    Tag: blume-20030926-wrappriv
282    Description:
283    
284    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
285    
286    ----------------------------------------------------------------------
287    Name: Matthias Blume (blume (at) tti - c (dot) org)
288    Date: 2003/09/26 15:00:00 CDT
289    Tag: blume-20030926-110_43_3
290    Description:
291    
292    - additional cleanup
293    - version number bump, NEW BOOTFILES
294    
295    ----------------------------------------------------------------------
296    Name: Matthias Blume (blume (at) tti - c (dot) org)
297    Date: 2003/09/26 12:00:00 CDT
298    Tag: blume-20030926-ppautoload
299    Description:
300    
301    I modified the read-eval-print loop so that the autoloader gets
302    invoked whenever the prettyprinter tries to look up a symbol that
303    is not currently defined in the toplevel environment but which
304    appears in CM's autoload registry.  As a result, we see far fewer of
305    those ?.Foo.Bar.xxx names in the prettyprinter's output.
306    
307    In addition to this I tried to clean up some pieces of the Basis
308    implementation (e.g., Socket, Word8Array) in order to prevent other
309    instances of these ?.Foo.Bar.xxx names from being printed.
310    
311    The mechanism that picks names for types still needs some work, though.
312    (Right now it seems that if there is a type A.t which is defined to
313    be B.u, but B is unavailable at toplevel, then A.t gets printed as
314    "?.B.u" although the perhaps more sensible solution would be to use
315    "A.t" in this case.  In other words, the prettyprinter should follow
316    a chain of DEFtycs not farther than there are corresponding toplevel
317    names in the current environment.)
318    
319    ----------------------------------------------------------------------
320    Name: Matthias Blume (blume (at) tti - c (dot) org)
321    Date: 2003/09/24 16:31:00 CDT
322    Tag: blume-20030924-installer
323    Description:
324    
325    Another installer tweak:  All the ML code for the installer is now
326    compiled during CMB.make and put into a little library called
327    $smlnj/installer.cm.  The installation then simply invokes
328    
329       sml -m $smlnj/installer.cm
330    
331    and everything happens automagically.
332    
333    Win32:  ML code senses value of environment variable SMLNJ_HOME.
334    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
335            and BINDIR.
336    
337    The new scheme guarantees that the ML code responsible for the installation
338    is in sync with the APIs of the main system.  Also, the installer is
339    somewhat faster because the installer script is precompiled.
340    
341    ----------------------------------------------------------------------
342    Name: Matthias Blume (blume (at) tti - c (dot) org)
343    Date: 2003/09/24 15:35:00 CDT
344    Tag: blume-20030924-synsock
345    Description:
346    
347    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
348    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
349    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
350    implementations of non-blocking operations from CML's Socket
351    structure.
352    
353    ----------------------------------------------------------------------
354    Name: Matthias Blume (blume (at) tti - c (dot) org)
355    Date: 2003/09/24 15:10:05 CDT
356    Tag: blume-20030924-sockets
357    Description:
358    
359    1. Fixed SOCKET API and implementation to match Basis spec.
360       This required changing the internal representation of sockets to one
361       that remembers (for each socket file descriptor) whether it is currently
362       blocking or non-blocking.  This state is maintained lazily (i.e., a system
363       call is made only if the state actually needs to change).
364    
365    2. OS-specific details of sockets were moved into separate files, thus
366       making it possible to unify the bulk of the socket implementations
367       between Unix and Win32.
368    
369    3. CML's socket API changed accordingly.
370       (Note that we need to remove non-blocking functions from this API
371       since they are redundant in the case of CML!)
372    
373    4. CML's socket implementation now makes use of non-blocking functions
374       provided by Basis, thus removing all OS-dependent code from this part
375       of CML.
376    
377    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
378    
379    ----------------------------------------------------------------------
380    Name: Matthias Blume (blume (at) tti - c (dot) org)
381    Date: 2003/09/22 12:10:00 CDT
382    Tag: blume-20030922-110_43_2
383    Description:
384    
385    Made a new interim version and bootfiles for developer's bootstrapping
386    convenience.
387    
388       110.43.2 -- NEW BOOTFILES
389    
390    ----------------------------------------------------------------------
391    Name: Matthias Blume (blume (at) tti - c (dot) org)
392    Date: 2003/09/19 15:55:00 CDT
393    Tag: blume-20030919-cmdir
394    Description:
395    
396    1. new-install.sh -> install.sh
397    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
398    3. tweaked installer so that another name instead of .cm can be chosen
399       at install time (by setting the CM_DIR_ARC environment variable
400       during installation); once installation is complete, the name is
401       fixed
402    
403    ----------------------------------------------------------------------
404    Name: Matthias Blume (blume (at) tti - c (dot) org)
405    Date: 2003/09/18 16:00:00 CDT
406    Tag: blume-20030918-110_43_1
407    Description:
408    
409    Made a new interim version and bootfiles for developer's bootstrapping
410    convenience.
411    
412       110.43.1 -- NEW BOOTFILES
413    
414    ----------------------------------------------------------------------
415    Name: Matthias Blume (blume (at) tti - c (dot) org)
416    Date: 2003/09/18 15:20:00 CDT
417    Tag: blume-20030918-misc
418    Description:
419    
420    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
421       this seems to be controversial at the moment) and used those in
422       Posix.ProcEnv.times.
423    
424    2. Added Time.{from,to}Nanoseconds to Time.
425    
426    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
427       For example, fromLargeInt never needs to look at more than 3 "big
428       digits" to get its 53 bits of precision.
429    
430    ----------------------------------------------------------------------
431    Name: Matthias Blume (blume (at) tti - c (dot) org)
432    Date: 2003/09/17 16:30:00 CDT
433    Tag: blume-20030917-real32-slices
434    Description:
435    
436    Added an entry to the primitive environment
437    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
438    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
439    
440    Removed some of the "magic" constants in real64.sml and replaced them
441    with code that generates these values from their corresponding
442    integer counterparts.
443    
444    Made all(?) the slice-related changes to the Basis and made everything
445    compile again...
446    
447    ----------------------------------------------------------------------
448    Name: Matthias Blume (blume (at) tti - c (dot) org)
449    Date: 2003/09/15 17:45:00 CDT
450    Tag: blume-20030915-rbase
451    Description:
452    
453    Fixed bug in Real.fromLargeInt.
454    
455    ----------------------------------------------------------------------
456    Name: Matthias Blume (blume (at) tti - c (dot) org)
457    Date: 2003/09/13 18:11:00 CDT
458    Tag: blume-20030913-libinstall
459    Description:
460    
461    Minor bugfix in config/libinstall (set anchor with path to
462    standalone tool after installing it, otherwise libraries that
463    need ml-lex or ml-yacc won't compile the first time the installer
464    runs).
465    
466    ----------------------------------------------------------------------
467    Name: Matthias Blume (blume (at) tti - c (dot) org)
468    Date: 2003/09/12 11:45:00 CDT
469    Tag: blume-20030912-various
470    Description:
471    
472     - fixed bug in Real.toLargeInt
473     - fixed bug in Posix.ProcEnv.times
474     - changed inputLine functions to return an option
475     - minor installer improvements / bugfixes
476     - changed default @SMLalloc parameter for x86/celeron to 64k
477    
478    ----------------------------------------------------------------------
479    Name: Matthias Blume (blume (at) tti - c (dot) org)
480    Date: 2003/09/09 22:00:00 CDT
481    Tag: Release_110_43
482    Description:
483    
484        New working release 110.43.  New bootfiles.
485    
486    ----------------------------------------------------------------------
487    Name: Matthias Blume (blume (at) tti - c (dot) org)
488    Date: 2003/09/09 19:20:00 CDT
489    Tag: blume-20030909-installer
490    Description:
491    
492    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
493    Modified config/install.bat to take advantage of it.  Also modified
494    config/install.sh (and called it config/new-install.sh) to take advantage
495    of it on Unix systems.  (The SML code is (supposed to be) platform-
496    independent.)
497    
498    The installer can now install everything under Win32
499    as well as under *nix as long as it compiles.
500    
501    Other changes:
502    
503      - made CML compile again under Win32
504      - made eXene compile under Win32 (by providing a fake structure UnixSock
505        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
506      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
507        same as type int;  under Win32 it isn't
508      - fixed some slice-related problems in the win32-specific parts of CML
509      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
510        (under Win32, the volume name is case-insensitive, and the
511        OS.Path code compares volume names for equality)
512    
513    ----------------------------------------------------------------------
514    Name: Matthias Blume (blume (at) tti - c (dot) org)
515    Date: 2003/09/08 11:55:00 CDT
516    Tag: blume-20030908-fullpath
517    Description:
518    
519    Made Win32 version of OS.FileSys.fullPath return current directory
520    when given an empty string.  This is what the spec says, and incidentally,
521    CM depends on it.  (CM otherwise goes into an infinite loop in certain
522    cases when presented with the name of a non-existing .cm file.)
523    
524    ----------------------------------------------------------------------
525    Name: Matthias Blume (blume (at) tti - c (dot) org)
526    Date: 2003/09/04 16:30:00 CDT
527    Tag: blume-20030905-slices-etc
528    Description:
529    
530      1. Changed interface to vectors and arrays in Basis to match
531         (draft) Basis spec.
532      2. Added signatures and implementations of slices according to
533         Basis spec.
534      3. Edited source code throughout the system to make it compile again
535         under 1. and 2.  (In some cases code had to be added to have it
536         match the new signatures.)
537      4. MLRISC should be backward-compatible: the copies of the originals
538         of files that needed to change under 3. were retained, the .cm files
539         check the compiler version number and use old versions when
540         appropriate.
541      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
542         dirstream -> string option (in accordance with Basis spec).
543      6. When generating code that counts lines, ml-lex used function
544         CharVector.foldli, taking advantage of its old interface.
545         This has been replaced with the corresponding code from
546         CharVectorSlice.  (html-lex must be re-lexed!)
547      7. BitArray in smlnj-lib/Util has been extended/modified to match the
548         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
549      8. Removed temporary additions (fromInternal, toInternal) from the
550         (now obsolete) IntInf in smlnj-lib/Util.
551      9. Cleaned up structure Byte.
552     10. Added localOffset, scan, and fromString to Date (according to spec).
553         Cleaned/corrected implementation of Date.
554         (Still need to check for correctness; implement better canonicalizeDate.)
555     11. Added "scan" to signature IEEE_REAL.
556     12. Some improvements to IntInf [in particular: efficiency-hack for
557         mod and rem when second operand is 2 (for parity checks).]
558     13. Changed representation of type Time.time, using a single IntInf.int
559         value counting microseconds.  This considerably simplified the
560         implementation of structure Time.  We now support negative time
561         values; scan and fromString handle signs.
562     14. Functor PrimIO now takes two additional arguments (VectorSlice and
563         ArraySlice).
564    
565    ----------------------------------------------------------------------
566    Name: Matthias Blume (blume (at) tti - c (dot) org)
567    Date: 2003/08/28 17:00:00 CDT
568    Tag: blume-20030828-intinf
569    Description:
570    
571    This is a major update which comes with a version number bump
572    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
573    and an implementation of IntInf in the Basis.
574    
575    There are a fairly large number of related changes and updates throughout
576    the system:
577    
578    Basis:
579            - Implemented IntInf.
580            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
581            - Added some missing Real64 operations, most notably Real.toLargeInt.
582            - Added FixedInt as a synonym for Int32.
583    
584    compiler:
585            * Added support for a built-in intinf type.
586              - literals
587              - pattern matching
588              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
589              - overloading on literals and operations
590    
591            This required adding a primitive type intinf, some additional
592            primops, and implementations for several non-trivial intinf
593            operations in Core.  (The intinf type is completely abstract
594            to the compiler; all operations get delegated back to the Core.)
595    
596            * Intinf equality is handled by polyequal.  However, the compiler
597            does not print its usual warning in this case (since polyequal
598            is the right thing to do there).
599    
600            * Improved the organization of structure InlineT.
601    
602            * A word about conversion primops:
603            If conversions involving intinf do not cancel out during
604            CPS contract, then the compiler must insert calls to Core functions.
605            Since all core access must be resolved already during the FLINT
606            translate phase, it would be too late a the time of CPS contract
607            to add new Core calls.  For this reason, conversion primops
608            for intinf carry two arguments:  1. the numeric argument that
609            they are supposed to convert, and 2. the Core function that
610            can help with this conversion if necessary.  If CPS contract
611            eliminates a primop, then the associated Core function becomes
612            dead and goes away.  Intinf conversion primops that do not get
613            eliminated by CPS contract get rewritten into calls of their
614            core functions by a separate, new phase.
615    
616    interactive system:
617            - Control.Print.intinfDepth controls max length of intinf constants
618            being printed. (Analogous to Control.Print.stringDepth.)
619            - Cleanup in printutil and pputil: got rid of unused stuff and
620            duplicates; replaced some of the code with code that makes better
621            use of library functionality.
622    
623    CM:
624            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
625            are no longer silent.
626    
627    CKIT:
628            Fixed mismatched uses of Int32 and LargeInt.  I always decided
629            in favor of LargeInt -- which is now the same as IntInf.
630            CKIT-knowledgable people should check whether this is what's
631            intended and otherwise change things back to using Int32 or
632            FixedInt.
633    
634    Throughout the code:
635            Started using IntInf.int literals and built-in operations
636            (e.g., comparison with 0) where this seems appropriate.
637    
638    
639    ----------------------------------------------------------------------
640    Name: Dave MacQueen (dbm@cs.uchicago.edu)
641    Date: 2003/08/13 11:36:00 CDT
642    Tag: dbm-20030813-mcz-merge1
643    Description:
644    
645    Merging changes from the mcz-branch development branch into trunk.
646    These changes involve replacement of the emulated old prettyprinter
647    interface with direct use of the SML/NJ Lib PP library, and fixing
648    of a couple of bugs (895, 1186) relating to error messages.  A new
649    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
650    has been added.
651    
652    ----------------------------------------------------------------------
653    Name: Matthias Blume (blume (at) tti - c (dot) org)
654    Date: 2003/08/11 15:45:00 CDT
655    Tag: blume-20030811-windows
656    Description:
657    
658    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
659    
660        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
661    
662    This patch restores SML/NJ's ability to run under win32.  There are a
663    number of changes, including fixes for several bugs that had gone
664    unnoticed until now:
665    
666      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
667        trying to run the win32 version from within cygwin.)
668    
669      - There are a number of simple .bat scripts that substitute for their
670        corresponding Unix shell-scripts.  (See below.)
671    
672      - The internals of ml-build have been modified slightly.  The main
673        difference is that instead of calling ".link-sml" (or link-sml.bat)
674        using OS.Process.system, the ML process delegates this task back
675        to the script.  Otherwise problems arise in mixed environments such
676        as Cygwin where scripts look and work like Unix scripts, but
677        where OS.Process.system cannot run them.
678    
679      - In CM, the srcpath pickler used native pathname syntax -- which
680        is incorrect in the case of cross-compilation.  The new pickle format
681        is independent of platform-specific naming conventions.
682    
683      - Path configuration files (such as lib/pathconfig) can now choose
684        between native and standard syntax.  Placing a line of the form
685    
686            standard!
687    
688        into the file causes all subsequent paths to be interpreted using
689        CM standard pathname syntax (= Unix conventions); a line
690    
691            native!
692    
693        switches back to native style.  This was needed so that
694        path config files can be written portably, see src/system/pathconfig.
695    
696      - Runtime system:
697    
698          - win32-filesys.c:  get_file_time and set_file_time now
699            access modification time, not creation time.
700    
701          - I/O code made aware of new array representation.
702    
703          - Bug fixes in X86.prim.masm.
704    
705       - src/system/makeml made aware of win32. (For use under cygwin
706            and other Unix-environments for windows.)
707    
708       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
709         which caused BinIO.inputAll to fail consistently.
710    
711    .bat scripts:
712    
713       Windows .bat scripts assume that SMLNJ_HOME is defined.
714    
715       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
716         applications (sml, ml-yacc, ml-lex).
717       - ml-build.bat: analogous to ml-build.
718       - config\install.bat:  Analogous to config/install.sh.  This requires
719         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
720         (nmake etc. must be on the path, and vcvars32 must have been run.)
721         Moreover, sources for ml-lex and ml-yacc need to exist under src,
722         and the bootfile hierarchy must have been unpacked under
723         sml.boot.x86-win32.
724         The script is very primitive and does a poor job at error checking.
725         It only installs the base system, ml-lex, and ml-yacc.  No other
726         libraries are being installed (i.e., you get only those that
727         are part of the compiler.)
728       - link-sml.bat:  analogous to .link-sml, but not currently used
729    
730    Unrelated bug fixes:
731    
732      - ml-nlffigen now exports structures ST_* corresponding to incomplete
733        types.
734      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
735        otherwise.)
736    
737    ----------------------------------------------------------------------
738    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
739    Date: 2003/06/17
740    Tag: macqueen-20030617-bug895
741    Description:
742    
743    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
744    Tag will be used for new development branch (mcz-branch) for
745    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
746    
747    ----------------------------------------------------------------------
748    Name: Matthias Blume (blume (at) tti - c (dot) org)
749    Date: 2003/05/27 16:55:00 CDT
750    Tag: blume-20030527-polyeq
751    Description:
752    
753    Tried to eliminated most cases of polymorphic equality.
754    
755    ----------------------------------------------------------------------
756    Name: Matthias Blume (blume (at) tti - c (dot) org)
757    Date: 2003/05/21 17:45:00 CDT
758    Tag: blume-20030517-complete
759    Description:
760    
761    Two changes:
762    
763      1. Added a flag for controlling whether non-exhaustive bindings will
764         be treated as errors (default is false).
765      2. Cleaned up the *entire* source tree so that CMB.make goes through
766         without a single non-exhaustive match- or bind warning.
767    
768    ----------------------------------------------------------------------
769    Name: Matthias Blume (blume (at) tti - c (dot) org)
770    Date: 2003/05/17 10:20:00 CDT
771    Tag: blume-20030517-absyn
772    Description:
773    
774    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
775    
776       This mainly affects the quality of error messages.  However, some
777       of the code is now more straightforward than before. (Treatment of
778       the above four constructs in translate.sml is much simpler than
779       the "macro-expansion" that was going on before.  Plus, the mach-
780       compiler no longer gets invoked just to be able to compile an
781       if-expression.)
782    
783    2. The ErrorMsg.Error exception is now caught and absorbed by the
784       interactive loop.
785    
786    ----------------------------------------------------------------------
787    Name: Allen Leung
788    Date: 2003/05/16 13:05:00 CDT
789    Tag: leunga-20030516-cygwin-runtime
790    Description:
791    
792        Ported the runtime system to cygwin, which uses the unix
793    x86-unix bin files.  Missing/buggy features:
794    
795       o getnetbyname, getnetbyaddr: these functions seem to be missing in
796         the Cygwin library.
797       o Ctrl-C handling may be flaky.
798       o Windows system calls and Windows I/O are not supported.
799    
800       A new set of binfiles is located at:
801    
802         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
803    
804       This is only needed for bootstrapping the cygwin version of smlnj.
805       Other x86 versions can use the existing binfiles.
806    
807    ----------------------------------------------------------------------
808    Name: Matthias Blume
809    Date: 2003/04/08 15:42:00 CDT
810    Tag: blume-20030408-listpair
811    Description:
812    
813    1. Added a target 'mlrisc' to installer.
814    
815    2. Added missing elements to structure ListPair.
816    
817    ----------------------------------------------------------------------
818    Name: Allen Leung
819    Date: 2003/01/07 10:40:00 EST
820    Tag: leunga-20030107-int-rem
821    Description:
822    
823    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
824    The arguments to the SUBL instruction were swapped.
825    
826    ----------------------------------------------------------------------
827    Name: Matthias Blume
828    Date: 2002/12/12 16:25:00 EST
829    Tag: blume-20021212-risc-ra
830    Description:
831    
832    Fixed a serious bug in the rewrite code for FP spilling/reloading that
833    sent the RA into an infinite loop when floating point registers get
834    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
835    110.37 and 110.38.)
836    There was another set of potential problems related to the handling of
837    MLRISC annotations (but those did not yet cause real problems, apparently).
838    
839    ----------------------------------------------------------------------
840    Name: Matthias Blume
841    Date: 2002/12/06 22:40:00 EST
842    Tag: blume-20021206-cm-fileid
843    Description:
844    
845    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
846    This fixes the problem of CM getting confused by files that suddenly
847    change their identity (e.g., by getting unlinked and recreated by some
848    text editor such as vi).  There might be a better/cheaper/cleaner way
849    of doing this, but for now this will have to do.
850    
851    ----------------------------------------------------------------------
852    Name: Matthias Blume
853    Date: 2002/10/28 09:50:00 EST
854    Tag: blume-20021028-typecheck
855    Description:
856    
857    Exported structure Typecheck from $smlnj/viscomp/core.cm.
858    
859    ----------------------------------------------------------------------
860    Name: Matthias Blume
861    Date: 2002/10/17 09:10:00 EDT
862    Tag: Release_110_42
863    Description:
864    
865    In good old tradition, there has been a slight hiccup so that we have
866    to patch 110.42 after the fact.   The old release tag has been replaced
867    (see below).
868    
869    The change solves a problem with two competing approaches the
870    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
871    each other's way.
872    
873    This change only affects the runtime system code and the installer script.
874    (No new bootfiles.)
875    
876    ----------------------------------------------------------------------
877    Name: Matthias Blume
878    Date: 2002/10/16 12:00:00 EDT
879    Tag: Release_110_42_removed
880    Description:
881    
882    New working release.  New bootfiles.
883    
884    ----------------------------------------------------------------------
885    Name: Matthias Blume
886    Date: 2002/10/10 13:10:00 EDT
887    Tag: blume-20021010-ppc-divs
888    Description:
889    
890    The mltree operator DIVS must be implemented with an overflow check on
891    the PPC because the hardware indicates divide-by-zero using "overflow" as
892    well.
893    
894    ----------------------------------------------------------------------
895    Name: Matthias Blume
896    Date: 2002/07/23 11:45:00
897    Tag: blume-20020723-smlnj-home
898    Description:
899    
900    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
901    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
902    CM_PATHCONFIG is also set), the path configuration file is assumed
903    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
904    the entire tree to some other place and everything will "just work".
905    
906    (Companion commands such as ml-build and ml-makedepend also sense this
907    variable.)
908    
909    ----------------------------------------------------------------------
910    Name: Matthias Blume
911    Date: 2002/07/12 21:19:00 EDT
912    Tag: blume-20020712-liveness
913    Description:
914    
915    Exported two useful "step" functions from liveness module (MLRISC).
916    
917    ----------------------------------------------------------------------
918    Name: Matthias Blume
919    Date: 2002/07/05 16:00 EDT
920    Tag: Release_110_41
921    Description:
922    
923    New working release.  New bootfiles.
924    
925    ----------------------------------------------------------------------
926    Name: Matthias Blume
927    Date: 2002/07/05 10:25:00 EDT
928    Tag: blume-20020705-btimp
929    Description:
930    
931    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
932    other clients can set up backtracing support.
933    
934    ----------------------------------------------------------------------
935    Name: Matthias Blume
936    Date: 2002/06/25 17:23:00 EDT
937    Tag: blume-20020625-fpmax
938    Description:
939    
940    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
941    case.  (The sense of the isNaN test was reversed -- which made min and
942    max always return their first argument.)
943    
944    ----------------------------------------------------------------------
945    Name: Matthias Blume
946    Date: 2002/06/11
947    Tag: blume-20020611-unixpath
948    Description:
949    
950    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
951    
952    ----------------------------------------------------------------------
953    Name: Matthias Blume
954    Date: 2002/06/10 16:35:00 EDT
955    Tag: blume-20020610-ieeereal
956    Description:
957    
958    I back-ported my implementation of IEEEReal.fromString from the
959    idlbasis-devel branch so that we can test it.
960    
961    Another small change is that ppDec tries to give more information
962    than just "<sig>" in the case of functors.  However, this code is
963    broken in some mysterious way if the functor's body's signature
964    has not been declared by ascription but gets inferred from the
965    implementation.  This needs fixing...
966    
967    ----------------------------------------------------------------------
968    Name: Matthias Blume
969    Date: 2002/05/31
970    Tag: blume-20020531-btrace-mode
971    Description:
972    
973    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
974    the wayside when I switched over to using Controls everywhere.)
975    
976    ----------------------------------------------------------------------
977    Name: Lal George
978    Date: 2002/05/23 12:21:40 EDT
979    Tag: george-20020523-visual-labels
980    Description:
981    
982    Labels are now displayed in the graphical output to make
983    the fall-through and target blocks obvious.
984    
985    ----------------------------------------------------------------------
986    Name: Matthias Blume
987    Date: 2002/05/22 11:03:00 EDT
988    Tag: blume-20020522-shrink
989    Description:
990    
991    John tweaked yesterday's fix for 1131 to handle an out-of-memory
992    situation that comes up when allocating huge arrays.
993    
994    ----------------------------------------------------------------------
995    Name: Matthias Blume
996    Date: 2002/05/21 16:00:00 EDT
997    Tag: Release_110_40
998    Description:
999    
1000    New working release (110.40).  New bootfiles.
1001    
1002    [Also: John Reppy fixed GC bug 1131.]
1003    
1004    ----------------------------------------------------------------------
1005    Name: Matthias Blume
1006    Date: 2002/05/21 12:35:00 EDT
1007    Tag: blume-20020521-cmdoc
1008    Description:
1009    
1010    CM documentation update.
1011    
1012    ----------------------------------------------------------------------
1013    Name: Matthias Blume
1014    Date: 2002/05/21 10:55:00 EDT
1015    Tag: blume-20020521-misc
1016    Description:
1017    
1018    - John tweaked runtime to be silent on heap export (except when
1019      GC messages are on).
1020    - I added a few more things (cross-compiling versions of CMB) to
1021      config/preloads (as suggestions).
1022    
1023    ----------------------------------------------------------------------
1024    Name: Matthias Blume
1025    Date: 2002/05/20 22:25:00 EDT
1026    Tag: blume-20020520-controls
1027    Description:
1028    
1029    - Added ControlUtil structure to control-lib.cm.
1030    - Use it throughout.
1031    - Used Controls facility to define MLRISC controls (as opposed to
1032      registering MLRISC control ref cells with Controls after the fact)
1033    - Fixed messed-up controls priorities.
1034    
1035    * Removed again all the stuff from config/preloads that one wouldn't
1036      be able to preload at the time the initial heap image is built.
1037      (Many libraries, e.g., CML, do not exist yet at this time.  The
1038       only libraries that can be preloaded via config/preloads are those
1039       that come bundled with the bootfiles.)
1040    
1041    ----------------------------------------------------------------------
1042    Name: Matthias Blume
1043    Date: 2002/05/20 10:59:00 EDT
1044    Tag: blume-20020520-preloads
1045    Description:
1046    
1047    Added a lot of commented-out suggestions for things to be included
1048    in config/preloads.
1049    
1050    ----------------------------------------------------------------------
1051    Name: Allen Leung
1052    Date: 2002/05/18 14:20:00 EDT
1053    Tag: leunga-20020518-mdl
1054    Description:
1055    
1056    o Made the mdl tool stuff compile and run again.
1057    o I've disabled all the stuff that depends on RTL specifications; they
1058      are all badly broken anyway.
1059    
1060    ----------------------------------------------------------------------
1061    Name: Matthias Blume
1062    Date: 2002/05/17 16:49:00 EDT
1063    Tag: blume-20020517-controls
1064    Description:
1065    
1066    0. John Reppy made several modifications to the SML/NJ library.
1067       In particular, there is a shiny new controls-lib.cm.
1068    
1069    1. Pushed new controls interface through compiler so that everything
1070       compiles again.
1071    
1072    2. Added FormatComb and FORMAT_COMB to the CML version of the
1073       SML/NJ library (so that CML compiles again).
1074    
1075    3. Modified init scripts because XXX_DEFAULT environment variables
1076       are no longer with us.  (Boot-time initialization is now done
1077       using the same environment variables that are also used for
1078       startup-time initialization of controls.)
1079    
1080    ----------------------------------------------------------------------
1081    Name: Lal George
1082    Date: 2002/05/15 09:20:10 EDT
1083    Tag: george-20020515-pseudo-op-decls
1084    Description:
1085    
1086            All pseudo-ops emitted before the first segment declaration
1087    such as TEXT, DATA, and BSS directives are assumed to be global
1088    declarations and are emitted first in the assembly file. This is
1089    useful in a number of situations where one has pseudo-ops that are not
1090    specific to any segment, and also works around the constraint that one
1091    cannot have client pseudo-ops in the TEXT segment.
1092    
1093            Because no segment is associated with these declarations it is
1094    an error to allocate any space or objects before the first segment
1095    directive and an exception will be raised. However, we cannot make
1096    this check for client pseudo-ops.
1097    
1098            These top level declarations are a field in the CFG graph_info.
1099    In theory you can continue to add to this field after the CFG has been
1100    built -- provided you know what you are doing;-)
1101    
1102    ----------------------------------------------------------------------
1103    Name: Matthias Blume
1104    Date: 2002/05/13 16:40:00 EDT
1105    Tag: blume-20020513-pp-etc
1106    Description:
1107    
1108    A few minor bugfixes:
1109    
1110      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1111        (Bogus printouts for redefined bindings still occur.  Compiler
1112        bug should no longer occur now.  We need to redo the prettyprinter
1113        from scratch.)
1114    
1115      - CM pathname printer now also adds escape sequences for ( and )
1116    
1117      - commend and docu fixes for ml-nlffi
1118    
1119    ----------------------------------------------------------------------
1120    Name: Matthias Blume
1121    Date: 2002/05/10 16:40:00 EDT
1122    Tag: blume-20020510-erg-textio
1123    Description:
1124    
1125    Applied the following bugfix provided by Emden Gansner:
1126    
1127        Output is corrupted when outputSubstr is used rather than output.
1128    
1129        The problem occurs when a substring
1130    
1131            ss = (s, dataStart, dataLen)
1132    
1133        where dataStart > 0, fills a stream buffer with avail bytes left.
1134        avail bytes of s, starting at index dataStart, are copied into the
1135        buffer, the buffer is flushed, and then the remaining dataLen-avail
1136        bytes of ss are copied into the beginning of the buffer. Instead of
1137        starting this copy at index dataStart+avail in s, the current code
1138        starts the copy at index avail.
1139    
1140        Fix:
1141        In text-io-fn.sml, change line 695 from
1142             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1143        to
1144             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1145    
1146    ----------------------------------------------------------------------
1147    Name: Matthias Blume
1148    Date: 2002/04/12 13:55:00 EDT
1149    Tag: blume-20020412-assyntax
1150    Description:
1151    
1152    1. Grabbed newer assyntax.h from the XFree86 project.
1153    2. Fiddled with how to compile X86.prim.asm without warnings.
1154    3. (Very) Minor cleanup in CM.
1155    
1156    ----------------------------------------------------------------------
1157    Name: Matthias Blume
1158    Date: 2002/04/01 (no joke!) 17:07:00 EST
1159    Tag: blume-20020401-x86div
1160    Description:
1161    
1162    Added full support for div/mod/rem/quot on the x86, using the machine
1163    instruction's two results (without clumsily recomputing the remainder)
1164    directly where appropriate.
1165    
1166    Some more extensive power-of-two support was added to the x86 instruction
1167    selector (avoiding expensive divs, mods, and muls where they can be
1168    replaced with cheaper shifts and masks).  However, this sort of thing
1169    ought to be done earlier, e.g., within the CPS optimizer so that
1170    all architectures benefit from it.
1171    
1172    The compiler compiles to a fixed point, but changes might be somewhat
1173    fragile nevertheless.  Please, report any strange things that you might
1174    see wrt. div/mod/quot/rem...
1175    
1176    ----------------------------------------------------------------------
1177    Name: Matthias Blume
1178    Date: 2002/03/29 17:22:00
1179    Tag: blume-20020329-div
1180    Description:
1181    
1182    Fixed my broken div/mod logic.  Unfortunately, this means that the
1183    inline code for div/mod now has one more comparison than before.
1184    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
1185    The problem was with quotient = 0, because that alone does not tell
1186    us which way the rounding went.  One then has to look at whether
1187    remainder and divisor have the same sign...  :(
1188    
1189    Anyway, I replaced the bootfiles with fresh ones...
1190    
1191    ----------------------------------------------------------------------
1192    Name: Matthias Blume
1193    Date: 2002/03/29 14:10:00 EST
1194    Tag: blume-20020329-inlprims
1195    Description:
1196    
1197    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
1198    
1199    Primops have changed. This means that the bin/boot-file formats have
1200    changed as well.
1201    
1202    To make sure that there is no confusion, I made a new version.
1203    
1204    
1205    CHANGES:
1206    
1207    * removed REMT from mltree (remainder should never overflow).
1208    
1209    * added primops to deal with divisions of all flavors to the frontend
1210    
1211    * handled these primops all the way through so they map to their respective
1212      MLRISC support
1213    
1214    * used these primops in the implementation of Int, Int32, Word, Word32
1215    
1216    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
1217    
1218    * parameterized INLMIN, INLMAX, and INLABS by a numkind
1219    
1220    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
1221      floating point
1222    
1223    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
1224      and Real (but Real.abs maps to a separate floating-point-only primop)
1225    
1226    
1227    TODO items:
1228    
1229    * Hacked Alpha32 instruction selection, disabling the selection of REMx
1230      instructions because the machine instruction encoder cannot handle
1231      them.  (Hppa, PPC, and Sparc instruction selection did not handle
1232      REM in the first place, and REM is supported by the x86 machine coder.)
1233    
1234    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
1235      selection phase.  (The two can be streamlined because the hardware
1236      delivers both quotient and remainder at the same time anyway.)
1237    
1238    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
1239      (Currently the behavior is inconsistent both across architectures and
1240      wrt. the draft Basis spec.)
1241    
1242    * Word8 should eventually be handled natively, too.
1243    
1244    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
1245      as if there currently is no execution path that could trigger it in
1246      SML/NJ.  (The assumptions underlying functions arith and promotable do not
1247      hold for things like multiplication and division.)
1248    
1249    ----------------------------------------------------------------------
1250    Name: Matthias Blume
1251    Date: 2002/03/27 16:27:00 EST
1252    Tag: blume-20020327-mlrisc-divisions
1253    Description:
1254    
1255    Added support for all four division operations (ML's div, mod, quot,
1256    and rem) to MLRISC.  In the course of doing so, I also rationalized
1257    the naming (no more annoying switch-around of DIV and QUOT), by
1258    parameterizing the operation by div_rounding_mode (which can be either
1259    DIV_TO_ZERO or DIV_TO_NEGINF).
1260    
1261    The generic MLTreeGen functor takes care of compiling all four
1262    operations down to only round-to-zero div.
1263    
1264    Missing pieces:
1265    
1266      * Doing something smarter than relying on MLTreeGen on architectures
1267        like, e.g., the x86 where hardware division delivers both quotient and
1268        remainder at the same time.  With this, the implementation of the
1269        round-to-neginf operations could be further streamlined.
1270    
1271      * Remove inlining support for div/mod/rem from the frontend and replace it
1272        with primops that get carried through to the backend.  Do this for all
1273        int and word types.
1274    
1275    ----------------------------------------------------------------------
1276    Name: Matthias Blume
1277    Date: 2002/03/25 17:25:00 EST
1278    Tag: blume-20020325-divmod
1279    Description:
1280    
1281    I improved (hopefully without breaking them) the implementation of Int.div,
1282    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
1283    advantage of the following observations:
1284    
1285      Let  q = x quot y      r = x rem y
1286           d = x div  y      m = x mod y
1287    
1288    where "quot" is the round-to-zero version of integer division that
1289    hardware usually provides.  Then we have:
1290    
1291         r = x - q * y        where neither the * nor the - will overflow
1292         d = if q >= 0 orelse x = q * y then q else q - 1
1293                              where neither the * nor the - will overflow
1294         m = if q >= 0 orelse r = 0 then r else r + y
1295                              where the + will not overflow
1296    
1297    This results in substantial simplification of the generated code.
1298    The following table shows the number of CFG nodes and edges generated
1299    for
1300            fun f (x, y) = x OPER y
1301            (* with OPER \in div, mod, quot, rem *)
1302    
1303    
1304        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
1305        --------------------------------------------------------
1306         div |         24 |         39 |         12 |         16
1307         mod |         41 |         71 |         12 |         16
1308        quot |          8 |         10 |          8 |         10
1309         rem |         10 |         14 |          8 |         10
1310    
1311    
1312    ----------------------------------------------------------------------
1313    Name: Matthias Blume
1314    Date: 2002/03/25 22:06:00 EST
1315    Tag: blume-20020325-cprotobug
1316    Description:
1317    
1318    Fixed a bug in cproto (c prototype decoder).
1319    
1320    ----------------------------------------------------------------------
1321    Name: Matthias Blume
1322    Date: 2002/03/25 16:00:00 EST
1323    Tag: blume-20020325-raw-primops
1324    Description:
1325    
1326    I did some cleanup to Allen's new primop code and
1327    replaced yesterday's bootfiles with new ones.
1328    (But they are stored in the same place.)
1329    
1330    ----------------------------------------------------------------------
1331    Name: Matthias Blume
1332    Date: 2002/03/24 22:40:00 EST
1333    Tag: blume-20020324-bootfiles
1334    Description:
1335    
1336    Made the bootfiles that Allen asked for.
1337    
1338    ----------------------------------------------------------------------
1339    Name: Allen Leung
1340    Date: 2002/03/23 15:50:00 EST
1341    Tag: leunga-20020323-flint-cps-rcc-primops
1342    Description:
1343    
1344      1. Changes to FLINT primops:
1345    
1346        (* make a call to a C-function;
1347         * The primop carries C function prototype information and specifies
1348         * which of its (ML-) arguments are floating point. C prototype
1349         * information is for use by the backend, ML information is for
1350         * use by the CPS converter. *)
1351      | RAW_CCALL of { c_proto: CTypes.c_proto,
1352                       ml_args: ccall_type list,
1353                       ml_res_opt: ccall_type option,
1354                       reentrant : bool
1355                     } option
1356       (* Allocate uninitialized storage on the heap.
1357        * The record is meant to hold short-lived C objects, i.e., they
1358        * are not ML pointers.  With the tag, the representation is
1359        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1360        *)
1361      | RAW_RECORD of {tag:bool,sz:int}
1362      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1363    
1364      2.  These CPS primops are now overloaded:
1365    
1366           rawload of {kind:numkind}
1367           rawstore of {kind:numkind}
1368    
1369          The one argument form is:
1370    
1371             rawload {kind} address
1372    
1373          The two argument form is:
1374    
1375             rawload {kind} [ml object, byte-offset]
1376    
1377      3. RAW_CCALL/RCC now takes two extra arguments:
1378    
1379         a. The first is whether the C call is reentrant, i.e., whether
1380            ML state should be saved and restored.
1381         b. The second argument is a string argument specifying the name of
1382            library and the C function.
1383    
1384         These things are currently not handled in the code generator, yet.
1385    
1386      4. In CProto,
1387    
1388         An encoding type of "bool" means "ml object" and is mapped into
1389         C prototype of PTR.  Note that "bool" is different than "string",
1390         even though "string" is also mapped into PTR, because "bool"
1391         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
1392    
1393      5. Pickler/unpicker
1394    
1395         Changed to handle RAW_RECORD and newest RAW_CCALL
1396    
1397      6. MLRiscGen,
1398    
1399         1. Changed to handle the new rawload/rawstore/rawrecord operators.
1400         2. Code for handling C Calls has been moved to a new module CPSCCalls,
1401            in the file CodeGen/cpscompile/cps-c-calls.sml
1402    
1403      7. Added the conditional move operator
1404    
1405             condmove of branch
1406    
1407         to cps.  Generation of this is still buggy so it is currently
1408         disabled.
1409    
1410    ----------------------------------------------------------------------
1411    Name: Lal George
1412    Date: 2002/03/22 14:18:25 EST
1413    Tag: george-20020322-cps-branch-prob
1414    Description:
1415    
1416    Implemented the Ball-Larus branch prediction-heuristics, and
1417    incorporated graphical viewers for control flow graphs.
1418    
1419    Ball-Larus Heuristics:
1420    ---------------------
1421    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
1422    
1423    By design it uses the Dempster-Shafer theory for combining
1424    probabilities.  For example, in the function:
1425    
1426        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
1427    
1428    the ball-larus heuristics predicts that the n=0 is unlikely
1429    (OH-heuristic), and the 'then' branch is unlikely because of the
1430    RH-heuristic -- giving the 'then' branch an even lower combined
1431    probability using the Dempster-Shafer theory.
1432    
1433    Finally, John Reppy's loop analysis in MLRISC, further lowers the
1434    probability of the 'then' branch because of the loop in the else
1435    branch.
1436    
1437    
1438    Graphical Viewing:
1439    ------------------
1440    I merely plugged in Allen's graphical viewers into the compiler. The
1441    additional code is not much. At the top level, saying:
1442    
1443            Control.MLRISC.getFlag "cfg-graphical-view" := true;
1444    
1445    will display the graphical view of the control flow graph just before
1446    back-patching.  daVinci must be in your path for this to work. If
1447    daVinci is not available, then the default viewer can be changed
1448    using:
1449    
1450            Control.MLRISC.getString "viewer"
1451    
1452    which can be set to "dot" or "vcg" for the corresponding viewers. Of
1453    course, these viewers must be in your path.
1454    
1455    The above will display the compilation unit at the level of clusters,
1456    many of which are small, boring, and un-interesting. Also setting:
1457    
1458            Control.MLRISC.getInt "cfg-graphical-view_size"
1459    
1460    will display clusters that are larger than the value set by the above.
1461    
1462    
1463    ----------------------------------------------------------------------
1464    Name: Matthias Blume
1465    Date: 2002/03/21 22:20:00 EST
1466    Tag: blume-20020321-kmp-bugfix
1467    Description:
1468    
1469    Changed the interface to the KMP routine in PreString and fixed
1470    a minor bug in one place where it was used.
1471    
1472    ----------------------------------------------------------------------
1473    Name: Allen Leung
1474    Date: 2002/03/21 20:30:00 EST
1475    Tag: leunga-20020321-cfg
1476    Description:
1477    
1478      Fixed a potential problem in cfg edge splitting.
1479    
1480    ----------------------------------------------------------------------
1481    Name: Allen Leung
1482    Date: 2002/03/21 17:15:00 EST
1483    Tag: leunga-20020321-x86-fp-cfg
1484    Description:
1485    
1486      1. Recoded the buggy parts of x86-fp.
1487    
1488         a. All the block reordering code has been removed.
1489            We now depend on the block placement phases to do this work.
1490    
1491         b. Critical edge splitting code has been simplified and moved into the
1492            CFG modules, as where they belong.
1493    
1494         Both of these were quite buggy and complex.  The code is now much, much
1495         simpler.
1496    
1497      2. X86 backend.
1498    
1499         a. Added instructions for 64-bit support.  Instruction selection for
1500            64-bit has not been committed, however, since that
1501            requires changes to MLTREE which haven't been approved by
1502            Lal and John.
1503    
1504         b. Added support for FUCOMI and FUCOMIP when generating code for
1505            PentiumPro and above.  We only generate these instructions in
1506            the fast-fp mode.
1507    
1508         c. Added cases for JP and JNP in X86FreqProps.
1509    
1510      3. CFG
1511    
1512         CFG now has a bunch of methods for edge splitting and merging.
1513    
1514      4. Machine description.
1515    
1516         John's simplification of MLTREE_BASIS.fcond broke a few machine
1517         description things:
1518    
1519         rtl-build.{sig,sml} and hppa.mdl fixed.
1520    
1521         NOTE: the machine description stuff in the repository is still broken.
1522               Again, I can't put my fixes in because that involves
1523               changes to MLTREE.
1524    
1525    ----------------------------------------------------------------------
1526    Name: Matthias Blume
1527    Date: 2002/03/20 15:55:00 EST
1528    Tag: blume-20020320-kmp
1529    Description:
1530    
1531    Implemented Knuth-Morris-Pratt string matching in PreString and used
1532    it for String.isSubstring, Substring.isSubstring, and
1533    Substring.position.
1534    
1535    (Might need some stress-testing.  Simple examples worked fine.)
1536    
1537    ----------------------------------------------------------------------
1538    Name: Matthias Blume
1539    Date: 2002/03/19 16:37:00 EST
1540    Tag: blume-20020319-witnesses
1541    Description:
1542    
1543    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
1544    
1545    This implements a generic mechanism for changing constness qualifiers
1546    anywhere within big C types without resorting to outright "casts".
1547    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
1548    modify the constness at the outermost level.)
1549    The implementation of "convert" is based on the idea of "witness"
1550    values -- values that are not used by the operation but whose types
1551    "testify" to their applicability.  On the implementation side, "convert"
1552    is simply a projection (returning its second curried argument).  With
1553    cross-module inlining, it should not result in any machine code being
1554    generated.
1555    
1556    ----------------------------------------------------------------------
1557    Name: Matthias Blume
1558    Date: 2002/03/15 16:40:00 EST
1559    Tag: blume-20020315-basis
1560    Description:
1561    
1562    Provided (preliminary?) implementations for
1563    
1564      {String,Substring}.{concatWith,isSuffix,isSubstring}
1565    
1566    and
1567    
1568      Substring.full
1569    
1570    Those are in the Basis spec but they were missing in SML/NJ.
1571    
1572    ----------------------------------------------------------------------
1573    Name: Matthias Blume
1574    Date: 2002/03/14 21:30:00 EST
1575    Tag: blume-20020314-controls
1576    Description:
1577    
1578    Controls:
1579    ---------
1580    
1581    1. Factored out the recently-added Controls : CONTROLS stuff and put
1582       it into its own library $/controls-lib.cm.  The source tree for
1583       this is under src/smlnj-lib/Controls.
1584    
1585    2. Changed the names of types and functions in this interface, so they
1586       make a bit more "sense":
1587    
1588          module -> registry
1589          'a registry -> 'a group
1590    
1591    3. The interface now deals in ref cells only.  The getter/setter interface
1592       is (mostly) gone.
1593    
1594    4. Added a function that lets one register an already-existing ref cell.
1595    
1596    5. Made the corresponding modifications to the rest of the code so that
1597       everything compiles again.
1598    
1599    6. Changed the implementation of Controls.MLRISC back to something closer
1600       to the original.  In particular, this module (and therefore MLRISC)
1601       does not depend on Controls.  There now is some link-time code in
1602       int-sys.sml that registers the MLRISC controls with the Controls
1603       module.
1604    
1605    CM:
1606    ---
1607    
1608      * One can now specify the lambda-split aggressiveness in init.cmi.
1609    
1610    ----------------------------------------------------------------------
1611    Name: Allen Leung
1612    Date: 2002/03/13 17:30:00 EST
1613    Tag: leunga-20020313-x86-fp-unary
1614    Description:
1615    
1616    Bug fix for:
1617    
1618    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
1619    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
1620    > - fun f(x,(y,z)) = Real.~ y;
1621    > [autoloading]
1622    > [autoloading done]
1623    >       fchsl   (%eax), 184(%esp)
1624    > Error: MLRisc bug: X86MCEmitter.emitInstr
1625    >
1626    > uncaught exception Error
1627    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
1628    
1629    The problem was that the code generator did not generate any fp registers
1630    in this case, and the ra didn't know that it needed to run the X86FP phase to
1631    translate the pseudo fp instruction.   This only happened with unary fp
1632    operators in certain situations.
1633    
1634    ----------------------------------------------------------------------
1635    Name: Matthias Blume
1636    Date: 2002/03/13 14:00:00 EST
1637    Tag: blume-20020313-overload-etc
1638    Description:
1639    
1640    1. Added _overload as a synonym for overload for backward compatibility.
1641       (Control.overloadKW must be true for either version to be accepted.)
1642    
1643    2. Fixed bug in install script that caused more things to be installed
1644       than what was requested in config/targets.
1645    
1646    3. Made CM aware of the (_)overload construct so that autoloading
1647       works.
1648    
1649    ----------------------------------------------------------------------
1650    Name: Matthias Blume
1651    Date: 2002/03/12 22:03:00 EST
1652    Tag: blume-20020312-url
1653    Description:
1654    
1655    Forgot to update BOOT and srcarchiveurl.
1656    
1657    ----------------------------------------------------------------------
1658    Name: Matthias Blume
1659    Date: 2002/03/12 17:30:00 EST
1660    Tag: blume-20020312-version110392
1661    Description:
1662    
1663    Yet another version number bump (because of small changes to the
1664    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
1665    
1666    Changes:
1667    
1668      The new pid generation scheme described a few weeks ago was overly
1669      complicated.  I implemented a new mechanism that is simpler and
1670      provides a bit more "stability":  Once CM has seen a compilation
1671      unit, it keeps its identity constant (as long as you do not delete
1672      those crucial CM/GUID/* files).  This means that when you change
1673      an interface, compile, then go back to the old interface, and
1674      compile again, you arrive at the original pid.
1675    
1676      There now also is a mechanism that instructs CM to use the plain
1677      environment hash as a module's pid (effectively making its GUID
1678      the empty string).  For this, "noguid" must be specified as an
1679      option to the .sml file in question within its .cm file.
1680      This is most useful for code that is being generated by tools such
1681      as ml-nlffigen (because during development programmers tend to
1682      erase the tool's entire output directory tree including CM's cached
1683      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
1684      revert to the old, broken behavior of SML/NJ, but in specific cases
1685      where there is no danger of interface confusion, its use is ok
1686      (I think).
1687    
1688      ml-nlffigen by default generates "noguid" annotations.  They can be
1689      turned off by specifying -guid in its command line.
1690    
1691    ----------------------------------------------------------------------
1692    Name: Lal George
1693    Date: 2002/03/12 12 14:42:36 EST
1694    Tag: george-20020312-frequency-computation
1695    Description:
1696    
1697    Integrated jump chaining and static block frequency into the
1698    compiler. More details and numbers later.
1699    
1700    ----------------------------------------------------------------------
1701    Name: Lal George
1702    Date: 2002/03/11 11 22:38:53 EST
1703    Tag: george-20020311-jump-chain-elim
1704    Description:
1705    
1706    Tested the jump chain elimination on all architectures (except the
1707    hppa).  This is on by default right now and is profitable for the
1708    alpha and x86, however, it may not be profitable for the sparc and ppc
1709    when compiling the compiler.
1710    
1711    The gc test will typically jump to a label at the end of the cluster,
1712    where there is another jump to an external cluster containing the actual
1713    code to invoke gc. This is to allow factoring of common gc invocation
1714    sequences. That is to say, we generate:
1715    
1716            f:
1717               testgc
1718               ja   L1      % jump if above to L1
1719    
1720            L1:
1721               jmp L2
1722    
1723    
1724    After jump chain elimination the 'ja L1' instructions is converted to
1725    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
1726    up being implemented in their long form (if L2 is far away) using:
1727    
1728            jbe     L3      % jump if below or equal to L3
1729            jmp     L2
1730         L3:
1731            ...
1732    
1733    
1734    For large compilation units L2  may be far away.
1735    
1736    
1737    ----------------------------------------------------------------------
1738    Name: Matthias Blume
1739    Date: 2002/03/11 13:30:00 EST
1740    Tag: blume-20020311-mltreeeval
1741    Description:
1742    
1743    A functor parameter was missing.
1744    
1745    ----------------------------------------------------------------------
1746    Name: Allen Leung
1747    Date: 2002/03/11 10:30:00 EST
1748    Tag: leunga-20020311-runtime-string0
1749    Description:
1750    
1751       The representation of the empty string now points to a
1752    legal null terminated C string instead of unit.  It is now possible
1753    to convert an ML string into C string with InlineT.CharVector.getData.
1754    This compiles into one single machine instruction.
1755    
1756    ----------------------------------------------------------------------
1757    Name: Allen Leung
1758    Date: 2002/03/10 23:55:00 EST
1759    Tag: leunga-20020310-x86-call
1760    Description:
1761    
1762       Added machine generation for CALL instruction (relative displacement mode)
1763    
1764    ----------------------------------------------------------------------
1765    Name: Matthias Blume
1766    Date: 2002/03/08 16:05:00
1767    Tag: blume-20020308-entrypoints
1768    Description:
1769    
1770    Version number bumped to 110.39.1.  NEW BOOTFILES!
1771    
1772    Entrypoints: non-zero offset into a code object where execution should begin.
1773    
1774    - Added the notion of an entrypoint to CodeObj.
1775    - Added reading/writing of entrypoint info to Binfile.
1776    - Made runtime system bootloader aware of entrypoints.
1777    - Use the address of the label of the first function given to mlriscGen
1778      as the entrypoint.  This address is currently always 0, but it will
1779      not be 0 once we turn on block placement.
1780    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
1781      with entry points) from mlriscGen.
1782    
1783    ----------------------------------------------------------------------
1784    Name: Allen Leung
1785    Date: 2002/03/07 20:45:00 EST
1786    Tag: leunga-20020307-x86-cmov
1787    Description:
1788    
1789       Bug fixes for CMOVcc on x86.
1790    
1791       1. Added machine code generation for CMOVcc
1792       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
1793       3. CMOVcc cannot have an immediate operand as argument.
1794    
1795    ----------------------------------------------------------------------
1796    Name: Matthias Blume
1797    Date: 2002/03/07 16:15:00 EST
1798    Tag: blume-20020307-controls
1799    Description:
1800    
1801    This is a very large but mostly boring patch which makes (almost)
1802    every tuneable compiler knob (i.e., pretty much everything under
1803    Control.* plus a few other things) configurable via both the command
1804    line and environment variables in the style CM did its configuration
1805    until now.
1806    
1807    Try starting sml with '-h' (or, if you are brave, '-H')
1808    
1809    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
1810    implements the underlying generic mechanism.
1811    
1812    The interface to some of the existing such facilities has changed somewhat.
1813    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
1814    (The getFoo interface is still there for backward-compatibility, but its
1815    use is deprecated.)
1816    
1817    The ml-build script passes -Cxxx=yyy command-line arguments through so
1818    that one can now twiddle the compiler settings when using this "batch"
1819    compiler.
1820    
1821    TODO items:
1822    
1823    We should go through and throw out all controls that are no longer
1824    connected to anything.  Moreover, we should go through and provide
1825    meaningful (and correct!) documentation strings for those controls
1826    that still are connected.
1827    
1828    Currently, multiple calls to Controls.new are accepted (only the first
1829    has any effect).  Eventually we should make sure that every control
1830    is being made (via Controls.new) exactly once.  Future access can then
1831    be done using Controls.acc.
1832    
1833    Finally, it would probably be a good idea to use the getter-setter
1834    interface to controls rather than ref cells.  For the time being, both
1835    styles are provided by the Controls module, but getter-setter pairs are
1836    better if thread-safety is of any concern because they can be wrapped.
1837    
1838    *****************************************
1839    
1840    One bug fix: The function blockPlacement in three of the MLRISC
1841    backpatch files used to be hard-wired to one of two possibilities at
1842    link time (according to the value of the placementFlag).  But (I
1843    think) it should rather sense the flag every time.
1844    
1845    *****************************************
1846    
1847    Other assorted changes (by other people who did not supply a HISTORY entry):
1848    
1849    1. the cross-module inliner now works much better (Monnier)
1850    2. representation of weights, frequencies, and probabilities in MLRISC
1851       changed in preparation of using those for weighted block placement
1852       (Reppy, George)
1853    
1854    ----------------------------------------------------------------------
1855    Name: Lal George
1856    Date: 2002/03/07 14:44:24 EST 2002
1857    Tag: george-20020307-weighted-block-placement
1858    
1859    Tested the weighted block placement optimization on all architectures
1860    (except the hppa) using AMPL to generate the block and edge frequencies.
1861    Changes were required in the machine properties to correctly
1862    categorize trap instructions. There is an MLRISC flag
1863    "weighted-block-placement" that can be used to enable weighted block
1864    placement, but this will be ineffective without block/edge
1865    frequencies (coming soon).
1866    
1867    
1868    ----------------------------------------------------------------------
1869    Name: Lal George
1870    Date: 2002/03/05 17:24:48 EST
1871    Tag: george-20020305-linkage-cluster
1872    
1873    In order to support the block placement optimization, a new cluster
1874    is generated as the very first cluster (called the linkage cluster).
1875    It contains a single jump to the 'real' entry point for the compilation
1876    unit. Block placement has no effect on the linkage cluster itself, but
1877    all the other clusters  have full freedom in the manner in which they
1878    reorder blocks or functions.
1879    
1880    On the x86 the typical linkage code that is generated is:
1881       ----------------------
1882            .align 2
1883       L0:
1884            addl    $L1-L0, 72(%esp)
1885            jmp     L1
1886    
1887    
1888            .align  2
1889       L1:
1890       ----------------------
1891    
1892    72(%esp) is the memory location for the stdlink register. This
1893    must contain the address of the CPS function being called. In the
1894    above example, it contains the address of  L0; before
1895    calling L1 (the real entry point for the compilation unit), it
1896    must contain the address for L1, and hence
1897    
1898            addl $L1-L0, 72(%esp)
1899    
1900    I have tested this on all architectures except the hppa.The increase
1901    in code size is of course negligible
1902    
1903    ----------------------------------------------------------------------
1904    Name: Allen Leung
1905    Date: 2002/03/03 13:20:00 EST
1906    Tag: leunga-20020303-mlrisc-tools
1907    
1908      Added #[ ... ] expressions to mlrisc tools
1909    
1910    ----------------------------------------------------------------------
1911    Name: Matthias Blume
1912    Date: 2002/02/27 12:29:00 EST
1913    Tag: blume-20020227-cdebug
1914    Description:
1915    
1916    - made types in structure C and C_Debug to be equal
1917    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
1918    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
1919    
1920    ----------------------------------------------------------------------
1921    Name: Matthias Blume
1922    Date: 2002/02/26 12:00:00 EST
1923    Tag: blume-20020226-ffi
1924    Description:
1925    
1926    1. Fixed a minor bug in CM's "noweb" tool:
1927       If numbering is turned off, then truly don't number (i.e., do not
1928       supply the -L option to noweb).  The previous behavior was to supply
1929       -L'' -- which caused noweb to use the "default" line numbering scheme.
1930       Thanks to Chris Richards for pointing this out (and supplying the fix).
1931    
1932    2. Once again, I reworked some aspects of the FFI:
1933    
1934       A. The incomplete/complete type business:
1935    
1936       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
1937         gone!
1938       - ML types representing an incomplete type are now *equal* to
1939         ML types representing their corresponding complete types (just like
1940         in C).  This is still safe because ml-nlffigen will not generate
1941         RTTI for incomplete types, nor will it generate functions that
1942         require access to such RTTI.   But when ML code generated from both
1943         incomplete and complete versions of the C type meet, the ML types
1944         are trivially interoperable.
1945    
1946         NOTE:  These changes restore the full generality of the translation
1947         (which was previously lost when I eliminated functorization)!
1948    
1949       B. Enum types:
1950    
1951       - Structure C now has a type constructor "enum" that is similar to
1952         how the "su" constructor works.  However, "enum" is not a phantom
1953         type because each "T enum" has values (and is isomorphic to
1954         MLRep.Signed.int).
1955       - There are generic access operations for enum objects (using
1956         MLRep.Signed.int).
1957       - ml-nlffigen will generate a structure E_foo for each "enum foo".
1958         * The structure contains the definition of type "mlrep" (the ML-side
1959         representation type of the enum).  Normally, mlrep is the same
1960         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
1961         then mlrep will be defined as a datatype -- thus facilitating
1962         pattern matching on mlrep values.
1963         ("-ec" will be suppressed if there are duplicate values in an
1964          enumeration.)
1965         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
1966         will be generated for each C enum constant xxx.
1967         * Conversion functions m2i and i2m convert between mlrep and
1968         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
1969         * Coversion functions c and ml convert between mlrep and "tag enum".
1970         * Access functions (get/set) fetch and store mlrep values.
1971       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
1972         enumerations are merged into one single enumeration represented by
1973         structure E_'.
1974    
1975    ----------------------------------------------------------------------
1976    Name: Allen Leung
1977    Date: 2002/02/25 04:45:00 EST
1978    Tag: leunga-20020225-cps-spill
1979    
1980    This is a new implementation of the CPS spill phase.
1981    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
1982    In case of problems, replace it with the old file spill.sml
1983    
1984    The current compiler runs into some serious performance problems when
1985    constructing a large record.  This can happen when we try to compile a
1986    structure with many items.  Even a very simple structure like the following
1987    makes the compiler slow down.
1988    
1989        structure Foo = struct
1990           val x_1 = 0w1 : Word32.int
1991           val x_2 = 0w2 : Word32.int
1992           val x_3 = 0w3 : Word32.int
1993           ...
1994           val x_N = 0wN : Word32.int
1995        end
1996    
1997    The following table shows the compile time, from N=1000 to N=4000,
1998    with the old compiler:
1999    
2000    N
2001    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2002           MLRISC ra                               0.06u  0.00s  0.05g
2003              (spills = 0 reloads = 0)
2004           TOTAL                                   0.63u  0.07s  0.21g
2005    
2006    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2007           MLRISC ra                               5.68u  0.59s  3.93g
2008              (spills = 0 reloads = 0)
2009           TOTAL                                   14.71u  0.99s  4.81g
2010    
2011    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2012           MLRISC ra                               5.54u  0.65s  3.91g
2013              (spills = 543 reloads = 1082)
2014           TOTAL                                   65.40u  3.13s  6.00g
2015    
2016    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2017           MLRISC ra                               0.80u  0.10s  0.55g
2018              (spills = 42 reloads = 84)
2019           TOTAL                                   129.42u  5.10s  4.13g
2020    
2021    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2022           MLRISC ra                               2.69u  0.27s  1.38g
2023              (spills = 62 reloads = 124)
2024           TOTAL                                   682.48u  19.61s  13.99g
2025    
2026    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2027           MLRISC ra                               4.96u  0.27s  2.72g
2028              (spills = 85 reloads = 170)
2029           TOTAL                                   2375.26u  57.21s  48.00g
2030    
2031    As you can see the old cps spill module suffers from some serious
2032    performance problem.  But since I cannot decipher the old code fully,
2033    instead of patching the problems up, I'm reimplementing it
2034    with a different algorithm.  The new code is more modular,
2035    smaller when compiled, and substantially faster
2036    (O(n log n) time and O(n) space).  Timing of the new spill module:
2037    
2038    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2039          MLRISC ra                               0.25u  0.02s  0.15g
2040             (spills=1 reloads=3)
2041          TOTAL                                   7.74u  0.34s  1.62g
2042    
2043    Implementation details:
2044    
2045    As far as I can tell, the purpose of the CPS spill module is to make sure the
2046    number of live variables at any program point (the bandwidth)
2047    does not exceed a certain limit, which is determined by the
2048    size of the spill area.
2049    
2050    When the bandwidth is too large, we decrease the register pressure by
2051    packing live variables into spill records.  How we achieve this is
2052    completely different than what we did in the old code.
2053    
2054    First, there is something about the MLRiscGen code generator
2055    that we should be aware of:
2056    
2057    o MLRiscGen performs code motion!
2058    
2059       In particular, it will move floating point computations and
2060       address computations involving only the heap pointer to
2061       their use sites (if there is only a single use).
2062       What this means is that if we have a CPS record construction
2063       statement
2064    
2065           RECORD(k,vl,w,e)
2066    
2067       we should never count the new record address w as live if w
2068       has only one use (which is often the case).
2069    
2070       We should do something similar to floating point, but the transformation
2071       there is much more complex, so I won't deal with that.
2072    
2073    Secondly, there are now two new cps primops at our disposal:
2074    
2075     1. rawrecord of record_kind option
2076        This pure operator allocates some uninitialized storage from the heap.
2077        There are two forms:
2078    
2079         rawrecord NONE [INT n]  allocates a tagless record of length n
2080         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2081                                     and initializes the tag.
2082    
2083     2. rawupdate of cty
2084          rawupdate cty (v,i,x)
2085          Assigns to x to the ith component of record v.
2086          The storelist is not updated.
2087    
2088    We use these new primops for both spilling and increment record construction.
2089    
2090     1. Spilling.
2091    
2092        This is implemented with a linear scan algorithm (but generalized
2093        to trees).  The algorithm will create a single spill record at the
2094        beginning of the cps function and use rawupdate to spill to it,
2095        and SELECT or SELp to reload from it.  So both spills and reloads
2096        are fine-grain operations.  In contrast, in the old algorithm
2097        "spills" have to be bundled together in records.
2098    
2099        Ideally, we should sink the spill record construction to where
2100        it is needed.  We can even split the spill record into multiple ones
2101        at the places where they are needed.  But CPS is not a good
2102        representation for global code motion, so I'll keep it simple and
2103        am not attempting this.
2104    
2105     2. Incremental record construction (aka record splitting).
2106    
2107        Long records with many component values which are simulatenously live
2108        (recall that single use record addresses are not considered to
2109         be live) are constructed with rawrecord and rawupdate.
2110        We allocate space on the heap with rawrecord first, then gradually
2111        fill it in with rawupdate.  This is the technique suggested to me
2112        by Matthias.
2113    
2114        Some restrictions on when this is applicable:
2115        1. It is not a VECTOR record.  The code generator currently does not handle
2116           this case. VECTOR record uses double indirection like arrays.
2117        2. All the record component values are defined in the same "basic block"
2118           as the record constructor.  This is to prevent speculative
2119           record construction.
2120    
2121    ----------------------------------------------------------------------
2122    Name: Allen Leung
2123    Date: 2002/02/22 01:02:00 EST
2124    Tag: leunga-20020222-mlrisc-tools
2125    
2126    Minor bug fixes in the parser and rewriter
2127    
2128    ----------------------------------------------------------------------
2129    Name: Allen Leung
2130    Date: 2002/02/21 20:20:00 EST
2131    Tag: leunga-20020221-peephole
2132    
2133    Regenerated the peephole files.  Some contained typos in the specification
2134    and some didn't compile because of pretty printing bugs in the old version
2135    of 'nowhere'.
2136    
2137    ----------------------------------------------------------------------
2138    Name: Allen Leung
2139    Date: 2002/02/19 20:20:00 EST
2140    Tag: leunga-20020219-mlrisc-tools
2141    Description:
2142    
2143       Minor bug fixes to the mlrisc-tools library:
2144    
2145       1.  Fixed up parsing colon suffixed keywords
2146       2.  Added the ability to shut the error messages up
2147       3.  Reimplemented the pretty printer and fixed up/improved
2148           the pretty printing of handle and -> types.
2149       4.  Fixed up generation of literal symbols in the nowhere tool.
2150       5.  Added some SML keywords to to sml.sty
2151    
2152    ----------------------------------------------------------------------
2153    Name: Matthias Blume
2154    Date: 2002/02/19 16:20:00 EST
2155    Tag: blume-20020219-cmffi
2156    Description:
2157    
2158    A wild mix of changes, some minor, some major:
2159    
2160    * All C FFI-related libraries are now anchored under $c:
2161        $/c.cm      --> $c/c.cm
2162        $/c-int.cm  --> $c/internals/c-int.cm
2163        $/memory.cm --> $c/memory/memory.cm
2164    
2165    * "make" tool (in CM) now treats its argument pathname slightly
2166      differently:
2167        1. If the native expansion is an absolute name, then before invoking
2168           the "make" command on it, CM will apply OS.Path.mkRelative
2169           (with relativeTo = OS.FileSys.getDir()) to it.
2170        2. The argument will be passed through to subsequent phases of CM
2171           processing without "going native".  In particular, if the argument
2172           was an anchored path, then "make" will not lose track of that anchor.
2173    
2174    * Compiler backends now "know" their respective C calling conventions
2175      instead of having to be told about it by ml-nlffigen.  This relieves
2176      ml-nlffigen from one of its burdens.
2177    
2178    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
2179    
2180    * Export C_DEBUG and C_Debug from $c/c.cm.
2181    
2182    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
2183      subtyping relationship between incomplete pointers and their complete
2184      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
2185      with the convention of instantiating 'o with ('t, 'c) obj whenever
2186      the pointer target type is complete.  In the incomplete case, 'o
2187      will be instantiated with some "'c iobj" -- a type obtained by
2188      using one of the functors PointerToIncompleteType or PointerToCompleteType.
2189    
2190      Operations that work on both incomplete and complete pointer types are
2191      typed as taking an 'o ptr while operations that require the target to
2192      be known are typed as taking some ('t, 'c) obj ptr.
2193    
2194      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
2195      where void is an eqtype without any values.  This makes it possible
2196      to work on voidptr values using functions meant to operate on light
2197      incomplete pointers.
2198    
2199    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
2200      been vastly simplified.
2201    
2202    ----------------------------------------------------------------------
2203    Name: Matthias Blume
2204    Date: 2002/02/19 10:48:00 EST
2205    Tag: blume-20020219-pqfix
2206    Description:
2207    
2208    Applied Chris Okasaki's bug fix for priority queues.
2209    
2210    ----------------------------------------------------------------------
2211    Name: Matthias Blume
2212    Date: 2002/02/15 17:05:00
2213    Tag: Release_110_39
2214    Description:
2215    
2216    Last-minute retagging is becoming a tradition... :-(
2217    
2218    This is the working release 110.39.
2219    
2220    ----------------------------------------------------------------------
2221    Name: Matthias Blume
2222    Date: 2002/02/15 16:00:00 EST
2223    Tag: Release_110_39-orig
2224    Description:
2225    
2226    Working release 110.39.  New bootfiles.
2227    
2228    (Update: There was a small bug in the installer so it wouldn't work
2229    with all shells.  So I retagged. -Matthias)
2230    
2231    ----------------------------------------------------------------------
2232    Name: Matthias Blume
2233    Date: 2002/02/15 14:17:00 EST
2234    Tag: blume-20020215-showbindings
2235    Description:
2236    
2237    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
2238    the latter can be useful for exploring what bindings are available at
2239    the interactive prompt.  (The first function returns only the list
2240    of symbols that are really bound, the second prints those but also the
2241    ones that CM's autoloading mechanism knows about.)
2242    
2243    ----------------------------------------------------------------------
2244    Name: Matthias Blume
2245    Date: 2002/02/15 12:08:00 EST
2246    Tag: blume-20020215-iptrs
2247    Description:
2248    
2249    Two improvements to ml-nlffigen:
2250    
2251      1. Write files only if they do not exist or if their current contents
2252         do not coincide with what's being written.  (That is, avoid messing
2253         with the time stamps unless absolutely necessary.)
2254    
2255      2. Implement a "repository" mechanism for generated files related
2256         to "incomplete pointer types".   See the README file for details.
2257    
2258    ----------------------------------------------------------------------
2259    Name: Matthias Blume
2260    Date: 2002/02/14 11:50:00 EST
2261    Tag: blume-20020214-quote
2262    Description:
2263    
2264    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
2265    because of the new and improved tag generation scheme.  (Thanks to Allen
2266    Leung for pointing it out.)
2267    
2268    ----------------------------------------------------------------------
2269    Name: Lal George
2270    Date: 2002/02/14 09:55:27 EST 2002
2271    Tag: george-20020214-isabelle-bug
2272    Description:
2273    
2274    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
2275    of Isabelle on the x86.
2276    
2277    From Allen:
2278    -----------
2279     I've found the problem:
2280    
2281         in ra-core.sml, I use the counter "blocked" to keep track of the
2282         true number of elements in the freeze queue.  When the counter goes
2283         to zero, I skip examining the queue.  But I've messed up the
2284         bookkeeping in combine():
2285    
2286             else ();
2287             case !ucol of
2288               PSEUDO => (if !cntv > 0 then
2289                     (if !cntu > 0 then blocked := !blocked - 1 else ();
2290                                        ^^^^^^^^^^^^^^^^^^^^^^^
2291                      moveu := mergeMoveList(!movev, !moveu)
2292                     )
2293                  else ();
2294    
2295         combine() is called to coalesce two nodes u and v.
2296         I think I was thinking that if the move counts of u and v are both
2297         greater than zero then after they are coalesced then one node is
2298         removed from the freeze queue.  Apparently I was thinking that
2299         both u and v are of low degree, but that's clearly not necessarily true.
2300    
2301    
2302    02/12/2002:
2303        Here's the patch.  HOL now compiles.
2304    
2305        I don't know how this impact on performance (compile
2306        time or runtime).  This bug caused the RA (especially on the x86)
2307        to go thru the potential spill phase when there are still nodes on the
2308        freeze queue.
2309    
2310    
2311    
2312    
2313    ----------------------------------------------------------------------
2314    Name: Matthias Blume
2315    Date: 2002/02/13 22:40:00 EST
2316    Tag: blume-20020213-fptr-rtti
2317    Description:
2318    
2319    Fixed a bug in ml-nlffigen that was introduced with one of the previous
2320    updates.
2321    
2322    ----------------------------------------------------------------------
2323    Name: Matthias Blume
2324    Date: 2002/02/13 16:41:00 EST
2325    Tag: blume-20020213-cmlpq
2326    Description:
2327    
2328    Added new priority queue export symbols (which have just been added to
2329    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
2330    not compile and the installer would choke.)
2331    
2332    ----------------------------------------------------------------------
2333    Name: Matthias Blume
2334    Date: 2002/02/13 16:15:00 EST
2335    Tag: blume-20020213-various
2336    Description:
2337    
2338    1. More tweaks to ml-nlffigen:
2339    
2340       - better internal datastructures (resulting in slight speedup)
2341       - "-match" option requires exact match
2342       - "localized" gensym counters (untagged structs/unions nested within
2343         other structs/unions or within typedefs get a fresh counter; their
2344         tag will be prefixed by a concatenation of their parents' tags)
2345       - bug fixes (related to calculation of transitive closure of types
2346         to be included in the output)
2347    
2348    2. Minor Basis updates:
2349    
2350       - added implementations for List.collate and Option.app
2351    
2352    ----------------------------------------------------------------------
2353    Name: Matthias Blume
2354    Date: 2002/02/11 15:55:00 EST
2355    Tag: blume-20020211-gensym
2356    Description:
2357    
2358    Added a "-gensym" option to command line of ml-nlffigen.  This can be
2359    used to specify a "stem" -- a string that is inserted in all "gensym'd"
2360    names (ML structure names that correspond to unnamed C structs, unions,
2361    and enums), so that separate runs of ml-nlffigen do not clash.
2362    
2363    ----------------------------------------------------------------------
2364    Name: Matthias Blume
2365    Date: 2002/02/11 12:05:00 EST
2366    Tag: blume-20020211-gensml
2367    Description:
2368    
2369    A quick fix for a problem with GenSML (in the pgraph-util library):
2370    Make generation of toplevel "local" optional.  (Strictly speaking,
2371    signature definitions within "local" are not legal SML.)
2372    
2373    Other than that: updates to INSTALL and cm/TODO.
2374    
2375    ----------------------------------------------------------------------
2376    Name: Matthias Blume
2377    Date: 2002/02/08 15:00:00 EST
2378    Tag: blume-20020208-uniquepid
2379    Description:
2380    
2381    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
2382    
2383    1. The installer (config/install.sh) has gotten smarter:
2384    
2385         - Configuration options are a bit easier to specify now
2386           (in config/targets).
2387         - Bug in recognizing .tar.bz2 files fixed.
2388         - Installer automatically resolves dependencies between
2389           configuration options (e.g., if you ask for eXene, you will
2390           also get cml -- regardless whether you asked for it or not).
2391         - Installer can run in "quieter mode" by setting the environment
2392           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
2393           "completely silent", though.
2394         - Build HashCons library as part of smlnj-lib.
2395    
2396    2. A new scheme for assigning persistent identifiers to compilation
2397       units (and, by extension, to types etc.) has been put into place.
2398       This fixes a long-standing bug where types and even dynamic values
2399       can get internally confused, thereby compromising type safety
2400       (abstraction) and dynamic correctness.  See
2401    
2402         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
2403    
2404       for an example of how things could go wrong until now.
2405    
2406       The downside of the new scheme is that pids are not quite as
2407       persistent as they used to be: CM will generate a fresh pid
2408       for every compilation unit that it thinks it sees for the first
2409       time.  That means that if you compile starting from a clean, fresh
2410       source tree at two different times, you end up with different
2411       binaries.
2412    
2413       Cutoff recompilation, however, has not been compromised because
2414       CM keeps pid information in special caches between runs.
2415    
2416    ----------------------------------------------------------------------
2417    Name: Lal George
2418    Date: 2002/02/07 15:34:13 EST 2002
2419    Tag: <none>
2420    Description:
2421    
2422    Compilers that generate assembly code may produce  global labels
2423    whose value is resolved at link time. The various peephole optimization
2424    modules did not take this in account.
2425    
2426    TODO. The Labels.addrOf function should really return an option
2427    type so that clients are forced to deal with this issue, rather
2428    than an exception being raised.
2429    
2430    ----------------------------------------------------------------------
2431    Name: Lal George
2432    Date: 2002/02/06 13:55:02 EST
2433    Tag: george-20020206-ra-breakup
2434    Description:
2435    
2436    1. A bug fix from Allen.
2437    
2438        A typo causes extra fstp %st(0)'s to be generated at compensation
2439        edges, which might cause stack underflow traps at runtime.  This
2440        occurs in fft where there are extraneous fstps right before the 'into'
2441        trap instruction (in this case they are harmless since none of the
2442        integers overflow.)
2443    
2444    2. Pulled out various utility modules that were embedded in the modules
2445       of the register allocator. I need these modules for other purposes, but
2446       they are not complete enough to put into a library (just yet).
2447    ----------------------------------------------------------------------
2448    Name: Matthias Blume
2449    Date: 2002/01/31 16:05:00 EST
2450    Tag: blume-20020131-sparc-ccalls
2451    Description:
2452    
2453    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
2454       of extra stack space by mistake.  Fixed.
2455    
2456    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
2457    
2458    ----------------------------------------------------------------------
2459    Name: Allen Leung
2460    Date: 2002/01/30
2461    Tag: leunga-20020130-nowhere-bug-fix
2462    Description:
2463    
2464       MLRISC bug fixes:
2465       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
2466       2. MachineInt.fromString was negating its value.
2467    
2468    ----------------------------------------------------------------------
2469    Name: Matthias Blume
2470    Date: 2002/01/29
2471    Tag: blume-20020129-INSTALL
2472    Description:
2473    
2474    - Added somewhat detailed installation instructions (file INSTALL).
2475    - Fixed curl-detection bug in config/install.sh.
2476    - It is now possible to select the URL getter using the URLGETTER
2477      environment variable:
2478    
2479          not set / "unknown"      --> automatic detection (script tries wget,
2480                                       curl, and lynx)
2481          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
2482                                       how to properly invoke them)
2483          other                    --> use $URLGETTER directly, it must take
2484                                       precisely two command-line arguments
2485                                       (source URL and destination file name)
2486    
2487    ----------------------------------------------------------------------
2488    Name: Matthias Blume
2489    Date: 2002/01/28
2490    Tag: blume-20020128-sparc-ccalls
2491    Description:
2492    
2493    - Fixed problem with calculation of "used" registers in sparc-c-calls.
2494    - Make use of the allocParam argument in sparc-c-calls.
2495    
2496    ----------------------------------------------------------------------
2497    Name: Matthias Blume
2498    Date: 2002/01/28
2499    Tag: blume-20020128-allocParam
2500    Description:
2501    
2502    John Reppy:  Changes c-calls API to accept client-callback for
2503    allocating extra stack space.
2504    me: Corresponding changes to mlriscGen (using a dummy argument that
2505        does not change the current behavior).
2506    
2507    ----------------------------------------------------------------------
2508    Name: Matthias Blume
2509    Date: 2002/01/28 12:00:00
2510    Tag: Release_110_38
2511    Description:
2512    
2513    This time for real!!!
2514    
2515    ----------------------------------------------------------------------
2516    Name: Matthias Blume
2517    Date: 2002/01/28 10:56:00 EST
2518    Tag: blume-20020128-retraction
2519    Description:
2520    
2521    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
2522       with blume-Release_110_38-retracted.)
2523    
2524    1. Fixed a problem with incorrect rounding modes in real64.sml.
2525       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
2526    
2527    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
2528       and enums fixed.  The naming of corresponding ML identifiers should
2529       now be consistent again.
2530    
2531    ----------------------------------------------------------------------
2532    Name: Allen Leung
2533    Date: 2002/01/27
2534    Tag: leunga-20020127-nowhere
2535    Description:
2536    
2537       Added a target called nowhere in the configuration scripts.
2538       Enabling this will build the MLRISC 'nowhere' tool (for translating
2539       programs with where-clauses into legal SML code) during installation.
2540    
2541    ----------------------------------------------------------------------
2542    Name: Matthias Blume
2543    Date: 2002/01/25 21:27:00 EST
2544    Tag: blume-Release_110_38-retracted
2545    Description:
2546    
2547    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
2548    
2549    README will be added later.
2550    
2551    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
2552    (Release_110_38) removed.  Reason: Last-minute bug fixes.
2553    
2554    ----------------------------------------------------------------------
2555    Name: Matthias Blume
2556    Date: 2002/01/25
2557    Tag: blume-20020125-ffi
2558    Description:
2559    
2560    A large number of tweaks and improvements to ml-nlffi-lib and
2561    ml-nlffigen:
2562    
2563       - ML represenation types have been streamlined
2564       - getter and setter functions work with concrete values, not abstract
2565         ones where possible
2566       - ml-nlffigen command line more flexible (see README file there)
2567       - some bugs have been fixed (hopefully)
2568    
2569    ----------------------------------------------------------------------
2570    Name: Lal George
2571    Date: 2002/01/24
2572    Tag: george-20020124-risc-ra-interface
2573    Description:
2574    
2575       There is a dramatic simplification in the interface to the
2576       register allocator for RISC architectures as a result of making
2577       parallel copy instructions explicit.
2578    
2579    ----------------------------------------------------------------------
2580    Name: Matthias Blume
2581    Date: 2002/01/22
2582    Tag: blume-20020122-x86-ccalls
2583    Description:
2584    
2585    Bug fix for c-calls on x86 (having to do with how char- and
2586    short-arguments are being handled).
2587    
2588    ----------------------------------------------------------------------
2589    Name: Matthias Blume
2590    Date: 2002/01/21
2591    Tag: blume-20020121-ff
2592    Description:
2593    
2594    Another day of fiddling with the FFI...
2595    
2596    1. Bug fix/workaround:  CKIT does not complain about negative array
2597       dimensions, so ml-nlffigen has to guard itself against this possibility.
2598       (Otherwise a negative dimension would send it into an infinite loop.)
2599    
2600    2. Some of the abstract types (light objects, light pointers, most "base"
2601       types) in structure C are now eqtypes.
2602    
2603    3. Added constructors and test functions for NULL function pointers.
2604    
2605    ----------------------------------------------------------------------
2606    Name: Matthias Blume
2607    Date: 2002/01/18
2608    Tag: blume-20020118-ready-for-new-release
2609    Description:
2610    
2611    Made config/srcarchiveurl point to a new place.  (Will provide boot
2612    files shortly.)
2613    
2614    Maybe we christen this to be 110.38?
2615    
2616    ----------------------------------------------------------------------
2617    Name: Matthias Blume
2618    Date: 2002/01/18
2619    Tag: blume-20020118-more-ffifiddle
2620    Description:
2621    
2622    Today's FFI fiddling:
2623    
2624      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
2625        C.Set.* that use concrete (MLRep.*) arguments and results instead
2626        of abstract ones.
2627    
2628      - Provided word-style bit operations etc. for "int" representation
2629        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
2630        and Long.
2631    
2632    ----------------------------------------------------------------------
2633    Name: Matthias Blume
2634    Date: 2002/01/18
2635    Tag: blume-20020118-use-x86-fp
2636    Description:
2637    
2638    Now that x86-fast-fp seems to be working, I turned it back on again
2639    by default.  (Seems to work fine now, even with the FFI.)
2640    
2641    Other than that, I added some documentation about the FFI to
2642    src/ml-nlffigen/README and updated the FFI test examples in
2643    src/ml-nlffi-lib/Tests/*.
2644    
2645    ----------------------------------------------------------------------
2646    Name: Allen Leung
2647    Date: 2002/01/17
2648    Tag: leunga-20020117-x86-fast-fp-call
2649    Description:
2650    
2651       1. Fixed a problem with handling return fp values when x86's fast fp
2652          mode is turned on.
2653    
2654       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
2655          of %f32.
2656    
2657       3. Added a constructor INT32lit to the ast of MLRISC tools.
2658    
2659    ----------------------------------------------------------------------
2660    Name: Matthias Blume
2661    Date: 2002/01/16
2662    Tag: blume-20020116-ffifiddle
2663    Description:
2664    
2665    More fiddling with the FFI interface:
2666    
2667     - Make constness 'c instead of rw wherever possible.  This eliminates
2668       the need for certain explicit coercions.  (However, due to ML's
2669       value polymorphism, there will still be many cases where explicit
2670       coercions are necessary.  Phantom types are not the whole answer
2671       to modeling a subtyping relationship in ML.)
2672    
2673     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
2674    
2675     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
2676       clumsy workaround for varargs problem.)
2677    
2678  ----------------------------------------------------------------------  ----------------------------------------------------------------------
2679  Name: Lal George  Name: Lal George
2680  Date: 2002/01/15  Date: 2002/01/15
# Line 401  Line 3065 
3065    
3066  ----------------------------------------------------------------------  ----------------------------------------------------------------------
3067  Name: Matthias Blume  Name: Matthias Blume
 >>>>>>> 1.169  
3068  Date: 2001/09/18 15:35:00 EDT  Date: 2001/09/18 15:35:00 EDT
3069  Tag: blume-20010918-readme11036  Tag: blume-20010918-readme11036
3070  Description:  Description:

Legend:
Removed from v.1016  
changed lines
  Added in v.1490

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