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 977, Wed Nov 14 16:53:16 2001 UTC revision 1614, Mon Aug 16 01:21:56 2004 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Allen Leung (leunga (at) reservoir (dot) com)
17    Date: 2004/08/15 21:21:00 EST
18    Tag: leunga-110_48-udgraph
19    Description:
20    
21    Another bug fix from Carl Hauser:
22    
23    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
24    > 48c48
25    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
26    > e::L)
27    > ---
28    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
29    > Without this, any deletion of an edge in an undirected graph does severe
30    > violence to the graph.
31    
32    ----------------------------------------------------------------------
33    Name: Allen Leung (leunga (at) reservoir (dot) com)
34    Date: 2004/08/10 23:35:00 EST
35    Tag: leunga-110_48-ppc
36    Description:
37    
38       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
39    
40    ----------------------------------------------------------------------
41    Name: Matthias Blume (blume (at) tti - c (dot) org)
42    Date: 2004/10/04 12:00:00 CDT
43    Tag: Release_110_48
44    Description:
45    
46    New working version (110.48).  NEW BOOTFILES!
47    
48    ----------------------------------------------------------------------
49    Name: Allen Leung (leunga (at) reservoir (dot) com)
50    Date: 2004/08/09 12:21:00 EST
51    Tag: leunga-110_47-dijsktra
52    Description:
53    
54    Bug fix from Carl Hauser:
55    
56    single_source_shortest_paths in dijkstra.sml was observed to get wrong
57    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
58    
59    The problem is that following the expression A.update(dist,s,Num.zero)
60    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
61    
62    ----------------------------------------------------------------------
63    Name: Matthias Blume (blume (at) tti - c (dot) org)
64    Date: 2004/08/06 18:10:00 CDT
65    Tag: blume-20040806-cmdline
66    Description:
67    
68    Fiddled with handling of command-line options:
69    
70      * sml now quits after processing the command line
71        if -H, -S, -h<n>, or -s<n> appears as the last
72        command-line argument
73      * a new option -q terminates the session when encountered on
74        the command line; subsequent arguments will be ignored
75      * bug fixes: short (erroneous) arguments are no longer ignored
76        completely
77    
78    ----------------------------------------------------------------------
79    Name: Allen Leung (leunga (at) reservoir (dot) com)
80    Date: 2004/08/04 18:17:00 EST
81    Tag: leunga-110_47-ppc-ibm-asm
82    Description:
83    
84      - Added minimal IBM assembly syntax support for PowerPC.
85    
86      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
87        been altered in the runtime.  We need an automatic way to keep the file
88        in sync.
89    
90    ----------------------------------------------------------------------
91    Name: Matthias Blume (blume (at) tti - c (dot) org)
92    Date: 2004/08/04 14:00:00 CDT
93    Tag: Release_110_47
94    Description:
95    
96    New working version (110.47).  NEW BOOTFILES!
97    
98    ----------------------------------------------------------------------
99    Name: Matthias Blume (blume (at) tti - c (dot) org)
100    Date: 2004/08/03 14:25:00 CDT
101    Tag: blume-20040803-callingconv
102    Description:
103    
104    Added low-level support for choosing C calling conventions by
105    twiddling the type of rawccall.  (See
106    src/compiler/Semant/types/cproto.sml for details.)
107    
108    ----------------------------------------------------------------------
109    Name: Matthias Blume (blume (at) tti - c (dot) org)
110    Date: 2004/08/02 15:55:00 CDT
111    Tag: blume-20040802-backout
112    Description:
113    
114    Backed out of change to win32-filesys.c.  The earlier patch to
115    get_file_time caused CM to produce files with the wrong time stamp.
116    
117    ----------------------------------------------------------------------
118    Name: Matthias Blume (blume (at) tti - c (dot) org)
119    Date: 2004/08/02 14:45:00 CDT
120    Tag: blume-20040802-nlffi-win32
121    Description:
122    
123    Added NLFFI support for Win32, adapted from a patch provided by David
124    Hansel.  This is currently completely untested.  Also, the issue
125    concerning stdcall vs. ccall is still unresolved.
126    
127    ----------------------------------------------------------------------
128    Name: Matthias Blume (blume (at) tti - c (dot) org)
129    Date: 2004/07/30 17:55:00 CDT
130    Tag: blume-20040730-various
131    Description:
132    
133    Gearing up towards 110.47...
134    
135    - various minor bugfixes to ml-nlffigen
136    - a beginning of a manual for nlffi
137    
138    - eliminated 'export name=value' in config/install.sh as this does
139      not work with certain versions of /bin/sh
140      (Thanks to David King at Motorola for catching this.)
141    
142    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
143      - added a test for tm==NULL to gmtime.c and localtime.c
144      - applied patch for incorrect GetFileTime under win32
145      - toSeconds -> toMilliseconds in Win32/win32-process.sml
146    
147    ----------------------------------------------------------------------
148    Name: Matthias Blume (blume (at) tti - c (dot) org)
149    Date: 2004/07/21 18:20:00 CDT
150    Tag: blume-20040721-nlffigen
151    Description:
152    
153    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
154      for a typedef to an incomplete type, but leave out the "typ" member.
155      (This is just for consistency.)
156    - Started to produce what is supposed to become better (i.e., comprehensive)
157      documentation of what ml-nlffigen does and produces.
158    
159    ----------------------------------------------------------------------
160    Name: Matthias Blume (blume (at) tti - c (dot) org)
161    Date: 2004/07/14 16:25:00 CDT
162    Tag: blume-20040714-union
163    Description:
164    
165    Added C_UNION to c-calls/c-types.sml and updated the machinery
166    (ml-nlffigen, cproto.sml) that conveys C function interface
167    information to the code generator.
168    
169    However, the actual architecture-specific implementation of function
170    arguments and results that are C unions is still not implemented.
171    
172    ----------------------------------------------------------------------
173    Name: Allen Leung (leunga (at) reservoir (dot) com)
174    Date: 2004/07/14 14:38:00 EST
175    Tag: leunga-110_46_1-ppc-lwzu
176    Description:
177    
178        Added these instructions to the PowerPC architecture:
179        LBZU(X), LHZU(X), LWZU(X),
180        STWU(X), STFDU, STFSU
181    
182        etc...
183    
184        Note: I haven't added their instruction encoding into the description.
185    
186    ----------------------------------------------------------------------
187    Name: Allen Leung (leunga (at) reservoir (dot) com)
188    Date: 2004/07/13 15:04:00 EST
189    Tag: leunga-110_46_1-ppc-lwarx
190    Description:
191    
192       Added the two instructions LWARX and STWCX to the PowerPC
193    instruction set.
194    
195       A (untested) rewrite of loop-structure.sml.  The old version
196    is completely broken.
197    
198    ----------------------------------------------------------------------
199    Name: Matthias Blume (blume (at) tti - c (dot) org)
200    Date: 2004/07/13 13:50:00 CDT
201    Tag: blume-20040713-nlffi
202    Description:
203    
204    - use paramAlloc to report c-calls with too many arguments
205      (for PPC version where parameter area is pre-allocated)
206    - added ccall_maxargspace to machspec (to implement the above)
207    - made "make" commend in CM's "make" tool configurable
208    - added option (default: on) for passing the name of the SML/NJ's "bin"
209      directory to "make"; the call looks like this:
210    
211         make <options> SMLNJ_BINDIR=<dir> <target>
212    
213      This can be used by the Makefile to, e.g., pick the "right" version
214      of ml-nlffigen.
215    - minor code tweaks
216    
217    ----------------------------------------------------------------------
218    Name: Matthias Blume (blume (at) tti - c (dot) org)
219    Date: 2004/07/12 22:50:00 CDT
220    Tag: blume-110_46_1-macosx-nlffi
221    Description:
222    
223    NLFFI under Mac OS X now working (sort of).  This is largely untested,
224    though.
225    
226    Note:  1. You have to make a new, clean build of the runtime system.
227           2. There are new BOOTFILES, you have to use them!
228              (Doing the bootstrap process yourself would be *very* painful!
229               If you absolutely have to do it, build the system under
230               a different architecture and then cross-compile.)
231    
232    Version bumped to 110.46.1 to account for runtime data format changes.
233    
234    ----------------------------------------------------------------------
235    Name: Matthias Blume (blume (at) tti - c (dot) org)
236    Date: 2004/06/18 14:30:00 CDT
237    Tag: blume-20040618-unix
238    Description:
239    
240    Changed the implementation of structure Unix so that the same stream
241    is returned every time one of the {text,bin}{In,Out}streamOf functions
242    is invoked on the same proc.  This is not what the spec currently
243    says -- although IMO it arguably should.  (See discussion below.)
244    
245    ----------------------------------------------------------------------
246    Name: Matthias Blume (blume (at) tti - c (dot) org)
247    Date: 2004/06/17 18:15:00 CDT
248    Tag: Release_110_46
249    Description:
250    
251    New working version (110.46).  NEW BOOTFILES!
252    
253    ----------------------------------------------------------------------
254    Name: Matthias Blume (blume (at) tti - c (dot) org)
255    Date: 2004/06/17 17:20:00 CDT
256    Tag: blume-20040617-timer-unix
257    Description:
258    
259    Changed the interface of structures Timer and Unix to match the most
260    recent Basis spec.
261    
262    In the case of Unix there still seems to be an open/weird issue:
263    
264       The {text,bin}{In,Out}streamOf functions are supposed to create
265       fresh streams whenever they are called -- as opposed to have them
266       return the same stream every time.  This design is supposed to
267       prevent space leaks caused by proc values hanging on to streams.
268    
269       The reap function, on the other hand, is supposed to close the
270       streams.  This cannot be done without having a handle on the
271       stream in proc after all...
272    
273    I took the liberty to implement the following stopgap solution:
274    
275      The proc value hangs on to the most recently created stream(s).
276      Reap closes those.  If either or both of the two streams hadn't
277      been created at all yet, then reap will close the corresponding
278      file descriptors directly.
279    
280    PS: I don't understand the original space leak argument anymore.  If
281    a proc hangs on to the imperative stream, then I/O operations on those
282    will advance the state of the cached stream and avoid the space leak.
283    
284    ----------------------------------------------------------------------
285    Name: Matthias Blume (blume (at) tti - c (dot) org)
286    Date: 2004/05/28 16:45:00 CDT
287    Tag: blume-20040528-basis
288    Description:
289    
290    Added signature PACK_REAL and exported functor PrimIO.
291    
292    ----------------------------------------------------------------------
293    Name: Matthias Blume (blume (at) tti - c (dot) org)
294    Date: 2004/05/25 16:00:00 CDT
295    Tag: blume-20040525-group-owner
296    Description:
297    
298    CM now ignores (but still accepts) the "owner" information in group
299    descriptions.  The owner of a group is its next enclosing
300    library. Each group must have a unique owner.  (There is a virtual
301    "toplevel" library that owns groups which are not nested within a real
302    library.)  Previously, each group had to explicitly declare its owner,
303    and CM would check that such a declaration is correct.  The new scheme
304    is to have CM check that for each group there is precisely one owning
305    library.
306    
307    The advantage of the new scheme is that the programmer no longer needs
308    to maintain the somewhat annoying owner information.  The downside is
309    that CM cannot enforce the ownership rule across multiple runs of
310    CM.make.  Fortunately, enclosing the same group in two different
311    libraries A and B which are not part of the same program does not
312    cause real problems.
313    
314    ----------------------------------------------------------------------
315    Name: Matthias Blume (blume (at) tti - c (dot) org)
316    Date: 2004/05/20 16:00:00 CDT
317    Tag: blume-20040520-win32
318    Description:
319    
320    Made the win32 version work again.  (Strangely, a misplaced comma had
321    slipped into win32-process.c which prevented the runtime from being
322    compiled correctly.)
323    
324    Also, included a minor addition to ml-build.bat analogous to what was
325    done in blume-20040519-ml-build.
326    
327    ----------------------------------------------------------------------
328    Name: Matthias Blume (blume (at) tti - c (dot) org)
329    Date: 2004/05/19 22:10:00 CDT
330    Tag: blume-20040519-ml-build
331    Description:
332    
333    Arranged for ml-build to clean up after itself a little bit better.
334    The script generates a temporary SML source file and compiles it using
335    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
336    gets rid of those at the end, so they don't accumulate under .cm.
337    
338    This required a minor change to install.sh because the name of the
339    metadata directory (default: .cm) is actually configurable at
340    installation time.
341    
342    ----------------------------------------------------------------------
343    Name: Matthias Blume (blume (at) tti - c (dot) org)
344    Date: 2004/05/18 15:50:00 CDT
345    Tag: blume-20040518-mkreader
346    Description:
347    
348    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
349    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
350    
351    ----------------------------------------------------------------------
352    Name: Matthias Blume (blume (at) tti - c (dot) org)
353    Date: 2004/05/11 14:35:00 CDT
354    Tag: blume-20040511-win32sock
355    Description:
356    
357    Added previously missing support for many socket-related functions
358    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
359    for the voluminous patch!
360    
361    (I have not tested this patch under win32 yet.)
362    
363    Here is David's e-mail:
364    
365    Hi,
366    
367    Attached to this email you find a diff against sml/nj 110.45
368    that will enable socket support under Windows.
369    
370    To apply the patch (using unix or cygwin)
371    1) gunzip runtime.diff.gz
372    2) "cd" into "src/runtime" in the source tree of a fresh
373       110.45 installation.
374    3) patch -p 1 < [your/path/to]runtime.diff
375    
376    The code compiles fine but has NOT yet been extensively tested.
377    I only ran a few tests for basic socket client functionality
378    (which worked fine).  Especially the functions that use ioctl
379    are not tested at all and might not work (see below).
380    
381    I implemented this since we want to move to a newer version of sml/nj
382    but need socket support in order to use it.  This is the first time I
383    even had a look at the sml/nj source,  so please review my changes
384    before making this part of the distribution!  Here are a few issues
385    that I think might be better for someone to solve who is more
386    familiar with the sml/nj source (and socket programming):
387    
388    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
389      exception since I could not figure out what the windows equivalent
390      of these functions is
391    
392    - In sockets-osdep.h there are a some #include statements that are
393      only used in a few files that include sockets-osdep.h
394    
395    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
396      does not process its return value since I don't know how to
397      report an error upwards.
398    
399    - It would probably be good to have a call to WSACleanup() when
400      the library is unloaded (if there is such a possibility).
401      Otherwise I think Windows will take care of this automatically
402      when the process finishes.
403    
404    - I used ioctlsocket() as a replacement for ioctl() but I have
405      no idea if that is actually the proper replacement on Windows.
406    
407    - All these issues are marked in the code by "FIXME" comments.
408    
409    We use sml/nj extensively in our products and are quite happy
410    with it.  I hope this contribution will help you.
411    
412    Keep up the good work!
413    
414    David
415    
416    ----------------------------------------------------------------------
417    Name: Matthias Blume (blume (at) tti - c (dot) org)
418    Date: 2004/05/11 14:20:00 CDT
419    Tag: blume-20040511-installml
420    Description:
421    
422    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
423    the patch.)
424    
425    ----------------------------------------------------------------------
426    Name: Matthias Blume (blume (at) tti - c (dot) org)
427    Date: 2004/05/11 14:05:00 CDT
428    Tag: blume-20040511-nlffi-netbsd
429    Description:
430    
431    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
432    the patch.)
433    
434    ----------------------------------------------------------------------
435    Name: Matthias Blume (blume (at) tti - c (dot) org)
436    Date: 2004/05/11 12:05:00 CDT
437    Tag: blume-20040511-exports
438    Description:
439    
440    As per request by Adam Chlipala <adam@hcoop.net>, extended various
441    export lists in compiler-related .cm-files.
442    
443    ----------------------------------------------------------------------
444    Name: Matthias Blume (blume (at) tti - c (dot) org)
445    Date: 2004/05/11 11:35:00 CDT
446    Tag: blume-20040511-allsource
447    Description:
448    
449    The installer now honors the "src-smlnj" target again, although its meaning
450    has changed from "all sources required for the compiler" to "all sources
451    the installer knows about".  In other words, if you enable "src-smlnj"
452    in the "targets" file, then the installer will pull in sources for
453    everything.  (Notice that this refers to source code only.  Compiled
454    code is still only installed for modules that were requested explicitly
455    or which are required for other modules that were requested explicitly.)
456    
457    ----------------------------------------------------------------------
458    Name: Matthias Blume (blume (at) tti - c (dot) org)
459    Date: 2004/04/23 17:40:00 CDT
460    Tag: blume-20040423-ieee-scan
461    Description:
462    
463    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
464    in the exponent calculation we get INF or ZERO (depending on the mantissa
465    and the sign of the exponent).
466    
467    ----------------------------------------------------------------------
468    Name: Matthias Blume (blume (at) tti - c (dot) org)
469    Date: 2004/04/23 10:40:00 CDT
470    Tag: blume-20040423-ml-build
471    Description:
472    
473    The ml-build script now terminates with a non-0 status when something
474    goes wrong.
475    
476    ----------------------------------------------------------------------
477    Name: Matthias Blume (blume (at) tti - c (dot) org)
478    Date: 2004/04/22 16:35:00 CDT
479    Tag: blume-20040422-Option
480    Description:
481    
482    Made exception Option to be the same as exception Option.Option
483    (as it should be).
484    
485    ----------------------------------------------------------------------
486    Name: Allen Leung (leunga (at) reservoir (dot) com)
487    Date: 2004/03/19 14:40:00 EST
488    Tag: leunga-20040319-cygwin-nlffi
489    Description:
490    
491    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
492    of SML/NJ.  The problem is that
493    
494       lib = dlopen(NULL, ...)
495       f   = dlsym(lib, "malloc");
496    
497    does not work on Windows unless we explicitly export symbols
498    such as 'malloc' during linking.  We fixed this by explicitly
499    exporting the required symbols with the magic gcc incantation:
500    
501        -Wl,--export-all cygwin.def
502    
503    where cygwin.def is a file containing all the symbols that
504    we wish to export.
505    
506    I suspect this is a Windows problem and we'll have to
507    do the same (somehow with windows compilers) when
508    we build the native win32 version with the system
509    calls LoadLibrary/GetProcAddress.
510    
511    ----------------------------------------------------------------------
512    Name: Matthias Blume (blume (at) tti - c (dot) org)
513    Date: 2004/03/04 16:35:00 CST
514    Tag: blume-20040304-intinf-fmt
515    Description:
516    
517    Fixed problem with IntInf.fmt (sign would show up on the right instead
518    of on the left for BIN, OCT, and HEX).
519    
520    ----------------------------------------------------------------------
521    Name: Matthias Blume (blume (at) tti - c (dot) org)
522    Date: 2004/03/04 11:25:00 CST
523    Tag: blume-20040304-symlinks
524    Description:
525    
526    Fixed problem with installer script (unix only) where bin/ml-yacc and
527    friends pointed (via symlinks) to absolute locations instead of just
528    .run-sml.  This was reported by Vesa A Norrman.
529    
530    ----------------------------------------------------------------------
531    Name: Matthias Blume (blume (at) tti - c (dot) org)
532    Date: 2004/02/13 14:50:00 CST
533    Tag: Release_110_45
534    Description:
535    
536    New working version (110.45). New bootfiles.
537    
538    ----------------------------------------------------------------------
539    Name: Matthias Blume (blume (at) tti - c (dot) org)
540    Date: 2004/01/26 15:15:15 CST
541    Tag: blume-20040126-toplevel
542    Description:
543    
544    Improved handling of exceptions at the interactive toplevel.
545    
546    ----------------------------------------------------------------------
547    Name: Matthias Blume (blume (at) tti - c (dot) org)
548    Date: 2004/01/26 11:25:00
549    Tag: blume-20040126-app
550    Description:
551    
552    Type of top-level "app" corrected.
553    Added code for setting vp_limitPtrMask to Win32-specific runtime.
554    
555    ----------------------------------------------------------------------
556    Name: Matthias Blume (blume (at) tti - c (dot) org)
557    Date: 2003/11/18 17:10 CST
558    Tag: blume-20031118-basis-fiddle
559    Description:
560    
561    - changed Timer interface to what might become the spec
562    - POSIX_FLAGS -> BIT_FLAGS according to spec
563    - some other minor discrepancies wrt. spec eliminated
564    
565    ----------------------------------------------------------------------
566    Name: Matthias Blume (blume (at) tti - c (dot) org)
567    Date: 2003/11/06 12:00:00 CST
568    Tag: Release_110_44
569    Description:
570    
571    New working version (110.44). New bootfiles.
572    
573    ----------------------------------------------------------------------
574    Name: Matthias Blume (blume (at) tti - c (dot) org)
575    Date: 2003/11/04 11:50:00 CST
576    Tag: blume-20031104-move-libraries
577    Description:
578    
579    Eliminated the "dont_move_libraries" directive in config/targets.
580    (The mechanism was broken and could not be fixed easily.  Moreover,
581    there does not seem to be any reason not to move all libraries into
582    lib during installation.  I originally implemented this directive as a
583    backward-compatibility feature when I first introduced the new CM.
584    Now that things have been stable for a long time and going back to the
585    old CM is not an option, there is no reason to keep it around.)
586    
587    ----------------------------------------------------------------------
588    Name: Matthias Blume (blume (at) tti - c (dot) org)
589    Date: 2003/11/03 16:00:00 CST
590    Tag: blume-20031103-installdir
591    Description:
592    
593    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
594    Richards for pointing out the problem and providing the solution.)
595    
596    ----------------------------------------------------------------------
597    Name: Matthias Blume (blume (at) tti - c (dot) org)
598    Date: 2003/10/01 17:05:00 CDT
599    Tag: blume-20031001-lal-mlrisc
600    Description:
601    
602    MLRISC bug fix from Lal.
603    
604    ----------------------------------------------------------------------
605    Name: Matthias Blume (blume (at) tti - c (dot) org)
606    Date: 2003/09/30 16:10:00 CDT
607    Tag: blume-20030930-primio-bat
608    Description:
609    
610    1. Added openVector, nullRd, and nullWr to PRIM_IO.
611    2. Improved .bat files (for Win32 port) to make things work under Win95.
612       (thanks to Aaron S. Hawley for this one)
613    
614    ----------------------------------------------------------------------
615    Name: Matthias Blume (blume (at) tti - c (dot) org)
616    Date: 2003/09/26 16:05:00 CDT
617    Tag: blume-20030926-wrappriv
618    Description:
619    
620    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
621    
622    ----------------------------------------------------------------------
623    Name: Matthias Blume (blume (at) tti - c (dot) org)
624    Date: 2003/09/26 15:00:00 CDT
625    Tag: blume-20030926-110_43_3
626    Description:
627    
628    - additional cleanup
629    - version number bump, NEW BOOTFILES
630    
631    ----------------------------------------------------------------------
632    Name: Matthias Blume (blume (at) tti - c (dot) org)
633    Date: 2003/09/26 12:00:00 CDT
634    Tag: blume-20030926-ppautoload
635    Description:
636    
637    I modified the read-eval-print loop so that the autoloader gets
638    invoked whenever the prettyprinter tries to look up a symbol that
639    is not currently defined in the toplevel environment but which
640    appears in CM's autoload registry.  As a result, we see far fewer of
641    those ?.Foo.Bar.xxx names in the prettyprinter's output.
642    
643    In addition to this I tried to clean up some pieces of the Basis
644    implementation (e.g., Socket, Word8Array) in order to prevent other
645    instances of these ?.Foo.Bar.xxx names from being printed.
646    
647    The mechanism that picks names for types still needs some work, though.
648    (Right now it seems that if there is a type A.t which is defined to
649    be B.u, but B is unavailable at toplevel, then A.t gets printed as
650    "?.B.u" although the perhaps more sensible solution would be to use
651    "A.t" in this case.  In other words, the prettyprinter should follow
652    a chain of DEFtycs not farther than there are corresponding toplevel
653    names in the current environment.)
654    
655    ----------------------------------------------------------------------
656    Name: Matthias Blume (blume (at) tti - c (dot) org)
657    Date: 2003/09/24 16:31:00 CDT
658    Tag: blume-20030924-installer
659    Description:
660    
661    Another installer tweak:  All the ML code for the installer is now
662    compiled during CMB.make and put into a little library called
663    $smlnj/installer.cm.  The installation then simply invokes
664    
665       sml -m $smlnj/installer.cm
666    
667    and everything happens automagically.
668    
669    Win32:  ML code senses value of environment variable SMLNJ_HOME.
670    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
671            and BINDIR.
672    
673    The new scheme guarantees that the ML code responsible for the installation
674    is in sync with the APIs of the main system.  Also, the installer is
675    somewhat faster because the installer script is precompiled.
676    
677    ----------------------------------------------------------------------
678    Name: Matthias Blume (blume (at) tti - c (dot) org)
679    Date: 2003/09/24 15:35:00 CDT
680    Tag: blume-20030924-synsock
681    Description:
682    
683    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
684    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
685    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
686    implementations of non-blocking operations from CML's Socket
687    structure.
688    
689    ----------------------------------------------------------------------
690    Name: Matthias Blume (blume (at) tti - c (dot) org)
691    Date: 2003/09/24 15:10:05 CDT
692    Tag: blume-20030924-sockets
693    Description:
694    
695    1. Fixed SOCKET API and implementation to match Basis spec.
696       This required changing the internal representation of sockets to one
697       that remembers (for each socket file descriptor) whether it is currently
698       blocking or non-blocking.  This state is maintained lazily (i.e., a system
699       call is made only if the state actually needs to change).
700    
701    2. OS-specific details of sockets were moved into separate files, thus
702       making it possible to unify the bulk of the socket implementations
703       between Unix and Win32.
704    
705    3. CML's socket API changed accordingly.
706       (Note that we need to remove non-blocking functions from this API
707       since they are redundant in the case of CML!)
708    
709    4. CML's socket implementation now makes use of non-blocking functions
710       provided by Basis, thus removing all OS-dependent code from this part
711       of CML.
712    
713    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
714    
715    ----------------------------------------------------------------------
716    Name: Matthias Blume (blume (at) tti - c (dot) org)
717    Date: 2003/09/22 12:10:00 CDT
718    Tag: blume-20030922-110_43_2
719    Description:
720    
721    Made a new interim version and bootfiles for developer's bootstrapping
722    convenience.
723    
724       110.43.2 -- NEW BOOTFILES
725    
726    ----------------------------------------------------------------------
727    Name: Matthias Blume (blume (at) tti - c (dot) org)
728    Date: 2003/09/19 15:55:00 CDT
729    Tag: blume-20030919-cmdir
730    Description:
731    
732    1. new-install.sh -> install.sh
733    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
734    3. tweaked installer so that another name instead of .cm can be chosen
735       at install time (by setting the CM_DIR_ARC environment variable
736       during installation); once installation is complete, the name is
737       fixed
738    
739    ----------------------------------------------------------------------
740    Name: Matthias Blume (blume (at) tti - c (dot) org)
741    Date: 2003/09/18 16:00:00 CDT
742    Tag: blume-20030918-110_43_1
743    Description:
744    
745    Made a new interim version and bootfiles for developer's bootstrapping
746    convenience.
747    
748       110.43.1 -- NEW BOOTFILES
749    
750    ----------------------------------------------------------------------
751    Name: Matthias Blume (blume (at) tti - c (dot) org)
752    Date: 2003/09/18 15:20:00 CDT
753    Tag: blume-20030918-misc
754    Description:
755    
756    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
757       this seems to be controversial at the moment) and used those in
758       Posix.ProcEnv.times.
759    
760    2. Added Time.{from,to}Nanoseconds to Time.
761    
762    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
763       For example, fromLargeInt never needs to look at more than 3 "big
764       digits" to get its 53 bits of precision.
765    
766    ----------------------------------------------------------------------
767    Name: Matthias Blume (blume (at) tti - c (dot) org)
768    Date: 2003/09/17 16:30:00 CDT
769    Tag: blume-20030917-real32-slices
770    Description:
771    
772    Added an entry to the primitive environment
773    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
774    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
775    
776    Removed some of the "magic" constants in real64.sml and replaced them
777    with code that generates these values from their corresponding
778    integer counterparts.
779    
780    Made all(?) the slice-related changes to the Basis and made everything
781    compile again...
782    
783    ----------------------------------------------------------------------
784    Name: Matthias Blume (blume (at) tti - c (dot) org)
785    Date: 2003/09/15 17:45:00 CDT
786    Tag: blume-20030915-rbase
787    Description:
788    
789    Fixed bug in Real.fromLargeInt.
790    
791    ----------------------------------------------------------------------
792    Name: Matthias Blume (blume (at) tti - c (dot) org)
793    Date: 2003/09/13 18:11:00 CDT
794    Tag: blume-20030913-libinstall
795    Description:
796    
797    Minor bugfix in config/libinstall (set anchor with path to
798    standalone tool after installing it, otherwise libraries that
799    need ml-lex or ml-yacc won't compile the first time the installer
800    runs).
801    
802    ----------------------------------------------------------------------
803    Name: Matthias Blume (blume (at) tti - c (dot) org)
804    Date: 2003/09/12 11:45:00 CDT
805    Tag: blume-20030912-various
806    Description:
807    
808     - fixed bug in Real.toLargeInt
809     - fixed bug in Posix.ProcEnv.times
810     - changed inputLine functions to return an option
811     - minor installer improvements / bugfixes
812     - changed default @SMLalloc parameter for x86/celeron to 64k
813    
814    ----------------------------------------------------------------------
815    Name: Matthias Blume (blume (at) tti - c (dot) org)
816    Date: 2003/09/09 22:00:00 CDT
817    Tag: Release_110_43
818    Description:
819    
820        New working release 110.43.  New bootfiles.
821    
822    ----------------------------------------------------------------------
823    Name: Matthias Blume (blume (at) tti - c (dot) org)
824    Date: 2003/09/09 19:20:00 CDT
825    Tag: blume-20030909-installer
826    Description:
827    
828    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
829    Modified config/install.bat to take advantage of it.  Also modified
830    config/install.sh (and called it config/new-install.sh) to take advantage
831    of it on Unix systems.  (The SML code is (supposed to be) platform-
832    independent.)
833    
834    The installer can now install everything under Win32
835    as well as under *nix as long as it compiles.
836    
837    Other changes:
838    
839      - made CML compile again under Win32
840      - made eXene compile under Win32 (by providing a fake structure UnixSock
841        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
842      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
843        same as type int;  under Win32 it isn't
844      - fixed some slice-related problems in the win32-specific parts of CML
845      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
846        (under Win32, the volume name is case-insensitive, and the
847        OS.Path code compares volume names for equality)
848    
849    ----------------------------------------------------------------------
850    Name: Matthias Blume (blume (at) tti - c (dot) org)
851    Date: 2003/09/08 11:55:00 CDT
852    Tag: blume-20030908-fullpath
853    Description:
854    
855    Made Win32 version of OS.FileSys.fullPath return current directory
856    when given an empty string.  This is what the spec says, and incidentally,
857    CM depends on it.  (CM otherwise goes into an infinite loop in certain
858    cases when presented with the name of a non-existing .cm file.)
859    
860    ----------------------------------------------------------------------
861    Name: Matthias Blume (blume (at) tti - c (dot) org)
862    Date: 2003/09/04 16:30:00 CDT
863    Tag: blume-20030905-slices-etc
864    Description:
865    
866      1. Changed interface to vectors and arrays in Basis to match
867         (draft) Basis spec.
868      2. Added signatures and implementations of slices according to
869         Basis spec.
870      3. Edited source code throughout the system to make it compile again
871         under 1. and 2.  (In some cases code had to be added to have it
872         match the new signatures.)
873      4. MLRISC should be backward-compatible: the copies of the originals
874         of files that needed to change under 3. were retained, the .cm files
875         check the compiler version number and use old versions when
876         appropriate.
877      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
878         dirstream -> string option (in accordance with Basis spec).
879      6. When generating code that counts lines, ml-lex used function
880         CharVector.foldli, taking advantage of its old interface.
881         This has been replaced with the corresponding code from
882         CharVectorSlice.  (html-lex must be re-lexed!)
883      7. BitArray in smlnj-lib/Util has been extended/modified to match the
884         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
885      8. Removed temporary additions (fromInternal, toInternal) from the
886         (now obsolete) IntInf in smlnj-lib/Util.
887      9. Cleaned up structure Byte.
888     10. Added localOffset, scan, and fromString to Date (according to spec).
889         Cleaned/corrected implementation of Date.
890         (Still need to check for correctness; implement better canonicalizeDate.)
891     11. Added "scan" to signature IEEE_REAL.
892     12. Some improvements to IntInf [in particular: efficiency-hack for
893         mod and rem when second operand is 2 (for parity checks).]
894     13. Changed representation of type Time.time, using a single IntInf.int
895         value counting microseconds.  This considerably simplified the
896         implementation of structure Time.  We now support negative time
897         values; scan and fromString handle signs.
898     14. Functor PrimIO now takes two additional arguments (VectorSlice and
899         ArraySlice).
900    
901    ----------------------------------------------------------------------
902    Name: Matthias Blume (blume (at) tti - c (dot) org)
903    Date: 2003/08/28 17:00:00 CDT
904    Tag: blume-20030828-intinf
905    Description:
906    
907    This is a major update which comes with a version number bump
908    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
909    and an implementation of IntInf in the Basis.
910    
911    There are a fairly large number of related changes and updates throughout
912    the system:
913    
914    Basis:
915            - Implemented IntInf.
916            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
917            - Added some missing Real64 operations, most notably Real.toLargeInt.
918            - Added FixedInt as a synonym for Int32.
919    
920    compiler:
921            * Added support for a built-in intinf type.
922              - literals
923              - pattern matching
924              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
925              - overloading on literals and operations
926    
927            This required adding a primitive type intinf, some additional
928            primops, and implementations for several non-trivial intinf
929            operations in Core.  (The intinf type is completely abstract
930            to the compiler; all operations get delegated back to the Core.)
931    
932            * Intinf equality is handled by polyequal.  However, the compiler
933            does not print its usual warning in this case (since polyequal
934            is the right thing to do there).
935    
936            * Improved the organization of structure InlineT.
937    
938            * A word about conversion primops:
939            If conversions involving intinf do not cancel out during
940            CPS contract, then the compiler must insert calls to Core functions.
941            Since all core access must be resolved already during the FLINT
942            translate phase, it would be too late a the time of CPS contract
943            to add new Core calls.  For this reason, conversion primops
944            for intinf carry two arguments:  1. the numeric argument that
945            they are supposed to convert, and 2. the Core function that
946            can help with this conversion if necessary.  If CPS contract
947            eliminates a primop, then the associated Core function becomes
948            dead and goes away.  Intinf conversion primops that do not get
949            eliminated by CPS contract get rewritten into calls of their
950            core functions by a separate, new phase.
951    
952    interactive system:
953            - Control.Print.intinfDepth controls max length of intinf constants
954            being printed. (Analogous to Control.Print.stringDepth.)
955            - Cleanup in printutil and pputil: got rid of unused stuff and
956            duplicates; replaced some of the code with code that makes better
957            use of library functionality.
958    
959    CM:
960            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
961            are no longer silent.
962    
963    CKIT:
964            Fixed mismatched uses of Int32 and LargeInt.  I always decided
965            in favor of LargeInt -- which is now the same as IntInf.
966            CKIT-knowledgable people should check whether this is what's
967            intended and otherwise change things back to using Int32 or
968            FixedInt.
969    
970    Throughout the code:
971            Started using IntInf.int literals and built-in operations
972            (e.g., comparison with 0) where this seems appropriate.
973    
974    
975    ----------------------------------------------------------------------
976    Name: Dave MacQueen (dbm@cs.uchicago.edu)
977    Date: 2003/08/13 11:36:00 CDT
978    Tag: dbm-20030813-mcz-merge1
979    Description:
980    
981    Merging changes from the mcz-branch development branch into trunk.
982    These changes involve replacement of the emulated old prettyprinter
983    interface with direct use of the SML/NJ Lib PP library, and fixing
984    of a couple of bugs (895, 1186) relating to error messages.  A new
985    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
986    has been added.
987    
988    ----------------------------------------------------------------------
989    Name: Matthias Blume (blume (at) tti - c (dot) org)
990    Date: 2003/08/11 15:45:00 CDT
991    Tag: blume-20030811-windows
992    Description:
993    
994    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
995    
996        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
997    
998    This patch restores SML/NJ's ability to run under win32.  There are a
999    number of changes, including fixes for several bugs that had gone
1000    unnoticed until now:
1001    
1002      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1003        trying to run the win32 version from within cygwin.)
1004    
1005      - There are a number of simple .bat scripts that substitute for their
1006        corresponding Unix shell-scripts.  (See below.)
1007    
1008      - The internals of ml-build have been modified slightly.  The main
1009        difference is that instead of calling ".link-sml" (or link-sml.bat)
1010        using OS.Process.system, the ML process delegates this task back
1011        to the script.  Otherwise problems arise in mixed environments such
1012        as Cygwin where scripts look and work like Unix scripts, but
1013        where OS.Process.system cannot run them.
1014    
1015      - In CM, the srcpath pickler used native pathname syntax -- which
1016        is incorrect in the case of cross-compilation.  The new pickle format
1017        is independent of platform-specific naming conventions.
1018    
1019      - Path configuration files (such as lib/pathconfig) can now choose
1020        between native and standard syntax.  Placing a line of the form
1021    
1022            standard!
1023    
1024        into the file causes all subsequent paths to be interpreted using
1025        CM standard pathname syntax (= Unix conventions); a line
1026    
1027            native!
1028    
1029        switches back to native style.  This was needed so that
1030        path config files can be written portably, see src/system/pathconfig.
1031    
1032      - Runtime system:
1033    
1034          - win32-filesys.c:  get_file_time and set_file_time now
1035            access modification time, not creation time.
1036    
1037          - I/O code made aware of new array representation.
1038    
1039          - Bug fixes in X86.prim.masm.
1040    
1041       - src/system/makeml made aware of win32. (For use under cygwin
1042            and other Unix-environments for windows.)
1043    
1044       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1045         which caused BinIO.inputAll to fail consistently.
1046    
1047    .bat scripts:
1048    
1049       Windows .bat scripts assume that SMLNJ_HOME is defined.
1050    
1051       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1052         applications (sml, ml-yacc, ml-lex).
1053       - ml-build.bat: analogous to ml-build.
1054       - config\install.bat:  Analogous to config/install.sh.  This requires
1055         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1056         (nmake etc. must be on the path, and vcvars32 must have been run.)
1057         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1058         and the bootfile hierarchy must have been unpacked under
1059         sml.boot.x86-win32.
1060         The script is very primitive and does a poor job at error checking.
1061         It only installs the base system, ml-lex, and ml-yacc.  No other
1062         libraries are being installed (i.e., you get only those that
1063         are part of the compiler.)
1064       - link-sml.bat:  analogous to .link-sml, but not currently used
1065    
1066    Unrelated bug fixes:
1067    
1068      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1069        types.
1070      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1071        otherwise.)
1072    
1073    ----------------------------------------------------------------------
1074    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1075    Date: 2003/06/17
1076    Tag: macqueen-20030617-bug895
1077    Description:
1078    
1079    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1080    Tag will be used for new development branch (mcz-branch) for
1081    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1082    
1083    ----------------------------------------------------------------------
1084    Name: Matthias Blume (blume (at) tti - c (dot) org)
1085    Date: 2003/05/27 16:55:00 CDT
1086    Tag: blume-20030527-polyeq
1087    Description:
1088    
1089    Tried to eliminated most cases of polymorphic equality.
1090    
1091    ----------------------------------------------------------------------
1092    Name: Matthias Blume (blume (at) tti - c (dot) org)
1093    Date: 2003/05/21 17:45:00 CDT
1094    Tag: blume-20030517-complete
1095    Description:
1096    
1097    Two changes:
1098    
1099      1. Added a flag for controlling whether non-exhaustive bindings will
1100         be treated as errors (default is false).
1101      2. Cleaned up the *entire* source tree so that CMB.make goes through
1102         without a single non-exhaustive match- or bind warning.
1103    
1104    ----------------------------------------------------------------------
1105    Name: Matthias Blume (blume (at) tti - c (dot) org)
1106    Date: 2003/05/17 10:20:00 CDT
1107    Tag: blume-20030517-absyn
1108    Description:
1109    
1110    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1111    
1112       This mainly affects the quality of error messages.  However, some
1113       of the code is now more straightforward than before. (Treatment of
1114       the above four constructs in translate.sml is much simpler than
1115       the "macro-expansion" that was going on before.  Plus, the mach-
1116       compiler no longer gets invoked just to be able to compile an
1117       if-expression.)
1118    
1119    2. The ErrorMsg.Error exception is now caught and absorbed by the
1120       interactive loop.
1121    
1122    ----------------------------------------------------------------------
1123    Name: Allen Leung
1124    Date: 2003/05/16 13:05:00 CDT
1125    Tag: leunga-20030516-cygwin-runtime
1126    Description:
1127    
1128        Ported the runtime system to cygwin, which uses the unix
1129    x86-unix bin files.  Missing/buggy features:
1130    
1131       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1132         the Cygwin library.
1133       o Ctrl-C handling may be flaky.
1134       o Windows system calls and Windows I/O are not supported.
1135    
1136       A new set of binfiles is located at:
1137    
1138         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1139    
1140       This is only needed for bootstrapping the cygwin version of smlnj.
1141       Other x86 versions can use the existing binfiles.
1142    
1143    ----------------------------------------------------------------------
1144    Name: Matthias Blume
1145    Date: 2003/04/08 15:42:00 CDT
1146    Tag: blume-20030408-listpair
1147    Description:
1148    
1149    1. Added a target 'mlrisc' to installer.
1150    
1151    2. Added missing elements to structure ListPair.
1152    
1153    ----------------------------------------------------------------------
1154    Name: Allen Leung
1155    Date: 2003/01/07 10:40:00 EST
1156    Tag: leunga-20030107-int-rem
1157    Description:
1158    
1159    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1160    The arguments to the SUBL instruction were swapped.
1161    
1162    ----------------------------------------------------------------------
1163    Name: Matthias Blume
1164    Date: 2002/12/12 16:25:00 EST
1165    Tag: blume-20021212-risc-ra
1166    Description:
1167    
1168    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1169    sent the RA into an infinite loop when floating point registers get
1170    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1171    110.37 and 110.38.)
1172    There was another set of potential problems related to the handling of
1173    MLRISC annotations (but those did not yet cause real problems, apparently).
1174    
1175    ----------------------------------------------------------------------
1176    Name: Matthias Blume
1177    Date: 2002/12/06 22:40:00 EST
1178    Tag: blume-20021206-cm-fileid
1179    Description:
1180    
1181    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1182    This fixes the problem of CM getting confused by files that suddenly
1183    change their identity (e.g., by getting unlinked and recreated by some
1184    text editor such as vi).  There might be a better/cheaper/cleaner way
1185    of doing this, but for now this will have to do.
1186    
1187    ----------------------------------------------------------------------
1188    Name: Matthias Blume
1189    Date: 2002/10/28 09:50:00 EST
1190    Tag: blume-20021028-typecheck
1191    Description:
1192    
1193    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1194    
1195    ----------------------------------------------------------------------
1196    Name: Matthias Blume
1197    Date: 2002/10/17 09:10:00 EDT
1198    Tag: Release_110_42
1199    Description:
1200    
1201    In good old tradition, there has been a slight hiccup so that we have
1202    to patch 110.42 after the fact.   The old release tag has been replaced
1203    (see below).
1204    
1205    The change solves a problem with two competing approaches the
1206    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1207    each other's way.
1208    
1209    This change only affects the runtime system code and the installer script.
1210    (No new bootfiles.)
1211    
1212    ----------------------------------------------------------------------
1213    Name: Matthias Blume
1214    Date: 2002/10/16 12:00:00 EDT
1215    Tag: Release_110_42_removed
1216    Description:
1217    
1218    New working release.  New bootfiles.
1219    
1220    ----------------------------------------------------------------------
1221    Name: Matthias Blume
1222    Date: 2002/10/10 13:10:00 EDT
1223    Tag: blume-20021010-ppc-divs
1224    Description:
1225    
1226    The mltree operator DIVS must be implemented with an overflow check on
1227    the PPC because the hardware indicates divide-by-zero using "overflow" as
1228    well.
1229    
1230    ----------------------------------------------------------------------
1231    Name: Matthias Blume
1232    Date: 2002/07/23 11:45:00
1233    Tag: blume-20020723-smlnj-home
1234    Description:
1235    
1236    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1237    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1238    CM_PATHCONFIG is also set), the path configuration file is assumed
1239    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1240    the entire tree to some other place and everything will "just work".
1241    
1242    (Companion commands such as ml-build and ml-makedepend also sense this
1243    variable.)
1244    
1245    ----------------------------------------------------------------------
1246    Name: Matthias Blume
1247    Date: 2002/07/12 21:19:00 EDT
1248    Tag: blume-20020712-liveness
1249    Description:
1250    
1251    Exported two useful "step" functions from liveness module (MLRISC).
1252    
1253    ----------------------------------------------------------------------
1254    Name: Matthias Blume
1255    Date: 2002/07/05 16:00 EDT
1256    Tag: Release_110_41
1257    Description:
1258    
1259    New working release.  New bootfiles.
1260    
1261    ----------------------------------------------------------------------
1262    Name: Matthias Blume
1263    Date: 2002/07/05 10:25:00 EDT
1264    Tag: blume-20020705-btimp
1265    Description:
1266    
1267    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1268    other clients can set up backtracing support.
1269    
1270    ----------------------------------------------------------------------
1271    Name: Matthias Blume
1272    Date: 2002/06/25 17:23:00 EDT
1273    Tag: blume-20020625-fpmax
1274    Description:
1275    
1276    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1277    case.  (The sense of the isNaN test was reversed -- which made min and
1278    max always return their first argument.)
1279    
1280    ----------------------------------------------------------------------
1281    Name: Matthias Blume
1282    Date: 2002/06/11
1283    Tag: blume-20020611-unixpath
1284    Description:
1285    
1286    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1287    
1288    ----------------------------------------------------------------------
1289    Name: Matthias Blume
1290    Date: 2002/06/10 16:35:00 EDT
1291    Tag: blume-20020610-ieeereal
1292    Description:
1293    
1294    I back-ported my implementation of IEEEReal.fromString from the
1295    idlbasis-devel branch so that we can test it.
1296    
1297    Another small change is that ppDec tries to give more information
1298    than just "<sig>" in the case of functors.  However, this code is
1299    broken in some mysterious way if the functor's body's signature
1300    has not been declared by ascription but gets inferred from the
1301    implementation.  This needs fixing...
1302    
1303    ----------------------------------------------------------------------
1304    Name: Matthias Blume
1305    Date: 2002/05/31
1306    Tag: blume-20020531-btrace-mode
1307    Description:
1308    
1309    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1310    the wayside when I switched over to using Controls everywhere.)
1311    
1312    ----------------------------------------------------------------------
1313    Name: Lal George
1314    Date: 2002/05/23 12:21:40 EDT
1315    Tag: george-20020523-visual-labels
1316    Description:
1317    
1318    Labels are now displayed in the graphical output to make
1319    the fall-through and target blocks obvious.
1320    
1321    ----------------------------------------------------------------------
1322    Name: Matthias Blume
1323    Date: 2002/05/22 11:03:00 EDT
1324    Tag: blume-20020522-shrink
1325    Description:
1326    
1327    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1328    situation that comes up when allocating huge arrays.
1329    
1330    ----------------------------------------------------------------------
1331    Name: Matthias Blume
1332    Date: 2002/05/21 16:00:00 EDT
1333    Tag: Release_110_40
1334    Description:
1335    
1336    New working release (110.40).  New bootfiles.
1337    
1338    [Also: John Reppy fixed GC bug 1131.]
1339    
1340    ----------------------------------------------------------------------
1341    Name: Matthias Blume
1342    Date: 2002/05/21 12:35:00 EDT
1343    Tag: blume-20020521-cmdoc
1344    Description:
1345    
1346    CM documentation update.
1347    
1348    ----------------------------------------------------------------------
1349    Name: Matthias Blume
1350    Date: 2002/05/21 10:55:00 EDT
1351    Tag: blume-20020521-misc
1352    Description:
1353    
1354    - John tweaked runtime to be silent on heap export (except when
1355      GC messages are on).
1356    - I added a few more things (cross-compiling versions of CMB) to
1357      config/preloads (as suggestions).
1358    
1359    ----------------------------------------------------------------------
1360    Name: Matthias Blume
1361    Date: 2002/05/20 22:25:00 EDT
1362    Tag: blume-20020520-controls
1363    Description:
1364    
1365    - Added ControlUtil structure to control-lib.cm.
1366    - Use it throughout.
1367    - Used Controls facility to define MLRISC controls (as opposed to
1368      registering MLRISC control ref cells with Controls after the fact)
1369    - Fixed messed-up controls priorities.
1370    
1371    * Removed again all the stuff from config/preloads that one wouldn't
1372      be able to preload at the time the initial heap image is built.
1373      (Many libraries, e.g., CML, do not exist yet at this time.  The
1374       only libraries that can be preloaded via config/preloads are those
1375       that come bundled with the bootfiles.)
1376    
1377    ----------------------------------------------------------------------
1378    Name: Matthias Blume
1379    Date: 2002/05/20 10:59:00 EDT
1380    Tag: blume-20020520-preloads
1381    Description:
1382    
1383    Added a lot of commented-out suggestions for things to be included
1384    in config/preloads.
1385    
1386    ----------------------------------------------------------------------
1387    Name: Allen Leung
1388    Date: 2002/05/18 14:20:00 EDT
1389    Tag: leunga-20020518-mdl
1390    Description:
1391    
1392    o Made the mdl tool stuff compile and run again.
1393    o I've disabled all the stuff that depends on RTL specifications; they
1394      are all badly broken anyway.
1395    
1396    ----------------------------------------------------------------------
1397    Name: Matthias Blume
1398    Date: 2002/05/17 16:49:00 EDT
1399    Tag: blume-20020517-controls
1400    Description:
1401    
1402    0. John Reppy made several modifications to the SML/NJ library.
1403       In particular, there is a shiny new controls-lib.cm.
1404    
1405    1. Pushed new controls interface through compiler so that everything
1406       compiles again.
1407    
1408    2. Added FormatComb and FORMAT_COMB to the CML version of the
1409       SML/NJ library (so that CML compiles again).
1410    
1411    3. Modified init scripts because XXX_DEFAULT environment variables
1412       are no longer with us.  (Boot-time initialization is now done
1413       using the same environment variables that are also used for
1414       startup-time initialization of controls.)
1415    
1416    ----------------------------------------------------------------------
1417    Name: Lal George
1418    Date: 2002/05/15 09:20:10 EDT
1419    Tag: george-20020515-pseudo-op-decls
1420    Description:
1421    
1422            All pseudo-ops emitted before the first segment declaration
1423    such as TEXT, DATA, and BSS directives are assumed to be global
1424    declarations and are emitted first in the assembly file. This is
1425    useful in a number of situations where one has pseudo-ops that are not
1426    specific to any segment, and also works around the constraint that one
1427    cannot have client pseudo-ops in the TEXT segment.
1428    
1429            Because no segment is associated with these declarations it is
1430    an error to allocate any space or objects before the first segment
1431    directive and an exception will be raised. However, we cannot make
1432    this check for client pseudo-ops.
1433    
1434            These top level declarations are a field in the CFG graph_info.
1435    In theory you can continue to add to this field after the CFG has been
1436    built -- provided you know what you are doing;-)
1437    
1438    ----------------------------------------------------------------------
1439    Name: Matthias Blume
1440    Date: 2002/05/13 16:40:00 EDT
1441    Tag: blume-20020513-pp-etc
1442    Description:
1443    
1444    A few minor bugfixes:
1445    
1446      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1447        (Bogus printouts for redefined bindings still occur.  Compiler
1448        bug should no longer occur now.  We need to redo the prettyprinter
1449        from scratch.)
1450    
1451      - CM pathname printer now also adds escape sequences for ( and )
1452    
1453      - commend and docu fixes for ml-nlffi
1454    
1455    ----------------------------------------------------------------------
1456    Name: Matthias Blume
1457    Date: 2002/05/10 16:40:00 EDT
1458    Tag: blume-20020510-erg-textio
1459    Description:
1460    
1461    Applied the following bugfix provided by Emden Gansner:
1462    
1463        Output is corrupted when outputSubstr is used rather than output.
1464    
1465        The problem occurs when a substring
1466    
1467            ss = (s, dataStart, dataLen)
1468    
1469        where dataStart > 0, fills a stream buffer with avail bytes left.
1470        avail bytes of s, starting at index dataStart, are copied into the
1471        buffer, the buffer is flushed, and then the remaining dataLen-avail
1472        bytes of ss are copied into the beginning of the buffer. Instead of
1473        starting this copy at index dataStart+avail in s, the current code
1474        starts the copy at index avail.
1475    
1476        Fix:
1477        In text-io-fn.sml, change line 695 from
1478             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1479        to
1480             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1481    
1482    ----------------------------------------------------------------------
1483    Name: Matthias Blume
1484    Date: 2002/04/12 13:55:00 EDT
1485    Tag: blume-20020412-assyntax
1486    Description:
1487    
1488    1. Grabbed newer assyntax.h from the XFree86 project.
1489    2. Fiddled with how to compile X86.prim.asm without warnings.
1490    3. (Very) Minor cleanup in CM.
1491    
1492    ----------------------------------------------------------------------
1493    Name: Matthias Blume
1494    Date: 2002/04/01 (no joke!) 17:07:00 EST
1495    Tag: blume-20020401-x86div
1496    Description:
1497    
1498    Added full support for div/mod/rem/quot on the x86, using the machine
1499    instruction's two results (without clumsily recomputing the remainder)
1500    directly where appropriate.
1501    
1502    Some more extensive power-of-two support was added to the x86 instruction
1503    selector (avoiding expensive divs, mods, and muls where they can be
1504    replaced with cheaper shifts and masks).  However, this sort of thing
1505    ought to be done earlier, e.g., within the CPS optimizer so that
1506    all architectures benefit from it.
1507    
1508    The compiler compiles to a fixed point, but changes might be somewhat
1509    fragile nevertheless.  Please, report any strange things that you might
1510    see wrt. div/mod/quot/rem...
1511    
1512    ----------------------------------------------------------------------
1513    Name: Matthias Blume
1514    Date: 2002/03/29 17:22:00
1515    Tag: blume-20020329-div
1516    Description:
1517    
1518    Fixed my broken div/mod logic.  Unfortunately, this means that the
1519    inline code for div/mod now has one more comparison than before.
1520    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
1521    The problem was with quotient = 0, because that alone does not tell
1522    us which way the rounding went.  One then has to look at whether
1523    remainder and divisor have the same sign...  :(
1524    
1525    Anyway, I replaced the bootfiles with fresh ones...
1526    
1527    ----------------------------------------------------------------------
1528    Name: Matthias Blume
1529    Date: 2002/03/29 14:10:00 EST
1530    Tag: blume-20020329-inlprims
1531    Description:
1532    
1533    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
1534    
1535    Primops have changed. This means that the bin/boot-file formats have
1536    changed as well.
1537    
1538    To make sure that there is no confusion, I made a new version.
1539    
1540    
1541    CHANGES:
1542    
1543    * removed REMT from mltree (remainder should never overflow).
1544    
1545    * added primops to deal with divisions of all flavors to the frontend
1546    
1547    * handled these primops all the way through so they map to their respective
1548      MLRISC support
1549    
1550    * used these primops in the implementation of Int, Int32, Word, Word32
1551    
1552    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
1553    
1554    * parameterized INLMIN, INLMAX, and INLABS by a numkind
1555    
1556    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
1557      floating point
1558    
1559    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
1560      and Real (but Real.abs maps to a separate floating-point-only primop)
1561    
1562    
1563    TODO items:
1564    
1565    * Hacked Alpha32 instruction selection, disabling the selection of REMx
1566      instructions because the machine instruction encoder cannot handle
1567      them.  (Hppa, PPC, and Sparc instruction selection did not handle
1568      REM in the first place, and REM is supported by the x86 machine coder.)
1569    
1570    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
1571      selection phase.  (The two can be streamlined because the hardware
1572      delivers both quotient and remainder at the same time anyway.)
1573    
1574    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
1575      (Currently the behavior is inconsistent both across architectures and
1576      wrt. the draft Basis spec.)
1577    
1578    * Word8 should eventually be handled natively, too.
1579    
1580    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
1581      as if there currently is no execution path that could trigger it in
1582      SML/NJ.  (The assumptions underlying functions arith and promotable do not
1583      hold for things like multiplication and division.)
1584    
1585    ----------------------------------------------------------------------
1586    Name: Matthias Blume
1587    Date: 2002/03/27 16:27:00 EST
1588    Tag: blume-20020327-mlrisc-divisions
1589    Description:
1590    
1591    Added support for all four division operations (ML's div, mod, quot,
1592    and rem) to MLRISC.  In the course of doing so, I also rationalized
1593    the naming (no more annoying switch-around of DIV and QUOT), by
1594    parameterizing the operation by div_rounding_mode (which can be either
1595    DIV_TO_ZERO or DIV_TO_NEGINF).
1596    
1597    The generic MLTreeGen functor takes care of compiling all four
1598    operations down to only round-to-zero div.
1599    
1600    Missing pieces:
1601    
1602      * Doing something smarter than relying on MLTreeGen on architectures
1603        like, e.g., the x86 where hardware division delivers both quotient and
1604        remainder at the same time.  With this, the implementation of the
1605        round-to-neginf operations could be further streamlined.
1606    
1607      * Remove inlining support for div/mod/rem from the frontend and replace it
1608        with primops that get carried through to the backend.  Do this for all
1609        int and word types.
1610    
1611    ----------------------------------------------------------------------
1612    Name: Matthias Blume
1613    Date: 2002/03/25 17:25:00 EST
1614    Tag: blume-20020325-divmod
1615    Description:
1616    
1617    I improved (hopefully without breaking them) the implementation of Int.div,
1618    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
1619    advantage of the following observations:
1620    
1621      Let  q = x quot y      r = x rem y
1622           d = x div  y      m = x mod y
1623    
1624    where "quot" is the round-to-zero version of integer division that
1625    hardware usually provides.  Then we have:
1626    
1627         r = x - q * y        where neither the * nor the - will overflow
1628         d = if q >= 0 orelse x = q * y then q else q - 1
1629                              where neither the * nor the - will overflow
1630         m = if q >= 0 orelse r = 0 then r else r + y
1631                              where the + will not overflow
1632    
1633    This results in substantial simplification of the generated code.
1634    The following table shows the number of CFG nodes and edges generated
1635    for
1636            fun f (x, y) = x OPER y
1637            (* with OPER \in div, mod, quot, rem *)
1638    
1639    
1640        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
1641        --------------------------------------------------------
1642         div |         24 |         39 |         12 |         16
1643         mod |         41 |         71 |         12 |         16
1644        quot |          8 |         10 |          8 |         10
1645         rem |         10 |         14 |          8 |         10
1646    
1647    
1648    ----------------------------------------------------------------------
1649    Name: Matthias Blume
1650    Date: 2002/03/25 22:06:00 EST
1651    Tag: blume-20020325-cprotobug
1652    Description:
1653    
1654    Fixed a bug in cproto (c prototype decoder).
1655    
1656    ----------------------------------------------------------------------
1657    Name: Matthias Blume
1658    Date: 2002/03/25 16:00:00 EST
1659    Tag: blume-20020325-raw-primops
1660    Description:
1661    
1662    I did some cleanup to Allen's new primop code and
1663    replaced yesterday's bootfiles with new ones.
1664    (But they are stored in the same place.)
1665    
1666    ----------------------------------------------------------------------
1667    Name: Matthias Blume
1668    Date: 2002/03/24 22:40:00 EST
1669    Tag: blume-20020324-bootfiles
1670    Description:
1671    
1672    Made the bootfiles that Allen asked for.
1673    
1674    ----------------------------------------------------------------------
1675    Name: Allen Leung
1676    Date: 2002/03/23 15:50:00 EST
1677    Tag: leunga-20020323-flint-cps-rcc-primops
1678    Description:
1679    
1680      1. Changes to FLINT primops:
1681    
1682        (* make a call to a C-function;
1683         * The primop carries C function prototype information and specifies
1684         * which of its (ML-) arguments are floating point. C prototype
1685         * information is for use by the backend, ML information is for
1686         * use by the CPS converter. *)
1687      | RAW_CCALL of { c_proto: CTypes.c_proto,
1688                       ml_args: ccall_type list,
1689                       ml_res_opt: ccall_type option,
1690                       reentrant : bool
1691                     } option
1692       (* Allocate uninitialized storage on the heap.
1693        * The record is meant to hold short-lived C objects, i.e., they
1694        * are not ML pointers.  With the tag, the representation is
1695        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1696        *)
1697      | RAW_RECORD of {tag:bool,sz:int}
1698      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1699    
1700      2.  These CPS primops are now overloaded:
1701    
1702           rawload of {kind:numkind}
1703           rawstore of {kind:numkind}
1704    
1705          The one argument form is:
1706    
1707             rawload {kind} address
1708    
1709          The two argument form is:
1710    
1711             rawload {kind} [ml object, byte-offset]
1712    
1713      3. RAW_CCALL/RCC now takes two extra arguments:
1714    
1715         a. The first is whether the C call is reentrant, i.e., whether
1716            ML state should be saved and restored.
1717         b. The second argument is a string argument specifying the name of
1718            library and the C function.
1719    
1720         These things are currently not handled in the code generator, yet.
1721    
1722      4. In CProto,
1723    
1724         An encoding type of "bool" means "ml object" and is mapped into
1725         C prototype of PTR.  Note that "bool" is different than "string",
1726         even though "string" is also mapped into PTR, because "bool"
1727         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
1728    
1729      5. Pickler/unpicker
1730    
1731         Changed to handle RAW_RECORD and newest RAW_CCALL
1732    
1733      6. MLRiscGen,
1734    
1735         1. Changed to handle the new rawload/rawstore/rawrecord operators.
1736         2. Code for handling C Calls has been moved to a new module CPSCCalls,
1737            in the file CodeGen/cpscompile/cps-c-calls.sml
1738    
1739      7. Added the conditional move operator
1740    
1741             condmove of branch
1742    
1743         to cps.  Generation of this is still buggy so it is currently
1744         disabled.
1745    
1746    ----------------------------------------------------------------------
1747    Name: Lal George
1748    Date: 2002/03/22 14:18:25 EST
1749    Tag: george-20020322-cps-branch-prob
1750    Description:
1751    
1752    Implemented the Ball-Larus branch prediction-heuristics, and
1753    incorporated graphical viewers for control flow graphs.
1754    
1755    Ball-Larus Heuristics:
1756    ---------------------
1757    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
1758    
1759    By design it uses the Dempster-Shafer theory for combining
1760    probabilities.  For example, in the function:
1761    
1762        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
1763    
1764    the ball-larus heuristics predicts that the n=0 is unlikely
1765    (OH-heuristic), and the 'then' branch is unlikely because of the
1766    RH-heuristic -- giving the 'then' branch an even lower combined
1767    probability using the Dempster-Shafer theory.
1768    
1769    Finally, John Reppy's loop analysis in MLRISC, further lowers the
1770    probability of the 'then' branch because of the loop in the else
1771    branch.
1772    
1773    
1774    Graphical Viewing:
1775    ------------------
1776    I merely plugged in Allen's graphical viewers into the compiler. The
1777    additional code is not much. At the top level, saying:
1778    
1779            Control.MLRISC.getFlag "cfg-graphical-view" := true;
1780    
1781    will display the graphical view of the control flow graph just before
1782    back-patching.  daVinci must be in your path for this to work. If
1783    daVinci is not available, then the default viewer can be changed
1784    using:
1785    
1786            Control.MLRISC.getString "viewer"
1787    
1788    which can be set to "dot" or "vcg" for the corresponding viewers. Of
1789    course, these viewers must be in your path.
1790    
1791    The above will display the compilation unit at the level of clusters,
1792    many of which are small, boring, and un-interesting. Also setting:
1793    
1794            Control.MLRISC.getInt "cfg-graphical-view_size"
1795    
1796    will display clusters that are larger than the value set by the above.
1797    
1798    
1799    ----------------------------------------------------------------------
1800    Name: Matthias Blume
1801    Date: 2002/03/21 22:20:00 EST
1802    Tag: blume-20020321-kmp-bugfix
1803    Description:
1804    
1805    Changed the interface to the KMP routine in PreString and fixed
1806    a minor bug in one place where it was used.
1807    
1808    ----------------------------------------------------------------------
1809    Name: Allen Leung
1810    Date: 2002/03/21 20:30:00 EST
1811    Tag: leunga-20020321-cfg
1812    Description:
1813    
1814      Fixed a potential problem in cfg edge splitting.
1815    
1816    ----------------------------------------------------------------------
1817    Name: Allen Leung
1818    Date: 2002/03/21 17:15:00 EST
1819    Tag: leunga-20020321-x86-fp-cfg
1820    Description:
1821    
1822      1. Recoded the buggy parts of x86-fp.
1823    
1824         a. All the block reordering code has been removed.
1825            We now depend on the block placement phases to do this work.
1826    
1827         b. Critical edge splitting code has been simplified and moved into the
1828            CFG modules, as where they belong.
1829    
1830         Both of these were quite buggy and complex.  The code is now much, much
1831         simpler.
1832    
1833      2. X86 backend.
1834    
1835         a. Added instructions for 64-bit support.  Instruction selection for
1836            64-bit has not been committed, however, since that
1837            requires changes to MLTREE which haven't been approved by
1838            Lal and John.
1839    
1840         b. Added support for FUCOMI and FUCOMIP when generating code for
1841            PentiumPro and above.  We only generate these instructions in
1842            the fast-fp mode.
1843    
1844         c. Added cases for JP and JNP in X86FreqProps.
1845    
1846      3. CFG
1847    
1848         CFG now has a bunch of methods for edge splitting and merging.
1849    
1850      4. Machine description.
1851    
1852         John's simplification of MLTREE_BASIS.fcond broke a few machine
1853         description things:
1854    
1855         rtl-build.{sig,sml} and hppa.mdl fixed.
1856    
1857         NOTE: the machine description stuff in the repository is still broken.
1858               Again, I can't put my fixes in because that involves
1859               changes to MLTREE.
1860    
1861    ----------------------------------------------------------------------
1862    Name: Matthias Blume
1863    Date: 2002/03/20 15:55:00 EST
1864    Tag: blume-20020320-kmp
1865    Description:
1866    
1867    Implemented Knuth-Morris-Pratt string matching in PreString and used
1868    it for String.isSubstring, Substring.isSubstring, and
1869    Substring.position.
1870    
1871    (Might need some stress-testing.  Simple examples worked fine.)
1872    
1873    ----------------------------------------------------------------------
1874    Name: Matthias Blume
1875    Date: 2002/03/19 16:37:00 EST
1876    Tag: blume-20020319-witnesses
1877    Description:
1878    
1879    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
1880    
1881    This implements a generic mechanism for changing constness qualifiers
1882    anywhere within big C types without resorting to outright "casts".
1883    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
1884    modify the constness at the outermost level.)
1885    The implementation of "convert" is based on the idea of "witness"
1886    values -- values that are not used by the operation but whose types
1887    "testify" to their applicability.  On the implementation side, "convert"
1888    is simply a projection (returning its second curried argument).  With
1889    cross-module inlining, it should not result in any machine code being
1890    generated.
1891    
1892    ----------------------------------------------------------------------
1893    Name: Matthias Blume
1894    Date: 2002/03/15 16:40:00 EST
1895    Tag: blume-20020315-basis
1896    Description:
1897    
1898    Provided (preliminary?) implementations for
1899    
1900      {String,Substring}.{concatWith,isSuffix,isSubstring}
1901    
1902    and
1903    
1904      Substring.full
1905    
1906    Those are in the Basis spec but they were missing in SML/NJ.
1907    
1908    ----------------------------------------------------------------------
1909    Name: Matthias Blume
1910    Date: 2002/03/14 21:30:00 EST
1911    Tag: blume-20020314-controls
1912    Description:
1913    
1914    Controls:
1915    ---------
1916    
1917    1. Factored out the recently-added Controls : CONTROLS stuff and put
1918       it into its own library $/controls-lib.cm.  The source tree for
1919       this is under src/smlnj-lib/Controls.
1920    
1921    2. Changed the names of types and functions in this interface, so they
1922       make a bit more "sense":
1923    
1924          module -> registry
1925          'a registry -> 'a group
1926    
1927    3. The interface now deals in ref cells only.  The getter/setter interface
1928       is (mostly) gone.
1929    
1930    4. Added a function that lets one register an already-existing ref cell.
1931    
1932    5. Made the corresponding modifications to the rest of the code so that
1933       everything compiles again.
1934    
1935    6. Changed the implementation of Controls.MLRISC back to something closer
1936       to the original.  In particular, this module (and therefore MLRISC)
1937       does not depend on Controls.  There now is some link-time code in
1938       int-sys.sml that registers the MLRISC controls with the Controls
1939       module.
1940    
1941    CM:
1942    ---
1943    
1944      * One can now specify the lambda-split aggressiveness in init.cmi.
1945    
1946    ----------------------------------------------------------------------
1947    Name: Allen Leung
1948    Date: 2002/03/13 17:30:00 EST
1949    Tag: leunga-20020313-x86-fp-unary
1950    Description:
1951    
1952    Bug fix for:
1953    
1954    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
1955    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
1956    > - fun f(x,(y,z)) = Real.~ y;
1957    > [autoloading]
1958    > [autoloading done]
1959    >       fchsl   (%eax), 184(%esp)
1960    > Error: MLRisc bug: X86MCEmitter.emitInstr
1961    >
1962    > uncaught exception Error
1963    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
1964    
1965    The problem was that the code generator did not generate any fp registers
1966    in this case, and the ra didn't know that it needed to run the X86FP phase to
1967    translate the pseudo fp instruction.   This only happened with unary fp
1968    operators in certain situations.
1969    
1970    ----------------------------------------------------------------------
1971    Name: Matthias Blume
1972    Date: 2002/03/13 14:00:00 EST
1973    Tag: blume-20020313-overload-etc
1974    Description:
1975    
1976    1. Added _overload as a synonym for overload for backward compatibility.
1977       (Control.overloadKW must be true for either version to be accepted.)
1978    
1979    2. Fixed bug in install script that caused more things to be installed
1980       than what was requested in config/targets.
1981    
1982    3. Made CM aware of the (_)overload construct so that autoloading
1983       works.
1984    
1985    ----------------------------------------------------------------------
1986    Name: Matthias Blume
1987    Date: 2002/03/12 22:03:00 EST
1988    Tag: blume-20020312-url
1989    Description:
1990    
1991    Forgot to update BOOT and srcarchiveurl.
1992    
1993    ----------------------------------------------------------------------
1994    Name: Matthias Blume
1995    Date: 2002/03/12 17:30:00 EST
1996    Tag: blume-20020312-version110392
1997    Description:
1998    
1999    Yet another version number bump (because of small changes to the
2000    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2001    
2002    Changes:
2003    
2004      The new pid generation scheme described a few weeks ago was overly
2005      complicated.  I implemented a new mechanism that is simpler and
2006      provides a bit more "stability":  Once CM has seen a compilation
2007      unit, it keeps its identity constant (as long as you do not delete
2008      those crucial CM/GUID/* files).  This means that when you change
2009      an interface, compile, then go back to the old interface, and
2010      compile again, you arrive at the original pid.
2011    
2012      There now also is a mechanism that instructs CM to use the plain
2013      environment hash as a module's pid (effectively making its GUID
2014      the empty string).  For this, "noguid" must be specified as an
2015      option to the .sml file in question within its .cm file.
2016      This is most useful for code that is being generated by tools such
2017      as ml-nlffigen (because during development programmers tend to
2018      erase the tool's entire output directory tree including CM's cached
2019      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2020      revert to the old, broken behavior of SML/NJ, but in specific cases
2021      where there is no danger of interface confusion, its use is ok
2022      (I think).
2023    
2024      ml-nlffigen by default generates "noguid" annotations.  They can be
2025      turned off by specifying -guid in its command line.
2026    
2027    ----------------------------------------------------------------------
2028    Name: Lal George
2029    Date: 2002/03/12 12 14:42:36 EST
2030    Tag: george-20020312-frequency-computation
2031    Description:
2032    
2033    Integrated jump chaining and static block frequency into the
2034    compiler. More details and numbers later.
2035    
2036    ----------------------------------------------------------------------
2037    Name: Lal George
2038    Date: 2002/03/11 11 22:38:53 EST
2039    Tag: george-20020311-jump-chain-elim
2040    Description:
2041    
2042    Tested the jump chain elimination on all architectures (except the
2043    hppa).  This is on by default right now and is profitable for the
2044    alpha and x86, however, it may not be profitable for the sparc and ppc
2045    when compiling the compiler.
2046    
2047    The gc test will typically jump to a label at the end of the cluster,
2048    where there is another jump to an external cluster containing the actual
2049    code to invoke gc. This is to allow factoring of common gc invocation
2050    sequences. That is to say, we generate:
2051    
2052            f:
2053               testgc
2054               ja   L1      % jump if above to L1
2055    
2056            L1:
2057               jmp L2
2058    
2059    
2060    After jump chain elimination the 'ja L1' instructions is converted to
2061    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2062    up being implemented in their long form (if L2 is far away) using:
2063    
2064            jbe     L3      % jump if below or equal to L3
2065            jmp     L2
2066         L3:
2067            ...
2068    
2069    
2070    For large compilation units L2  may be far away.
2071    
2072    
2073    ----------------------------------------------------------------------
2074    Name: Matthias Blume
2075    Date: 2002/03/11 13:30:00 EST
2076    Tag: blume-20020311-mltreeeval
2077    Description:
2078    
2079    A functor parameter was missing.
2080    
2081    ----------------------------------------------------------------------
2082    Name: Allen Leung
2083    Date: 2002/03/11 10:30:00 EST
2084    Tag: leunga-20020311-runtime-string0
2085    Description:
2086    
2087       The representation of the empty string now points to a
2088    legal null terminated C string instead of unit.  It is now possible
2089    to convert an ML string into C string with InlineT.CharVector.getData.
2090    This compiles into one single machine instruction.
2091    
2092    ----------------------------------------------------------------------
2093    Name: Allen Leung
2094    Date: 2002/03/10 23:55:00 EST
2095    Tag: leunga-20020310-x86-call
2096    Description:
2097    
2098       Added machine generation for CALL instruction (relative displacement mode)
2099    
2100    ----------------------------------------------------------------------
2101    Name: Matthias Blume
2102    Date: 2002/03/08 16:05:00
2103    Tag: blume-20020308-entrypoints
2104    Description:
2105    
2106    Version number bumped to 110.39.1.  NEW BOOTFILES!
2107    
2108    Entrypoints: non-zero offset into a code object where execution should begin.
2109    
2110    - Added the notion of an entrypoint to CodeObj.
2111    - Added reading/writing of entrypoint info to Binfile.
2112    - Made runtime system bootloader aware of entrypoints.
2113    - Use the address of the label of the first function given to mlriscGen
2114      as the entrypoint.  This address is currently always 0, but it will
2115      not be 0 once we turn on block placement.
2116    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2117      with entry points) from mlriscGen.
2118    
2119    ----------------------------------------------------------------------
2120    Name: Allen Leung
2121    Date: 2002/03/07 20:45:00 EST
2122    Tag: leunga-20020307-x86-cmov
2123    Description:
2124    
2125       Bug fixes for CMOVcc on x86.
2126    
2127       1. Added machine code generation for CMOVcc
2128       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2129       3. CMOVcc cannot have an immediate operand as argument.
2130    
2131    ----------------------------------------------------------------------
2132    Name: Matthias Blume
2133    Date: 2002/03/07 16:15:00 EST
2134    Tag: blume-20020307-controls
2135    Description:
2136    
2137    This is a very large but mostly boring patch which makes (almost)
2138    every tuneable compiler knob (i.e., pretty much everything under
2139    Control.* plus a few other things) configurable via both the command
2140    line and environment variables in the style CM did its configuration
2141    until now.
2142    
2143    Try starting sml with '-h' (or, if you are brave, '-H')
2144    
2145    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2146    implements the underlying generic mechanism.
2147    
2148    The interface to some of the existing such facilities has changed somewhat.
2149    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2150    (The getFoo interface is still there for backward-compatibility, but its
2151    use is deprecated.)
2152    
2153    The ml-build script passes -Cxxx=yyy command-line arguments through so
2154    that one can now twiddle the compiler settings when using this "batch"
2155    compiler.
2156    
2157    TODO items:
2158    
2159    We should go through and throw out all controls that are no longer
2160    connected to anything.  Moreover, we should go through and provide
2161    meaningful (and correct!) documentation strings for those controls
2162    that still are connected.
2163    
2164    Currently, multiple calls to Controls.new are accepted (only the first
2165    has any effect).  Eventually we should make sure that every control
2166    is being made (via Controls.new) exactly once.  Future access can then
2167    be done using Controls.acc.
2168    
2169    Finally, it would probably be a good idea to use the getter-setter
2170    interface to controls rather than ref cells.  For the time being, both
2171    styles are provided by the Controls module, but getter-setter pairs are
2172    better if thread-safety is of any concern because they can be wrapped.
2173    
2174    *****************************************
2175    
2176    One bug fix: The function blockPlacement in three of the MLRISC
2177    backpatch files used to be hard-wired to one of two possibilities at
2178    link time (according to the value of the placementFlag).  But (I
2179    think) it should rather sense the flag every time.
2180    
2181    *****************************************
2182    
2183    Other assorted changes (by other people who did not supply a HISTORY entry):
2184    
2185    1. the cross-module inliner now works much better (Monnier)
2186    2. representation of weights, frequencies, and probabilities in MLRISC
2187       changed in preparation of using those for weighted block placement
2188       (Reppy, George)
2189    
2190    ----------------------------------------------------------------------
2191    Name: Lal George
2192    Date: 2002/03/07 14:44:24 EST 2002
2193    Tag: george-20020307-weighted-block-placement
2194    
2195    Tested the weighted block placement optimization on all architectures
2196    (except the hppa) using AMPL to generate the block and edge frequencies.
2197    Changes were required in the machine properties to correctly
2198    categorize trap instructions. There is an MLRISC flag
2199    "weighted-block-placement" that can be used to enable weighted block
2200    placement, but this will be ineffective without block/edge
2201    frequencies (coming soon).
2202    
2203    
2204    ----------------------------------------------------------------------
2205    Name: Lal George
2206    Date: 2002/03/05 17:24:48 EST
2207    Tag: george-20020305-linkage-cluster
2208    
2209    In order to support the block placement optimization, a new cluster
2210    is generated as the very first cluster (called the linkage cluster).
2211    It contains a single jump to the 'real' entry point for the compilation
2212    unit. Block placement has no effect on the linkage cluster itself, but
2213    all the other clusters  have full freedom in the manner in which they
2214    reorder blocks or functions.
2215    
2216    On the x86 the typical linkage code that is generated is:
2217       ----------------------
2218            .align 2
2219       L0:
2220            addl    $L1-L0, 72(%esp)
2221            jmp     L1
2222    
2223    
2224            .align  2
2225       L1:
2226       ----------------------
2227    
2228    72(%esp) is the memory location for the stdlink register. This
2229    must contain the address of the CPS function being called. In the
2230    above example, it contains the address of  L0; before
2231    calling L1 (the real entry point for the compilation unit), it
2232    must contain the address for L1, and hence
2233    
2234            addl $L1-L0, 72(%esp)
2235    
2236    I have tested this on all architectures except the hppa.The increase
2237    in code size is of course negligible
2238    
2239    ----------------------------------------------------------------------
2240    Name: Allen Leung
2241    Date: 2002/03/03 13:20:00 EST
2242    Tag: leunga-20020303-mlrisc-tools
2243    
2244      Added #[ ... ] expressions to mlrisc tools
2245    
2246    ----------------------------------------------------------------------
2247    Name: Matthias Blume
2248    Date: 2002/02/27 12:29:00 EST
2249    Tag: blume-20020227-cdebug
2250    Description:
2251    
2252    - made types in structure C and C_Debug to be equal
2253    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2254    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2255    
2256    ----------------------------------------------------------------------
2257    Name: Matthias Blume
2258    Date: 2002/02/26 12:00:00 EST
2259    Tag: blume-20020226-ffi
2260    Description:
2261    
2262    1. Fixed a minor bug in CM's "noweb" tool:
2263       If numbering is turned off, then truly don't number (i.e., do not
2264       supply the -L option to noweb).  The previous behavior was to supply
2265       -L'' -- which caused noweb to use the "default" line numbering scheme.
2266       Thanks to Chris Richards for pointing this out (and supplying the fix).
2267    
2268    2. Once again, I reworked some aspects of the FFI:
2269    
2270       A. The incomplete/complete type business:
2271    
2272       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2273         gone!
2274       - ML types representing an incomplete type are now *equal* to
2275         ML types representing their corresponding complete types (just like
2276         in C).  This is still safe because ml-nlffigen will not generate
2277         RTTI for incomplete types, nor will it generate functions that
2278         require access to such RTTI.   But when ML code generated from both
2279         incomplete and complete versions of the C type meet, the ML types
2280         are trivially interoperable.
2281    
2282         NOTE:  These changes restore the full generality of the translation
2283         (which was previously lost when I eliminated functorization)!
2284    
2285       B. Enum types:
2286    
2287       - Structure C now has a type constructor "enum" that is similar to
2288         how the "su" constructor works.  However, "enum" is not a phantom
2289         type because each "T enum" has values (and is isomorphic to
2290         MLRep.Signed.int).
2291       - There are generic access operations for enum objects (using
2292         MLRep.Signed.int).
2293       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2294         * The structure contains the definition of type "mlrep" (the ML-side
2295         representation type of the enum).  Normally, mlrep is the same
2296         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2297         then mlrep will be defined as a datatype -- thus facilitating
2298         pattern matching on mlrep values.
2299         ("-ec" will be suppressed if there are duplicate values in an
2300          enumeration.)
2301         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2302         will be generated for each C enum constant xxx.
2303         * Conversion functions m2i and i2m convert between mlrep and
2304         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2305         * Coversion functions c and ml convert between mlrep and "tag enum".
2306         * Access functions (get/set) fetch and store mlrep values.
2307       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2308         enumerations are merged into one single enumeration represented by
2309         structure E_'.
2310    
2311    ----------------------------------------------------------------------
2312    Name: Allen Leung
2313    Date: 2002/02/25 04:45:00 EST
2314    Tag: leunga-20020225-cps-spill
2315    
2316    This is a new implementation of the CPS spill phase.
2317    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2318    In case of problems, replace it with the old file spill.sml
2319    
2320    The current compiler runs into some serious performance problems when
2321    constructing a large record.  This can happen when we try to compile a
2322    structure with many items.  Even a very simple structure like the following
2323    makes the compiler slow down.
2324    
2325        structure Foo = struct
2326           val x_1 = 0w1 : Word32.int
2327           val x_2 = 0w2 : Word32.int
2328           val x_3 = 0w3 : Word32.int
2329           ...
2330           val x_N = 0wN : Word32.int
2331        end
2332    
2333    The following table shows the compile time, from N=1000 to N=4000,
2334    with the old compiler:
2335    
2336    N
2337    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2338           MLRISC ra                               0.06u  0.00s  0.05g
2339              (spills = 0 reloads = 0)
2340           TOTAL                                   0.63u  0.07s  0.21g
2341    
2342    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2343           MLRISC ra                               5.68u  0.59s  3.93g
2344              (spills = 0 reloads = 0)
2345           TOTAL                                   14.71u  0.99s  4.81g
2346    
2347    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2348           MLRISC ra                               5.54u  0.65s  3.91g
2349              (spills = 543 reloads = 1082)
2350           TOTAL                                   65.40u  3.13s  6.00g
2351    
2352    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2353           MLRISC ra                               0.80u  0.10s  0.55g
2354              (spills = 42 reloads = 84)
2355           TOTAL                                   129.42u  5.10s  4.13g
2356    
2357    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2358           MLRISC ra                               2.69u  0.27s  1.38g
2359              (spills = 62 reloads = 124)
2360           TOTAL                                   682.48u  19.61s  13.99g
2361    
2362    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2363           MLRISC ra                               4.96u  0.27s  2.72g
2364              (spills = 85 reloads = 170)
2365           TOTAL                                   2375.26u  57.21s  48.00g
2366    
2367    As you can see the old cps spill module suffers from some serious
2368    performance problem.  But since I cannot decipher the old code fully,
2369    instead of patching the problems up, I'm reimplementing it
2370    with a different algorithm.  The new code is more modular,
2371    smaller when compiled, and substantially faster
2372    (O(n log n) time and O(n) space).  Timing of the new spill module:
2373    
2374    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2375          MLRISC ra                               0.25u  0.02s  0.15g
2376             (spills=1 reloads=3)
2377          TOTAL                                   7.74u  0.34s  1.62g
2378    
2379    Implementation details:
2380    
2381    As far as I can tell, the purpose of the CPS spill module is to make sure the
2382    number of live variables at any program point (the bandwidth)
2383    does not exceed a certain limit, which is determined by the
2384    size of the spill area.
2385    
2386    When the bandwidth is too large, we decrease the register pressure by
2387    packing live variables into spill records.  How we achieve this is
2388    completely different than what we did in the old code.
2389    
2390    First, there is something about the MLRiscGen code generator
2391    that we should be aware of:
2392    
2393    o MLRiscGen performs code motion!
2394    
2395       In particular, it will move floating point computations and
2396       address computations involving only the heap pointer to
2397       their use sites (if there is only a single use).
2398       What this means is that if we have a CPS record construction
2399       statement
2400    
2401           RECORD(k,vl,w,e)
2402    
2403       we should never count the new record address w as live if w
2404       has only one use (which is often the case).
2405    
2406       We should do something similar to floating point, but the transformation
2407       there is much more complex, so I won't deal with that.
2408    
2409    Secondly, there are now two new cps primops at our disposal:
2410    
2411     1. rawrecord of record_kind option
2412        This pure operator allocates some uninitialized storage from the heap.
2413        There are two forms:
2414    
2415         rawrecord NONE [INT n]  allocates a tagless record of length n
2416         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2417                                     and initializes the tag.
2418    
2419     2. rawupdate of cty
2420          rawupdate cty (v,i,x)
2421          Assigns to x to the ith component of record v.
2422          The storelist is not updated.
2423    
2424    We use these new primops for both spilling and increment record construction.
2425    
2426     1. Spilling.
2427    
2428        This is implemented with a linear scan algorithm (but generalized
2429        to trees).  The algorithm will create a single spill record at the
2430        beginning of the cps function and use rawupdate to spill to it,
2431        and SELECT or SELp to reload from it.  So both spills and reloads
2432        are fine-grain operations.  In contrast, in the old algorithm
2433        "spills" have to be bundled together in records.
2434    
2435        Ideally, we should sink the spill record construction to where
2436        it is needed.  We can even split the spill record into multiple ones
2437        at the places where they are needed.  But CPS is not a good
2438        representation for global code motion, so I'll keep it simple and
2439        am not attempting this.
2440    
2441     2. Incremental record construction (aka record splitting).
2442    
2443        Long records with many component values which are simulatenously live
2444        (recall that single use record addresses are not considered to
2445         be live) are constructed with rawrecord and rawupdate.
2446        We allocate space on the heap with rawrecord first, then gradually
2447        fill it in with rawupdate.  This is the technique suggested to me
2448        by Matthias.
2449    
2450        Some restrictions on when this is applicable:
2451        1. It is not a VECTOR record.  The code generator currently does not handle
2452           this case. VECTOR record uses double indirection like arrays.
2453        2. All the record component values are defined in the same "basic block"
2454           as the record constructor.  This is to prevent speculative
2455           record construction.
2456    
2457    ----------------------------------------------------------------------
2458    Name: Allen Leung
2459    Date: 2002/02/22 01:02:00 EST
2460    Tag: leunga-20020222-mlrisc-tools
2461    
2462    Minor bug fixes in the parser and rewriter
2463    
2464    ----------------------------------------------------------------------
2465    Name: Allen Leung
2466    Date: 2002/02/21 20:20:00 EST
2467    Tag: leunga-20020221-peephole
2468    
2469    Regenerated the peephole files.  Some contained typos in the specification
2470    and some didn't compile because of pretty printing bugs in the old version
2471    of 'nowhere'.
2472    
2473    ----------------------------------------------------------------------
2474    Name: Allen Leung
2475    Date: 2002/02/19 20:20:00 EST
2476    Tag: leunga-20020219-mlrisc-tools
2477    Description:
2478    
2479       Minor bug fixes to the mlrisc-tools library:
2480    
2481       1.  Fixed up parsing colon suffixed keywords
2482       2.  Added the ability to shut the error messages up
2483       3.  Reimplemented the pretty printer and fixed up/improved
2484           the pretty printing of handle and -> types.
2485       4.  Fixed up generation of literal symbols in the nowhere tool.
2486       5.  Added some SML keywords to to sml.sty
2487    
2488    ----------------------------------------------------------------------
2489    Name: Matthias Blume
2490    Date: 2002/02/19 16:20:00 EST
2491    Tag: blume-20020219-cmffi
2492    Description:
2493    
2494    A wild mix of changes, some minor, some major:
2495    
2496    * All C FFI-related libraries are now anchored under $c:
2497        $/c.cm      --> $c/c.cm
2498        $/c-int.cm  --> $c/internals/c-int.cm
2499        $/memory.cm --> $c/memory/memory.cm
2500    
2501    * "make" tool (in CM) now treats its argument pathname slightly
2502      differently:
2503        1. If the native expansion is an absolute name, then before invoking
2504           the "make" command on it, CM will apply OS.Path.mkRelative
2505           (with relativeTo = OS.FileSys.getDir()) to it.
2506        2. The argument will be passed through to subsequent phases of CM
2507           processing without "going native".  In particular, if the argument
2508           was an anchored path, then "make" will not lose track of that anchor.
2509    
2510    * Compiler backends now "know" their respective C calling conventions
2511      instead of having to be told about it by ml-nlffigen.  This relieves
2512      ml-nlffigen from one of its burdens.
2513    
2514    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
2515    
2516    * Export C_DEBUG and C_Debug from $c/c.cm.
2517    
2518    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
2519      subtyping relationship between incomplete pointers and their complete
2520      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
2521      with the convention of instantiating 'o with ('t, 'c) obj whenever
2522      the pointer target type is complete.  In the incomplete case, 'o
2523      will be instantiated with some "'c iobj" -- a type obtained by
2524      using one of the functors PointerToIncompleteType or PointerToCompleteType.
2525    
2526      Operations that work on both incomplete and complete pointer types are
2527      typed as taking an 'o ptr while operations that require the target to
2528      be known are typed as taking some ('t, 'c) obj ptr.
2529    
2530      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
2531      where void is an eqtype without any values.  This makes it possible
2532      to work on voidptr values using functions meant to operate on light
2533      incomplete pointers.
2534    
2535    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
2536      been vastly simplified.
2537    
2538    ----------------------------------------------------------------------
2539    Name: Matthias Blume
2540    Date: 2002/02/19 10:48:00 EST
2541    Tag: blume-20020219-pqfix
2542    Description:
2543    
2544    Applied Chris Okasaki's bug fix for priority queues.
2545    
2546    ----------------------------------------------------------------------
2547    Name: Matthias Blume
2548    Date: 2002/02/15 17:05:00
2549    Tag: Release_110_39
2550    Description:
2551    
2552    Last-minute retagging is becoming a tradition... :-(
2553    
2554    This is the working release 110.39.
2555    
2556    ----------------------------------------------------------------------
2557    Name: Matthias Blume
2558    Date: 2002/02/15 16:00:00 EST
2559    Tag: Release_110_39-orig
2560    Description:
2561    
2562    Working release 110.39.  New bootfiles.
2563    
2564    (Update: There was a small bug in the installer so it wouldn't work
2565    with all shells.  So I retagged. -Matthias)
2566    
2567    ----------------------------------------------------------------------
2568    Name: Matthias Blume
2569    Date: 2002/02/15 14:17:00 EST
2570    Tag: blume-20020215-showbindings
2571    Description:
2572    
2573    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
2574    the latter can be useful for exploring what bindings are available at
2575    the interactive prompt.  (The first function returns only the list
2576    of symbols that are really bound, the second prints those but also the
2577    ones that CM's autoloading mechanism knows about.)
2578    
2579    ----------------------------------------------------------------------
2580    Name: Matthias Blume
2581    Date: 2002/02/15 12:08:00 EST
2582    Tag: blume-20020215-iptrs
2583    Description:
2584    
2585    Two improvements to ml-nlffigen:
2586    
2587      1. Write files only if they do not exist or if their current contents
2588         do not coincide with what's being written.  (That is, avoid messing
2589         with the time stamps unless absolutely necessary.)
2590    
2591      2. Implement a "repository" mechanism for generated files related
2592         to "incomplete pointer types".   See the README file for details.
2593    
2594    ----------------------------------------------------------------------
2595    Name: Matthias Blume
2596    Date: 2002/02/14 11:50:00 EST
2597    Tag: blume-20020214-quote
2598    Description:
2599    
2600    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
2601    because of the new and improved tag generation scheme.  (Thanks to Allen
2602    Leung for pointing it out.)
2603    
2604    ----------------------------------------------------------------------
2605    Name: Lal George
2606    Date: 2002/02/14 09:55:27 EST 2002
2607    Tag: george-20020214-isabelle-bug
2608    Description:
2609    
2610    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
2611    of Isabelle on the x86.
2612    
2613    From Allen:
2614    -----------
2615     I've found the problem:
2616    
2617         in ra-core.sml, I use the counter "blocked" to keep track of the
2618         true number of elements in the freeze queue.  When the counter goes
2619         to zero, I skip examining the queue.  But I've messed up the
2620         bookkeeping in combine():
2621    
2622             else ();
2623             case !ucol of
2624               PSEUDO => (if !cntv > 0 then
2625                     (if !cntu > 0 then blocked := !blocked - 1 else ();
2626                                        ^^^^^^^^^^^^^^^^^^^^^^^
2627                      moveu := mergeMoveList(!movev, !moveu)
2628                     )
2629                  else ();
2630    
2631         combine() is called to coalesce two nodes u and v.
2632         I think I was thinking that if the move counts of u and v are both
2633         greater than zero then after they are coalesced then one node is
2634         removed from the freeze queue.  Apparently I was thinking that
2635         both u and v are of low degree, but that's clearly not necessarily true.
2636    
2637    
2638    02/12/2002:
2639        Here's the patch.  HOL now compiles.
2640    
2641        I don't know how this impact on performance (compile
2642        time or runtime).  This bug caused the RA (especially on the x86)
2643        to go thru the potential spill phase when there are still nodes on the
2644        freeze queue.
2645    
2646    
2647    
2648    
2649    ----------------------------------------------------------------------
2650    Name: Matthias Blume
2651    Date: 2002/02/13 22:40:00 EST
2652    Tag: blume-20020213-fptr-rtti
2653    Description:
2654    
2655    Fixed a bug in ml-nlffigen that was introduced with one of the previous
2656    updates.
2657    
2658    ----------------------------------------------------------------------
2659    Name: Matthias Blume
2660    Date: 2002/02/13 16:41:00 EST
2661    Tag: blume-20020213-cmlpq
2662    Description:
2663    
2664    Added new priority queue export symbols (which have just been added to
2665    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
2666    not compile and the installer would choke.)
2667    
2668    ----------------------------------------------------------------------
2669    Name: Matthias Blume
2670    Date: 2002/02/13 16:15:00 EST
2671    Tag: blume-20020213-various
2672    Description:
2673    
2674    1. More tweaks to ml-nlffigen:
2675    
2676       - better internal datastructures (resulting in slight speedup)
2677       - "-match" option requires exact match
2678       - "localized" gensym counters (untagged structs/unions nested within
2679         other structs/unions or within typedefs get a fresh counter; their
2680         tag will be prefixed by a concatenation of their parents' tags)
2681       - bug fixes (related to calculation of transitive closure of types
2682         to be included in the output)
2683    
2684    2. Minor Basis updates:
2685    
2686       - added implementations for List.collate and Option.app
2687    
2688    ----------------------------------------------------------------------
2689    Name: Matthias Blume
2690    Date: 2002/02/11 15:55:00 EST
2691    Tag: blume-20020211-gensym
2692    Description:
2693    
2694    Added a "-gensym" option to command line of ml-nlffigen.  This can be
2695    used to specify a "stem" -- a string that is inserted in all "gensym'd"
2696    names (ML structure names that correspond to unnamed C structs, unions,
2697    and enums), so that separate runs of ml-nlffigen do not clash.
2698    
2699    ----------------------------------------------------------------------
2700    Name: Matthias Blume
2701    Date: 2002/02/11 12:05:00 EST
2702    Tag: blume-20020211-gensml
2703    Description:
2704    
2705    A quick fix for a problem with GenSML (in the pgraph-util library):
2706    Make generation of toplevel "local" optional.  (Strictly speaking,
2707    signature definitions within "local" are not legal SML.)
2708    
2709    Other than that: updates to INSTALL and cm/TODO.
2710    
2711    ----------------------------------------------------------------------
2712    Name: Matthias Blume
2713    Date: 2002/02/08 15:00:00 EST
2714    Tag: blume-20020208-uniquepid
2715    Description:
2716    
2717    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
2718    
2719    1. The installer (config/install.sh) has gotten smarter:
2720    
2721         - Configuration options are a bit easier to specify now
2722           (in config/targets).
2723         - Bug in recognizing .tar.bz2 files fixed.
2724         - Installer automatically resolves dependencies between
2725           configuration options (e.g., if you ask for eXene, you will
2726           also get cml -- regardless whether you asked for it or not).
2727         - Installer can run in "quieter mode" by setting the environment
2728           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
2729           "completely silent", though.
2730         - Build HashCons library as part of smlnj-lib.
2731    
2732    2. A new scheme for assigning persistent identifiers to compilation
2733       units (and, by extension, to types etc.) has been put into place.
2734       This fixes a long-standing bug where types and even dynamic values
2735       can get internally confused, thereby compromising type safety
2736       (abstraction) and dynamic correctness.  See
2737    
2738         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
2739    
2740       for an example of how things could go wrong until now.
2741    
2742       The downside of the new scheme is that pids are not quite as
2743       persistent as they used to be: CM will generate a fresh pid
2744       for every compilation unit that it thinks it sees for the first
2745       time.  That means that if you compile starting from a clean, fresh
2746       source tree at two different times, you end up with different
2747       binaries.
2748    
2749       Cutoff recompilation, however, has not been compromised because
2750       CM keeps pid information in special caches between runs.
2751    
2752    ----------------------------------------------------------------------
2753    Name: Lal George
2754    Date: 2002/02/07 15:34:13 EST 2002
2755    Tag: <none>
2756    Description:
2757    
2758    Compilers that generate assembly code may produce  global labels
2759    whose value is resolved at link time. The various peephole optimization
2760    modules did not take this in account.
2761    
2762    TODO. The Labels.addrOf function should really return an option
2763    type so that clients are forced to deal with this issue, rather
2764    than an exception being raised.
2765    
2766    ----------------------------------------------------------------------
2767    Name: Lal George
2768    Date: 2002/02/06 13:55:02 EST
2769    Tag: george-20020206-ra-breakup
2770    Description:
2771    
2772    1. A bug fix from Allen.
2773    
2774        A typo causes extra fstp %st(0)'s to be generated at compensation
2775        edges, which might cause stack underflow traps at runtime.  This
2776        occurs in fft where there are extraneous fstps right before the 'into'
2777        trap instruction (in this case they are harmless since none of the
2778        integers overflow.)
2779    
2780    2. Pulled out various utility modules that were embedded in the modules
2781       of the register allocator. I need these modules for other purposes, but
2782       they are not complete enough to put into a library (just yet).
2783    ----------------------------------------------------------------------
2784    Name: Matthias Blume
2785    Date: 2002/01/31 16:05:00 EST
2786    Tag: blume-20020131-sparc-ccalls
2787    Description:
2788    
2789    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
2790       of extra stack space by mistake.  Fixed.
2791    
2792    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
2793    
2794    ----------------------------------------------------------------------
2795    Name: Allen Leung
2796    Date: 2002/01/30
2797    Tag: leunga-20020130-nowhere-bug-fix
2798    Description:
2799    
2800       MLRISC bug fixes:
2801       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
2802       2. MachineInt.fromString was negating its value.
2803    
2804    ----------------------------------------------------------------------
2805    Name: Matthias Blume
2806    Date: 2002/01/29
2807    Tag: blume-20020129-INSTALL
2808    Description:
2809    
2810    - Added somewhat detailed installation instructions (file INSTALL).
2811    - Fixed curl-detection bug in config/install.sh.
2812    - It is now possible to select the URL getter using the URLGETTER
2813      environment variable:
2814    
2815          not set / "unknown"      --> automatic detection (script tries wget,
2816                                       curl, and lynx)
2817          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
2818                                       how to properly invoke them)
2819          other                    --> use $URLGETTER directly, it must take
2820                                       precisely two command-line arguments
2821                                       (source URL and destination file name)
2822    
2823    ----------------------------------------------------------------------
2824    Name: Matthias Blume
2825    Date: 2002/01/28
2826    Tag: blume-20020128-sparc-ccalls
2827    Description:
2828    
2829    - Fixed problem with calculation of "used" registers in sparc-c-calls.
2830    - Make use of the allocParam argument in sparc-c-calls.
2831    
2832    ----------------------------------------------------------------------
2833    Name: Matthias Blume
2834    Date: 2002/01/28
2835    Tag: blume-20020128-allocParam
2836    Description:
2837    
2838    John Reppy:  Changes c-calls API to accept client-callback for
2839    allocating extra stack space.
2840    me: Corresponding changes to mlriscGen (using a dummy argument that
2841        does not change the current behavior).
2842    
2843    ----------------------------------------------------------------------
2844    Name: Matthias Blume
2845    Date: 2002/01/28 12:00:00
2846    Tag: Release_110_38
2847    Description:
2848    
2849    This time for real!!!
2850    
2851    ----------------------------------------------------------------------
2852    Name: Matthias Blume
2853    Date: 2002/01/28 10:56:00 EST
2854    Tag: blume-20020128-retraction
2855    Description:
2856    
2857    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
2858       with blume-Release_110_38-retracted.)
2859    
2860    1. Fixed a problem with incorrect rounding modes in real64.sml.
2861       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
2862    
2863    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
2864       and enums fixed.  The naming of corresponding ML identifiers should
2865       now be consistent again.
2866    
2867    ----------------------------------------------------------------------
2868    Name: Allen Leung
2869    Date: 2002/01/27
2870    Tag: leunga-20020127-nowhere
2871    Description:
2872    
2873       Added a target called nowhere in the configuration scripts.
2874       Enabling this will build the MLRISC 'nowhere' tool (for translating
2875       programs with where-clauses into legal SML code) during installation.
2876    
2877    ----------------------------------------------------------------------
2878    Name: Matthias Blume
2879    Date: 2002/01/25 21:27:00 EST
2880    Tag: blume-Release_110_38-retracted
2881    Description:
2882    
2883    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
2884    
2885    README will be added later.
2886    
2887    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
2888    (Release_110_38) removed.  Reason: Last-minute bug fixes.
2889    
2890    ----------------------------------------------------------------------
2891    Name: Matthias Blume
2892    Date: 2002/01/25
2893    Tag: blume-20020125-ffi
2894    Description:
2895    
2896    A large number of tweaks and improvements to ml-nlffi-lib and
2897    ml-nlffigen:
2898    
2899       - ML represenation types have been streamlined
2900       - getter and setter functions work with concrete values, not abstract
2901         ones where possible
2902       - ml-nlffigen command line more flexible (see README file there)
2903       - some bugs have been fixed (hopefully)
2904    
2905    ----------------------------------------------------------------------
2906    Name: Lal George
2907    Date: 2002/01/24
2908    Tag: george-20020124-risc-ra-interface
2909    Description:
2910    
2911       There is a dramatic simplification in the interface to the
2912       register allocator for RISC architectures as a result of making
2913       parallel copy instructions explicit.
2914    
2915    ----------------------------------------------------------------------
2916    Name: Matthias Blume
2917    Date: 2002/01/22
2918    Tag: blume-20020122-x86-ccalls
2919    Description:
2920    
2921    Bug fix for c-calls on x86 (having to do with how char- and
2922    short-arguments are being handled).
2923    
2924    ----------------------------------------------------------------------
2925    Name: Matthias Blume
2926    Date: 2002/01/21
2927    Tag: blume-20020121-ff
2928    Description:
2929    
2930    Another day of fiddling with the FFI...
2931    
2932    1. Bug fix/workaround:  CKIT does not complain about negative array
2933       dimensions, so ml-nlffigen has to guard itself against this possibility.
2934       (Otherwise a negative dimension would send it into an infinite loop.)
2935    
2936    2. Some of the abstract types (light objects, light pointers, most "base"
2937       types) in structure C are now eqtypes.
2938    
2939    3. Added constructors and test functions for NULL function pointers.
2940    
2941    ----------------------------------------------------------------------
2942    Name: Matthias Blume
2943    Date: 2002/01/18
2944    Tag: blume-20020118-ready-for-new-release
2945    Description:
2946    
2947    Made config/srcarchiveurl point to a new place.  (Will provide boot
2948    files shortly.)
2949    
2950    Maybe we christen this to be 110.38?
2951    
2952    ----------------------------------------------------------------------
2953    Name: Matthias Blume
2954    Date: 2002/01/18
2955    Tag: blume-20020118-more-ffifiddle
2956    Description:
2957    
2958    Today's FFI fiddling:
2959    
2960      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
2961        C.Set.* that use concrete (MLRep.*) arguments and results instead
2962        of abstract ones.
2963    
2964      - Provided word-style bit operations etc. for "int" representation
2965        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
2966        and Long.
2967    
2968    ----------------------------------------------------------------------
2969    Name: Matthias Blume
2970    Date: 2002/01/18
2971    Tag: blume-20020118-use-x86-fp
2972    Description:
2973    
2974    Now that x86-fast-fp seems to be working, I turned it back on again
2975    by default.  (Seems to work fine now, even with the FFI.)
2976    
2977    Other than that, I added some documentation about the FFI to
2978    src/ml-nlffigen/README and updated the FFI test examples in
2979    src/ml-nlffi-lib/Tests/*.
2980    
2981    ----------------------------------------------------------------------
2982    Name: Allen Leung
2983    Date: 2002/01/17
2984    Tag: leunga-20020117-x86-fast-fp-call
2985    Description:
2986    
2987       1. Fixed a problem with handling return fp values when x86's fast fp
2988          mode is turned on.
2989    
2990       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
2991          of %f32.
2992    
2993       3. Added a constructor INT32lit to the ast of MLRISC tools.
2994    
2995    ----------------------------------------------------------------------
2996    Name: Matthias Blume
2997    Date: 2002/01/16
2998    Tag: blume-20020116-ffifiddle
2999    Description:
3000    
3001    More fiddling with the FFI interface:
3002    
3003     - Make constness 'c instead of rw wherever possible.  This eliminates
3004       the need for certain explicit coercions.  (However, due to ML's
3005       value polymorphism, there will still be many cases where explicit
3006       coercions are necessary.  Phantom types are not the whole answer
3007       to modeling a subtyping relationship in ML.)
3008    
3009     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3010    
3011     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3012       clumsy workaround for varargs problem.)
3013    
3014    ----------------------------------------------------------------------
3015    Name: Lal George
3016    Date: 2002/01/15
3017    Tag: <none>
3018    Description:
3019    
3020    1. Since COPY instructions are no longer native to the architecture,
3021       a generic functor can be used to implement the expandCopies function.
3022    
3023    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3024       TEXT segment.
3025    
3026    ----------------------------------------------------------------------
3027    Name: Matthias Blume
3028    Date: 2002/01/15
3029    Tag: blume-20020115-ffiupdates
3030    Description:
3031    
3032    1. Fix for bug resulting in single-precision float values being returned
3033       incorrectly from FFI calls.
3034    
3035    2. Small modifications to C FFI API:
3036    
3037        - memory-allocation routines return straight objects (no options)
3038          and raise an exception in out-of-memory situations
3039        - unsafe extensions to cast between function pointers and pointers
3040          from/to ints
3041        - added structure C_Debug as an alternative to structure C where
3042          pointer-dereferencing (|*| and |*!) always check for null-pointers
3043        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3044          but also takes a (possibly empty) list of existing library handles
3045          that the current library depends on
3046    
3047    ----------------------------------------------------------------------
3048    Name: Matthias Blume
3049    Date: 2002/01/10
3050    Tag: blume-20020110-newffigen
3051    Description:
3052    
3053    1. Updates to portable graph code.
3054    
3055    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3056       more scalable now so that even huge interfaces such as the one
3057       for GTK compile in finite time and space. :-)
3058       See src/ml-nlffigen/README for details on what's new.
3059    
3060    ----------------------------------------------------------------------
3061    Name: Lal George
3062    Date: 2001/01/09 14:31:35 EST 2002
3063    Tag: george-20011206-rm-native-copy
3064    Description:
3065    
3066            Removed the native COPY and FCOPY instructions
3067            from all the architectures and replaced it with the
3068            explicit COPY instruction from the previous commit.
3069    
3070            It is now possible to simplify many of the optimizations
3071            modules that manipulate copies. This has not been
3072            done in this change.
3073    
3074    ----------------------------------------------------------------------
3075    Name: Lal George
3076    Date: 2001/12/06 16:50:13 EST 2001
3077    Tag: george-20011206-mlrisc-instruction
3078    Description:
3079    
3080    Changed the representation of instructions from being fully abstract
3081    to being partially concrete. That is to say:
3082    
3083      from
3084            type instruction
3085    
3086      to
3087            type instr                              (* machine instruction *)
3088    
3089            datatype instruction =
3090                LIVE of {regs: C.cellset, spilled: C.cellset}
3091              | KILL of {regs: C.cellset, spilled: C.cellset}
3092              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3093              | ANNOTATION of {i: instruction, a: Annotations.annotation}
3094              | INSTR of instr
3095    
3096    This makes the handling of certain special instructions that appear on
3097    all architectures easier and uniform.
3098    
3099    LIVE and KILL say that a list of registers are live or killed at the
3100    program point where they appear. No spill code is generated when an
3101    element of the 'regs' field is spilled, but the register is moved to
3102    the 'spilled' (which is present, more for debugging than anything else).
3103    
3104    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3105    We used to generate:
3106    
3107            DEFFREG f1
3108            f1 := f2 + f3
3109            trapb
3110    
3111    but now generate:
3112    
3113            f1 := f2 + f3
3114            trapb
3115            LIVE {regs=[f1,f2,f3], spilled=[]}
3116    
3117    Furthermore, the DEFFREG (hack) required that all floating point instruction
3118    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3119    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3120    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3121    this architecture).
3122    
3123    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3124    all the architectures. This will result in further simplification of the
3125    register allocator that must be aware of them for coalescing purposes, and
3126    will also simplify certain aspects of the machine description that provides
3127    callbacks related to parallel copies.
3128    
3129    ANNOTATION should be obvious, and now INSTR represents the honest to God
3130    machine instruction set!
3131    
3132    The <arch>/instructions/<arch>Instr.sml files define certain utility
3133    functions for making porting easier -- essentially converting upper case
3134    to lower case. All machine instructions (of type instr) are in upper case,
3135    and the lower case form generates an MLRISC instruction. For example on
3136    the alpha we have:
3137    
3138      datatype instr =
3139         LDA of {r:cell, b:cell, d:operand}
3140       | ...
3141    
3142      val lda : {r:cell, b:cell, d:operand} -> instruction
3143        ...
3144    
3145    where lda is just (INSTR o LDA), etc.
3146    
3147    ----------------------------------------------------------------------
3148    Name: Matthias Blume
3149    Date: 2001/11/22 21:40:00 EST
3150    Tag: Release_110_37
3151    Description:
3152    
3153    Release 110.37.  This time for real.
3154    
3155    ----------------------------------------------------------------------
3156    Name: Matthias Blume
3157    Date: 2001/11/21 16:35:00 EST
3158    Tag: blume-20011121-foot-in-mouth
3159    Description:
3160    
3161    Removed the "Release_110_37" tag because of a serious bug.
3162    This will be re-tagged once the bug is fixed.
3163    
3164    ----------------------------------------------------------------------
3165    Name: Matthias Blume
3166    Date: 2001/11/21 16:14:00 EST
3167    Tag: blume-20011121-forgottenfile
3168    Description:
3169    
3170    Forgot to add a file.  (Just a .tex-file -- part of
3171    the CM manual source.)
3172    
3173    ----------------------------------------------------------------------
3174    Name: Matthias Blume
3175    Date: 2001/11/21 16:10:00 EST
3176    Tag: blume-20011121-invalid_110_37
3177    Description:
3178    
3179    Note: I removed the original tag "Release_110_37" from this commit
3180          because we found a serious bug in all non-x86 backends.
3181          - Matthias
3182    
3183    1. Modifications to the SML/NJ code generator and to the runtime system
3184       so that code object name strings are directly inserted into code
3185       objects at code generation time.  The only business the runtime system
3186       has with this is now to read the name strings on occasions.
3187       (The encoding of the name string has also changed somewhat.)
3188    
3189    2. CM now implements a simple "set calculus" for specifying export lists.
3190       In particular, it is now possible to refer to the export lists of
3191       other libraries/groups/sources and form unions as well as differences.
3192       See the latest CM manual for details.
3193    
3194    3. An separate notion of "proxy" libraries has again be eliminated from
3195       CM's model.  (Proxy libraries are now simply a special case of using
3196       the export list calculus.)
3197    
3198    4. Some of the existing libraries now take advantage of the new set
3199       calculus.
3200       (Notice that not all libraries have been converted because some
3201       of the existing .cm-files are supposed to be backward compatible
3202       with 110.0.x.)
3203    
3204    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3205       "General.exnMessage"!  The former relies on a certain hook to be
3206       initialized, and that often does not happen in the stand-alone case.)
3207    
3208    ----------------------------------------------------------------------
3209    Name: Lal George
3210    Date: 2001/11/21  13:56:18 EST
3211    Tag: george-2001121-pseudo-ops
3212    Description:
3213    
3214      Implemented a complete redesign of MLRISC pseudo-ops. Now there
3215      ought to never be any question of incompatabilities with
3216      pseudo-op syntax expected by host assemblers.
3217    
3218      For now, only modules supporting GAS syntax are implemented
3219      but more should follow, such as MASM, and vendor assembler
3220      syntax, e.g. IBM as, Sun as, etc.
3221    
3222    ----------------------------------------------------------------------
3223  Name: Matthias Blume  Name: Matthias Blume
3224  Date: 2001/11/14 11:52:00 EST  Date: 2001/11/14 11:52:00 EST
3225  Tag: blume-20011114-srcname  Tag: blume-20011114-srcname
# Line 194  Line 3401 
3401    
3402  ----------------------------------------------------------------------  ----------------------------------------------------------------------
3403  Name: Matthias Blume  Name: Matthias Blume
 >>>>>>> 1.169  
3404  Date: 2001/09/18 15:35:00 EDT  Date: 2001/09/18 15:35:00 EDT
3405  Tag: blume-20010918-readme11036  Tag: blume-20010918-readme11036
3406  Description:  Description:

Legend:
Removed from v.977  
changed lines
  Added in v.1614

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