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 958, Fri Oct 12 03:54:43 2001 UTC revision 1572, Fri Jul 30 22:55:23 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: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/07/30 17:55:00 CDT
18    Tag: blume-20040730-various
19    Description:
20    
21    Gearing up towards 110.47...
22    
23    - various minor bugfixes to ml-nlffigen
24    - a beginning of a manual for nlffi
25    
26    - eliminated 'export name=value' in config/install.sh as this does
27      not work with certain versions of /bin/sh
28      (Thanks to David King at Motorola for catching this.)
29    
30    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
31      - added a test for tm==NULL to gmtime.c and localtime.c
32      - applied patch for incorrect GetFileTime under win32
33      - toSeconds -> toMilliseconds in Win32/win32-process.sml
34    
35    ----------------------------------------------------------------------
36    Name: Matthias Blume (blume (at) tti - c (dot) org)
37    Date: 2004/07/21 18:20:00 CDT
38    Tag: blume-20040721-nlffigen
39    Description:
40    
41    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
42      for a typedef to an incomplete type, but leave out the "typ" member.
43      (This is just for consistency.)
44    - Started to produce what is supposed to become better (i.e., comprehensive)
45      documentation of what ml-nlffigen does and produces.
46    
47    ----------------------------------------------------------------------
48    Name: Matthias Blume (blume (at) tti - c (dot) org)
49    Date: 2004/07/14 16:25:00 CDT
50    Tag: blume-20040714-union
51    Description:
52    
53    Added C_UNION to c-calls/c-types.sml and updated the machinery
54    (ml-nlffigen, cproto.sml) that conveys C function interface
55    information to the code generator.
56    
57    However, the actual architecture-specific implementation of function
58    arguments and results that are C unions is still not implemented.
59    
60    ----------------------------------------------------------------------
61    Name: Allen Leung (leunga (at) reservoir (dot) com)
62    Date: 2004/07/14 14:38:00 EST
63    Tag: leunga-110_46_1-ppc-lwzu
64    Description:
65    
66        Added these instructions to the PowerPC architecture:
67        LBZU(X), LHZU(X), LWZU(X),
68        STWU(X), STFDU, STFSU
69    
70        etc...
71    
72        Note: I haven't added their instruction encoding into the description.
73    
74    ----------------------------------------------------------------------
75    Name: Allen Leung (leunga (at) reservoir (dot) com)
76    Date: 2004/07/13 15:04:00 EST
77    Tag: leunga-110_46_1-ppc-lwarx
78    Description:
79    
80       Added the two instructions LWARX and STWCX to the PowerPC
81    instruction set.
82    
83       A (untested) rewrite of loop-structure.sml.  The old version
84    is completely broken.
85    
86    ----------------------------------------------------------------------
87    Name: Matthias Blume (blume (at) tti - c (dot) org)
88    Date: 2004/07/13 13:50:00 CDT
89    Tag: blume-20040713-nlffi
90    Description:
91    
92    - use paramAlloc to report c-calls with too many arguments
93      (for PPC version where parameter area is pre-allocated)
94    - added ccall_maxargspace to machspec (to implement the above)
95    - made "make" commend in CM's "make" tool configurable
96    - added option (default: on) for passing the name of the SML/NJ's "bin"
97      directory to "make"; the call looks like this:
98    
99         make <options> SMLNJ_BINDIR=<dir> <target>
100    
101      This can be used by the Makefile to, e.g., pick the "right" version
102      of ml-nlffigen.
103    - minor code tweaks
104    
105    ----------------------------------------------------------------------
106    Name: Matthias Blume (blume (at) tti - c (dot) org)
107    Date: 2004/07/12 22:50:00 CDT
108    Tag: blume-110_46_1-macosx-nlffi
109    Description:
110    
111    NLFFI under Mac OS X now working (sort of).  This is largely untested,
112    though.
113    
114    Note:  1. You have to make a new, clean build of the runtime system.
115           2. There are new BOOTFILES, you have to use them!
116              (Doing the bootstrap process yourself would be *very* painful!
117               If you absolutely have to do it, build the system under
118               a different architecture and then cross-compile.)
119    
120    Version bumped to 110.46.1 to account for runtime data format changes.
121    
122    ----------------------------------------------------------------------
123    Name: Matthias Blume (blume (at) tti - c (dot) org)
124    Date: 2004/06/18 14:30:00 CDT
125    Tag: blume-20040618-unix
126    Description:
127    
128    Changed the implementation of structure Unix so that the same stream
129    is returned every time one of the {text,bin}{In,Out}streamOf functions
130    is invoked on the same proc.  This is not what the spec currently
131    says -- although IMO it arguably should.  (See discussion below.)
132    
133    ----------------------------------------------------------------------
134    Name: Matthias Blume (blume (at) tti - c (dot) org)
135    Date: 2004/06/17 18:15:00 CDT
136    Tag: Release_110_46
137    Description:
138    
139    New working version (110.46).  NEW BOOTFILES!
140    
141    ----------------------------------------------------------------------
142    Name: Matthias Blume (blume (at) tti - c (dot) org)
143    Date: 2004/06/17 17:20:00 CDT
144    Tag: blume-20040617-timer-unix
145    Description:
146    
147    Changed the interface of structures Timer and Unix to match the most
148    recent Basis spec.
149    
150    In the case of Unix there still seems to be an open/weird issue:
151    
152       The {text,bin}{In,Out}streamOf functions are supposed to create
153       fresh streams whenever they are called -- as opposed to have them
154       return the same stream every time.  This design is supposed to
155       prevent space leaks caused by proc values hanging on to streams.
156    
157       The reap function, on the other hand, is supposed to close the
158       streams.  This cannot be done without having a handle on the
159       stream in proc after all...
160    
161    I took the liberty to implement the following stopgap solution:
162    
163      The proc value hangs on to the most recently created stream(s).
164      Reap closes those.  If either or both of the two streams hadn't
165      been created at all yet, then reap will close the corresponding
166      file descriptors directly.
167    
168    PS: I don't understand the original space leak argument anymore.  If
169    a proc hangs on to the imperative stream, then I/O operations on those
170    will advance the state of the cached stream and avoid the space leak.
171    
172    ----------------------------------------------------------------------
173    Name: Matthias Blume (blume (at) tti - c (dot) org)
174    Date: 2004/05/28 16:45:00 CDT
175    Tag: blume-20040528-basis
176    Description:
177    
178    Added signature PACK_REAL and exported functor PrimIO.
179    
180    ----------------------------------------------------------------------
181    Name: Matthias Blume (blume (at) tti - c (dot) org)
182    Date: 2004/05/25 16:00:00 CDT
183    Tag: blume-20040525-group-owner
184    Description:
185    
186    CM now ignores (but still accepts) the "owner" information in group
187    descriptions.  The owner of a group is its next enclosing
188    library. Each group must have a unique owner.  (There is a virtual
189    "toplevel" library that owns groups which are not nested within a real
190    library.)  Previously, each group had to explicitly declare its owner,
191    and CM would check that such a declaration is correct.  The new scheme
192    is to have CM check that for each group there is precisely one owning
193    library.
194    
195    The advantage of the new scheme is that the programmer no longer needs
196    to maintain the somewhat annoying owner information.  The downside is
197    that CM cannot enforce the ownership rule across multiple runs of
198    CM.make.  Fortunately, enclosing the same group in two different
199    libraries A and B which are not part of the same program does not
200    cause real problems.
201    
202    ----------------------------------------------------------------------
203    Name: Matthias Blume (blume (at) tti - c (dot) org)
204    Date: 2004/05/20 16:00:00 CDT
205    Tag: blume-20040520-win32
206    Description:
207    
208    Made the win32 version work again.  (Strangely, a misplaced comma had
209    slipped into win32-process.c which prevented the runtime from being
210    compiled correctly.)
211    
212    Also, included a minor addition to ml-build.bat analogous to what was
213    done in blume-20040519-ml-build.
214    
215    ----------------------------------------------------------------------
216    Name: Matthias Blume (blume (at) tti - c (dot) org)
217    Date: 2004/05/19 22:10:00 CDT
218    Tag: blume-20040519-ml-build
219    Description:
220    
221    Arranged for ml-build to clean up after itself a little bit better.
222    The script generates a temporary SML source file and compiles it using
223    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
224    gets rid of those at the end, so they don't accumulate under .cm.
225    
226    This required a minor change to install.sh because the name of the
227    metadata directory (default: .cm) is actually configurable at
228    installation time.
229    
230    ----------------------------------------------------------------------
231    Name: Matthias Blume (blume (at) tti - c (dot) org)
232    Date: 2004/05/18 15:50:00 CDT
233    Tag: blume-20040518-mkreader
234    Description:
235    
236    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
237    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
238    
239    ----------------------------------------------------------------------
240    Name: Matthias Blume (blume (at) tti - c (dot) org)
241    Date: 2004/05/11 14:35:00 CDT
242    Tag: blume-20040511-win32sock
243    Description:
244    
245    Added previously missing support for many socket-related functions
246    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
247    for the voluminous patch!
248    
249    (I have not tested this patch under win32 yet.)
250    
251    Here is David's e-mail:
252    
253    Hi,
254    
255    Attached to this email you find a diff against sml/nj 110.45
256    that will enable socket support under Windows.
257    
258    To apply the patch (using unix or cygwin)
259    1) gunzip runtime.diff.gz
260    2) "cd" into "src/runtime" in the source tree of a fresh
261       110.45 installation.
262    3) patch -p 1 < [your/path/to]runtime.diff
263    
264    The code compiles fine but has NOT yet been extensively tested.
265    I only ran a few tests for basic socket client functionality
266    (which worked fine).  Especially the functions that use ioctl
267    are not tested at all and might not work (see below).
268    
269    I implemented this since we want to move to a newer version of sml/nj
270    but need socket support in order to use it.  This is the first time I
271    even had a look at the sml/nj source,  so please review my changes
272    before making this part of the distribution!  Here are a few issues
273    that I think might be better for someone to solve who is more
274    familiar with the sml/nj source (and socket programming):
275    
276    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
277      exception since I could not figure out what the windows equivalent
278      of these functions is
279    
280    - In sockets-osdep.h there are a some #include statements that are
281      only used in a few files that include sockets-osdep.h
282    
283    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
284      does not process its return value since I don't know how to
285      report an error upwards.
286    
287    - It would probably be good to have a call to WSACleanup() when
288      the library is unloaded (if there is such a possibility).
289      Otherwise I think Windows will take care of this automatically
290      when the process finishes.
291    
292    - I used ioctlsocket() as a replacement for ioctl() but I have
293      no idea if that is actually the proper replacement on Windows.
294    
295    - All these issues are marked in the code by "FIXME" comments.
296    
297    We use sml/nj extensively in our products and are quite happy
298    with it.  I hope this contribution will help you.
299    
300    Keep up the good work!
301    
302    David
303    
304    ----------------------------------------------------------------------
305    Name: Matthias Blume (blume (at) tti - c (dot) org)
306    Date: 2004/05/11 14:20:00 CDT
307    Tag: blume-20040511-installml
308    Description:
309    
310    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
311    the patch.)
312    
313    ----------------------------------------------------------------------
314    Name: Matthias Blume (blume (at) tti - c (dot) org)
315    Date: 2004/05/11 14:05:00 CDT
316    Tag: blume-20040511-nlffi-netbsd
317    Description:
318    
319    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
320    the patch.)
321    
322    ----------------------------------------------------------------------
323    Name: Matthias Blume (blume (at) tti - c (dot) org)
324    Date: 2004/05/11 12:05:00 CDT
325    Tag: blume-20040511-exports
326    Description:
327    
328    As per request by Adam Chlipala <adam@hcoop.net>, extended various
329    export lists in compiler-related .cm-files.
330    
331    ----------------------------------------------------------------------
332    Name: Matthias Blume (blume (at) tti - c (dot) org)
333    Date: 2004/05/11 11:35:00 CDT
334    Tag: blume-20040511-allsource
335    Description:
336    
337    The installer now honors the "src-smlnj" target again, although its meaning
338    has changed from "all sources required for the compiler" to "all sources
339    the installer knows about".  In other words, if you enable "src-smlnj"
340    in the "targets" file, then the installer will pull in sources for
341    everything.  (Notice that this refers to source code only.  Compiled
342    code is still only installed for modules that were requested explicitly
343    or which are required for other modules that were requested explicitly.)
344    
345    ----------------------------------------------------------------------
346    Name: Matthias Blume (blume (at) tti - c (dot) org)
347    Date: 2004/04/23 17:40:00 CDT
348    Tag: blume-20040423-ieee-scan
349    Description:
350    
351    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
352    in the exponent calculation we get INF or ZERO (depending on the mantissa
353    and the sign of the exponent).
354    
355    ----------------------------------------------------------------------
356    Name: Matthias Blume (blume (at) tti - c (dot) org)
357    Date: 2004/04/23 10:40:00 CDT
358    Tag: blume-20040423-ml-build
359    Description:
360    
361    The ml-build script now terminates with a non-0 status when something
362    goes wrong.
363    
364    ----------------------------------------------------------------------
365    Name: Matthias Blume (blume (at) tti - c (dot) org)
366    Date: 2004/04/22 16:35:00 CDT
367    Tag: blume-20040422-Option
368    Description:
369    
370    Made exception Option to be the same as exception Option.Option
371    (as it should be).
372    
373    ----------------------------------------------------------------------
374    Name: Allen Leung (leunga (at) reservoir (dot) com)
375    Date: 2004/03/19 14:40:00 EST
376    Tag: leunga-20040319-cygwin-nlffi
377    Description:
378    
379    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
380    of SML/NJ.  The problem is that
381    
382       lib = dlopen(NULL, ...)
383       f   = dlsym(lib, "malloc");
384    
385    does not work on Windows unless we explicitly export symbols
386    such as 'malloc' during linking.  We fixed this by explicitly
387    exporting the required symbols with the magic gcc incantation:
388    
389        -Wl,--export-all cygwin.def
390    
391    where cygwin.def is a file containing all the symbols that
392    we wish to export.
393    
394    I suspect this is a Windows problem and we'll have to
395    do the same (somehow with windows compilers) when
396    we build the native win32 version with the system
397    calls LoadLibrary/GetProcAddress.
398    
399    ----------------------------------------------------------------------
400    Name: Matthias Blume (blume (at) tti - c (dot) org)
401    Date: 2004/03/04 16:35:00 CST
402    Tag: blume-20040304-intinf-fmt
403    Description:
404    
405    Fixed problem with IntInf.fmt (sign would show up on the right instead
406    of on the left for BIN, OCT, and HEX).
407    
408    ----------------------------------------------------------------------
409    Name: Matthias Blume (blume (at) tti - c (dot) org)
410    Date: 2004/03/04 11:25:00 CST
411    Tag: blume-20040304-symlinks
412    Description:
413    
414    Fixed problem with installer script (unix only) where bin/ml-yacc and
415    friends pointed (via symlinks) to absolute locations instead of just
416    .run-sml.  This was reported by Vesa A Norrman.
417    
418    ----------------------------------------------------------------------
419    Name: Matthias Blume (blume (at) tti - c (dot) org)
420    Date: 2004/02/13 14:50:00 CST
421    Tag: Release_110_45
422    Description:
423    
424    New working version (110.45). New bootfiles.
425    
426    ----------------------------------------------------------------------
427    Name: Matthias Blume (blume (at) tti - c (dot) org)
428    Date: 2004/01/26 15:15:15 CST
429    Tag: blume-20040126-toplevel
430    Description:
431    
432    Improved handling of exceptions at the interactive toplevel.
433    
434    ----------------------------------------------------------------------
435    Name: Matthias Blume (blume (at) tti - c (dot) org)
436    Date: 2004/01/26 11:25:00
437    Tag: blume-20040126-app
438    Description:
439    
440    Type of top-level "app" corrected.
441    Added code for setting vp_limitPtrMask to Win32-specific runtime.
442    
443    ----------------------------------------------------------------------
444    Name: Matthias Blume (blume (at) tti - c (dot) org)
445    Date: 2003/11/18 17:10 CST
446    Tag: blume-20031118-basis-fiddle
447    Description:
448    
449    - changed Timer interface to what might become the spec
450    - POSIX_FLAGS -> BIT_FLAGS according to spec
451    - some other minor discrepancies wrt. spec eliminated
452    
453    ----------------------------------------------------------------------
454    Name: Matthias Blume (blume (at) tti - c (dot) org)
455    Date: 2003/11/06 12:00:00 CST
456    Tag: Release_110_44
457    Description:
458    
459    New working version (110.44). New bootfiles.
460    
461    ----------------------------------------------------------------------
462    Name: Matthias Blume (blume (at) tti - c (dot) org)
463    Date: 2003/11/04 11:50:00 CST
464    Tag: blume-20031104-move-libraries
465    Description:
466    
467    Eliminated the "dont_move_libraries" directive in config/targets.
468    (The mechanism was broken and could not be fixed easily.  Moreover,
469    there does not seem to be any reason not to move all libraries into
470    lib during installation.  I originally implemented this directive as a
471    backward-compatibility feature when I first introduced the new CM.
472    Now that things have been stable for a long time and going back to the
473    old CM is not an option, there is no reason to keep it around.)
474    
475    ----------------------------------------------------------------------
476    Name: Matthias Blume (blume (at) tti - c (dot) org)
477    Date: 2003/11/03 16:00:00 CST
478    Tag: blume-20031103-installdir
479    Description:
480    
481    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
482    Richards for pointing out the problem and providing the solution.)
483    
484    ----------------------------------------------------------------------
485    Name: Matthias Blume (blume (at) tti - c (dot) org)
486    Date: 2003/10/01 17:05:00 CDT
487    Tag: blume-20031001-lal-mlrisc
488    Description:
489    
490    MLRISC bug fix from Lal.
491    
492    ----------------------------------------------------------------------
493    Name: Matthias Blume (blume (at) tti - c (dot) org)
494    Date: 2003/09/30 16:10:00 CDT
495    Tag: blume-20030930-primio-bat
496    Description:
497    
498    1. Added openVector, nullRd, and nullWr to PRIM_IO.
499    2. Improved .bat files (for Win32 port) to make things work under Win95.
500       (thanks to Aaron S. Hawley for this one)
501    
502    ----------------------------------------------------------------------
503    Name: Matthias Blume (blume (at) tti - c (dot) org)
504    Date: 2003/09/26 16:05:00 CDT
505    Tag: blume-20030926-wrappriv
506    Description:
507    
508    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
509    
510    ----------------------------------------------------------------------
511    Name: Matthias Blume (blume (at) tti - c (dot) org)
512    Date: 2003/09/26 15:00:00 CDT
513    Tag: blume-20030926-110_43_3
514    Description:
515    
516    - additional cleanup
517    - version number bump, NEW BOOTFILES
518    
519    ----------------------------------------------------------------------
520    Name: Matthias Blume (blume (at) tti - c (dot) org)
521    Date: 2003/09/26 12:00:00 CDT
522    Tag: blume-20030926-ppautoload
523    Description:
524    
525    I modified the read-eval-print loop so that the autoloader gets
526    invoked whenever the prettyprinter tries to look up a symbol that
527    is not currently defined in the toplevel environment but which
528    appears in CM's autoload registry.  As a result, we see far fewer of
529    those ?.Foo.Bar.xxx names in the prettyprinter's output.
530    
531    In addition to this I tried to clean up some pieces of the Basis
532    implementation (e.g., Socket, Word8Array) in order to prevent other
533    instances of these ?.Foo.Bar.xxx names from being printed.
534    
535    The mechanism that picks names for types still needs some work, though.
536    (Right now it seems that if there is a type A.t which is defined to
537    be B.u, but B is unavailable at toplevel, then A.t gets printed as
538    "?.B.u" although the perhaps more sensible solution would be to use
539    "A.t" in this case.  In other words, the prettyprinter should follow
540    a chain of DEFtycs not farther than there are corresponding toplevel
541    names in the current environment.)
542    
543    ----------------------------------------------------------------------
544    Name: Matthias Blume (blume (at) tti - c (dot) org)
545    Date: 2003/09/24 16:31:00 CDT
546    Tag: blume-20030924-installer
547    Description:
548    
549    Another installer tweak:  All the ML code for the installer is now
550    compiled during CMB.make and put into a little library called
551    $smlnj/installer.cm.  The installation then simply invokes
552    
553       sml -m $smlnj/installer.cm
554    
555    and everything happens automagically.
556    
557    Win32:  ML code senses value of environment variable SMLNJ_HOME.
558    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
559            and BINDIR.
560    
561    The new scheme guarantees that the ML code responsible for the installation
562    is in sync with the APIs of the main system.  Also, the installer is
563    somewhat faster because the installer script is precompiled.
564    
565    ----------------------------------------------------------------------
566    Name: Matthias Blume (blume (at) tti - c (dot) org)
567    Date: 2003/09/24 15:35:00 CDT
568    Tag: blume-20030924-synsock
569    Description:
570    
571    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
572    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
573    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
574    implementations of non-blocking operations from CML's Socket
575    structure.
576    
577    ----------------------------------------------------------------------
578    Name: Matthias Blume (blume (at) tti - c (dot) org)
579    Date: 2003/09/24 15:10:05 CDT
580    Tag: blume-20030924-sockets
581    Description:
582    
583    1. Fixed SOCKET API and implementation to match Basis spec.
584       This required changing the internal representation of sockets to one
585       that remembers (for each socket file descriptor) whether it is currently
586       blocking or non-blocking.  This state is maintained lazily (i.e., a system
587       call is made only if the state actually needs to change).
588    
589    2. OS-specific details of sockets were moved into separate files, thus
590       making it possible to unify the bulk of the socket implementations
591       between Unix and Win32.
592    
593    3. CML's socket API changed accordingly.
594       (Note that we need to remove non-blocking functions from this API
595       since they are redundant in the case of CML!)
596    
597    4. CML's socket implementation now makes use of non-blocking functions
598       provided by Basis, thus removing all OS-dependent code from this part
599       of CML.
600    
601    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
602    
603    ----------------------------------------------------------------------
604    Name: Matthias Blume (blume (at) tti - c (dot) org)
605    Date: 2003/09/22 12:10:00 CDT
606    Tag: blume-20030922-110_43_2
607    Description:
608    
609    Made a new interim version and bootfiles for developer's bootstrapping
610    convenience.
611    
612       110.43.2 -- NEW BOOTFILES
613    
614    ----------------------------------------------------------------------
615    Name: Matthias Blume (blume (at) tti - c (dot) org)
616    Date: 2003/09/19 15:55:00 CDT
617    Tag: blume-20030919-cmdir
618    Description:
619    
620    1. new-install.sh -> install.sh
621    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
622    3. tweaked installer so that another name instead of .cm can be chosen
623       at install time (by setting the CM_DIR_ARC environment variable
624       during installation); once installation is complete, the name is
625       fixed
626    
627    ----------------------------------------------------------------------
628    Name: Matthias Blume (blume (at) tti - c (dot) org)
629    Date: 2003/09/18 16:00:00 CDT
630    Tag: blume-20030918-110_43_1
631    Description:
632    
633    Made a new interim version and bootfiles for developer's bootstrapping
634    convenience.
635    
636       110.43.1 -- NEW BOOTFILES
637    
638    ----------------------------------------------------------------------
639    Name: Matthias Blume (blume (at) tti - c (dot) org)
640    Date: 2003/09/18 15:20:00 CDT
641    Tag: blume-20030918-misc
642    Description:
643    
644    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
645       this seems to be controversial at the moment) and used those in
646       Posix.ProcEnv.times.
647    
648    2. Added Time.{from,to}Nanoseconds to Time.
649    
650    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
651       For example, fromLargeInt never needs to look at more than 3 "big
652       digits" to get its 53 bits of precision.
653    
654    ----------------------------------------------------------------------
655    Name: Matthias Blume (blume (at) tti - c (dot) org)
656    Date: 2003/09/17 16:30:00 CDT
657    Tag: blume-20030917-real32-slices
658    Description:
659    
660    Added an entry to the primitive environment
661    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
662    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
663    
664    Removed some of the "magic" constants in real64.sml and replaced them
665    with code that generates these values from their corresponding
666    integer counterparts.
667    
668    Made all(?) the slice-related changes to the Basis and made everything
669    compile again...
670    
671    ----------------------------------------------------------------------
672    Name: Matthias Blume (blume (at) tti - c (dot) org)
673    Date: 2003/09/15 17:45:00 CDT
674    Tag: blume-20030915-rbase
675    Description:
676    
677    Fixed bug in Real.fromLargeInt.
678    
679    ----------------------------------------------------------------------
680    Name: Matthias Blume (blume (at) tti - c (dot) org)
681    Date: 2003/09/13 18:11:00 CDT
682    Tag: blume-20030913-libinstall
683    Description:
684    
685    Minor bugfix in config/libinstall (set anchor with path to
686    standalone tool after installing it, otherwise libraries that
687    need ml-lex or ml-yacc won't compile the first time the installer
688    runs).
689    
690    ----------------------------------------------------------------------
691    Name: Matthias Blume (blume (at) tti - c (dot) org)
692    Date: 2003/09/12 11:45:00 CDT
693    Tag: blume-20030912-various
694    Description:
695    
696     - fixed bug in Real.toLargeInt
697     - fixed bug in Posix.ProcEnv.times
698     - changed inputLine functions to return an option
699     - minor installer improvements / bugfixes
700     - changed default @SMLalloc parameter for x86/celeron to 64k
701    
702    ----------------------------------------------------------------------
703    Name: Matthias Blume (blume (at) tti - c (dot) org)
704    Date: 2003/09/09 22:00:00 CDT
705    Tag: Release_110_43
706    Description:
707    
708        New working release 110.43.  New bootfiles.
709    
710    ----------------------------------------------------------------------
711    Name: Matthias Blume (blume (at) tti - c (dot) org)
712    Date: 2003/09/09 19:20:00 CDT
713    Tag: blume-20030909-installer
714    Description:
715    
716    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
717    Modified config/install.bat to take advantage of it.  Also modified
718    config/install.sh (and called it config/new-install.sh) to take advantage
719    of it on Unix systems.  (The SML code is (supposed to be) platform-
720    independent.)
721    
722    The installer can now install everything under Win32
723    as well as under *nix as long as it compiles.
724    
725    Other changes:
726    
727      - made CML compile again under Win32
728      - made eXene compile under Win32 (by providing a fake structure UnixSock
729        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
730      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
731        same as type int;  under Win32 it isn't
732      - fixed some slice-related problems in the win32-specific parts of CML
733      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
734        (under Win32, the volume name is case-insensitive, and the
735        OS.Path code compares volume names for equality)
736    
737    ----------------------------------------------------------------------
738    Name: Matthias Blume (blume (at) tti - c (dot) org)
739    Date: 2003/09/08 11:55:00 CDT
740    Tag: blume-20030908-fullpath
741    Description:
742    
743    Made Win32 version of OS.FileSys.fullPath return current directory
744    when given an empty string.  This is what the spec says, and incidentally,
745    CM depends on it.  (CM otherwise goes into an infinite loop in certain
746    cases when presented with the name of a non-existing .cm file.)
747    
748    ----------------------------------------------------------------------
749    Name: Matthias Blume (blume (at) tti - c (dot) org)
750    Date: 2003/09/04 16:30:00 CDT
751    Tag: blume-20030905-slices-etc
752    Description:
753    
754      1. Changed interface to vectors and arrays in Basis to match
755         (draft) Basis spec.
756      2. Added signatures and implementations of slices according to
757         Basis spec.
758      3. Edited source code throughout the system to make it compile again
759         under 1. and 2.  (In some cases code had to be added to have it
760         match the new signatures.)
761      4. MLRISC should be backward-compatible: the copies of the originals
762         of files that needed to change under 3. were retained, the .cm files
763         check the compiler version number and use old versions when
764         appropriate.
765      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
766         dirstream -> string option (in accordance with Basis spec).
767      6. When generating code that counts lines, ml-lex used function
768         CharVector.foldli, taking advantage of its old interface.
769         This has been replaced with the corresponding code from
770         CharVectorSlice.  (html-lex must be re-lexed!)
771      7. BitArray in smlnj-lib/Util has been extended/modified to match the
772         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
773      8. Removed temporary additions (fromInternal, toInternal) from the
774         (now obsolete) IntInf in smlnj-lib/Util.
775      9. Cleaned up structure Byte.
776     10. Added localOffset, scan, and fromString to Date (according to spec).
777         Cleaned/corrected implementation of Date.
778         (Still need to check for correctness; implement better canonicalizeDate.)
779     11. Added "scan" to signature IEEE_REAL.
780     12. Some improvements to IntInf [in particular: efficiency-hack for
781         mod and rem when second operand is 2 (for parity checks).]
782     13. Changed representation of type Time.time, using a single IntInf.int
783         value counting microseconds.  This considerably simplified the
784         implementation of structure Time.  We now support negative time
785         values; scan and fromString handle signs.
786     14. Functor PrimIO now takes two additional arguments (VectorSlice and
787         ArraySlice).
788    
789    ----------------------------------------------------------------------
790    Name: Matthias Blume (blume (at) tti - c (dot) org)
791    Date: 2003/08/28 17:00:00 CDT
792    Tag: blume-20030828-intinf
793    Description:
794    
795    This is a major update which comes with a version number bump
796    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
797    and an implementation of IntInf in the Basis.
798    
799    There are a fairly large number of related changes and updates throughout
800    the system:
801    
802    Basis:
803            - Implemented IntInf.
804            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
805            - Added some missing Real64 operations, most notably Real.toLargeInt.
806            - Added FixedInt as a synonym for Int32.
807    
808    compiler:
809            * Added support for a built-in intinf type.
810              - literals
811              - pattern matching
812              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
813              - overloading on literals and operations
814    
815            This required adding a primitive type intinf, some additional
816            primops, and implementations for several non-trivial intinf
817            operations in Core.  (The intinf type is completely abstract
818            to the compiler; all operations get delegated back to the Core.)
819    
820            * Intinf equality is handled by polyequal.  However, the compiler
821            does not print its usual warning in this case (since polyequal
822            is the right thing to do there).
823    
824            * Improved the organization of structure InlineT.
825    
826            * A word about conversion primops:
827            If conversions involving intinf do not cancel out during
828            CPS contract, then the compiler must insert calls to Core functions.
829            Since all core access must be resolved already during the FLINT
830            translate phase, it would be too late a the time of CPS contract
831            to add new Core calls.  For this reason, conversion primops
832            for intinf carry two arguments:  1. the numeric argument that
833            they are supposed to convert, and 2. the Core function that
834            can help with this conversion if necessary.  If CPS contract
835            eliminates a primop, then the associated Core function becomes
836            dead and goes away.  Intinf conversion primops that do not get
837            eliminated by CPS contract get rewritten into calls of their
838            core functions by a separate, new phase.
839    
840    interactive system:
841            - Control.Print.intinfDepth controls max length of intinf constants
842            being printed. (Analogous to Control.Print.stringDepth.)
843            - Cleanup in printutil and pputil: got rid of unused stuff and
844            duplicates; replaced some of the code with code that makes better
845            use of library functionality.
846    
847    CM:
848            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
849            are no longer silent.
850    
851    CKIT:
852            Fixed mismatched uses of Int32 and LargeInt.  I always decided
853            in favor of LargeInt -- which is now the same as IntInf.
854            CKIT-knowledgable people should check whether this is what's
855            intended and otherwise change things back to using Int32 or
856            FixedInt.
857    
858    Throughout the code:
859            Started using IntInf.int literals and built-in operations
860            (e.g., comparison with 0) where this seems appropriate.
861    
862    
863    ----------------------------------------------------------------------
864    Name: Dave MacQueen (dbm@cs.uchicago.edu)
865    Date: 2003/08/13 11:36:00 CDT
866    Tag: dbm-20030813-mcz-merge1
867    Description:
868    
869    Merging changes from the mcz-branch development branch into trunk.
870    These changes involve replacement of the emulated old prettyprinter
871    interface with direct use of the SML/NJ Lib PP library, and fixing
872    of a couple of bugs (895, 1186) relating to error messages.  A new
873    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
874    has been added.
875    
876    ----------------------------------------------------------------------
877    Name: Matthias Blume (blume (at) tti - c (dot) org)
878    Date: 2003/08/11 15:45:00 CDT
879    Tag: blume-20030811-windows
880    Description:
881    
882    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
883    
884        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
885    
886    This patch restores SML/NJ's ability to run under win32.  There are a
887    number of changes, including fixes for several bugs that had gone
888    unnoticed until now:
889    
890      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
891        trying to run the win32 version from within cygwin.)
892    
893      - There are a number of simple .bat scripts that substitute for their
894        corresponding Unix shell-scripts.  (See below.)
895    
896      - The internals of ml-build have been modified slightly.  The main
897        difference is that instead of calling ".link-sml" (or link-sml.bat)
898        using OS.Process.system, the ML process delegates this task back
899        to the script.  Otherwise problems arise in mixed environments such
900        as Cygwin where scripts look and work like Unix scripts, but
901        where OS.Process.system cannot run them.
902    
903      - In CM, the srcpath pickler used native pathname syntax -- which
904        is incorrect in the case of cross-compilation.  The new pickle format
905        is independent of platform-specific naming conventions.
906    
907      - Path configuration files (such as lib/pathconfig) can now choose
908        between native and standard syntax.  Placing a line of the form
909    
910            standard!
911    
912        into the file causes all subsequent paths to be interpreted using
913        CM standard pathname syntax (= Unix conventions); a line
914    
915            native!
916    
917        switches back to native style.  This was needed so that
918        path config files can be written portably, see src/system/pathconfig.
919    
920      - Runtime system:
921    
922          - win32-filesys.c:  get_file_time and set_file_time now
923            access modification time, not creation time.
924    
925          - I/O code made aware of new array representation.
926    
927          - Bug fixes in X86.prim.masm.
928    
929       - src/system/makeml made aware of win32. (For use under cygwin
930            and other Unix-environments for windows.)
931    
932       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
933         which caused BinIO.inputAll to fail consistently.
934    
935    .bat scripts:
936    
937       Windows .bat scripts assume that SMLNJ_HOME is defined.
938    
939       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
940         applications (sml, ml-yacc, ml-lex).
941       - ml-build.bat: analogous to ml-build.
942       - config\install.bat:  Analogous to config/install.sh.  This requires
943         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
944         (nmake etc. must be on the path, and vcvars32 must have been run.)
945         Moreover, sources for ml-lex and ml-yacc need to exist under src,
946         and the bootfile hierarchy must have been unpacked under
947         sml.boot.x86-win32.
948         The script is very primitive and does a poor job at error checking.
949         It only installs the base system, ml-lex, and ml-yacc.  No other
950         libraries are being installed (i.e., you get only those that
951         are part of the compiler.)
952       - link-sml.bat:  analogous to .link-sml, but not currently used
953    
954    Unrelated bug fixes:
955    
956      - ml-nlffigen now exports structures ST_* corresponding to incomplete
957        types.
958      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
959        otherwise.)
960    
961    ----------------------------------------------------------------------
962    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
963    Date: 2003/06/17
964    Tag: macqueen-20030617-bug895
965    Description:
966    
967    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
968    Tag will be used for new development branch (mcz-branch) for
969    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
970    
971    ----------------------------------------------------------------------
972    Name: Matthias Blume (blume (at) tti - c (dot) org)
973    Date: 2003/05/27 16:55:00 CDT
974    Tag: blume-20030527-polyeq
975    Description:
976    
977    Tried to eliminated most cases of polymorphic equality.
978    
979    ----------------------------------------------------------------------
980    Name: Matthias Blume (blume (at) tti - c (dot) org)
981    Date: 2003/05/21 17:45:00 CDT
982    Tag: blume-20030517-complete
983    Description:
984    
985    Two changes:
986    
987      1. Added a flag for controlling whether non-exhaustive bindings will
988         be treated as errors (default is false).
989      2. Cleaned up the *entire* source tree so that CMB.make goes through
990         without a single non-exhaustive match- or bind warning.
991    
992    ----------------------------------------------------------------------
993    Name: Matthias Blume (blume (at) tti - c (dot) org)
994    Date: 2003/05/17 10:20:00 CDT
995    Tag: blume-20030517-absyn
996    Description:
997    
998    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
999    
1000       This mainly affects the quality of error messages.  However, some
1001       of the code is now more straightforward than before. (Treatment of
1002       the above four constructs in translate.sml is much simpler than
1003       the "macro-expansion" that was going on before.  Plus, the mach-
1004       compiler no longer gets invoked just to be able to compile an
1005       if-expression.)
1006    
1007    2. The ErrorMsg.Error exception is now caught and absorbed by the
1008       interactive loop.
1009    
1010    ----------------------------------------------------------------------
1011    Name: Allen Leung
1012    Date: 2003/05/16 13:05:00 CDT
1013    Tag: leunga-20030516-cygwin-runtime
1014    Description:
1015    
1016        Ported the runtime system to cygwin, which uses the unix
1017    x86-unix bin files.  Missing/buggy features:
1018    
1019       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1020         the Cygwin library.
1021       o Ctrl-C handling may be flaky.
1022       o Windows system calls and Windows I/O are not supported.
1023    
1024       A new set of binfiles is located at:
1025    
1026         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1027    
1028       This is only needed for bootstrapping the cygwin version of smlnj.
1029       Other x86 versions can use the existing binfiles.
1030    
1031    ----------------------------------------------------------------------
1032    Name: Matthias Blume
1033    Date: 2003/04/08 15:42:00 CDT
1034    Tag: blume-20030408-listpair
1035    Description:
1036    
1037    1. Added a target 'mlrisc' to installer.
1038    
1039    2. Added missing elements to structure ListPair.
1040    
1041    ----------------------------------------------------------------------
1042    Name: Allen Leung
1043    Date: 2003/01/07 10:40:00 EST
1044    Tag: leunga-20030107-int-rem
1045    Description:
1046    
1047    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1048    The arguments to the SUBL instruction were swapped.
1049    
1050    ----------------------------------------------------------------------
1051    Name: Matthias Blume
1052    Date: 2002/12/12 16:25:00 EST
1053    Tag: blume-20021212-risc-ra
1054    Description:
1055    
1056    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1057    sent the RA into an infinite loop when floating point registers get
1058    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1059    110.37 and 110.38.)
1060    There was another set of potential problems related to the handling of
1061    MLRISC annotations (but those did not yet cause real problems, apparently).
1062    
1063    ----------------------------------------------------------------------
1064    Name: Matthias Blume
1065    Date: 2002/12/06 22:40:00 EST
1066    Tag: blume-20021206-cm-fileid
1067    Description:
1068    
1069    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1070    This fixes the problem of CM getting confused by files that suddenly
1071    change their identity (e.g., by getting unlinked and recreated by some
1072    text editor such as vi).  There might be a better/cheaper/cleaner way
1073    of doing this, but for now this will have to do.
1074    
1075    ----------------------------------------------------------------------
1076    Name: Matthias Blume
1077    Date: 2002/10/28 09:50:00 EST
1078    Tag: blume-20021028-typecheck
1079    Description:
1080    
1081    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1082    
1083    ----------------------------------------------------------------------
1084    Name: Matthias Blume
1085    Date: 2002/10/17 09:10:00 EDT
1086    Tag: Release_110_42
1087    Description:
1088    
1089    In good old tradition, there has been a slight hiccup so that we have
1090    to patch 110.42 after the fact.   The old release tag has been replaced
1091    (see below).
1092    
1093    The change solves a problem with two competing approaches the
1094    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1095    each other's way.
1096    
1097    This change only affects the runtime system code and the installer script.
1098    (No new bootfiles.)
1099    
1100    ----------------------------------------------------------------------
1101    Name: Matthias Blume
1102    Date: 2002/10/16 12:00:00 EDT
1103    Tag: Release_110_42_removed
1104    Description:
1105    
1106    New working release.  New bootfiles.
1107    
1108    ----------------------------------------------------------------------
1109    Name: Matthias Blume
1110    Date: 2002/10/10 13:10:00 EDT
1111    Tag: blume-20021010-ppc-divs
1112    Description:
1113    
1114    The mltree operator DIVS must be implemented with an overflow check on
1115    the PPC because the hardware indicates divide-by-zero using "overflow" as
1116    well.
1117    
1118    ----------------------------------------------------------------------
1119    Name: Matthias Blume
1120    Date: 2002/07/23 11:45:00
1121    Tag: blume-20020723-smlnj-home
1122    Description:
1123    
1124    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1125    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1126    CM_PATHCONFIG is also set), the path configuration file is assumed
1127    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1128    the entire tree to some other place and everything will "just work".
1129    
1130    (Companion commands such as ml-build and ml-makedepend also sense this
1131    variable.)
1132    
1133    ----------------------------------------------------------------------
1134    Name: Matthias Blume
1135    Date: 2002/07/12 21:19:00 EDT
1136    Tag: blume-20020712-liveness
1137    Description:
1138    
1139    Exported two useful "step" functions from liveness module (MLRISC).
1140    
1141    ----------------------------------------------------------------------
1142    Name: Matthias Blume
1143    Date: 2002/07/05 16:00 EDT
1144    Tag: Release_110_41
1145    Description:
1146    
1147    New working release.  New bootfiles.
1148    
1149    ----------------------------------------------------------------------
1150    Name: Matthias Blume
1151    Date: 2002/07/05 10:25:00 EDT
1152    Tag: blume-20020705-btimp
1153    Description:
1154    
1155    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1156    other clients can set up backtracing support.
1157    
1158    ----------------------------------------------------------------------
1159    Name: Matthias Blume
1160    Date: 2002/06/25 17:23:00 EDT
1161    Tag: blume-20020625-fpmax
1162    Description:
1163    
1164    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1165    case.  (The sense of the isNaN test was reversed -- which made min and
1166    max always return their first argument.)
1167    
1168    ----------------------------------------------------------------------
1169    Name: Matthias Blume
1170    Date: 2002/06/11
1171    Tag: blume-20020611-unixpath
1172    Description:
1173    
1174    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1175    
1176    ----------------------------------------------------------------------
1177    Name: Matthias Blume
1178    Date: 2002/06/10 16:35:00 EDT
1179    Tag: blume-20020610-ieeereal
1180    Description:
1181    
1182    I back-ported my implementation of IEEEReal.fromString from the
1183    idlbasis-devel branch so that we can test it.
1184    
1185    Another small change is that ppDec tries to give more information
1186    than just "<sig>" in the case of functors.  However, this code is
1187    broken in some mysterious way if the functor's body's signature
1188    has not been declared by ascription but gets inferred from the
1189    implementation.  This needs fixing...
1190    
1191    ----------------------------------------------------------------------
1192    Name: Matthias Blume
1193    Date: 2002/05/31
1194    Tag: blume-20020531-btrace-mode
1195    Description:
1196    
1197    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1198    the wayside when I switched over to using Controls everywhere.)
1199    
1200    ----------------------------------------------------------------------
1201    Name: Lal George
1202    Date: 2002/05/23 12:21:40 EDT
1203    Tag: george-20020523-visual-labels
1204    Description:
1205    
1206    Labels are now displayed in the graphical output to make
1207    the fall-through and target blocks obvious.
1208    
1209    ----------------------------------------------------------------------
1210    Name: Matthias Blume
1211    Date: 2002/05/22 11:03:00 EDT
1212    Tag: blume-20020522-shrink
1213    Description:
1214    
1215    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1216    situation that comes up when allocating huge arrays.
1217    
1218    ----------------------------------------------------------------------
1219    Name: Matthias Blume
1220    Date: 2002/05/21 16:00:00 EDT
1221    Tag: Release_110_40
1222    Description:
1223    
1224    New working release (110.40).  New bootfiles.
1225    
1226    [Also: John Reppy fixed GC bug 1131.]
1227    
1228    ----------------------------------------------------------------------
1229    Name: Matthias Blume
1230    Date: 2002/05/21 12:35:00 EDT
1231    Tag: blume-20020521-cmdoc
1232    Description:
1233    
1234    CM documentation update.
1235    
1236    ----------------------------------------------------------------------
1237    Name: Matthias Blume
1238    Date: 2002/05/21 10:55:00 EDT
1239    Tag: blume-20020521-misc
1240    Description:
1241    
1242    - John tweaked runtime to be silent on heap export (except when
1243      GC messages are on).
1244    - I added a few more things (cross-compiling versions of CMB) to
1245      config/preloads (as suggestions).
1246    
1247    ----------------------------------------------------------------------
1248    Name: Matthias Blume
1249    Date: 2002/05/20 22:25:00 EDT
1250    Tag: blume-20020520-controls
1251    Description:
1252    
1253    - Added ControlUtil structure to control-lib.cm.
1254    - Use it throughout.
1255    - Used Controls facility to define MLRISC controls (as opposed to
1256      registering MLRISC control ref cells with Controls after the fact)
1257    - Fixed messed-up controls priorities.
1258    
1259    * Removed again all the stuff from config/preloads that one wouldn't
1260      be able to preload at the time the initial heap image is built.
1261      (Many libraries, e.g., CML, do not exist yet at this time.  The
1262       only libraries that can be preloaded via config/preloads are those
1263       that come bundled with the bootfiles.)
1264    
1265    ----------------------------------------------------------------------
1266    Name: Matthias Blume
1267    Date: 2002/05/20 10:59:00 EDT
1268    Tag: blume-20020520-preloads
1269    Description:
1270    
1271    Added a lot of commented-out suggestions for things to be included
1272    in config/preloads.
1273    
1274    ----------------------------------------------------------------------
1275    Name: Allen Leung
1276    Date: 2002/05/18 14:20:00 EDT
1277    Tag: leunga-20020518-mdl
1278    Description:
1279    
1280    o Made the mdl tool stuff compile and run again.
1281    o I've disabled all the stuff that depends on RTL specifications; they
1282      are all badly broken anyway.
1283    
1284    ----------------------------------------------------------------------
1285    Name: Matthias Blume
1286    Date: 2002/05/17 16:49:00 EDT
1287    Tag: blume-20020517-controls
1288    Description:
1289    
1290    0. John Reppy made several modifications to the SML/NJ library.
1291       In particular, there is a shiny new controls-lib.cm.
1292    
1293    1. Pushed new controls interface through compiler so that everything
1294       compiles again.
1295    
1296    2. Added FormatComb and FORMAT_COMB to the CML version of the
1297       SML/NJ library (so that CML compiles again).
1298    
1299    3. Modified init scripts because XXX_DEFAULT environment variables
1300       are no longer with us.  (Boot-time initialization is now done
1301       using the same environment variables that are also used for
1302       startup-time initialization of controls.)
1303    
1304    ----------------------------------------------------------------------
1305    Name: Lal George
1306    Date: 2002/05/15 09:20:10 EDT
1307    Tag: george-20020515-pseudo-op-decls
1308    Description:
1309    
1310            All pseudo-ops emitted before the first segment declaration
1311    such as TEXT, DATA, and BSS directives are assumed to be global
1312    declarations and are emitted first in the assembly file. This is
1313    useful in a number of situations where one has pseudo-ops that are not
1314    specific to any segment, and also works around the constraint that one
1315    cannot have client pseudo-ops in the TEXT segment.
1316    
1317            Because no segment is associated with these declarations it is
1318    an error to allocate any space or objects before the first segment
1319    directive and an exception will be raised. However, we cannot make
1320    this check for client pseudo-ops.
1321    
1322            These top level declarations are a field in the CFG graph_info.
1323    In theory you can continue to add to this field after the CFG has been
1324    built -- provided you know what you are doing;-)
1325    
1326    ----------------------------------------------------------------------
1327    Name: Matthias Blume
1328    Date: 2002/05/13 16:40:00 EDT
1329    Tag: blume-20020513-pp-etc
1330    Description:
1331    
1332    A few minor bugfixes:
1333    
1334      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1335        (Bogus printouts for redefined bindings still occur.  Compiler
1336        bug should no longer occur now.  We need to redo the prettyprinter
1337        from scratch.)
1338    
1339      - CM pathname printer now also adds escape sequences for ( and )
1340    
1341      - commend and docu fixes for ml-nlffi
1342    
1343    ----------------------------------------------------------------------
1344    Name: Matthias Blume
1345    Date: 2002/05/10 16:40:00 EDT
1346    Tag: blume-20020510-erg-textio
1347    Description:
1348    
1349    Applied the following bugfix provided by Emden Gansner:
1350    
1351        Output is corrupted when outputSubstr is used rather than output.
1352    
1353        The problem occurs when a substring
1354    
1355            ss = (s, dataStart, dataLen)
1356    
1357        where dataStart > 0, fills a stream buffer with avail bytes left.
1358        avail bytes of s, starting at index dataStart, are copied into the
1359        buffer, the buffer is flushed, and then the remaining dataLen-avail
1360        bytes of ss are copied into the beginning of the buffer. Instead of
1361        starting this copy at index dataStart+avail in s, the current code
1362        starts the copy at index avail.
1363    
1364        Fix:
1365        In text-io-fn.sml, change line 695 from
1366             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1367        to
1368             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1369    
1370    ----------------------------------------------------------------------
1371    Name: Matthias Blume
1372    Date: 2002/04/12 13:55:00 EDT
1373    Tag: blume-20020412-assyntax
1374    Description:
1375    
1376    1. Grabbed newer assyntax.h from the XFree86 project.
1377    2. Fiddled with how to compile X86.prim.asm without warnings.
1378    3. (Very) Minor cleanup in CM.
1379    
1380    ----------------------------------------------------------------------
1381    Name: Matthias Blume
1382    Date: 2002/04/01 (no joke!) 17:07:00 EST
1383    Tag: blume-20020401-x86div
1384    Description:
1385    
1386    Added full support for div/mod/rem/quot on the x86, using the machine
1387    instruction's two results (without clumsily recomputing the remainder)
1388    directly where appropriate.
1389    
1390    Some more extensive power-of-two support was added to the x86 instruction
1391    selector (avoiding expensive divs, mods, and muls where they can be
1392    replaced with cheaper shifts and masks).  However, this sort of thing
1393    ought to be done earlier, e.g., within the CPS optimizer so that
1394    all architectures benefit from it.
1395    
1396    The compiler compiles to a fixed point, but changes might be somewhat
1397    fragile nevertheless.  Please, report any strange things that you might
1398    see wrt. div/mod/quot/rem...
1399    
1400    ----------------------------------------------------------------------
1401    Name: Matthias Blume
1402    Date: 2002/03/29 17:22:00
1403    Tag: blume-20020329-div
1404    Description:
1405    
1406    Fixed my broken div/mod logic.  Unfortunately, this means that the
1407    inline code for div/mod now has one more comparison than before.
1408    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
1409    The problem was with quotient = 0, because that alone does not tell
1410    us which way the rounding went.  One then has to look at whether
1411    remainder and divisor have the same sign...  :(
1412    
1413    Anyway, I replaced the bootfiles with fresh ones...
1414    
1415    ----------------------------------------------------------------------
1416    Name: Matthias Blume
1417    Date: 2002/03/29 14:10:00 EST
1418    Tag: blume-20020329-inlprims
1419    Description:
1420    
1421    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
1422    
1423    Primops have changed. This means that the bin/boot-file formats have
1424    changed as well.
1425    
1426    To make sure that there is no confusion, I made a new version.
1427    
1428    
1429    CHANGES:
1430    
1431    * removed REMT from mltree (remainder should never overflow).
1432    
1433    * added primops to deal with divisions of all flavors to the frontend
1434    
1435    * handled these primops all the way through so they map to their respective
1436      MLRISC support
1437    
1438    * used these primops in the implementation of Int, Int32, Word, Word32
1439    
1440    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
1441    
1442    * parameterized INLMIN, INLMAX, and INLABS by a numkind
1443    
1444    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
1445      floating point
1446    
1447    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
1448      and Real (but Real.abs maps to a separate floating-point-only primop)
1449    
1450    
1451    TODO items:
1452    
1453    * Hacked Alpha32 instruction selection, disabling the selection of REMx
1454      instructions because the machine instruction encoder cannot handle
1455      them.  (Hppa, PPC, and Sparc instruction selection did not handle
1456      REM in the first place, and REM is supported by the x86 machine coder.)
1457    
1458    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
1459      selection phase.  (The two can be streamlined because the hardware
1460      delivers both quotient and remainder at the same time anyway.)
1461    
1462    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
1463      (Currently the behavior is inconsistent both across architectures and
1464      wrt. the draft Basis spec.)
1465    
1466    * Word8 should eventually be handled natively, too.
1467    
1468    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
1469      as if there currently is no execution path that could trigger it in
1470      SML/NJ.  (The assumptions underlying functions arith and promotable do not
1471      hold for things like multiplication and division.)
1472    
1473    ----------------------------------------------------------------------
1474    Name: Matthias Blume
1475    Date: 2002/03/27 16:27:00 EST
1476    Tag: blume-20020327-mlrisc-divisions
1477    Description:
1478    
1479    Added support for all four division operations (ML's div, mod, quot,
1480    and rem) to MLRISC.  In the course of doing so, I also rationalized
1481    the naming (no more annoying switch-around of DIV and QUOT), by
1482    parameterizing the operation by div_rounding_mode (which can be either
1483    DIV_TO_ZERO or DIV_TO_NEGINF).
1484    
1485    The generic MLTreeGen functor takes care of compiling all four
1486    operations down to only round-to-zero div.
1487    
1488    Missing pieces:
1489    
1490      * Doing something smarter than relying on MLTreeGen on architectures
1491        like, e.g., the x86 where hardware division delivers both quotient and
1492        remainder at the same time.  With this, the implementation of the
1493        round-to-neginf operations could be further streamlined.
1494    
1495      * Remove inlining support for div/mod/rem from the frontend and replace it
1496        with primops that get carried through to the backend.  Do this for all
1497        int and word types.
1498    
1499    ----------------------------------------------------------------------
1500    Name: Matthias Blume
1501    Date: 2002/03/25 17:25:00 EST
1502    Tag: blume-20020325-divmod
1503    Description:
1504    
1505    I improved (hopefully without breaking them) the implementation of Int.div,
1506    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
1507    advantage of the following observations:
1508    
1509      Let  q = x quot y      r = x rem y
1510           d = x div  y      m = x mod y
1511    
1512    where "quot" is the round-to-zero version of integer division that
1513    hardware usually provides.  Then we have:
1514    
1515         r = x - q * y        where neither the * nor the - will overflow
1516         d = if q >= 0 orelse x = q * y then q else q - 1
1517                              where neither the * nor the - will overflow
1518         m = if q >= 0 orelse r = 0 then r else r + y
1519                              where the + will not overflow
1520    
1521    This results in substantial simplification of the generated code.
1522    The following table shows the number of CFG nodes and edges generated
1523    for
1524            fun f (x, y) = x OPER y
1525            (* with OPER \in div, mod, quot, rem *)
1526    
1527    
1528        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
1529        --------------------------------------------------------
1530         div |         24 |         39 |         12 |         16
1531         mod |         41 |         71 |         12 |         16
1532        quot |          8 |         10 |          8 |         10
1533         rem |         10 |         14 |          8 |         10
1534    
1535    
1536    ----------------------------------------------------------------------
1537    Name: Matthias Blume
1538    Date: 2002/03/25 22:06:00 EST
1539    Tag: blume-20020325-cprotobug
1540    Description:
1541    
1542    Fixed a bug in cproto (c prototype decoder).
1543    
1544    ----------------------------------------------------------------------
1545    Name: Matthias Blume
1546    Date: 2002/03/25 16:00:00 EST
1547    Tag: blume-20020325-raw-primops
1548    Description:
1549    
1550    I did some cleanup to Allen's new primop code and
1551    replaced yesterday's bootfiles with new ones.
1552    (But they are stored in the same place.)
1553    
1554    ----------------------------------------------------------------------
1555    Name: Matthias Blume
1556    Date: 2002/03/24 22:40:00 EST
1557    Tag: blume-20020324-bootfiles
1558    Description:
1559    
1560    Made the bootfiles that Allen asked for.
1561    
1562    ----------------------------------------------------------------------
1563    Name: Allen Leung
1564    Date: 2002/03/23 15:50:00 EST
1565    Tag: leunga-20020323-flint-cps-rcc-primops
1566    Description:
1567    
1568      1. Changes to FLINT primops:
1569    
1570        (* make a call to a C-function;
1571         * The primop carries C function prototype information and specifies
1572         * which of its (ML-) arguments are floating point. C prototype
1573         * information is for use by the backend, ML information is for
1574         * use by the CPS converter. *)
1575      | RAW_CCALL of { c_proto: CTypes.c_proto,
1576                       ml_args: ccall_type list,
1577                       ml_res_opt: ccall_type option,
1578                       reentrant : bool
1579                     } option
1580       (* Allocate uninitialized storage on the heap.
1581        * The record is meant to hold short-lived C objects, i.e., they
1582        * are not ML pointers.  With the tag, the representation is
1583        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1584        *)
1585      | RAW_RECORD of {tag:bool,sz:int}
1586      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1587    
1588      2.  These CPS primops are now overloaded:
1589    
1590           rawload of {kind:numkind}
1591           rawstore of {kind:numkind}
1592    
1593          The one argument form is:
1594    
1595             rawload {kind} address
1596    
1597          The two argument form is:
1598    
1599             rawload {kind} [ml object, byte-offset]
1600    
1601      3. RAW_CCALL/RCC now takes two extra arguments:
1602    
1603         a. The first is whether the C call is reentrant, i.e., whether
1604            ML state should be saved and restored.
1605         b. The second argument is a string argument specifying the name of
1606            library and the C function.
1607    
1608         These things are currently not handled in the code generator, yet.
1609    
1610      4. In CProto,
1611    
1612         An encoding type of "bool" means "ml object" and is mapped into
1613         C prototype of PTR.  Note that "bool" is different than "string",
1614         even though "string" is also mapped into PTR, because "bool"
1615         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
1616    
1617      5. Pickler/unpicker
1618    
1619         Changed to handle RAW_RECORD and newest RAW_CCALL
1620    
1621      6. MLRiscGen,
1622    
1623         1. Changed to handle the new rawload/rawstore/rawrecord operators.
1624         2. Code for handling C Calls has been moved to a new module CPSCCalls,
1625            in the file CodeGen/cpscompile/cps-c-calls.sml
1626    
1627      7. Added the conditional move operator
1628    
1629             condmove of branch
1630    
1631         to cps.  Generation of this is still buggy so it is currently
1632         disabled.
1633    
1634    ----------------------------------------------------------------------
1635    Name: Lal George
1636    Date: 2002/03/22 14:18:25 EST
1637    Tag: george-20020322-cps-branch-prob
1638    Description:
1639    
1640    Implemented the Ball-Larus branch prediction-heuristics, and
1641    incorporated graphical viewers for control flow graphs.
1642    
1643    Ball-Larus Heuristics:
1644    ---------------------
1645    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
1646    
1647    By design it uses the Dempster-Shafer theory for combining
1648    probabilities.  For example, in the function:
1649    
1650        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
1651    
1652    the ball-larus heuristics predicts that the n=0 is unlikely
1653    (OH-heuristic), and the 'then' branch is unlikely because of the
1654    RH-heuristic -- giving the 'then' branch an even lower combined
1655    probability using the Dempster-Shafer theory.
1656    
1657    Finally, John Reppy's loop analysis in MLRISC, further lowers the
1658    probability of the 'then' branch because of the loop in the else
1659    branch.
1660    
1661    
1662    Graphical Viewing:
1663    ------------------
1664    I merely plugged in Allen's graphical viewers into the compiler. The
1665    additional code is not much. At the top level, saying:
1666    
1667            Control.MLRISC.getFlag "cfg-graphical-view" := true;
1668    
1669    will display the graphical view of the control flow graph just before
1670    back-patching.  daVinci must be in your path for this to work. If
1671    daVinci is not available, then the default viewer can be changed
1672    using:
1673    
1674            Control.MLRISC.getString "viewer"
1675    
1676    which can be set to "dot" or "vcg" for the corresponding viewers. Of
1677    course, these viewers must be in your path.
1678    
1679    The above will display the compilation unit at the level of clusters,
1680    many of which are small, boring, and un-interesting. Also setting:
1681    
1682            Control.MLRISC.getInt "cfg-graphical-view_size"
1683    
1684    will display clusters that are larger than the value set by the above.
1685    
1686    
1687    ----------------------------------------------------------------------
1688    Name: Matthias Blume
1689    Date: 2002/03/21 22:20:00 EST
1690    Tag: blume-20020321-kmp-bugfix
1691    Description:
1692    
1693    Changed the interface to the KMP routine in PreString and fixed
1694    a minor bug in one place where it was used.
1695    
1696    ----------------------------------------------------------------------
1697    Name: Allen Leung
1698    Date: 2002/03/21 20:30:00 EST
1699    Tag: leunga-20020321-cfg
1700    Description:
1701    
1702      Fixed a potential problem in cfg edge splitting.
1703    
1704    ----------------------------------------------------------------------
1705    Name: Allen Leung
1706    Date: 2002/03/21 17:15:00 EST
1707    Tag: leunga-20020321-x86-fp-cfg
1708    Description:
1709    
1710      1. Recoded the buggy parts of x86-fp.
1711    
1712         a. All the block reordering code has been removed.
1713            We now depend on the block placement phases to do this work.
1714    
1715         b. Critical edge splitting code has been simplified and moved into the
1716            CFG modules, as where they belong.
1717    
1718         Both of these were quite buggy and complex.  The code is now much, much
1719         simpler.
1720    
1721      2. X86 backend.
1722    
1723         a. Added instructions for 64-bit support.  Instruction selection for
1724            64-bit has not been committed, however, since that
1725            requires changes to MLTREE which haven't been approved by
1726            Lal and John.
1727    
1728         b. Added support for FUCOMI and FUCOMIP when generating code for
1729            PentiumPro and above.  We only generate these instructions in
1730            the fast-fp mode.
1731    
1732         c. Added cases for JP and JNP in X86FreqProps.
1733    
1734      3. CFG
1735    
1736         CFG now has a bunch of methods for edge splitting and merging.
1737    
1738      4. Machine description.
1739    
1740         John's simplification of MLTREE_BASIS.fcond broke a few machine
1741         description things:
1742    
1743         rtl-build.{sig,sml} and hppa.mdl fixed.
1744    
1745         NOTE: the machine description stuff in the repository is still broken.
1746               Again, I can't put my fixes in because that involves
1747               changes to MLTREE.
1748    
1749    ----------------------------------------------------------------------
1750    Name: Matthias Blume
1751    Date: 2002/03/20 15:55:00 EST
1752    Tag: blume-20020320-kmp
1753    Description:
1754    
1755    Implemented Knuth-Morris-Pratt string matching in PreString and used
1756    it for String.isSubstring, Substring.isSubstring, and
1757    Substring.position.
1758    
1759    (Might need some stress-testing.  Simple examples worked fine.)
1760    
1761    ----------------------------------------------------------------------
1762    Name: Matthias Blume
1763    Date: 2002/03/19 16:37:00 EST
1764    Tag: blume-20020319-witnesses
1765    Description:
1766    
1767    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
1768    
1769    This implements a generic mechanism for changing constness qualifiers
1770    anywhere within big C types without resorting to outright "casts".
1771    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
1772    modify the constness at the outermost level.)
1773    The implementation of "convert" is based on the idea of "witness"
1774    values -- values that are not used by the operation but whose types
1775    "testify" to their applicability.  On the implementation side, "convert"
1776    is simply a projection (returning its second curried argument).  With
1777    cross-module inlining, it should not result in any machine code being
1778    generated.
1779    
1780    ----------------------------------------------------------------------
1781    Name: Matthias Blume
1782    Date: 2002/03/15 16:40:00 EST
1783    Tag: blume-20020315-basis
1784    Description:
1785    
1786    Provided (preliminary?) implementations for
1787    
1788      {String,Substring}.{concatWith,isSuffix,isSubstring}
1789    
1790    and
1791    
1792      Substring.full
1793    
1794    Those are in the Basis spec but they were missing in SML/NJ.
1795    
1796    ----------------------------------------------------------------------
1797    Name: Matthias Blume
1798    Date: 2002/03/14 21:30:00 EST
1799    Tag: blume-20020314-controls
1800    Description:
1801    
1802    Controls:
1803    ---------
1804    
1805    1. Factored out the recently-added Controls : CONTROLS stuff and put
1806       it into its own library $/controls-lib.cm.  The source tree for
1807       this is under src/smlnj-lib/Controls.
1808    
1809    2. Changed the names of types and functions in this interface, so they
1810       make a bit more "sense":
1811    
1812          module -> registry
1813          'a registry -> 'a group
1814    
1815    3. The interface now deals in ref cells only.  The getter/setter interface
1816       is (mostly) gone.
1817    
1818    4. Added a function that lets one register an already-existing ref cell.
1819    
1820    5. Made the corresponding modifications to the rest of the code so that
1821       everything compiles again.
1822    
1823    6. Changed the implementation of Controls.MLRISC back to something closer
1824       to the original.  In particular, this module (and therefore MLRISC)
1825       does not depend on Controls.  There now is some link-time code in
1826       int-sys.sml that registers the MLRISC controls with the Controls
1827       module.
1828    
1829    CM:
1830    ---
1831    
1832      * One can now specify the lambda-split aggressiveness in init.cmi.
1833    
1834    ----------------------------------------------------------------------
1835    Name: Allen Leung
1836    Date: 2002/03/13 17:30:00 EST
1837    Tag: leunga-20020313-x86-fp-unary
1838    Description:
1839    
1840    Bug fix for:
1841    
1842    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
1843    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
1844    > - fun f(x,(y,z)) = Real.~ y;
1845    > [autoloading]
1846    > [autoloading done]
1847    >       fchsl   (%eax), 184(%esp)
1848    > Error: MLRisc bug: X86MCEmitter.emitInstr
1849    >
1850    > uncaught exception Error
1851    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
1852    
1853    The problem was that the code generator did not generate any fp registers
1854    in this case, and the ra didn't know that it needed to run the X86FP phase to
1855    translate the pseudo fp instruction.   This only happened with unary fp
1856    operators in certain situations.
1857    
1858    ----------------------------------------------------------------------
1859    Name: Matthias Blume
1860    Date: 2002/03/13 14:00:00 EST
1861    Tag: blume-20020313-overload-etc
1862    Description:
1863    
1864    1. Added _overload as a synonym for overload for backward compatibility.
1865       (Control.overloadKW must be true for either version to be accepted.)
1866    
1867    2. Fixed bug in install script that caused more things to be installed
1868       than what was requested in config/targets.
1869    
1870    3. Made CM aware of the (_)overload construct so that autoloading
1871       works.
1872    
1873    ----------------------------------------------------------------------
1874    Name: Matthias Blume
1875    Date: 2002/03/12 22:03:00 EST
1876    Tag: blume-20020312-url
1877    Description:
1878    
1879    Forgot to update BOOT and srcarchiveurl.
1880    
1881    ----------------------------------------------------------------------
1882    Name: Matthias Blume
1883    Date: 2002/03/12 17:30:00 EST
1884    Tag: blume-20020312-version110392
1885    Description:
1886    
1887    Yet another version number bump (because of small changes to the
1888    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
1889    
1890    Changes:
1891    
1892      The new pid generation scheme described a few weeks ago was overly
1893      complicated.  I implemented a new mechanism that is simpler and
1894      provides a bit more "stability":  Once CM has seen a compilation
1895      unit, it keeps its identity constant (as long as you do not delete
1896      those crucial CM/GUID/* files).  This means that when you change
1897      an interface, compile, then go back to the old interface, and
1898      compile again, you arrive at the original pid.
1899    
1900      There now also is a mechanism that instructs CM to use the plain
1901      environment hash as a module's pid (effectively making its GUID
1902      the empty string).  For this, "noguid" must be specified as an
1903      option to the .sml file in question within its .cm file.
1904      This is most useful for code that is being generated by tools such
1905      as ml-nlffigen (because during development programmers tend to
1906      erase the tool's entire output directory tree including CM's cached
1907      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
1908      revert to the old, broken behavior of SML/NJ, but in specific cases
1909      where there is no danger of interface confusion, its use is ok
1910      (I think).
1911    
1912      ml-nlffigen by default generates "noguid" annotations.  They can be
1913      turned off by specifying -guid in its command line.
1914    
1915    ----------------------------------------------------------------------
1916    Name: Lal George
1917    Date: 2002/03/12 12 14:42:36 EST
1918    Tag: george-20020312-frequency-computation
1919    Description:
1920    
1921    Integrated jump chaining and static block frequency into the
1922    compiler. More details and numbers later.
1923    
1924    ----------------------------------------------------------------------
1925    Name: Lal George
1926    Date: 2002/03/11 11 22:38:53 EST
1927    Tag: george-20020311-jump-chain-elim
1928    Description:
1929    
1930    Tested the jump chain elimination on all architectures (except the
1931    hppa).  This is on by default right now and is profitable for the
1932    alpha and x86, however, it may not be profitable for the sparc and ppc
1933    when compiling the compiler.
1934    
1935    The gc test will typically jump to a label at the end of the cluster,
1936    where there is another jump to an external cluster containing the actual
1937    code to invoke gc. This is to allow factoring of common gc invocation
1938    sequences. That is to say, we generate:
1939    
1940            f:
1941               testgc
1942               ja   L1      % jump if above to L1
1943    
1944            L1:
1945               jmp L2
1946    
1947    
1948    After jump chain elimination the 'ja L1' instructions is converted to
1949    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
1950    up being implemented in their long form (if L2 is far away) using:
1951    
1952            jbe     L3      % jump if below or equal to L3
1953            jmp     L2
1954         L3:
1955            ...
1956    
1957    
1958    For large compilation units L2  may be far away.
1959    
1960    
1961    ----------------------------------------------------------------------
1962    Name: Matthias Blume
1963    Date: 2002/03/11 13:30:00 EST
1964    Tag: blume-20020311-mltreeeval
1965    Description:
1966    
1967    A functor parameter was missing.
1968    
1969    ----------------------------------------------------------------------
1970    Name: Allen Leung
1971    Date: 2002/03/11 10:30:00 EST
1972    Tag: leunga-20020311-runtime-string0
1973    Description:
1974    
1975       The representation of the empty string now points to a
1976    legal null terminated C string instead of unit.  It is now possible
1977    to convert an ML string into C string with InlineT.CharVector.getData.
1978    This compiles into one single machine instruction.
1979    
1980    ----------------------------------------------------------------------
1981    Name: Allen Leung
1982    Date: 2002/03/10 23:55:00 EST
1983    Tag: leunga-20020310-x86-call
1984    Description:
1985    
1986       Added machine generation for CALL instruction (relative displacement mode)
1987    
1988    ----------------------------------------------------------------------
1989    Name: Matthias Blume
1990    Date: 2002/03/08 16:05:00
1991    Tag: blume-20020308-entrypoints
1992    Description:
1993    
1994    Version number bumped to 110.39.1.  NEW BOOTFILES!
1995    
1996    Entrypoints: non-zero offset into a code object where execution should begin.
1997    
1998    - Added the notion of an entrypoint to CodeObj.
1999    - Added reading/writing of entrypoint info to Binfile.
2000    - Made runtime system bootloader aware of entrypoints.
2001    - Use the address of the label of the first function given to mlriscGen
2002      as the entrypoint.  This address is currently always 0, but it will
2003      not be 0 once we turn on block placement.
2004    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2005      with entry points) from mlriscGen.
2006    
2007    ----------------------------------------------------------------------
2008    Name: Allen Leung
2009    Date: 2002/03/07 20:45:00 EST
2010    Tag: leunga-20020307-x86-cmov
2011    Description:
2012    
2013       Bug fixes for CMOVcc on x86.
2014    
2015       1. Added machine code generation for CMOVcc
2016       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2017       3. CMOVcc cannot have an immediate operand as argument.
2018    
2019    ----------------------------------------------------------------------
2020    Name: Matthias Blume
2021    Date: 2002/03/07 16:15:00 EST
2022    Tag: blume-20020307-controls
2023    Description:
2024    
2025    This is a very large but mostly boring patch which makes (almost)
2026    every tuneable compiler knob (i.e., pretty much everything under
2027    Control.* plus a few other things) configurable via both the command
2028    line and environment variables in the style CM did its configuration
2029    until now.
2030    
2031    Try starting sml with '-h' (or, if you are brave, '-H')
2032    
2033    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2034    implements the underlying generic mechanism.
2035    
2036    The interface to some of the existing such facilities has changed somewhat.
2037    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2038    (The getFoo interface is still there for backward-compatibility, but its
2039    use is deprecated.)
2040    
2041    The ml-build script passes -Cxxx=yyy command-line arguments through so
2042    that one can now twiddle the compiler settings when using this "batch"
2043    compiler.
2044    
2045    TODO items:
2046    
2047    We should go through and throw out all controls that are no longer
2048    connected to anything.  Moreover, we should go through and provide
2049    meaningful (and correct!) documentation strings for those controls
2050    that still are connected.
2051    
2052    Currently, multiple calls to Controls.new are accepted (only the first
2053    has any effect).  Eventually we should make sure that every control
2054    is being made (via Controls.new) exactly once.  Future access can then
2055    be done using Controls.acc.
2056    
2057    Finally, it would probably be a good idea to use the getter-setter
2058    interface to controls rather than ref cells.  For the time being, both
2059    styles are provided by the Controls module, but getter-setter pairs are
2060    better if thread-safety is of any concern because they can be wrapped.
2061    
2062    *****************************************
2063    
2064    One bug fix: The function blockPlacement in three of the MLRISC
2065    backpatch files used to be hard-wired to one of two possibilities at
2066    link time (according to the value of the placementFlag).  But (I
2067    think) it should rather sense the flag every time.
2068    
2069    *****************************************
2070    
2071    Other assorted changes (by other people who did not supply a HISTORY entry):
2072    
2073    1. the cross-module inliner now works much better (Monnier)
2074    2. representation of weights, frequencies, and probabilities in MLRISC
2075       changed in preparation of using those for weighted block placement
2076       (Reppy, George)
2077    
2078    ----------------------------------------------------------------------
2079    Name: Lal George
2080    Date: 2002/03/07 14:44:24 EST 2002
2081    Tag: george-20020307-weighted-block-placement
2082    
2083    Tested the weighted block placement optimization on all architectures
2084    (except the hppa) using AMPL to generate the block and edge frequencies.
2085    Changes were required in the machine properties to correctly
2086    categorize trap instructions. There is an MLRISC flag
2087    "weighted-block-placement" that can be used to enable weighted block
2088    placement, but this will be ineffective without block/edge
2089    frequencies (coming soon).
2090    
2091    
2092    ----------------------------------------------------------------------
2093    Name: Lal George
2094    Date: 2002/03/05 17:24:48 EST
2095    Tag: george-20020305-linkage-cluster
2096    
2097    In order to support the block placement optimization, a new cluster
2098    is generated as the very first cluster (called the linkage cluster).
2099    It contains a single jump to the 'real' entry point for the compilation
2100    unit. Block placement has no effect on the linkage cluster itself, but
2101    all the other clusters  have full freedom in the manner in which they
2102    reorder blocks or functions.
2103    
2104    On the x86 the typical linkage code that is generated is:
2105       ----------------------
2106            .align 2
2107       L0:
2108            addl    $L1-L0, 72(%esp)
2109            jmp     L1
2110    
2111    
2112            .align  2
2113       L1:
2114       ----------------------
2115    
2116    72(%esp) is the memory location for the stdlink register. This
2117    must contain the address of the CPS function being called. In the
2118    above example, it contains the address of  L0; before
2119    calling L1 (the real entry point for the compilation unit), it
2120    must contain the address for L1, and hence
2121    
2122            addl $L1-L0, 72(%esp)
2123    
2124    I have tested this on all architectures except the hppa.The increase
2125    in code size is of course negligible
2126    
2127    ----------------------------------------------------------------------
2128    Name: Allen Leung
2129    Date: 2002/03/03 13:20:00 EST
2130    Tag: leunga-20020303-mlrisc-tools
2131    
2132      Added #[ ... ] expressions to mlrisc tools
2133    
2134    ----------------------------------------------------------------------
2135    Name: Matthias Blume
2136    Date: 2002/02/27 12:29:00 EST
2137    Tag: blume-20020227-cdebug
2138    Description:
2139    
2140    - made types in structure C and C_Debug to be equal
2141    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2142    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2143    
2144    ----------------------------------------------------------------------
2145    Name: Matthias Blume
2146    Date: 2002/02/26 12:00:00 EST
2147    Tag: blume-20020226-ffi
2148    Description:
2149    
2150    1. Fixed a minor bug in CM's "noweb" tool:
2151       If numbering is turned off, then truly don't number (i.e., do not
2152       supply the -L option to noweb).  The previous behavior was to supply
2153       -L'' -- which caused noweb to use the "default" line numbering scheme.
2154       Thanks to Chris Richards for pointing this out (and supplying the fix).
2155    
2156    2. Once again, I reworked some aspects of the FFI:
2157    
2158       A. The incomplete/complete type business:
2159    
2160       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2161         gone!
2162       - ML types representing an incomplete type are now *equal* to
2163         ML types representing their corresponding complete types (just like
2164         in C).  This is still safe because ml-nlffigen will not generate
2165         RTTI for incomplete types, nor will it generate functions that
2166         require access to such RTTI.   But when ML code generated from both
2167         incomplete and complete versions of the C type meet, the ML types
2168         are trivially interoperable.
2169    
2170         NOTE:  These changes restore the full generality of the translation
2171         (which was previously lost when I eliminated functorization)!
2172    
2173       B. Enum types:
2174    
2175       - Structure C now has a type constructor "enum" that is similar to
2176         how the "su" constructor works.  However, "enum" is not a phantom
2177         type because each "T enum" has values (and is isomorphic to
2178         MLRep.Signed.int).
2179       - There are generic access operations for enum objects (using
2180         MLRep.Signed.int).
2181       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2182         * The structure contains the definition of type "mlrep" (the ML-side
2183         representation type of the enum).  Normally, mlrep is the same
2184         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2185         then mlrep will be defined as a datatype -- thus facilitating
2186         pattern matching on mlrep values.
2187         ("-ec" will be suppressed if there are duplicate values in an
2188          enumeration.)
2189         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2190         will be generated for each C enum constant xxx.
2191         * Conversion functions m2i and i2m convert between mlrep and
2192         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2193         * Coversion functions c and ml convert between mlrep and "tag enum".
2194         * Access functions (get/set) fetch and store mlrep values.
2195       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2196         enumerations are merged into one single enumeration represented by
2197         structure E_'.
2198    
2199    ----------------------------------------------------------------------
2200    Name: Allen Leung
2201    Date: 2002/02/25 04:45:00 EST
2202    Tag: leunga-20020225-cps-spill
2203    
2204    This is a new implementation of the CPS spill phase.
2205    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2206    In case of problems, replace it with the old file spill.sml
2207    
2208    The current compiler runs into some serious performance problems when
2209    constructing a large record.  This can happen when we try to compile a
2210    structure with many items.  Even a very simple structure like the following
2211    makes the compiler slow down.
2212    
2213        structure Foo = struct
2214           val x_1 = 0w1 : Word32.int
2215           val x_2 = 0w2 : Word32.int
2216           val x_3 = 0w3 : Word32.int
2217           ...
2218           val x_N = 0wN : Word32.int
2219        end
2220    
2221    The following table shows the compile time, from N=1000 to N=4000,
2222    with the old compiler:
2223    
2224    N
2225    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2226           MLRISC ra                               0.06u  0.00s  0.05g
2227              (spills = 0 reloads = 0)
2228           TOTAL                                   0.63u  0.07s  0.21g
2229    
2230    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2231           MLRISC ra                               5.68u  0.59s  3.93g
2232              (spills = 0 reloads = 0)
2233           TOTAL                                   14.71u  0.99s  4.81g
2234    
2235    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2236           MLRISC ra                               5.54u  0.65s  3.91g
2237              (spills = 543 reloads = 1082)
2238           TOTAL                                   65.40u  3.13s  6.00g
2239    
2240    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2241           MLRISC ra                               0.80u  0.10s  0.55g
2242              (spills = 42 reloads = 84)
2243           TOTAL                                   129.42u  5.10s  4.13g
2244    
2245    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2246           MLRISC ra                               2.69u  0.27s  1.38g
2247              (spills = 62 reloads = 124)
2248           TOTAL                                   682.48u  19.61s  13.99g
2249    
2250    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2251           MLRISC ra                               4.96u  0.27s  2.72g
2252              (spills = 85 reloads = 170)
2253           TOTAL                                   2375.26u  57.21s  48.00g
2254    
2255    As you can see the old cps spill module suffers from some serious
2256    performance problem.  But since I cannot decipher the old code fully,
2257    instead of patching the problems up, I'm reimplementing it
2258    with a different algorithm.  The new code is more modular,
2259    smaller when compiled, and substantially faster
2260    (O(n log n) time and O(n) space).  Timing of the new spill module:
2261    
2262    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2263          MLRISC ra                               0.25u  0.02s  0.15g
2264             (spills=1 reloads=3)
2265          TOTAL                                   7.74u  0.34s  1.62g
2266    
2267    Implementation details:
2268    
2269    As far as I can tell, the purpose of the CPS spill module is to make sure the
2270    number of live variables at any program point (the bandwidth)
2271    does not exceed a certain limit, which is determined by the
2272    size of the spill area.
2273    
2274    When the bandwidth is too large, we decrease the register pressure by
2275    packing live variables into spill records.  How we achieve this is
2276    completely different than what we did in the old code.
2277    
2278    First, there is something about the MLRiscGen code generator
2279    that we should be aware of:
2280    
2281    o MLRiscGen performs code motion!
2282    
2283       In particular, it will move floating point computations and
2284       address computations involving only the heap pointer to
2285       their use sites (if there is only a single use).
2286       What this means is that if we have a CPS record construction
2287       statement
2288    
2289           RECORD(k,vl,w,e)
2290    
2291       we should never count the new record address w as live if w
2292       has only one use (which is often the case).
2293    
2294       We should do something similar to floating point, but the transformation
2295       there is much more complex, so I won't deal with that.
2296    
2297    Secondly, there are now two new cps primops at our disposal:
2298    
2299     1. rawrecord of record_kind option
2300        This pure operator allocates some uninitialized storage from the heap.
2301        There are two forms:
2302    
2303         rawrecord NONE [INT n]  allocates a tagless record of length n
2304         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2305                                     and initializes the tag.
2306    
2307     2. rawupdate of cty
2308          rawupdate cty (v,i,x)
2309          Assigns to x to the ith component of record v.
2310          The storelist is not updated.
2311    
2312    We use these new primops for both spilling and increment record construction.
2313    
2314     1. Spilling.
2315    
2316        This is implemented with a linear scan algorithm (but generalized
2317        to trees).  The algorithm will create a single spill record at the
2318        beginning of the cps function and use rawupdate to spill to it,
2319        and SELECT or SELp to reload from it.  So both spills and reloads
2320        are fine-grain operations.  In contrast, in the old algorithm
2321        "spills" have to be bundled together in records.
2322    
2323        Ideally, we should sink the spill record construction to where
2324        it is needed.  We can even split the spill record into multiple ones
2325        at the places where they are needed.  But CPS is not a good
2326        representation for global code motion, so I'll keep it simple and
2327        am not attempting this.
2328    
2329     2. Incremental record construction (aka record splitting).
2330    
2331        Long records with many component values which are simulatenously live
2332        (recall that single use record addresses are not considered to
2333         be live) are constructed with rawrecord and rawupdate.
2334        We allocate space on the heap with rawrecord first, then gradually
2335        fill it in with rawupdate.  This is the technique suggested to me
2336        by Matthias.
2337    
2338        Some restrictions on when this is applicable:
2339        1. It is not a VECTOR record.  The code generator currently does not handle
2340           this case. VECTOR record uses double indirection like arrays.
2341        2. All the record component values are defined in the same "basic block"
2342           as the record constructor.  This is to prevent speculative
2343           record construction.
2344    
2345    ----------------------------------------------------------------------
2346    Name: Allen Leung
2347    Date: 2002/02/22 01:02:00 EST
2348    Tag: leunga-20020222-mlrisc-tools
2349    
2350    Minor bug fixes in the parser and rewriter
2351    
2352    ----------------------------------------------------------------------
2353    Name: Allen Leung
2354    Date: 2002/02/21 20:20:00 EST
2355    Tag: leunga-20020221-peephole
2356    
2357    Regenerated the peephole files.  Some contained typos in the specification
2358    and some didn't compile because of pretty printing bugs in the old version
2359    of 'nowhere'.
2360    
2361    ----------------------------------------------------------------------
2362    Name: Allen Leung
2363    Date: 2002/02/19 20:20:00 EST
2364    Tag: leunga-20020219-mlrisc-tools
2365    Description:
2366    
2367       Minor bug fixes to the mlrisc-tools library:
2368    
2369       1.  Fixed up parsing colon suffixed keywords
2370       2.  Added the ability to shut the error messages up
2371       3.  Reimplemented the pretty printer and fixed up/improved
2372           the pretty printing of handle and -> types.
2373       4.  Fixed up generation of literal symbols in the nowhere tool.
2374       5.  Added some SML keywords to to sml.sty
2375    
2376    ----------------------------------------------------------------------
2377    Name: Matthias Blume
2378    Date: 2002/02/19 16:20:00 EST
2379    Tag: blume-20020219-cmffi
2380    Description:
2381    
2382    A wild mix of changes, some minor, some major:
2383    
2384    * All C FFI-related libraries are now anchored under $c:
2385        $/c.cm      --> $c/c.cm
2386        $/c-int.cm  --> $c/internals/c-int.cm
2387        $/memory.cm --> $c/memory/memory.cm
2388    
2389    * "make" tool (in CM) now treats its argument pathname slightly
2390      differently:
2391        1. If the native expansion is an absolute name, then before invoking
2392           the "make" command on it, CM will apply OS.Path.mkRelative
2393           (with relativeTo = OS.FileSys.getDir()) to it.
2394        2. The argument will be passed through to subsequent phases of CM
2395           processing without "going native".  In particular, if the argument
2396           was an anchored path, then "make" will not lose track of that anchor.
2397    
2398    * Compiler backends now "know" their respective C calling conventions
2399      instead of having to be told about it by ml-nlffigen.  This relieves
2400      ml-nlffigen from one of its burdens.
2401    
2402    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
2403    
2404    * Export C_DEBUG and C_Debug from $c/c.cm.
2405    
2406    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
2407      subtyping relationship between incomplete pointers and their complete
2408      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
2409      with the convention of instantiating 'o with ('t, 'c) obj whenever
2410      the pointer target type is complete.  In the incomplete case, 'o
2411      will be instantiated with some "'c iobj" -- a type obtained by
2412      using one of the functors PointerToIncompleteType or PointerToCompleteType.
2413    
2414      Operations that work on both incomplete and complete pointer types are
2415      typed as taking an 'o ptr while operations that require the target to
2416      be known are typed as taking some ('t, 'c) obj ptr.
2417    
2418      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
2419      where void is an eqtype without any values.  This makes it possible
2420      to work on voidptr values using functions meant to operate on light
2421      incomplete pointers.
2422    
2423    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
2424      been vastly simplified.
2425    
2426    ----------------------------------------------------------------------
2427    Name: Matthias Blume
2428    Date: 2002/02/19 10:48:00 EST
2429    Tag: blume-20020219-pqfix
2430    Description:
2431    
2432    Applied Chris Okasaki's bug fix for priority queues.
2433    
2434    ----------------------------------------------------------------------
2435    Name: Matthias Blume
2436    Date: 2002/02/15 17:05:00
2437    Tag: Release_110_39
2438    Description:
2439    
2440    Last-minute retagging is becoming a tradition... :-(
2441    
2442    This is the working release 110.39.
2443    
2444    ----------------------------------------------------------------------
2445    Name: Matthias Blume
2446    Date: 2002/02/15 16:00:00 EST
2447    Tag: Release_110_39-orig
2448    Description:
2449    
2450    Working release 110.39.  New bootfiles.
2451    
2452    (Update: There was a small bug in the installer so it wouldn't work
2453    with all shells.  So I retagged. -Matthias)
2454    
2455    ----------------------------------------------------------------------
2456    Name: Matthias Blume
2457    Date: 2002/02/15 14:17:00 EST
2458    Tag: blume-20020215-showbindings
2459    Description:
2460    
2461    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
2462    the latter can be useful for exploring what bindings are available at
2463    the interactive prompt.  (The first function returns only the list
2464    of symbols that are really bound, the second prints those but also the
2465    ones that CM's autoloading mechanism knows about.)
2466    
2467    ----------------------------------------------------------------------
2468    Name: Matthias Blume
2469    Date: 2002/02/15 12:08:00 EST
2470    Tag: blume-20020215-iptrs
2471    Description:
2472    
2473    Two improvements to ml-nlffigen:
2474    
2475      1. Write files only if they do not exist or if their current contents
2476         do not coincide with what's being written.  (That is, avoid messing
2477         with the time stamps unless absolutely necessary.)
2478    
2479      2. Implement a "repository" mechanism for generated files related
2480         to "incomplete pointer types".   See the README file for details.
2481    
2482    ----------------------------------------------------------------------
2483    Name: Matthias Blume
2484    Date: 2002/02/14 11:50:00 EST
2485    Tag: blume-20020214-quote
2486    Description:
2487    
2488    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
2489    because of the new and improved tag generation scheme.  (Thanks to Allen
2490    Leung for pointing it out.)
2491    
2492    ----------------------------------------------------------------------
2493    Name: Lal George
2494    Date: 2002/02/14 09:55:27 EST 2002
2495    Tag: george-20020214-isabelle-bug
2496    Description:
2497    
2498    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
2499    of Isabelle on the x86.
2500    
2501    From Allen:
2502    -----------
2503     I've found the problem:
2504    
2505         in ra-core.sml, I use the counter "blocked" to keep track of the
2506         true number of elements in the freeze queue.  When the counter goes
2507         to zero, I skip examining the queue.  But I've messed up the
2508         bookkeeping in combine():
2509    
2510             else ();
2511             case !ucol of
2512               PSEUDO => (if !cntv > 0 then
2513                     (if !cntu > 0 then blocked := !blocked - 1 else ();
2514                                        ^^^^^^^^^^^^^^^^^^^^^^^
2515                      moveu := mergeMoveList(!movev, !moveu)
2516                     )
2517                  else ();
2518    
2519         combine() is called to coalesce two nodes u and v.
2520         I think I was thinking that if the move counts of u and v are both
2521         greater than zero then after they are coalesced then one node is
2522         removed from the freeze queue.  Apparently I was thinking that
2523         both u and v are of low degree, but that's clearly not necessarily true.
2524    
2525    
2526    02/12/2002:
2527        Here's the patch.  HOL now compiles.
2528    
2529        I don't know how this impact on performance (compile
2530        time or runtime).  This bug caused the RA (especially on the x86)
2531        to go thru the potential spill phase when there are still nodes on the
2532        freeze queue.
2533    
2534    
2535    
2536    
2537    ----------------------------------------------------------------------
2538    Name: Matthias Blume
2539    Date: 2002/02/13 22:40:00 EST
2540    Tag: blume-20020213-fptr-rtti
2541    Description:
2542    
2543    Fixed a bug in ml-nlffigen that was introduced with one of the previous
2544    updates.
2545    
2546    ----------------------------------------------------------------------
2547    Name: Matthias Blume
2548    Date: 2002/02/13 16:41:00 EST
2549    Tag: blume-20020213-cmlpq
2550    Description:
2551    
2552    Added new priority queue export symbols (which have just been added to
2553    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
2554    not compile and the installer would choke.)
2555    
2556    ----------------------------------------------------------------------
2557    Name: Matthias Blume
2558    Date: 2002/02/13 16:15:00 EST
2559    Tag: blume-20020213-various
2560    Description:
2561    
2562    1. More tweaks to ml-nlffigen:
2563    
2564       - better internal datastructures (resulting in slight speedup)
2565       - "-match" option requires exact match
2566       - "localized" gensym counters (untagged structs/unions nested within
2567         other structs/unions or within typedefs get a fresh counter; their
2568         tag will be prefixed by a concatenation of their parents' tags)
2569       - bug fixes (related to calculation of transitive closure of types
2570         to be included in the output)
2571    
2572    2. Minor Basis updates:
2573    
2574       - added implementations for List.collate and Option.app
2575    
2576    ----------------------------------------------------------------------
2577    Name: Matthias Blume
2578    Date: 2002/02/11 15:55:00 EST
2579    Tag: blume-20020211-gensym
2580    Description:
2581    
2582    Added a "-gensym" option to command line of ml-nlffigen.  This can be
2583    used to specify a "stem" -- a string that is inserted in all "gensym'd"
2584    names (ML structure names that correspond to unnamed C structs, unions,
2585    and enums), so that separate runs of ml-nlffigen do not clash.
2586    
2587    ----------------------------------------------------------------------
2588    Name: Matthias Blume
2589    Date: 2002/02/11 12:05:00 EST
2590    Tag: blume-20020211-gensml
2591    Description:
2592    
2593    A quick fix for a problem with GenSML (in the pgraph-util library):
2594    Make generation of toplevel "local" optional.  (Strictly speaking,
2595    signature definitions within "local" are not legal SML.)
2596    
2597    Other than that: updates to INSTALL and cm/TODO.
2598    
2599    ----------------------------------------------------------------------
2600    Name: Matthias Blume
2601    Date: 2002/02/08 15:00:00 EST
2602    Tag: blume-20020208-uniquepid
2603    Description:
2604    
2605    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
2606    
2607    1. The installer (config/install.sh) has gotten smarter:
2608    
2609         - Configuration options are a bit easier to specify now
2610           (in config/targets).
2611         - Bug in recognizing .tar.bz2 files fixed.
2612         - Installer automatically resolves dependencies between
2613           configuration options (e.g., if you ask for eXene, you will
2614           also get cml -- regardless whether you asked for it or not).
2615         - Installer can run in "quieter mode" by setting the environment
2616           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
2617           "completely silent", though.
2618         - Build HashCons library as part of smlnj-lib.
2619    
2620    2. A new scheme for assigning persistent identifiers to compilation
2621       units (and, by extension, to types etc.) has been put into place.
2622       This fixes a long-standing bug where types and even dynamic values
2623       can get internally confused, thereby compromising type safety
2624       (abstraction) and dynamic correctness.  See
2625    
2626         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
2627    
2628       for an example of how things could go wrong until now.
2629    
2630       The downside of the new scheme is that pids are not quite as
2631       persistent as they used to be: CM will generate a fresh pid
2632       for every compilation unit that it thinks it sees for the first
2633       time.  That means that if you compile starting from a clean, fresh
2634       source tree at two different times, you end up with different
2635       binaries.
2636    
2637       Cutoff recompilation, however, has not been compromised because
2638       CM keeps pid information in special caches between runs.
2639    
2640    ----------------------------------------------------------------------
2641    Name: Lal George
2642    Date: 2002/02/07 15:34:13 EST 2002
2643    Tag: <none>
2644    Description:
2645    
2646    Compilers that generate assembly code may produce  global labels
2647    whose value is resolved at link time. The various peephole optimization
2648    modules did not take this in account.
2649    
2650    TODO. The Labels.addrOf function should really return an option
2651    type so that clients are forced to deal with this issue, rather
2652    than an exception being raised.
2653    
2654    ----------------------------------------------------------------------
2655    Name: Lal George
2656    Date: 2002/02/06 13:55:02 EST
2657    Tag: george-20020206-ra-breakup
2658    Description:
2659    
2660    1. A bug fix from Allen.
2661    
2662        A typo causes extra fstp %st(0)'s to be generated at compensation
2663        edges, which might cause stack underflow traps at runtime.  This
2664        occurs in fft where there are extraneous fstps right before the 'into'
2665        trap instruction (in this case they are harmless since none of the
2666        integers overflow.)
2667    
2668    2. Pulled out various utility modules that were embedded in the modules
2669       of the register allocator. I need these modules for other purposes, but
2670       they are not complete enough to put into a library (just yet).
2671    ----------------------------------------------------------------------
2672    Name: Matthias Blume
2673    Date: 2002/01/31 16:05:00 EST
2674    Tag: blume-20020131-sparc-ccalls
2675    Description:
2676    
2677    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
2678       of extra stack space by mistake.  Fixed.
2679    
2680    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
2681    
2682    ----------------------------------------------------------------------
2683    Name: Allen Leung
2684    Date: 2002/01/30
2685    Tag: leunga-20020130-nowhere-bug-fix
2686    Description:
2687    
2688       MLRISC bug fixes:
2689       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
2690       2. MachineInt.fromString was negating its value.
2691    
2692    ----------------------------------------------------------------------
2693    Name: Matthias Blume
2694    Date: 2002/01/29
2695    Tag: blume-20020129-INSTALL
2696    Description:
2697    
2698    - Added somewhat detailed installation instructions (file INSTALL).
2699    - Fixed curl-detection bug in config/install.sh.
2700    - It is now possible to select the URL getter using the URLGETTER
2701      environment variable:
2702    
2703          not set / "unknown"      --> automatic detection (script tries wget,
2704                                       curl, and lynx)
2705          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
2706                                       how to properly invoke them)
2707          other                    --> use $URLGETTER directly, it must take
2708                                       precisely two command-line arguments
2709                                       (source URL and destination file name)
2710    
2711    ----------------------------------------------------------------------
2712    Name: Matthias Blume
2713    Date: 2002/01/28
2714    Tag: blume-20020128-sparc-ccalls
2715    Description:
2716    
2717    - Fixed problem with calculation of "used" registers in sparc-c-calls.
2718    - Make use of the allocParam argument in sparc-c-calls.
2719    
2720    ----------------------------------------------------------------------
2721    Name: Matthias Blume
2722    Date: 2002/01/28
2723    Tag: blume-20020128-allocParam
2724    Description:
2725    
2726    John Reppy:  Changes c-calls API to accept client-callback for
2727    allocating extra stack space.
2728    me: Corresponding changes to mlriscGen (using a dummy argument that
2729        does not change the current behavior).
2730    
2731    ----------------------------------------------------------------------
2732    Name: Matthias Blume
2733    Date: 2002/01/28 12:00:00
2734    Tag: Release_110_38
2735    Description:
2736    
2737    This time for real!!!
2738    
2739    ----------------------------------------------------------------------
2740    Name: Matthias Blume
2741    Date: 2002/01/28 10:56:00 EST
2742    Tag: blume-20020128-retraction
2743    Description:
2744    
2745    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
2746       with blume-Release_110_38-retracted.)
2747    
2748    1. Fixed a problem with incorrect rounding modes in real64.sml.
2749       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
2750    
2751    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
2752       and enums fixed.  The naming of corresponding ML identifiers should
2753       now be consistent again.
2754    
2755    ----------------------------------------------------------------------
2756    Name: Allen Leung
2757    Date: 2002/01/27
2758    Tag: leunga-20020127-nowhere
2759    Description:
2760    
2761       Added a target called nowhere in the configuration scripts.
2762       Enabling this will build the MLRISC 'nowhere' tool (for translating
2763       programs with where-clauses into legal SML code) during installation.
2764    
2765    ----------------------------------------------------------------------
2766    Name: Matthias Blume
2767    Date: 2002/01/25 21:27:00 EST
2768    Tag: blume-Release_110_38-retracted
2769    Description:
2770    
2771    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
2772    
2773    README will be added later.
2774    
2775    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
2776    (Release_110_38) removed.  Reason: Last-minute bug fixes.
2777    
2778    ----------------------------------------------------------------------
2779    Name: Matthias Blume
2780    Date: 2002/01/25
2781    Tag: blume-20020125-ffi
2782    Description:
2783    
2784    A large number of tweaks and improvements to ml-nlffi-lib and
2785    ml-nlffigen:
2786    
2787       - ML represenation types have been streamlined
2788       - getter and setter functions work with concrete values, not abstract
2789         ones where possible
2790       - ml-nlffigen command line more flexible (see README file there)
2791       - some bugs have been fixed (hopefully)
2792    
2793    ----------------------------------------------------------------------
2794    Name: Lal George
2795    Date: 2002/01/24
2796    Tag: george-20020124-risc-ra-interface
2797    Description:
2798    
2799       There is a dramatic simplification in the interface to the
2800       register allocator for RISC architectures as a result of making
2801       parallel copy instructions explicit.
2802    
2803    ----------------------------------------------------------------------
2804    Name: Matthias Blume
2805    Date: 2002/01/22
2806    Tag: blume-20020122-x86-ccalls
2807    Description:
2808    
2809    Bug fix for c-calls on x86 (having to do with how char- and
2810    short-arguments are being handled).
2811    
2812    ----------------------------------------------------------------------
2813    Name: Matthias Blume
2814    Date: 2002/01/21
2815    Tag: blume-20020121-ff
2816    Description:
2817    
2818    Another day of fiddling with the FFI...
2819    
2820    1. Bug fix/workaround:  CKIT does not complain about negative array
2821       dimensions, so ml-nlffigen has to guard itself against this possibility.
2822       (Otherwise a negative dimension would send it into an infinite loop.)
2823    
2824    2. Some of the abstract types (light objects, light pointers, most "base"
2825       types) in structure C are now eqtypes.
2826    
2827    3. Added constructors and test functions for NULL function pointers.
2828    
2829    ----------------------------------------------------------------------
2830    Name: Matthias Blume
2831    Date: 2002/01/18
2832    Tag: blume-20020118-ready-for-new-release
2833    Description:
2834    
2835    Made config/srcarchiveurl point to a new place.  (Will provide boot
2836    files shortly.)
2837    
2838    Maybe we christen this to be 110.38?
2839    
2840    ----------------------------------------------------------------------
2841    Name: Matthias Blume
2842    Date: 2002/01/18
2843    Tag: blume-20020118-more-ffifiddle
2844    Description:
2845    
2846    Today's FFI fiddling:
2847    
2848      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
2849        C.Set.* that use concrete (MLRep.*) arguments and results instead
2850        of abstract ones.
2851    
2852      - Provided word-style bit operations etc. for "int" representation
2853        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
2854        and Long.
2855    
2856    ----------------------------------------------------------------------
2857    Name: Matthias Blume
2858    Date: 2002/01/18
2859    Tag: blume-20020118-use-x86-fp
2860    Description:
2861    
2862    Now that x86-fast-fp seems to be working, I turned it back on again
2863    by default.  (Seems to work fine now, even with the FFI.)
2864    
2865    Other than that, I added some documentation about the FFI to
2866    src/ml-nlffigen/README and updated the FFI test examples in
2867    src/ml-nlffi-lib/Tests/*.
2868    
2869    ----------------------------------------------------------------------
2870    Name: Allen Leung
2871    Date: 2002/01/17
2872    Tag: leunga-20020117-x86-fast-fp-call
2873    Description:
2874    
2875       1. Fixed a problem with handling return fp values when x86's fast fp
2876          mode is turned on.
2877    
2878       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
2879          of %f32.
2880    
2881       3. Added a constructor INT32lit to the ast of MLRISC tools.
2882    
2883    ----------------------------------------------------------------------
2884    Name: Matthias Blume
2885    Date: 2002/01/16
2886    Tag: blume-20020116-ffifiddle
2887    Description:
2888    
2889    More fiddling with the FFI interface:
2890    
2891     - Make constness 'c instead of rw wherever possible.  This eliminates
2892       the need for certain explicit coercions.  (However, due to ML's
2893       value polymorphism, there will still be many cases where explicit
2894       coercions are necessary.  Phantom types are not the whole answer
2895       to modeling a subtyping relationship in ML.)
2896    
2897     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
2898    
2899     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
2900       clumsy workaround for varargs problem.)
2901    
2902    ----------------------------------------------------------------------
2903    Name: Lal George
2904    Date: 2002/01/15
2905    Tag: <none>
2906    Description:
2907    
2908    1. Since COPY instructions are no longer native to the architecture,
2909       a generic functor can be used to implement the expandCopies function.
2910    
2911    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
2912       TEXT segment.
2913    
2914    ----------------------------------------------------------------------
2915    Name: Matthias Blume
2916    Date: 2002/01/15
2917    Tag: blume-20020115-ffiupdates
2918    Description:
2919    
2920    1. Fix for bug resulting in single-precision float values being returned
2921       incorrectly from FFI calls.
2922    
2923    2. Small modifications to C FFI API:
2924    
2925        - memory-allocation routines return straight objects (no options)
2926          and raise an exception in out-of-memory situations
2927        - unsafe extensions to cast between function pointers and pointers
2928          from/to ints
2929        - added structure C_Debug as an alternative to structure C where
2930          pointer-dereferencing (|*| and |*!) always check for null-pointers
2931        - added open_lib' to DynLinkage;  open_lib' works like open_lib
2932          but also takes a (possibly empty) list of existing library handles
2933          that the current library depends on
2934    
2935    ----------------------------------------------------------------------
2936    Name: Matthias Blume
2937    Date: 2002/01/10
2938    Tag: blume-20020110-newffigen
2939    Description:
2940    
2941    1. Updates to portable graph code.
2942    
2943    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
2944       more scalable now so that even huge interfaces such as the one
2945       for GTK compile in finite time and space. :-)
2946       See src/ml-nlffigen/README for details on what's new.
2947    
2948    ----------------------------------------------------------------------
2949    Name: Lal George
2950    Date: 2001/01/09 14:31:35 EST 2002
2951    Tag: george-20011206-rm-native-copy
2952    Description:
2953    
2954            Removed the native COPY and FCOPY instructions
2955            from all the architectures and replaced it with the
2956            explicit COPY instruction from the previous commit.
2957    
2958            It is now possible to simplify many of the optimizations
2959            modules that manipulate copies. This has not been
2960            done in this change.
2961    
2962    ----------------------------------------------------------------------
2963    Name: Lal George
2964    Date: 2001/12/06 16:50:13 EST 2001
2965    Tag: george-20011206-mlrisc-instruction
2966    Description:
2967    
2968    Changed the representation of instructions from being fully abstract
2969    to being partially concrete. That is to say:
2970    
2971      from
2972            type instruction
2973    
2974      to
2975            type instr                              (* machine instruction *)
2976    
2977            datatype instruction =
2978                LIVE of {regs: C.cellset, spilled: C.cellset}
2979              | KILL of {regs: C.cellset, spilled: C.cellset}
2980              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
2981              | ANNOTATION of {i: instruction, a: Annotations.annotation}
2982              | INSTR of instr
2983    
2984    This makes the handling of certain special instructions that appear on
2985    all architectures easier and uniform.
2986    
2987    LIVE and KILL say that a list of registers are live or killed at the
2988    program point where they appear. No spill code is generated when an
2989    element of the 'regs' field is spilled, but the register is moved to
2990    the 'spilled' (which is present, more for debugging than anything else).
2991    
2992    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
2993    We used to generate:
2994    
2995            DEFFREG f1
2996            f1 := f2 + f3
2997            trapb
2998    
2999    but now generate:
3000    
3001            f1 := f2 + f3
3002            trapb
3003            LIVE {regs=[f1,f2,f3], spilled=[]}
3004    
3005    Furthermore, the DEFFREG (hack) required that all floating point instruction
3006    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3007    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3008    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3009    this architecture).
3010    
3011    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3012    all the architectures. This will result in further simplification of the
3013    register allocator that must be aware of them for coalescing purposes, and
3014    will also simplify certain aspects of the machine description that provides
3015    callbacks related to parallel copies.
3016    
3017    ANNOTATION should be obvious, and now INSTR represents the honest to God
3018    machine instruction set!
3019    
3020    The <arch>/instructions/<arch>Instr.sml files define certain utility
3021    functions for making porting easier -- essentially converting upper case
3022    to lower case. All machine instructions (of type instr) are in upper case,
3023    and the lower case form generates an MLRISC instruction. For example on
3024    the alpha we have:
3025    
3026      datatype instr =
3027         LDA of {r:cell, b:cell, d:operand}
3028       | ...
3029    
3030      val lda : {r:cell, b:cell, d:operand} -> instruction
3031        ...
3032    
3033    where lda is just (INSTR o LDA), etc.
3034    
3035    ----------------------------------------------------------------------
3036    Name: Matthias Blume
3037    Date: 2001/11/22 21:40:00 EST
3038    Tag: Release_110_37
3039    Description:
3040    
3041    Release 110.37.  This time for real.
3042    
3043    ----------------------------------------------------------------------
3044    Name: Matthias Blume
3045    Date: 2001/11/21 16:35:00 EST
3046    Tag: blume-20011121-foot-in-mouth
3047    Description:
3048    
3049    Removed the "Release_110_37" tag because of a serious bug.
3050    This will be re-tagged once the bug is fixed.
3051    
3052    ----------------------------------------------------------------------
3053    Name: Matthias Blume
3054    Date: 2001/11/21 16:14:00 EST
3055    Tag: blume-20011121-forgottenfile
3056    Description:
3057    
3058    Forgot to add a file.  (Just a .tex-file -- part of
3059    the CM manual source.)
3060    
3061    ----------------------------------------------------------------------
3062    Name: Matthias Blume
3063    Date: 2001/11/21 16:10:00 EST
3064    Tag: blume-20011121-invalid_110_37
3065    Description:
3066    
3067    Note: I removed the original tag "Release_110_37" from this commit
3068          because we found a serious bug in all non-x86 backends.
3069          - Matthias
3070    
3071    1. Modifications to the SML/NJ code generator and to the runtime system
3072       so that code object name strings are directly inserted into code
3073       objects at code generation time.  The only business the runtime system
3074       has with this is now to read the name strings on occasions.
3075       (The encoding of the name string has also changed somewhat.)
3076    
3077    2. CM now implements a simple "set calculus" for specifying export lists.
3078       In particular, it is now possible to refer to the export lists of
3079       other libraries/groups/sources and form unions as well as differences.
3080       See the latest CM manual for details.
3081    
3082    3. An separate notion of "proxy" libraries has again be eliminated from
3083       CM's model.  (Proxy libraries are now simply a special case of using
3084       the export list calculus.)
3085    
3086    4. Some of the existing libraries now take advantage of the new set
3087       calculus.
3088       (Notice that not all libraries have been converted because some
3089       of the existing .cm-files are supposed to be backward compatible
3090       with 110.0.x.)
3091    
3092    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3093       "General.exnMessage"!  The former relies on a certain hook to be
3094       initialized, and that often does not happen in the stand-alone case.)
3095    
3096    ----------------------------------------------------------------------
3097    Name: Lal George
3098    Date: 2001/11/21  13:56:18 EST
3099    Tag: george-2001121-pseudo-ops
3100    Description:
3101    
3102      Implemented a complete redesign of MLRISC pseudo-ops. Now there
3103      ought to never be any question of incompatabilities with
3104      pseudo-op syntax expected by host assemblers.
3105    
3106      For now, only modules supporting GAS syntax are implemented
3107      but more should follow, such as MASM, and vendor assembler
3108      syntax, e.g. IBM as, Sun as, etc.
3109    
3110    ----------------------------------------------------------------------
3111    Name: Matthias Blume
3112    Date: 2001/11/14 11:52:00 EST
3113    Tag: blume-20011114-srcname
3114    Description:
3115    
3116    1. Routed the name of the current source file to mlriscgen where it
3117       should be directly emitted into the code object.  (This last part
3118       is yet to be done.)
3119    
3120    2. Some cleanup of the pgraph code to make it match the proposal that
3121       I put out the other day.  (The proposal notwithstanding, things are
3122       still in flux here.)
3123    
3124    ----------------------------------------------------------------------
3125    Name: Lal George
3126    Date: 2001/11/14 09:44:04 EST
3127    Tag:
3128    Description:
3129    
3130      Fix for a backpatching bug reported by Allen.
3131    
3132      Because the boundary between short and long span-dependent
3133      instructions is +/- 128, there are an astounding number of
3134      span-dependent instructions whose size is over estimated.
3135    
3136      Allen came up with the idea of letting the size of span
3137      dependent instructions be non-monotonic, for a maxIter
3138      number of times, after which the size must be monotonically
3139      increasing.
3140    
3141      This table shows the number of span-dependent instructions
3142      whose size was over-estimated as a function of maxIter, for the
3143      file Parse/parse/ml.grm.sml:
3144    
3145         maxIter            # of instructions:
3146            10                      687
3147            20                      438
3148            30                      198
3149            40                        0
3150    
3151      In compiling the compiler, there is no significant difference in
3152      compilation speed between maxIter=10 and maxIter=40. Actually,
3153      my measurements showed that maxIter=40 was a tad faster than
3154      maxIter=10! Also 96% of the  files in the compiler reach a fix
3155      point within 13 iterations, so fixing maxIter at 40, while high,
3156      is okay.
3157    
3158    ----------------------------------------------------------------------
3159    Name: Matthias Blume
3160    Date: 2001/10/31 15:25:00 EST
3161    Tag: blume-20011031-pgraph
3162    Description:
3163    
3164    CKIT:
3165    * Changed the "Function" constructor of type Ast.ctype to carry optional
3166      argument identifiers.
3167    * Changed the return type of TypeUtil.getFunction accordingly.
3168    * Type equality ignores the argument names.
3169    * TypeUtil.composite tries to preserve argument names but gives up quickly
3170      if there is a mismatch.
3171    
3172    installation script:
3173    * attempts to use "curl" if available (unless "wget" is available as well)
3174    
3175    CM:
3176    * has an experimental implementation of "portable graphs" which I will
3177      soon propose as an implementation-independent library format
3178    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
3179    
3180    NLFFI-LIB:
3181    * some cleanup (all cosmetic)
3182    
3183    NLFFIGEN:
3184    * temporarily disabled the mechanism that suppresses ML output for
3185      C definitions whose identifiers start with an underscore character
3186    * generate val bindings for enum constants
3187    * user can request that only one style (light or heavy) is being used;
3188      default is to use both (command-line arguments: -heavy and -light)
3189    * fixed bug in handling of function types involving incomplete pointers
3190    * generate ML entry points that take record arguments (i.e., using
3191      named arguments) for C functions that have a prototype with named
3192      arguments
3193      (see changes to CKIT)
3194    
3195    ----------------------------------------------------------------------
3196    Name: Allen Leung
3197    Date: 2001/10/27 20:34:00 EDT
3198    Tag: leunga-20011027-x86-fast-fp-call
3199    Description:
3200    
3201       Fixed the bug described in blume-20010920-slowfp.
3202    
3203       The fix involves
3204          1. generating FCOPYs in FSTP in ia32-svid
3205          2. marking a CALL with the appropriate annotation
3206    
3207    ----------------------------------------------------------------------
3208    Name: Matthias Blume
3209    Date: 2001/10/16 11:32:00 EDT
3210    Tag: blume-20011016-netbsd
3211    Description:
3212    
3213    Underscore patch from Chris Richards (fixing problem with compiling
3214    runtime system under recent NetBSD).
3215    
3216    ----------------------------------------------------------------------
3217    Name: Allen Leung
3218    Date: 2001/10/12 17:18:32 EDT 2001
3219    Tag: leung-20011012-x86-printflowgraph
3220    Description:
3221    
3222    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
3223    
3224    ----------------------------------------------------------------------
3225  Name: Lal George  Name: Lal George
3226  Date: 2001/10/11 23:51:34 EDT  Date: 2001/10/11 23:51:34 EDT
3227  Tag: george-20011011-too-many-instrs  Tag: george-20011011-too-many-instrs
# Line 80  Line 3289 
3289    
3290  ----------------------------------------------------------------------  ----------------------------------------------------------------------
3291  Name: Matthias Blume  Name: Matthias Blume
 >>>>>>> 1.169  
3292  Date: 2001/09/18 15:35:00 EDT  Date: 2001/09/18 15:35:00 EDT
3293  Tag: blume-20010918-readme11036  Tag: blume-20010918-readme11036
3294  Description:  Description:

Legend:
Removed from v.958  
changed lines
  Added in v.1572

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