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 716, Sat Nov 4 05:42:35 2000 UTC revision 1557, Wed Jul 21 23:19:18 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: 2004/07/21 18:20:00 CDT
18    Tag: blume-20040721-nlffigen
19    Description:
20    
21    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
22      for a typedef to an incomplete type, but leave out the "typ" member.
23      (This is just for consistency.)
24    - Started to produce what is supposed to become better (i.e., comprehensive)
25      documentation of what ml-nlffigen does and produces.
26    
27    ----------------------------------------------------------------------
28    Name: Matthias Blume (blume (at) tti - c (dot) org)
29    Date: 2004/07/14 16:25:00 CDT
30    Tag: blume-20040714-union
31    Description:
32    
33    Added C_UNION to c-calls/c-types.sml and updated the machinery
34    (ml-nlffigen, cproto.sml) that conveys C function interface
35    information to the code generator.
36    
37    However, the actual architecture-specific implementation of function
38    arguments and results that are C unions is still not implemented.
39    
40    ----------------------------------------------------------------------
41    Name: Allen Leung (leunga (at) reservoir (dot) com)
42    Date: 2004/07/14 14:38:00 EST
43    Tag: leunga-110_46_1-ppc-lwzu
44    Description:
45    
46        Added these instructions to the PowerPC architecture:
47        LBZU(X), LHZU(X), LWZU(X),
48        STWU(X), STFDU, STFSU
49    
50        etc...
51    
52        Note: I haven't added their instruction encoding into the description.
53    
54    ----------------------------------------------------------------------
55    Name: Allen Leung (leunga (at) reservoir (dot) com)
56    Date: 2004/07/13 15:04:00 EST
57    Tag: leunga-110_46_1-ppc-lwarx
58    Description:
59    
60       Added the two instructions LWARX and STWCX to the PowerPC
61    instruction set.
62    
63       A (untested) rewrite of loop-structure.sml.  The old version
64    is completely broken.
65    
66    ----------------------------------------------------------------------
67    Name: Matthias Blume (blume (at) tti - c (dot) org)
68    Date: 2004/07/13 13:50:00 CDT
69    Tag: blume-20040713-nlffi
70    Description:
71    
72    - use paramAlloc to report c-calls with too many arguments
73      (for PPC version where parameter area is pre-allocated)
74    - added ccall_maxargspace to machspec (to implement the above)
75    - made "make" commend in CM's "make" tool configurable
76    - added option (default: on) for passing the name of the SML/NJ's "bin"
77      directory to "make"; the call looks like this:
78    
79         make <options> SMLNJ_BINDIR=<dir> <target>
80    
81      This can be used by the Makefile to, e.g., pick the "right" version
82      of ml-nlffigen.
83    - minor code tweaks
84    
85    ----------------------------------------------------------------------
86    Name: Matthias Blume (blume (at) tti - c (dot) org)
87    Date: 2004/07/12 22:50:00 CDT
88    Tag: blume-110_46_1-macosx-nlffi
89    Description:
90    
91    NLFFI under Mac OS X now working (sort of).  This is largely untested,
92    though.
93    
94    Note:  1. You have to make a new, clean build of the runtime system.
95           2. There are new BOOTFILES, you have to use them!
96              (Doing the bootstrap process yourself would be *very* painful!
97               If you absolutely have to do it, build the system under
98               a different architecture and then cross-compile.)
99    
100    Version bumped to 110.46.1 to account for runtime data format changes.
101    
102    ----------------------------------------------------------------------
103    Name: Matthias Blume (blume (at) tti - c (dot) org)
104    Date: 2004/06/18 14:30:00 CDT
105    Tag: blume-20040618-unix
106    Description:
107    
108    Changed the implementation of structure Unix so that the same stream
109    is returned every time one of the {text,bin}{In,Out}streamOf functions
110    is invoked on the same proc.  This is not what the spec currently
111    says -- although IMO it arguably should.  (See discussion below.)
112    
113    ----------------------------------------------------------------------
114    Name: Matthias Blume (blume (at) tti - c (dot) org)
115    Date: 2004/06/17 18:15:00 CDT
116    Tag: Release_110_46
117    Description:
118    
119    New working version (110.46).  NEW BOOTFILES!
120    
121    ----------------------------------------------------------------------
122    Name: Matthias Blume (blume (at) tti - c (dot) org)
123    Date: 2004/06/17 17:20:00 CDT
124    Tag: blume-20040617-timer-unix
125    Description:
126    
127    Changed the interface of structures Timer and Unix to match the most
128    recent Basis spec.
129    
130    In the case of Unix there still seems to be an open/weird issue:
131    
132       The {text,bin}{In,Out}streamOf functions are supposed to create
133       fresh streams whenever they are called -- as opposed to have them
134       return the same stream every time.  This design is supposed to
135       prevent space leaks caused by proc values hanging on to streams.
136    
137       The reap function, on the other hand, is supposed to close the
138       streams.  This cannot be done without having a handle on the
139       stream in proc after all...
140    
141    I took the liberty to implement the following stopgap solution:
142    
143      The proc value hangs on to the most recently created stream(s).
144      Reap closes those.  If either or both of the two streams hadn't
145      been created at all yet, then reap will close the corresponding
146      file descriptors directly.
147    
148    PS: I don't understand the original space leak argument anymore.  If
149    a proc hangs on to the imperative stream, then I/O operations on those
150    will advance the state of the cached stream and avoid the space leak.
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume (blume (at) tti - c (dot) org)
154    Date: 2004/05/28 16:45:00 CDT
155    Tag: blume-20040528-basis
156    Description:
157    
158    Added signature PACK_REAL and exported functor PrimIO.
159    
160    ----------------------------------------------------------------------
161    Name: Matthias Blume (blume (at) tti - c (dot) org)
162    Date: 2004/05/25 16:00:00 CDT
163    Tag: blume-20040525-group-owner
164    Description:
165    
166    CM now ignores (but still accepts) the "owner" information in group
167    descriptions.  The owner of a group is its next enclosing
168    library. Each group must have a unique owner.  (There is a virtual
169    "toplevel" library that owns groups which are not nested within a real
170    library.)  Previously, each group had to explicitly declare its owner,
171    and CM would check that such a declaration is correct.  The new scheme
172    is to have CM check that for each group there is precisely one owning
173    library.
174    
175    The advantage of the new scheme is that the programmer no longer needs
176    to maintain the somewhat annoying owner information.  The downside is
177    that CM cannot enforce the ownership rule across multiple runs of
178    CM.make.  Fortunately, enclosing the same group in two different
179    libraries A and B which are not part of the same program does not
180    cause real problems.
181    
182    ----------------------------------------------------------------------
183    Name: Matthias Blume (blume (at) tti - c (dot) org)
184    Date: 2004/05/20 16:00:00 CDT
185    Tag: blume-20040520-win32
186    Description:
187    
188    Made the win32 version work again.  (Strangely, a misplaced comma had
189    slipped into win32-process.c which prevented the runtime from being
190    compiled correctly.)
191    
192    Also, included a minor addition to ml-build.bat analogous to what was
193    done in blume-20040519-ml-build.
194    
195    ----------------------------------------------------------------------
196    Name: Matthias Blume (blume (at) tti - c (dot) org)
197    Date: 2004/05/19 22:10:00 CDT
198    Tag: blume-20040519-ml-build
199    Description:
200    
201    Arranged for ml-build to clean up after itself a little bit better.
202    The script generates a temporary SML source file and compiles it using
203    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
204    gets rid of those at the end, so they don't accumulate under .cm.
205    
206    This required a minor change to install.sh because the name of the
207    metadata directory (default: .cm) is actually configurable at
208    installation time.
209    
210    ----------------------------------------------------------------------
211    Name: Matthias Blume (blume (at) tti - c (dot) org)
212    Date: 2004/05/18 15:50:00 CDT
213    Tag: blume-20040518-mkreader
214    Description:
215    
216    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
217    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
218    
219    ----------------------------------------------------------------------
220    Name: Matthias Blume (blume (at) tti - c (dot) org)
221    Date: 2004/05/11 14:35:00 CDT
222    Tag: blume-20040511-win32sock
223    Description:
224    
225    Added previously missing support for many socket-related functions
226    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
227    for the voluminous patch!
228    
229    (I have not tested this patch under win32 yet.)
230    
231    Here is David's e-mail:
232    
233    Hi,
234    
235    Attached to this email you find a diff against sml/nj 110.45
236    that will enable socket support under Windows.
237    
238    To apply the patch (using unix or cygwin)
239    1) gunzip runtime.diff.gz
240    2) "cd" into "src/runtime" in the source tree of a fresh
241       110.45 installation.
242    3) patch -p 1 < [your/path/to]runtime.diff
243    
244    The code compiles fine but has NOT yet been extensively tested.
245    I only ran a few tests for basic socket client functionality
246    (which worked fine).  Especially the functions that use ioctl
247    are not tested at all and might not work (see below).
248    
249    I implemented this since we want to move to a newer version of sml/nj
250    but need socket support in order to use it.  This is the first time I
251    even had a look at the sml/nj source,  so please review my changes
252    before making this part of the distribution!  Here are a few issues
253    that I think might be better for someone to solve who is more
254    familiar with the sml/nj source (and socket programming):
255    
256    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
257      exception since I could not figure out what the windows equivalent
258      of these functions is
259    
260    - In sockets-osdep.h there are a some #include statements that are
261      only used in a few files that include sockets-osdep.h
262    
263    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
264      does not process its return value since I don't know how to
265      report an error upwards.
266    
267    - It would probably be good to have a call to WSACleanup() when
268      the library is unloaded (if there is such a possibility).
269      Otherwise I think Windows will take care of this automatically
270      when the process finishes.
271    
272    - I used ioctlsocket() as a replacement for ioctl() but I have
273      no idea if that is actually the proper replacement on Windows.
274    
275    - All these issues are marked in the code by "FIXME" comments.
276    
277    We use sml/nj extensively in our products and are quite happy
278    with it.  I hope this contribution will help you.
279    
280    Keep up the good work!
281    
282    David
283    
284    ----------------------------------------------------------------------
285    Name: Matthias Blume (blume (at) tti - c (dot) org)
286    Date: 2004/05/11 14:20:00 CDT
287    Tag: blume-20040511-installml
288    Description:
289    
290    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
291    the patch.)
292    
293    ----------------------------------------------------------------------
294    Name: Matthias Blume (blume (at) tti - c (dot) org)
295    Date: 2004/05/11 14:05:00 CDT
296    Tag: blume-20040511-nlffi-netbsd
297    Description:
298    
299    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
300    the patch.)
301    
302    ----------------------------------------------------------------------
303    Name: Matthias Blume (blume (at) tti - c (dot) org)
304    Date: 2004/05/11 12:05:00 CDT
305    Tag: blume-20040511-exports
306    Description:
307    
308    As per request by Adam Chlipala <adam@hcoop.net>, extended various
309    export lists in compiler-related .cm-files.
310    
311    ----------------------------------------------------------------------
312    Name: Matthias Blume (blume (at) tti - c (dot) org)
313    Date: 2004/05/11 11:35:00 CDT
314    Tag: blume-20040511-allsource
315    Description:
316    
317    The installer now honors the "src-smlnj" target again, although its meaning
318    has changed from "all sources required for the compiler" to "all sources
319    the installer knows about".  In other words, if you enable "src-smlnj"
320    in the "targets" file, then the installer will pull in sources for
321    everything.  (Notice that this refers to source code only.  Compiled
322    code is still only installed for modules that were requested explicitly
323    or which are required for other modules that were requested explicitly.)
324    
325    ----------------------------------------------------------------------
326    Name: Matthias Blume (blume (at) tti - c (dot) org)
327    Date: 2004/04/23 17:40:00 CDT
328    Tag: blume-20040423-ieee-scan
329    Description:
330    
331    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
332    in the exponent calculation we get INF or ZERO (depending on the mantissa
333    and the sign of the exponent).
334    
335    ----------------------------------------------------------------------
336    Name: Matthias Blume (blume (at) tti - c (dot) org)
337    Date: 2004/04/23 10:40:00 CDT
338    Tag: blume-20040423-ml-build
339    Description:
340    
341    The ml-build script now terminates with a non-0 status when something
342    goes wrong.
343    
344    ----------------------------------------------------------------------
345    Name: Matthias Blume (blume (at) tti - c (dot) org)
346    Date: 2004/04/22 16:35:00 CDT
347    Tag: blume-20040422-Option
348    Description:
349    
350    Made exception Option to be the same as exception Option.Option
351    (as it should be).
352    
353    ----------------------------------------------------------------------
354    Name: Allen Leung (leunga (at) reservoir (dot) com)
355    Date: 2004/03/19 14:40:00 EST
356    Tag: leunga-20040319-cygwin-nlffi
357    Description:
358    
359    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
360    of SML/NJ.  The problem is that
361    
362       lib = dlopen(NULL, ...)
363       f   = dlsym(lib, "malloc");
364    
365    does not work on Windows unless we explicitly export symbols
366    such as 'malloc' during linking.  We fixed this by explicitly
367    exporting the required symbols with the magic gcc incantation:
368    
369        -Wl,--export-all cygwin.def
370    
371    where cygwin.def is a file containing all the symbols that
372    we wish to export.
373    
374    I suspect this is a Windows problem and we'll have to
375    do the same (somehow with windows compilers) when
376    we build the native win32 version with the system
377    calls LoadLibrary/GetProcAddress.
378    
379    ----------------------------------------------------------------------
380    Name: Matthias Blume (blume (at) tti - c (dot) org)
381    Date: 2004/03/04 16:35:00 CST
382    Tag: blume-20040304-intinf-fmt
383    Description:
384    
385    Fixed problem with IntInf.fmt (sign would show up on the right instead
386    of on the left for BIN, OCT, and HEX).
387    
388    ----------------------------------------------------------------------
389    Name: Matthias Blume (blume (at) tti - c (dot) org)
390    Date: 2004/03/04 11:25:00 CST
391    Tag: blume-20040304-symlinks
392    Description:
393    
394    Fixed problem with installer script (unix only) where bin/ml-yacc and
395    friends pointed (via symlinks) to absolute locations instead of just
396    .run-sml.  This was reported by Vesa A Norrman.
397    
398    ----------------------------------------------------------------------
399    Name: Matthias Blume (blume (at) tti - c (dot) org)
400    Date: 2004/02/13 14:50:00 CST
401    Tag: Release_110_45
402    Description:
403    
404    New working version (110.45). New bootfiles.
405    
406    ----------------------------------------------------------------------
407    Name: Matthias Blume (blume (at) tti - c (dot) org)
408    Date: 2004/01/26 15:15:15 CST
409    Tag: blume-20040126-toplevel
410    Description:
411    
412    Improved handling of exceptions at the interactive toplevel.
413    
414    ----------------------------------------------------------------------
415    Name: Matthias Blume (blume (at) tti - c (dot) org)
416    Date: 2004/01/26 11:25:00
417    Tag: blume-20040126-app
418    Description:
419    
420    Type of top-level "app" corrected.
421    Added code for setting vp_limitPtrMask to Win32-specific runtime.
422    
423    ----------------------------------------------------------------------
424    Name: Matthias Blume (blume (at) tti - c (dot) org)
425    Date: 2003/11/18 17:10 CST
426    Tag: blume-20031118-basis-fiddle
427    Description:
428    
429    - changed Timer interface to what might become the spec
430    - POSIX_FLAGS -> BIT_FLAGS according to spec
431    - some other minor discrepancies wrt. spec eliminated
432    
433    ----------------------------------------------------------------------
434    Name: Matthias Blume (blume (at) tti - c (dot) org)
435    Date: 2003/11/06 12:00:00 CST
436    Tag: Release_110_44
437    Description:
438    
439    New working version (110.44). New bootfiles.
440    
441    ----------------------------------------------------------------------
442    Name: Matthias Blume (blume (at) tti - c (dot) org)
443    Date: 2003/11/04 11:50:00 CST
444    Tag: blume-20031104-move-libraries
445    Description:
446    
447    Eliminated the "dont_move_libraries" directive in config/targets.
448    (The mechanism was broken and could not be fixed easily.  Moreover,
449    there does not seem to be any reason not to move all libraries into
450    lib during installation.  I originally implemented this directive as a
451    backward-compatibility feature when I first introduced the new CM.
452    Now that things have been stable for a long time and going back to the
453    old CM is not an option, there is no reason to keep it around.)
454    
455    ----------------------------------------------------------------------
456    Name: Matthias Blume (blume (at) tti - c (dot) org)
457    Date: 2003/11/03 16:00:00 CST
458    Tag: blume-20031103-installdir
459    Description:
460    
461    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
462    Richards for pointing out the problem and providing the solution.)
463    
464    ----------------------------------------------------------------------
465    Name: Matthias Blume (blume (at) tti - c (dot) org)
466    Date: 2003/10/01 17:05:00 CDT
467    Tag: blume-20031001-lal-mlrisc
468    Description:
469    
470    MLRISC bug fix from Lal.
471    
472    ----------------------------------------------------------------------
473    Name: Matthias Blume (blume (at) tti - c (dot) org)
474    Date: 2003/09/30 16:10:00 CDT
475    Tag: blume-20030930-primio-bat
476    Description:
477    
478    1. Added openVector, nullRd, and nullWr to PRIM_IO.
479    2. Improved .bat files (for Win32 port) to make things work under Win95.
480       (thanks to Aaron S. Hawley for this one)
481    
482    ----------------------------------------------------------------------
483    Name: Matthias Blume (blume (at) tti - c (dot) org)
484    Date: 2003/09/26 16:05:00 CDT
485    Tag: blume-20030926-wrappriv
486    Description:
487    
488    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
489    
490    ----------------------------------------------------------------------
491    Name: Matthias Blume (blume (at) tti - c (dot) org)
492    Date: 2003/09/26 15:00:00 CDT
493    Tag: blume-20030926-110_43_3
494    Description:
495    
496    - additional cleanup
497    - version number bump, NEW BOOTFILES
498    
499    ----------------------------------------------------------------------
500    Name: Matthias Blume (blume (at) tti - c (dot) org)
501    Date: 2003/09/26 12:00:00 CDT
502    Tag: blume-20030926-ppautoload
503    Description:
504    
505    I modified the read-eval-print loop so that the autoloader gets
506    invoked whenever the prettyprinter tries to look up a symbol that
507    is not currently defined in the toplevel environment but which
508    appears in CM's autoload registry.  As a result, we see far fewer of
509    those ?.Foo.Bar.xxx names in the prettyprinter's output.
510    
511    In addition to this I tried to clean up some pieces of the Basis
512    implementation (e.g., Socket, Word8Array) in order to prevent other
513    instances of these ?.Foo.Bar.xxx names from being printed.
514    
515    The mechanism that picks names for types still needs some work, though.
516    (Right now it seems that if there is a type A.t which is defined to
517    be B.u, but B is unavailable at toplevel, then A.t gets printed as
518    "?.B.u" although the perhaps more sensible solution would be to use
519    "A.t" in this case.  In other words, the prettyprinter should follow
520    a chain of DEFtycs not farther than there are corresponding toplevel
521    names in the current environment.)
522    
523    ----------------------------------------------------------------------
524    Name: Matthias Blume (blume (at) tti - c (dot) org)
525    Date: 2003/09/24 16:31:00 CDT
526    Tag: blume-20030924-installer
527    Description:
528    
529    Another installer tweak:  All the ML code for the installer is now
530    compiled during CMB.make and put into a little library called
531    $smlnj/installer.cm.  The installation then simply invokes
532    
533       sml -m $smlnj/installer.cm
534    
535    and everything happens automagically.
536    
537    Win32:  ML code senses value of environment variable SMLNJ_HOME.
538    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
539            and BINDIR.
540    
541    The new scheme guarantees that the ML code responsible for the installation
542    is in sync with the APIs of the main system.  Also, the installer is
543    somewhat faster because the installer script is precompiled.
544    
545    ----------------------------------------------------------------------
546    Name: Matthias Blume (blume (at) tti - c (dot) org)
547    Date: 2003/09/24 15:35:00 CDT
548    Tag: blume-20030924-synsock
549    Description:
550    
551    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
552    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
553    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
554    implementations of non-blocking operations from CML's Socket
555    structure.
556    
557    ----------------------------------------------------------------------
558    Name: Matthias Blume (blume (at) tti - c (dot) org)
559    Date: 2003/09/24 15:10:05 CDT
560    Tag: blume-20030924-sockets
561    Description:
562    
563    1. Fixed SOCKET API and implementation to match Basis spec.
564       This required changing the internal representation of sockets to one
565       that remembers (for each socket file descriptor) whether it is currently
566       blocking or non-blocking.  This state is maintained lazily (i.e., a system
567       call is made only if the state actually needs to change).
568    
569    2. OS-specific details of sockets were moved into separate files, thus
570       making it possible to unify the bulk of the socket implementations
571       between Unix and Win32.
572    
573    3. CML's socket API changed accordingly.
574       (Note that we need to remove non-blocking functions from this API
575       since they are redundant in the case of CML!)
576    
577    4. CML's socket implementation now makes use of non-blocking functions
578       provided by Basis, thus removing all OS-dependent code from this part
579       of CML.
580    
581    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
582    
583    ----------------------------------------------------------------------
584    Name: Matthias Blume (blume (at) tti - c (dot) org)
585    Date: 2003/09/22 12:10:00 CDT
586    Tag: blume-20030922-110_43_2
587    Description:
588    
589    Made a new interim version and bootfiles for developer's bootstrapping
590    convenience.
591    
592       110.43.2 -- NEW BOOTFILES
593    
594    ----------------------------------------------------------------------
595    Name: Matthias Blume (blume (at) tti - c (dot) org)
596    Date: 2003/09/19 15:55:00 CDT
597    Tag: blume-20030919-cmdir
598    Description:
599    
600    1. new-install.sh -> install.sh
601    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
602    3. tweaked installer so that another name instead of .cm can be chosen
603       at install time (by setting the CM_DIR_ARC environment variable
604       during installation); once installation is complete, the name is
605       fixed
606    
607    ----------------------------------------------------------------------
608    Name: Matthias Blume (blume (at) tti - c (dot) org)
609    Date: 2003/09/18 16:00:00 CDT
610    Tag: blume-20030918-110_43_1
611    Description:
612    
613    Made a new interim version and bootfiles for developer's bootstrapping
614    convenience.
615    
616       110.43.1 -- NEW BOOTFILES
617    
618    ----------------------------------------------------------------------
619    Name: Matthias Blume (blume (at) tti - c (dot) org)
620    Date: 2003/09/18 15:20:00 CDT
621    Tag: blume-20030918-misc
622    Description:
623    
624    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
625       this seems to be controversial at the moment) and used those in
626       Posix.ProcEnv.times.
627    
628    2. Added Time.{from,to}Nanoseconds to Time.
629    
630    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
631       For example, fromLargeInt never needs to look at more than 3 "big
632       digits" to get its 53 bits of precision.
633    
634    ----------------------------------------------------------------------
635    Name: Matthias Blume (blume (at) tti - c (dot) org)
636    Date: 2003/09/17 16:30:00 CDT
637    Tag: blume-20030917-real32-slices
638    Description:
639    
640    Added an entry to the primitive environment
641    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
642    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
643    
644    Removed some of the "magic" constants in real64.sml and replaced them
645    with code that generates these values from their corresponding
646    integer counterparts.
647    
648    Made all(?) the slice-related changes to the Basis and made everything
649    compile again...
650    
651    ----------------------------------------------------------------------
652    Name: Matthias Blume (blume (at) tti - c (dot) org)
653    Date: 2003/09/15 17:45:00 CDT
654    Tag: blume-20030915-rbase
655    Description:
656    
657    Fixed bug in Real.fromLargeInt.
658    
659    ----------------------------------------------------------------------
660    Name: Matthias Blume (blume (at) tti - c (dot) org)
661    Date: 2003/09/13 18:11:00 CDT
662    Tag: blume-20030913-libinstall
663    Description:
664    
665    Minor bugfix in config/libinstall (set anchor with path to
666    standalone tool after installing it, otherwise libraries that
667    need ml-lex or ml-yacc won't compile the first time the installer
668    runs).
669    
670    ----------------------------------------------------------------------
671    Name: Matthias Blume (blume (at) tti - c (dot) org)
672    Date: 2003/09/12 11:45:00 CDT
673    Tag: blume-20030912-various
674    Description:
675    
676     - fixed bug in Real.toLargeInt
677     - fixed bug in Posix.ProcEnv.times
678     - changed inputLine functions to return an option
679     - minor installer improvements / bugfixes
680     - changed default @SMLalloc parameter for x86/celeron to 64k
681    
682    ----------------------------------------------------------------------
683    Name: Matthias Blume (blume (at) tti - c (dot) org)
684    Date: 2003/09/09 22:00:00 CDT
685    Tag: Release_110_43
686    Description:
687    
688        New working release 110.43.  New bootfiles.
689    
690    ----------------------------------------------------------------------
691    Name: Matthias Blume (blume (at) tti - c (dot) org)
692    Date: 2003/09/09 19:20:00 CDT
693    Tag: blume-20030909-installer
694    Description:
695    
696    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
697    Modified config/install.bat to take advantage of it.  Also modified
698    config/install.sh (and called it config/new-install.sh) to take advantage
699    of it on Unix systems.  (The SML code is (supposed to be) platform-
700    independent.)
701    
702    The installer can now install everything under Win32
703    as well as under *nix as long as it compiles.
704    
705    Other changes:
706    
707      - made CML compile again under Win32
708      - made eXene compile under Win32 (by providing a fake structure UnixSock
709        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
710      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
711        same as type int;  under Win32 it isn't
712      - fixed some slice-related problems in the win32-specific parts of CML
713      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
714        (under Win32, the volume name is case-insensitive, and the
715        OS.Path code compares volume names for equality)
716    
717    ----------------------------------------------------------------------
718    Name: Matthias Blume (blume (at) tti - c (dot) org)
719    Date: 2003/09/08 11:55:00 CDT
720    Tag: blume-20030908-fullpath
721    Description:
722    
723    Made Win32 version of OS.FileSys.fullPath return current directory
724    when given an empty string.  This is what the spec says, and incidentally,
725    CM depends on it.  (CM otherwise goes into an infinite loop in certain
726    cases when presented with the name of a non-existing .cm file.)
727    
728    ----------------------------------------------------------------------
729    Name: Matthias Blume (blume (at) tti - c (dot) org)
730    Date: 2003/09/04 16:30:00 CDT
731    Tag: blume-20030905-slices-etc
732    Description:
733    
734      1. Changed interface to vectors and arrays in Basis to match
735         (draft) Basis spec.
736      2. Added signatures and implementations of slices according to
737         Basis spec.
738      3. Edited source code throughout the system to make it compile again
739         under 1. and 2.  (In some cases code had to be added to have it
740         match the new signatures.)
741      4. MLRISC should be backward-compatible: the copies of the originals
742         of files that needed to change under 3. were retained, the .cm files
743         check the compiler version number and use old versions when
744         appropriate.
745      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
746         dirstream -> string option (in accordance with Basis spec).
747      6. When generating code that counts lines, ml-lex used function
748         CharVector.foldli, taking advantage of its old interface.
749         This has been replaced with the corresponding code from
750         CharVectorSlice.  (html-lex must be re-lexed!)
751      7. BitArray in smlnj-lib/Util has been extended/modified to match the
752         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
753      8. Removed temporary additions (fromInternal, toInternal) from the
754         (now obsolete) IntInf in smlnj-lib/Util.
755      9. Cleaned up structure Byte.
756     10. Added localOffset, scan, and fromString to Date (according to spec).
757         Cleaned/corrected implementation of Date.
758         (Still need to check for correctness; implement better canonicalizeDate.)
759     11. Added "scan" to signature IEEE_REAL.
760     12. Some improvements to IntInf [in particular: efficiency-hack for
761         mod and rem when second operand is 2 (for parity checks).]
762     13. Changed representation of type Time.time, using a single IntInf.int
763         value counting microseconds.  This considerably simplified the
764         implementation of structure Time.  We now support negative time
765         values; scan and fromString handle signs.
766     14. Functor PrimIO now takes two additional arguments (VectorSlice and
767         ArraySlice).
768    
769    ----------------------------------------------------------------------
770    Name: Matthias Blume (blume (at) tti - c (dot) org)
771    Date: 2003/08/28 17:00:00 CDT
772    Tag: blume-20030828-intinf
773    Description:
774    
775    This is a major update which comes with a version number bump
776    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
777    and an implementation of IntInf in the Basis.
778    
779    There are a fairly large number of related changes and updates throughout
780    the system:
781    
782    Basis:
783            - Implemented IntInf.
784            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
785            - Added some missing Real64 operations, most notably Real.toLargeInt.
786            - Added FixedInt as a synonym for Int32.
787    
788    compiler:
789            * Added support for a built-in intinf type.
790              - literals
791              - pattern matching
792              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
793              - overloading on literals and operations
794    
795            This required adding a primitive type intinf, some additional
796            primops, and implementations for several non-trivial intinf
797            operations in Core.  (The intinf type is completely abstract
798            to the compiler; all operations get delegated back to the Core.)
799    
800            * Intinf equality is handled by polyequal.  However, the compiler
801            does not print its usual warning in this case (since polyequal
802            is the right thing to do there).
803    
804            * Improved the organization of structure InlineT.
805    
806            * A word about conversion primops:
807            If conversions involving intinf do not cancel out during
808            CPS contract, then the compiler must insert calls to Core functions.
809            Since all core access must be resolved already during the FLINT
810            translate phase, it would be too late a the time of CPS contract
811            to add new Core calls.  For this reason, conversion primops
812            for intinf carry two arguments:  1. the numeric argument that
813            they are supposed to convert, and 2. the Core function that
814            can help with this conversion if necessary.  If CPS contract
815            eliminates a primop, then the associated Core function becomes
816            dead and goes away.  Intinf conversion primops that do not get
817            eliminated by CPS contract get rewritten into calls of their
818            core functions by a separate, new phase.
819    
820    interactive system:
821            - Control.Print.intinfDepth controls max length of intinf constants
822            being printed. (Analogous to Control.Print.stringDepth.)
823            - Cleanup in printutil and pputil: got rid of unused stuff and
824            duplicates; replaced some of the code with code that makes better
825            use of library functionality.
826    
827    CM:
828            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
829            are no longer silent.
830    
831    CKIT:
832            Fixed mismatched uses of Int32 and LargeInt.  I always decided
833            in favor of LargeInt -- which is now the same as IntInf.
834            CKIT-knowledgable people should check whether this is what's
835            intended and otherwise change things back to using Int32 or
836            FixedInt.
837    
838    Throughout the code:
839            Started using IntInf.int literals and built-in operations
840            (e.g., comparison with 0) where this seems appropriate.
841    
842    
843    ----------------------------------------------------------------------
844    Name: Dave MacQueen (dbm@cs.uchicago.edu)
845    Date: 2003/08/13 11:36:00 CDT
846    Tag: dbm-20030813-mcz-merge1
847    Description:
848    
849    Merging changes from the mcz-branch development branch into trunk.
850    These changes involve replacement of the emulated old prettyprinter
851    interface with direct use of the SML/NJ Lib PP library, and fixing
852    of a couple of bugs (895, 1186) relating to error messages.  A new
853    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
854    has been added.
855    
856    ----------------------------------------------------------------------
857    Name: Matthias Blume (blume (at) tti - c (dot) org)
858    Date: 2003/08/11 15:45:00 CDT
859    Tag: blume-20030811-windows
860    Description:
861    
862    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
863    
864        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
865    
866    This patch restores SML/NJ's ability to run under win32.  There are a
867    number of changes, including fixes for several bugs that had gone
868    unnoticed until now:
869    
870      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
871        trying to run the win32 version from within cygwin.)
872    
873      - There are a number of simple .bat scripts that substitute for their
874        corresponding Unix shell-scripts.  (See below.)
875    
876      - The internals of ml-build have been modified slightly.  The main
877        difference is that instead of calling ".link-sml" (or link-sml.bat)
878        using OS.Process.system, the ML process delegates this task back
879        to the script.  Otherwise problems arise in mixed environments such
880        as Cygwin where scripts look and work like Unix scripts, but
881        where OS.Process.system cannot run them.
882    
883      - In CM, the srcpath pickler used native pathname syntax -- which
884        is incorrect in the case of cross-compilation.  The new pickle format
885        is independent of platform-specific naming conventions.
886    
887      - Path configuration files (such as lib/pathconfig) can now choose
888        between native and standard syntax.  Placing a line of the form
889    
890            standard!
891    
892        into the file causes all subsequent paths to be interpreted using
893        CM standard pathname syntax (= Unix conventions); a line
894    
895            native!
896    
897        switches back to native style.  This was needed so that
898        path config files can be written portably, see src/system/pathconfig.
899    
900      - Runtime system:
901    
902          - win32-filesys.c:  get_file_time and set_file_time now
903            access modification time, not creation time.
904    
905          - I/O code made aware of new array representation.
906    
907          - Bug fixes in X86.prim.masm.
908    
909       - src/system/makeml made aware of win32. (For use under cygwin
910            and other Unix-environments for windows.)
911    
912       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
913         which caused BinIO.inputAll to fail consistently.
914    
915    .bat scripts:
916    
917       Windows .bat scripts assume that SMLNJ_HOME is defined.
918    
919       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
920         applications (sml, ml-yacc, ml-lex).
921       - ml-build.bat: analogous to ml-build.
922       - config\install.bat:  Analogous to config/install.sh.  This requires
923         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
924         (nmake etc. must be on the path, and vcvars32 must have been run.)
925         Moreover, sources for ml-lex and ml-yacc need to exist under src,
926         and the bootfile hierarchy must have been unpacked under
927         sml.boot.x86-win32.
928         The script is very primitive and does a poor job at error checking.
929         It only installs the base system, ml-lex, and ml-yacc.  No other
930         libraries are being installed (i.e., you get only those that
931         are part of the compiler.)
932       - link-sml.bat:  analogous to .link-sml, but not currently used
933    
934    Unrelated bug fixes:
935    
936      - ml-nlffigen now exports structures ST_* corresponding to incomplete
937        types.
938      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
939        otherwise.)
940    
941    ----------------------------------------------------------------------
942    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
943    Date: 2003/06/17
944    Tag: macqueen-20030617-bug895
945    Description:
946    
947    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
948    Tag will be used for new development branch (mcz-branch) for
949    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
950    
951    ----------------------------------------------------------------------
952    Name: Matthias Blume (blume (at) tti - c (dot) org)
953    Date: 2003/05/27 16:55:00 CDT
954    Tag: blume-20030527-polyeq
955    Description:
956    
957    Tried to eliminated most cases of polymorphic equality.
958    
959    ----------------------------------------------------------------------
960    Name: Matthias Blume (blume (at) tti - c (dot) org)
961    Date: 2003/05/21 17:45:00 CDT
962    Tag: blume-20030517-complete
963    Description:
964    
965    Two changes:
966    
967      1. Added a flag for controlling whether non-exhaustive bindings will
968         be treated as errors (default is false).
969      2. Cleaned up the *entire* source tree so that CMB.make goes through
970         without a single non-exhaustive match- or bind warning.
971    
972    ----------------------------------------------------------------------
973    Name: Matthias Blume (blume (at) tti - c (dot) org)
974    Date: 2003/05/17 10:20:00 CDT
975    Tag: blume-20030517-absyn
976    Description:
977    
978    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
979    
980       This mainly affects the quality of error messages.  However, some
981       of the code is now more straightforward than before. (Treatment of
982       the above four constructs in translate.sml is much simpler than
983       the "macro-expansion" that was going on before.  Plus, the mach-
984       compiler no longer gets invoked just to be able to compile an
985       if-expression.)
986    
987    2. The ErrorMsg.Error exception is now caught and absorbed by the
988       interactive loop.
989    
990    ----------------------------------------------------------------------
991    Name: Allen Leung
992    Date: 2003/05/16 13:05:00 CDT
993    Tag: leunga-20030516-cygwin-runtime
994    Description:
995    
996        Ported the runtime system to cygwin, which uses the unix
997    x86-unix bin files.  Missing/buggy features:
998    
999       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1000         the Cygwin library.
1001       o Ctrl-C handling may be flaky.
1002       o Windows system calls and Windows I/O are not supported.
1003    
1004       A new set of binfiles is located at:
1005    
1006         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1007    
1008       This is only needed for bootstrapping the cygwin version of smlnj.
1009       Other x86 versions can use the existing binfiles.
1010    
1011    ----------------------------------------------------------------------
1012    Name: Matthias Blume
1013    Date: 2003/04/08 15:42:00 CDT
1014    Tag: blume-20030408-listpair
1015    Description:
1016    
1017    1. Added a target 'mlrisc' to installer.
1018    
1019    2. Added missing elements to structure ListPair.
1020    
1021    ----------------------------------------------------------------------
1022    Name: Allen Leung
1023    Date: 2003/01/07 10:40:00 EST
1024    Tag: leunga-20030107-int-rem
1025    Description:
1026    
1027    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1028    The arguments to the SUBL instruction were swapped.
1029    
1030    ----------------------------------------------------------------------
1031    Name: Matthias Blume
1032    Date: 2002/12/12 16:25:00 EST
1033    Tag: blume-20021212-risc-ra
1034    Description:
1035    
1036    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1037    sent the RA into an infinite loop when floating point registers get
1038    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1039    110.37 and 110.38.)
1040    There was another set of potential problems related to the handling of
1041    MLRISC annotations (but those did not yet cause real problems, apparently).
1042    
1043    ----------------------------------------------------------------------
1044    Name: Matthias Blume
1045    Date: 2002/12/06 22:40:00 EST
1046    Tag: blume-20021206-cm-fileid
1047    Description:
1048    
1049    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1050    This fixes the problem of CM getting confused by files that suddenly
1051    change their identity (e.g., by getting unlinked and recreated by some
1052    text editor such as vi).  There might be a better/cheaper/cleaner way
1053    of doing this, but for now this will have to do.
1054    
1055    ----------------------------------------------------------------------
1056    Name: Matthias Blume
1057    Date: 2002/10/28 09:50:00 EST
1058    Tag: blume-20021028-typecheck
1059    Description:
1060    
1061    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1062    
1063    ----------------------------------------------------------------------
1064    Name: Matthias Blume
1065    Date: 2002/10/17 09:10:00 EDT
1066    Tag: Release_110_42
1067    Description:
1068    
1069    In good old tradition, there has been a slight hiccup so that we have
1070    to patch 110.42 after the fact.   The old release tag has been replaced
1071    (see below).
1072    
1073    The change solves a problem with two competing approaches the
1074    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1075    each other's way.
1076    
1077    This change only affects the runtime system code and the installer script.
1078    (No new bootfiles.)
1079    
1080    ----------------------------------------------------------------------
1081    Name: Matthias Blume
1082    Date: 2002/10/16 12:00:00 EDT
1083    Tag: Release_110_42_removed
1084    Description:
1085    
1086    New working release.  New bootfiles.
1087    
1088    ----------------------------------------------------------------------
1089    Name: Matthias Blume
1090    Date: 2002/10/10 13:10:00 EDT
1091    Tag: blume-20021010-ppc-divs
1092    Description:
1093    
1094    The mltree operator DIVS must be implemented with an overflow check on
1095    the PPC because the hardware indicates divide-by-zero using "overflow" as
1096    well.
1097    
1098    ----------------------------------------------------------------------
1099    Name: Matthias Blume
1100    Date: 2002/07/23 11:45:00
1101    Tag: blume-20020723-smlnj-home
1102    Description:
1103    
1104    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1105    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1106    CM_PATHCONFIG is also set), the path configuration file is assumed
1107    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1108    the entire tree to some other place and everything will "just work".
1109    
1110    (Companion commands such as ml-build and ml-makedepend also sense this
1111    variable.)
1112    
1113    ----------------------------------------------------------------------
1114    Name: Matthias Blume
1115    Date: 2002/07/12 21:19:00 EDT
1116    Tag: blume-20020712-liveness
1117    Description:
1118    
1119    Exported two useful "step" functions from liveness module (MLRISC).
1120    
1121    ----------------------------------------------------------------------
1122    Name: Matthias Blume
1123    Date: 2002/07/05 16:00 EDT
1124    Tag: Release_110_41
1125    Description:
1126    
1127    New working release.  New bootfiles.
1128    
1129    ----------------------------------------------------------------------
1130    Name: Matthias Blume
1131    Date: 2002/07/05 10:25:00 EDT
1132    Tag: blume-20020705-btimp
1133    Description:
1134    
1135    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1136    other clients can set up backtracing support.
1137    
1138    ----------------------------------------------------------------------
1139    Name: Matthias Blume
1140    Date: 2002/06/25 17:23:00 EDT
1141    Tag: blume-20020625-fpmax
1142    Description:
1143    
1144    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1145    case.  (The sense of the isNaN test was reversed -- which made min and
1146    max always return their first argument.)
1147    
1148    ----------------------------------------------------------------------
1149    Name: Matthias Blume
1150    Date: 2002/06/11
1151    Tag: blume-20020611-unixpath
1152    Description:
1153    
1154    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1155    
1156    ----------------------------------------------------------------------
1157    Name: Matthias Blume
1158    Date: 2002/06/10 16:35:00 EDT
1159    Tag: blume-20020610-ieeereal
1160    Description:
1161    
1162    I back-ported my implementation of IEEEReal.fromString from the
1163    idlbasis-devel branch so that we can test it.
1164    
1165    Another small change is that ppDec tries to give more information
1166    than just "<sig>" in the case of functors.  However, this code is
1167    broken in some mysterious way if the functor's body's signature
1168    has not been declared by ascription but gets inferred from the
1169    implementation.  This needs fixing...
1170    
1171    ----------------------------------------------------------------------
1172    Name: Matthias Blume
1173    Date: 2002/05/31
1174    Tag: blume-20020531-btrace-mode
1175    Description:
1176    
1177    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1178    the wayside when I switched over to using Controls everywhere.)
1179    
1180    ----------------------------------------------------------------------
1181    Name: Lal George
1182    Date: 2002/05/23 12:21:40 EDT
1183    Tag: george-20020523-visual-labels
1184    Description:
1185    
1186    Labels are now displayed in the graphical output to make
1187    the fall-through and target blocks obvious.
1188    
1189    ----------------------------------------------------------------------
1190    Name: Matthias Blume
1191    Date: 2002/05/22 11:03:00 EDT
1192    Tag: blume-20020522-shrink
1193    Description:
1194    
1195    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1196    situation that comes up when allocating huge arrays.
1197    
1198    ----------------------------------------------------------------------
1199    Name: Matthias Blume
1200    Date: 2002/05/21 16:00:00 EDT
1201    Tag: Release_110_40
1202    Description:
1203    
1204    New working release (110.40).  New bootfiles.
1205    
1206    [Also: John Reppy fixed GC bug 1131.]
1207    
1208    ----------------------------------------------------------------------
1209    Name: Matthias Blume
1210    Date: 2002/05/21 12:35:00 EDT
1211    Tag: blume-20020521-cmdoc
1212    Description:
1213    
1214    CM documentation update.
1215    
1216    ----------------------------------------------------------------------
1217    Name: Matthias Blume
1218    Date: 2002/05/21 10:55:00 EDT
1219    Tag: blume-20020521-misc
1220    Description:
1221    
1222    - John tweaked runtime to be silent on heap export (except when
1223      GC messages are on).
1224    - I added a few more things (cross-compiling versions of CMB) to
1225      config/preloads (as suggestions).
1226    
1227    ----------------------------------------------------------------------
1228    Name: Matthias Blume
1229    Date: 2002/05/20 22:25:00 EDT
1230    Tag: blume-20020520-controls
1231    Description:
1232    
1233    - Added ControlUtil structure to control-lib.cm.
1234    - Use it throughout.
1235    - Used Controls facility to define MLRISC controls (as opposed to
1236      registering MLRISC control ref cells with Controls after the fact)
1237    - Fixed messed-up controls priorities.
1238    
1239    * Removed again all the stuff from config/preloads that one wouldn't
1240      be able to preload at the time the initial heap image is built.
1241      (Many libraries, e.g., CML, do not exist yet at this time.  The
1242       only libraries that can be preloaded via config/preloads are those
1243       that come bundled with the bootfiles.)
1244    
1245    ----------------------------------------------------------------------
1246    Name: Matthias Blume
1247    Date: 2002/05/20 10:59:00 EDT
1248    Tag: blume-20020520-preloads
1249    Description:
1250    
1251    Added a lot of commented-out suggestions for things to be included
1252    in config/preloads.
1253    
1254    ----------------------------------------------------------------------
1255    Name: Allen Leung
1256    Date: 2002/05/18 14:20:00 EDT
1257    Tag: leunga-20020518-mdl
1258    Description:
1259    
1260    o Made the mdl tool stuff compile and run again.
1261    o I've disabled all the stuff that depends on RTL specifications; they
1262      are all badly broken anyway.
1263    
1264    ----------------------------------------------------------------------
1265    Name: Matthias Blume
1266    Date: 2002/05/17 16:49:00 EDT
1267    Tag: blume-20020517-controls
1268    Description:
1269    
1270    0. John Reppy made several modifications to the SML/NJ library.
1271       In particular, there is a shiny new controls-lib.cm.
1272    
1273    1. Pushed new controls interface through compiler so that everything
1274       compiles again.
1275    
1276    2. Added FormatComb and FORMAT_COMB to the CML version of the
1277       SML/NJ library (so that CML compiles again).
1278    
1279    3. Modified init scripts because XXX_DEFAULT environment variables
1280       are no longer with us.  (Boot-time initialization is now done
1281       using the same environment variables that are also used for
1282       startup-time initialization of controls.)
1283    
1284    ----------------------------------------------------------------------
1285    Name: Lal George
1286    Date: 2002/05/15 09:20:10 EDT
1287    Tag: george-20020515-pseudo-op-decls
1288    Description:
1289    
1290            All pseudo-ops emitted before the first segment declaration
1291    such as TEXT, DATA, and BSS directives are assumed to be global
1292    declarations and are emitted first in the assembly file. This is
1293    useful in a number of situations where one has pseudo-ops that are not
1294    specific to any segment, and also works around the constraint that one
1295    cannot have client pseudo-ops in the TEXT segment.
1296    
1297            Because no segment is associated with these declarations it is
1298    an error to allocate any space or objects before the first segment
1299    directive and an exception will be raised. However, we cannot make
1300    this check for client pseudo-ops.
1301    
1302            These top level declarations are a field in the CFG graph_info.
1303    In theory you can continue to add to this field after the CFG has been
1304    built -- provided you know what you are doing;-)
1305    
1306    ----------------------------------------------------------------------
1307    Name: Matthias Blume
1308    Date: 2002/05/13 16:40:00 EDT
1309    Tag: blume-20020513-pp-etc
1310    Description:
1311    
1312    A few minor bugfixes:
1313    
1314      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1315        (Bogus printouts for redefined bindings still occur.  Compiler
1316        bug should no longer occur now.  We need to redo the prettyprinter
1317        from scratch.)
1318    
1319      - CM pathname printer now also adds escape sequences for ( and )
1320    
1321      - commend and docu fixes for ml-nlffi
1322    
1323    ----------------------------------------------------------------------
1324    Name: Matthias Blume
1325    Date: 2002/05/10 16:40:00 EDT
1326    Tag: blume-20020510-erg-textio
1327    Description:
1328    
1329    Applied the following bugfix provided by Emden Gansner:
1330    
1331        Output is corrupted when outputSubstr is used rather than output.
1332    
1333        The problem occurs when a substring
1334    
1335            ss = (s, dataStart, dataLen)
1336    
1337        where dataStart > 0, fills a stream buffer with avail bytes left.
1338        avail bytes of s, starting at index dataStart, are copied into the
1339        buffer, the buffer is flushed, and then the remaining dataLen-avail
1340        bytes of ss are copied into the beginning of the buffer. Instead of
1341        starting this copy at index dataStart+avail in s, the current code
1342        starts the copy at index avail.
1343    
1344        Fix:
1345        In text-io-fn.sml, change line 695 from
1346             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1347        to
1348             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1349    
1350    ----------------------------------------------------------------------
1351    Name: Matthias Blume
1352    Date: 2002/04/12 13:55:00 EDT
1353    Tag: blume-20020412-assyntax
1354    Description:
1355    
1356    1. Grabbed newer assyntax.h from the XFree86 project.
1357    2. Fiddled with how to compile X86.prim.asm without warnings.
1358    3. (Very) Minor cleanup in CM.
1359    
1360    ----------------------------------------------------------------------
1361    Name: Matthias Blume
1362    Date: 2002/04/01 (no joke!) 17:07:00 EST
1363    Tag: blume-20020401-x86div
1364    Description:
1365    
1366    Added full support for div/mod/rem/quot on the x86, using the machine
1367    instruction's two results (without clumsily recomputing the remainder)
1368    directly where appropriate.
1369    
1370    Some more extensive power-of-two support was added to the x86 instruction
1371    selector (avoiding expensive divs, mods, and muls where they can be
1372    replaced with cheaper shifts and masks).  However, this sort of thing
1373    ought to be done earlier, e.g., within the CPS optimizer so that
1374    all architectures benefit from it.
1375    
1376    The compiler compiles to a fixed point, but changes might be somewhat
1377    fragile nevertheless.  Please, report any strange things that you might
1378    see wrt. div/mod/quot/rem...
1379    
1380    ----------------------------------------------------------------------
1381    Name: Matthias Blume
1382    Date: 2002/03/29 17:22:00
1383    Tag: blume-20020329-div
1384    Description:
1385    
1386    Fixed my broken div/mod logic.  Unfortunately, this means that the
1387    inline code for div/mod now has one more comparison than before.
1388    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
1389    The problem was with quotient = 0, because that alone does not tell
1390    us which way the rounding went.  One then has to look at whether
1391    remainder and divisor have the same sign...  :(
1392    
1393    Anyway, I replaced the bootfiles with fresh ones...
1394    
1395    ----------------------------------------------------------------------
1396    Name: Matthias Blume
1397    Date: 2002/03/29 14:10:00 EST
1398    Tag: blume-20020329-inlprims
1399    Description:
1400    
1401    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
1402    
1403    Primops have changed. This means that the bin/boot-file formats have
1404    changed as well.
1405    
1406    To make sure that there is no confusion, I made a new version.
1407    
1408    
1409    CHANGES:
1410    
1411    * removed REMT from mltree (remainder should never overflow).
1412    
1413    * added primops to deal with divisions of all flavors to the frontend
1414    
1415    * handled these primops all the way through so they map to their respective
1416      MLRISC support
1417    
1418    * used these primops in the implementation of Int, Int32, Word, Word32
1419    
1420    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
1421    
1422    * parameterized INLMIN, INLMAX, and INLABS by a numkind
1423    
1424    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
1425      floating point
1426    
1427    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
1428      and Real (but Real.abs maps to a separate floating-point-only primop)
1429    
1430    
1431    TODO items:
1432    
1433    * Hacked Alpha32 instruction selection, disabling the selection of REMx
1434      instructions because the machine instruction encoder cannot handle
1435      them.  (Hppa, PPC, and Sparc instruction selection did not handle
1436      REM in the first place, and REM is supported by the x86 machine coder.)
1437    
1438    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
1439      selection phase.  (The two can be streamlined because the hardware
1440      delivers both quotient and remainder at the same time anyway.)
1441    
1442    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
1443      (Currently the behavior is inconsistent both across architectures and
1444      wrt. the draft Basis spec.)
1445    
1446    * Word8 should eventually be handled natively, too.
1447    
1448    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
1449      as if there currently is no execution path that could trigger it in
1450      SML/NJ.  (The assumptions underlying functions arith and promotable do not
1451      hold for things like multiplication and division.)
1452    
1453    ----------------------------------------------------------------------
1454    Name: Matthias Blume
1455    Date: 2002/03/27 16:27:00 EST
1456    Tag: blume-20020327-mlrisc-divisions
1457    Description:
1458    
1459    Added support for all four division operations (ML's div, mod, quot,
1460    and rem) to MLRISC.  In the course of doing so, I also rationalized
1461    the naming (no more annoying switch-around of DIV and QUOT), by
1462    parameterizing the operation by div_rounding_mode (which can be either
1463    DIV_TO_ZERO or DIV_TO_NEGINF).
1464    
1465    The generic MLTreeGen functor takes care of compiling all four
1466    operations down to only round-to-zero div.
1467    
1468    Missing pieces:
1469    
1470      * Doing something smarter than relying on MLTreeGen on architectures
1471        like, e.g., the x86 where hardware division delivers both quotient and
1472        remainder at the same time.  With this, the implementation of the
1473        round-to-neginf operations could be further streamlined.
1474    
1475      * Remove inlining support for div/mod/rem from the frontend and replace it
1476        with primops that get carried through to the backend.  Do this for all
1477        int and word types.
1478    
1479    ----------------------------------------------------------------------
1480    Name: Matthias Blume
1481    Date: 2002/03/25 17:25:00 EST
1482    Tag: blume-20020325-divmod
1483    Description:
1484    
1485    I improved (hopefully without breaking them) the implementation of Int.div,
1486    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
1487    advantage of the following observations:
1488    
1489      Let  q = x quot y      r = x rem y
1490           d = x div  y      m = x mod y
1491    
1492    where "quot" is the round-to-zero version of integer division that
1493    hardware usually provides.  Then we have:
1494    
1495         r = x - q * y        where neither the * nor the - will overflow
1496         d = if q >= 0 orelse x = q * y then q else q - 1
1497                              where neither the * nor the - will overflow
1498         m = if q >= 0 orelse r = 0 then r else r + y
1499                              where the + will not overflow
1500    
1501    This results in substantial simplification of the generated code.
1502    The following table shows the number of CFG nodes and edges generated
1503    for
1504            fun f (x, y) = x OPER y
1505            (* with OPER \in div, mod, quot, rem *)
1506    
1507    
1508        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
1509        --------------------------------------------------------
1510         div |         24 |         39 |         12 |         16
1511         mod |         41 |         71 |         12 |         16
1512        quot |          8 |         10 |          8 |         10
1513         rem |         10 |         14 |          8 |         10
1514    
1515    
1516    ----------------------------------------------------------------------
1517    Name: Matthias Blume
1518    Date: 2002/03/25 22:06:00 EST
1519    Tag: blume-20020325-cprotobug
1520    Description:
1521    
1522    Fixed a bug in cproto (c prototype decoder).
1523    
1524    ----------------------------------------------------------------------
1525    Name: Matthias Blume
1526    Date: 2002/03/25 16:00:00 EST
1527    Tag: blume-20020325-raw-primops
1528    Description:
1529    
1530    I did some cleanup to Allen's new primop code and
1531    replaced yesterday's bootfiles with new ones.
1532    (But they are stored in the same place.)
1533    
1534    ----------------------------------------------------------------------
1535    Name: Matthias Blume
1536    Date: 2002/03/24 22:40:00 EST
1537    Tag: blume-20020324-bootfiles
1538    Description:
1539    
1540    Made the bootfiles that Allen asked for.
1541    
1542    ----------------------------------------------------------------------
1543    Name: Allen Leung
1544    Date: 2002/03/23 15:50:00 EST
1545    Tag: leunga-20020323-flint-cps-rcc-primops
1546    Description:
1547    
1548      1. Changes to FLINT primops:
1549    
1550        (* make a call to a C-function;
1551         * The primop carries C function prototype information and specifies
1552         * which of its (ML-) arguments are floating point. C prototype
1553         * information is for use by the backend, ML information is for
1554         * use by the CPS converter. *)
1555      | RAW_CCALL of { c_proto: CTypes.c_proto,
1556                       ml_args: ccall_type list,
1557                       ml_res_opt: ccall_type option,
1558                       reentrant : bool
1559                     } option
1560       (* Allocate uninitialized storage on the heap.
1561        * The record is meant to hold short-lived C objects, i.e., they
1562        * are not ML pointers.  With the tag, the representation is
1563        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1564        *)
1565      | RAW_RECORD of {tag:bool,sz:int}
1566      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1567    
1568      2.  These CPS primops are now overloaded:
1569    
1570           rawload of {kind:numkind}
1571           rawstore of {kind:numkind}
1572    
1573          The one argument form is:
1574    
1575             rawload {kind} address
1576    
1577          The two argument form is:
1578    
1579             rawload {kind} [ml object, byte-offset]
1580    
1581      3. RAW_CCALL/RCC now takes two extra arguments:
1582    
1583         a. The first is whether the C call is reentrant, i.e., whether
1584            ML state should be saved and restored.
1585         b. The second argument is a string argument specifying the name of
1586            library and the C function.
1587    
1588         These things are currently not handled in the code generator, yet.
1589    
1590      4. In CProto,
1591    
1592         An encoding type of "bool" means "ml object" and is mapped into
1593         C prototype of PTR.  Note that "bool" is different than "string",
1594         even though "string" is also mapped into PTR, because "bool"
1595         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
1596    
1597      5. Pickler/unpicker
1598    
1599         Changed to handle RAW_RECORD and newest RAW_CCALL
1600    
1601      6. MLRiscGen,
1602    
1603         1. Changed to handle the new rawload/rawstore/rawrecord operators.
1604         2. Code for handling C Calls has been moved to a new module CPSCCalls,
1605            in the file CodeGen/cpscompile/cps-c-calls.sml
1606    
1607      7. Added the conditional move operator
1608    
1609             condmove of branch
1610    
1611         to cps.  Generation of this is still buggy so it is currently
1612         disabled.
1613    
1614    ----------------------------------------------------------------------
1615    Name: Lal George
1616    Date: 2002/03/22 14:18:25 EST
1617    Tag: george-20020322-cps-branch-prob
1618    Description:
1619    
1620    Implemented the Ball-Larus branch prediction-heuristics, and
1621    incorporated graphical viewers for control flow graphs.
1622    
1623    Ball-Larus Heuristics:
1624    ---------------------
1625    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
1626    
1627    By design it uses the Dempster-Shafer theory for combining
1628    probabilities.  For example, in the function:
1629    
1630        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
1631    
1632    the ball-larus heuristics predicts that the n=0 is unlikely
1633    (OH-heuristic), and the 'then' branch is unlikely because of the
1634    RH-heuristic -- giving the 'then' branch an even lower combined
1635    probability using the Dempster-Shafer theory.
1636    
1637    Finally, John Reppy's loop analysis in MLRISC, further lowers the
1638    probability of the 'then' branch because of the loop in the else
1639    branch.
1640    
1641    
1642    Graphical Viewing:
1643    ------------------
1644    I merely plugged in Allen's graphical viewers into the compiler. The
1645    additional code is not much. At the top level, saying:
1646    
1647            Control.MLRISC.getFlag "cfg-graphical-view" := true;
1648    
1649    will display the graphical view of the control flow graph just before
1650    back-patching.  daVinci must be in your path for this to work. If
1651    daVinci is not available, then the default viewer can be changed
1652    using:
1653    
1654            Control.MLRISC.getString "viewer"
1655    
1656    which can be set to "dot" or "vcg" for the corresponding viewers. Of
1657    course, these viewers must be in your path.
1658    
1659    The above will display the compilation unit at the level of clusters,
1660    many of which are small, boring, and un-interesting. Also setting:
1661    
1662            Control.MLRISC.getInt "cfg-graphical-view_size"
1663    
1664    will display clusters that are larger than the value set by the above.
1665    
1666    
1667    ----------------------------------------------------------------------
1668    Name: Matthias Blume
1669    Date: 2002/03/21 22:20:00 EST
1670    Tag: blume-20020321-kmp-bugfix
1671    Description:
1672    
1673    Changed the interface to the KMP routine in PreString and fixed
1674    a minor bug in one place where it was used.
1675    
1676    ----------------------------------------------------------------------
1677    Name: Allen Leung
1678    Date: 2002/03/21 20:30:00 EST
1679    Tag: leunga-20020321-cfg
1680    Description:
1681    
1682      Fixed a potential problem in cfg edge splitting.
1683    
1684    ----------------------------------------------------------------------
1685    Name: Allen Leung
1686    Date: 2002/03/21 17:15:00 EST
1687    Tag: leunga-20020321-x86-fp-cfg
1688    Description:
1689    
1690      1. Recoded the buggy parts of x86-fp.
1691    
1692         a. All the block reordering code has been removed.
1693            We now depend on the block placement phases to do this work.
1694    
1695         b. Critical edge splitting code has been simplified and moved into the
1696            CFG modules, as where they belong.
1697    
1698         Both of these were quite buggy and complex.  The code is now much, much
1699         simpler.
1700    
1701      2. X86 backend.
1702    
1703         a. Added instructions for 64-bit support.  Instruction selection for
1704            64-bit has not been committed, however, since that
1705            requires changes to MLTREE which haven't been approved by
1706            Lal and John.
1707    
1708         b. Added support for FUCOMI and FUCOMIP when generating code for
1709            PentiumPro and above.  We only generate these instructions in
1710            the fast-fp mode.
1711    
1712         c. Added cases for JP and JNP in X86FreqProps.
1713    
1714      3. CFG
1715    
1716         CFG now has a bunch of methods for edge splitting and merging.
1717    
1718      4. Machine description.
1719    
1720         John's simplification of MLTREE_BASIS.fcond broke a few machine
1721         description things:
1722    
1723         rtl-build.{sig,sml} and hppa.mdl fixed.
1724    
1725         NOTE: the machine description stuff in the repository is still broken.
1726               Again, I can't put my fixes in because that involves
1727               changes to MLTREE.
1728    
1729    ----------------------------------------------------------------------
1730    Name: Matthias Blume
1731    Date: 2002/03/20 15:55:00 EST
1732    Tag: blume-20020320-kmp
1733    Description:
1734    
1735    Implemented Knuth-Morris-Pratt string matching in PreString and used
1736    it for String.isSubstring, Substring.isSubstring, and
1737    Substring.position.
1738    
1739    (Might need some stress-testing.  Simple examples worked fine.)
1740    
1741    ----------------------------------------------------------------------
1742    Name: Matthias Blume
1743    Date: 2002/03/19 16:37:00 EST
1744    Tag: blume-20020319-witnesses
1745    Description:
1746    
1747    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
1748    
1749    This implements a generic mechanism for changing constness qualifiers
1750    anywhere within big C types without resorting to outright "casts".
1751    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
1752    modify the constness at the outermost level.)
1753    The implementation of "convert" is based on the idea of "witness"
1754    values -- values that are not used by the operation but whose types
1755    "testify" to their applicability.  On the implementation side, "convert"
1756    is simply a projection (returning its second curried argument).  With
1757    cross-module inlining, it should not result in any machine code being
1758    generated.
1759    
1760    ----------------------------------------------------------------------
1761    Name: Matthias Blume
1762    Date: 2002/03/15 16:40:00 EST
1763    Tag: blume-20020315-basis
1764    Description:
1765    
1766    Provided (preliminary?) implementations for
1767    
1768      {String,Substring}.{concatWith,isSuffix,isSubstring}
1769    
1770    and
1771    
1772      Substring.full
1773    
1774    Those are in the Basis spec but they were missing in SML/NJ.
1775    
1776    ----------------------------------------------------------------------
1777    Name: Matthias Blume
1778    Date: 2002/03/14 21:30:00 EST
1779    Tag: blume-20020314-controls
1780    Description:
1781    
1782    Controls:
1783    ---------
1784    
1785    1. Factored out the recently-added Controls : CONTROLS stuff and put
1786       it into its own library $/controls-lib.cm.  The source tree for
1787       this is under src/smlnj-lib/Controls.
1788    
1789    2. Changed the names of types and functions in this interface, so they
1790       make a bit more "sense":
1791    
1792          module -> registry
1793          'a registry -> 'a group
1794    
1795    3. The interface now deals in ref cells only.  The getter/setter interface
1796       is (mostly) gone.
1797    
1798    4. Added a function that lets one register an already-existing ref cell.
1799    
1800    5. Made the corresponding modifications to the rest of the code so that
1801       everything compiles again.
1802    
1803    6. Changed the implementation of Controls.MLRISC back to something closer
1804       to the original.  In particular, this module (and therefore MLRISC)
1805       does not depend on Controls.  There now is some link-time code in
1806       int-sys.sml that registers the MLRISC controls with the Controls
1807       module.
1808    
1809    CM:
1810    ---
1811    
1812      * One can now specify the lambda-split aggressiveness in init.cmi.
1813    
1814    ----------------------------------------------------------------------
1815    Name: Allen Leung
1816    Date: 2002/03/13 17:30:00 EST
1817    Tag: leunga-20020313-x86-fp-unary
1818    Description:
1819    
1820    Bug fix for:
1821    
1822    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
1823    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
1824    > - fun f(x,(y,z)) = Real.~ y;
1825    > [autoloading]
1826    > [autoloading done]
1827    >       fchsl   (%eax), 184(%esp)
1828    > Error: MLRisc bug: X86MCEmitter.emitInstr
1829    >
1830    > uncaught exception Error
1831    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
1832    
1833    The problem was that the code generator did not generate any fp registers
1834    in this case, and the ra didn't know that it needed to run the X86FP phase to
1835    translate the pseudo fp instruction.   This only happened with unary fp
1836    operators in certain situations.
1837    
1838    ----------------------------------------------------------------------
1839    Name: Matthias Blume
1840    Date: 2002/03/13 14:00:00 EST
1841    Tag: blume-20020313-overload-etc
1842    Description:
1843    
1844    1. Added _overload as a synonym for overload for backward compatibility.
1845       (Control.overloadKW must be true for either version to be accepted.)
1846    
1847    2. Fixed bug in install script that caused more things to be installed
1848       than what was requested in config/targets.
1849    
1850    3. Made CM aware of the (_)overload construct so that autoloading
1851       works.
1852    
1853    ----------------------------------------------------------------------
1854    Name: Matthias Blume
1855    Date: 2002/03/12 22:03:00 EST
1856    Tag: blume-20020312-url
1857    Description:
1858    
1859    Forgot to update BOOT and srcarchiveurl.
1860    
1861    ----------------------------------------------------------------------
1862    Name: Matthias Blume
1863    Date: 2002/03/12 17:30:00 EST
1864    Tag: blume-20020312-version110392
1865    Description:
1866    
1867    Yet another version number bump (because of small changes to the
1868    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
1869    
1870    Changes:
1871    
1872      The new pid generation scheme described a few weeks ago was overly
1873      complicated.  I implemented a new mechanism that is simpler and
1874      provides a bit more "stability":  Once CM has seen a compilation
1875      unit, it keeps its identity constant (as long as you do not delete
1876      those crucial CM/GUID/* files).  This means that when you change
1877      an interface, compile, then go back to the old interface, and
1878      compile again, you arrive at the original pid.
1879    
1880      There now also is a mechanism that instructs CM to use the plain
1881      environment hash as a module's pid (effectively making its GUID
1882      the empty string).  For this, "noguid" must be specified as an
1883      option to the .sml file in question within its .cm file.
1884      This is most useful for code that is being generated by tools such
1885      as ml-nlffigen (because during development programmers tend to
1886      erase the tool's entire output directory tree including CM's cached
1887      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
1888      revert to the old, broken behavior of SML/NJ, but in specific cases
1889      where there is no danger of interface confusion, its use is ok
1890      (I think).
1891    
1892      ml-nlffigen by default generates "noguid" annotations.  They can be
1893      turned off by specifying -guid in its command line.
1894    
1895    ----------------------------------------------------------------------
1896    Name: Lal George
1897    Date: 2002/03/12 12 14:42:36 EST
1898    Tag: george-20020312-frequency-computation
1899    Description:
1900    
1901    Integrated jump chaining and static block frequency into the
1902    compiler. More details and numbers later.
1903    
1904    ----------------------------------------------------------------------
1905    Name: Lal George
1906    Date: 2002/03/11 11 22:38:53 EST
1907    Tag: george-20020311-jump-chain-elim
1908    Description:
1909    
1910    Tested the jump chain elimination on all architectures (except the
1911    hppa).  This is on by default right now and is profitable for the
1912    alpha and x86, however, it may not be profitable for the sparc and ppc
1913    when compiling the compiler.
1914    
1915    The gc test will typically jump to a label at the end of the cluster,
1916    where there is another jump to an external cluster containing the actual
1917    code to invoke gc. This is to allow factoring of common gc invocation
1918    sequences. That is to say, we generate:
1919    
1920            f:
1921               testgc
1922               ja   L1      % jump if above to L1
1923    
1924            L1:
1925               jmp L2
1926    
1927    
1928    After jump chain elimination the 'ja L1' instructions is converted to
1929    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
1930    up being implemented in their long form (if L2 is far away) using:
1931    
1932            jbe     L3      % jump if below or equal to L3
1933            jmp     L2
1934         L3:
1935            ...
1936    
1937    
1938    For large compilation units L2  may be far away.
1939    
1940    
1941    ----------------------------------------------------------------------
1942    Name: Matthias Blume
1943    Date: 2002/03/11 13:30:00 EST
1944    Tag: blume-20020311-mltreeeval
1945    Description:
1946    
1947    A functor parameter was missing.
1948    
1949    ----------------------------------------------------------------------
1950    Name: Allen Leung
1951    Date: 2002/03/11 10:30:00 EST
1952    Tag: leunga-20020311-runtime-string0
1953    Description:
1954    
1955       The representation of the empty string now points to a
1956    legal null terminated C string instead of unit.  It is now possible
1957    to convert an ML string into C string with InlineT.CharVector.getData.
1958    This compiles into one single machine instruction.
1959    
1960    ----------------------------------------------------------------------
1961    Name: Allen Leung
1962    Date: 2002/03/10 23:55:00 EST
1963    Tag: leunga-20020310-x86-call
1964    Description:
1965    
1966       Added machine generation for CALL instruction (relative displacement mode)
1967    
1968    ----------------------------------------------------------------------
1969    Name: Matthias Blume
1970    Date: 2002/03/08 16:05:00
1971    Tag: blume-20020308-entrypoints
1972    Description:
1973    
1974    Version number bumped to 110.39.1.  NEW BOOTFILES!
1975    
1976    Entrypoints: non-zero offset into a code object where execution should begin.
1977    
1978    - Added the notion of an entrypoint to CodeObj.
1979    - Added reading/writing of entrypoint info to Binfile.
1980    - Made runtime system bootloader aware of entrypoints.
1981    - Use the address of the label of the first function given to mlriscGen
1982      as the entrypoint.  This address is currently always 0, but it will
1983      not be 0 once we turn on block placement.
1984    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
1985      with entry points) from mlriscGen.
1986    
1987    ----------------------------------------------------------------------
1988    Name: Allen Leung
1989    Date: 2002/03/07 20:45:00 EST
1990    Tag: leunga-20020307-x86-cmov
1991    Description:
1992    
1993       Bug fixes for CMOVcc on x86.
1994    
1995       1. Added machine code generation for CMOVcc
1996       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
1997       3. CMOVcc cannot have an immediate operand as argument.
1998    
1999    ----------------------------------------------------------------------
2000    Name: Matthias Blume
2001    Date: 2002/03/07 16:15:00 EST
2002    Tag: blume-20020307-controls
2003    Description:
2004    
2005    This is a very large but mostly boring patch which makes (almost)
2006    every tuneable compiler knob (i.e., pretty much everything under
2007    Control.* plus a few other things) configurable via both the command
2008    line and environment variables in the style CM did its configuration
2009    until now.
2010    
2011    Try starting sml with '-h' (or, if you are brave, '-H')
2012    
2013    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2014    implements the underlying generic mechanism.
2015    
2016    The interface to some of the existing such facilities has changed somewhat.
2017    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2018    (The getFoo interface is still there for backward-compatibility, but its
2019    use is deprecated.)
2020    
2021    The ml-build script passes -Cxxx=yyy command-line arguments through so
2022    that one can now twiddle the compiler settings when using this "batch"
2023    compiler.
2024    
2025    TODO items:
2026    
2027    We should go through and throw out all controls that are no longer
2028    connected to anything.  Moreover, we should go through and provide
2029    meaningful (and correct!) documentation strings for those controls
2030    that still are connected.
2031    
2032    Currently, multiple calls to Controls.new are accepted (only the first
2033    has any effect).  Eventually we should make sure that every control
2034    is being made (via Controls.new) exactly once.  Future access can then
2035    be done using Controls.acc.
2036    
2037    Finally, it would probably be a good idea to use the getter-setter
2038    interface to controls rather than ref cells.  For the time being, both
2039    styles are provided by the Controls module, but getter-setter pairs are
2040    better if thread-safety is of any concern because they can be wrapped.
2041    
2042    *****************************************
2043    
2044    One bug fix: The function blockPlacement in three of the MLRISC
2045    backpatch files used to be hard-wired to one of two possibilities at
2046    link time (according to the value of the placementFlag).  But (I
2047    think) it should rather sense the flag every time.
2048    
2049    *****************************************
2050    
2051    Other assorted changes (by other people who did not supply a HISTORY entry):
2052    
2053    1. the cross-module inliner now works much better (Monnier)
2054    2. representation of weights, frequencies, and probabilities in MLRISC
2055       changed in preparation of using those for weighted block placement
2056       (Reppy, George)
2057    
2058    ----------------------------------------------------------------------
2059    Name: Lal George
2060    Date: 2002/03/07 14:44:24 EST 2002
2061    Tag: george-20020307-weighted-block-placement
2062    
2063    Tested the weighted block placement optimization on all architectures
2064    (except the hppa) using AMPL to generate the block and edge frequencies.
2065    Changes were required in the machine properties to correctly
2066    categorize trap instructions. There is an MLRISC flag
2067    "weighted-block-placement" that can be used to enable weighted block
2068    placement, but this will be ineffective without block/edge
2069    frequencies (coming soon).
2070    
2071    
2072    ----------------------------------------------------------------------
2073    Name: Lal George
2074    Date: 2002/03/05 17:24:48 EST
2075    Tag: george-20020305-linkage-cluster
2076    
2077    In order to support the block placement optimization, a new cluster
2078    is generated as the very first cluster (called the linkage cluster).
2079    It contains a single jump to the 'real' entry point for the compilation
2080    unit. Block placement has no effect on the linkage cluster itself, but
2081    all the other clusters  have full freedom in the manner in which they
2082    reorder blocks or functions.
2083    
2084    On the x86 the typical linkage code that is generated is:
2085       ----------------------
2086            .align 2
2087       L0:
2088            addl    $L1-L0, 72(%esp)
2089            jmp     L1
2090    
2091    
2092            .align  2
2093       L1:
2094       ----------------------
2095    
2096    72(%esp) is the memory location for the stdlink register. This
2097    must contain the address of the CPS function being called. In the
2098    above example, it contains the address of  L0; before
2099    calling L1 (the real entry point for the compilation unit), it
2100    must contain the address for L1, and hence
2101    
2102            addl $L1-L0, 72(%esp)
2103    
2104    I have tested this on all architectures except the hppa.The increase
2105    in code size is of course negligible
2106    
2107    ----------------------------------------------------------------------
2108    Name: Allen Leung
2109    Date: 2002/03/03 13:20:00 EST
2110    Tag: leunga-20020303-mlrisc-tools
2111    
2112      Added #[ ... ] expressions to mlrisc tools
2113    
2114    ----------------------------------------------------------------------
2115    Name: Matthias Blume
2116    Date: 2002/02/27 12:29:00 EST
2117    Tag: blume-20020227-cdebug
2118    Description:
2119    
2120    - made types in structure C and C_Debug to be equal
2121    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2122    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2123    
2124    ----------------------------------------------------------------------
2125    Name: Matthias Blume
2126    Date: 2002/02/26 12:00:00 EST
2127    Tag: blume-20020226-ffi
2128    Description:
2129    
2130    1. Fixed a minor bug in CM's "noweb" tool:
2131       If numbering is turned off, then truly don't number (i.e., do not
2132       supply the -L option to noweb).  The previous behavior was to supply
2133       -L'' -- which caused noweb to use the "default" line numbering scheme.
2134       Thanks to Chris Richards for pointing this out (and supplying the fix).
2135    
2136    2. Once again, I reworked some aspects of the FFI:
2137    
2138       A. The incomplete/complete type business:
2139    
2140       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2141         gone!
2142       - ML types representing an incomplete type are now *equal* to
2143         ML types representing their corresponding complete types (just like
2144         in C).  This is still safe because ml-nlffigen will not generate
2145         RTTI for incomplete types, nor will it generate functions that
2146         require access to such RTTI.   But when ML code generated from both
2147         incomplete and complete versions of the C type meet, the ML types
2148         are trivially interoperable.
2149    
2150         NOTE:  These changes restore the full generality of the translation
2151         (which was previously lost when I eliminated functorization)!
2152    
2153       B. Enum types:
2154    
2155       - Structure C now has a type constructor "enum" that is similar to
2156         how the "su" constructor works.  However, "enum" is not a phantom
2157         type because each "T enum" has values (and is isomorphic to
2158         MLRep.Signed.int).
2159       - There are generic access operations for enum objects (using
2160         MLRep.Signed.int).
2161       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2162         * The structure contains the definition of type "mlrep" (the ML-side
2163         representation type of the enum).  Normally, mlrep is the same
2164         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2165         then mlrep will be defined as a datatype -- thus facilitating
2166         pattern matching on mlrep values.
2167         ("-ec" will be suppressed if there are duplicate values in an
2168          enumeration.)
2169         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2170         will be generated for each C enum constant xxx.
2171         * Conversion functions m2i and i2m convert between mlrep and
2172         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2173         * Coversion functions c and ml convert between mlrep and "tag enum".
2174         * Access functions (get/set) fetch and store mlrep values.
2175       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2176         enumerations are merged into one single enumeration represented by
2177         structure E_'.
2178    
2179    ----------------------------------------------------------------------
2180    Name: Allen Leung
2181    Date: 2002/02/25 04:45:00 EST
2182    Tag: leunga-20020225-cps-spill
2183    
2184    This is a new implementation of the CPS spill phase.
2185    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2186    In case of problems, replace it with the old file spill.sml
2187    
2188    The current compiler runs into some serious performance problems when
2189    constructing a large record.  This can happen when we try to compile a
2190    structure with many items.  Even a very simple structure like the following
2191    makes the compiler slow down.
2192    
2193        structure Foo = struct
2194           val x_1 = 0w1 : Word32.int
2195           val x_2 = 0w2 : Word32.int
2196           val x_3 = 0w3 : Word32.int
2197           ...
2198           val x_N = 0wN : Word32.int
2199        end
2200    
2201    The following table shows the compile time, from N=1000 to N=4000,
2202    with the old compiler:
2203    
2204    N
2205    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2206           MLRISC ra                               0.06u  0.00s  0.05g
2207              (spills = 0 reloads = 0)
2208           TOTAL                                   0.63u  0.07s  0.21g
2209    
2210    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2211           MLRISC ra                               5.68u  0.59s  3.93g
2212              (spills = 0 reloads = 0)
2213           TOTAL                                   14.71u  0.99s  4.81g
2214    
2215    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2216           MLRISC ra                               5.54u  0.65s  3.91g
2217              (spills = 543 reloads = 1082)
2218           TOTAL                                   65.40u  3.13s  6.00g
2219    
2220    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2221           MLRISC ra                               0.80u  0.10s  0.55g
2222              (spills = 42 reloads = 84)
2223           TOTAL                                   129.42u  5.10s  4.13g
2224    
2225    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2226           MLRISC ra                               2.69u  0.27s  1.38g
2227              (spills = 62 reloads = 124)
2228           TOTAL                                   682.48u  19.61s  13.99g
2229    
2230    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2231           MLRISC ra                               4.96u  0.27s  2.72g
2232              (spills = 85 reloads = 170)
2233           TOTAL                                   2375.26u  57.21s  48.00g
2234    
2235    As you can see the old cps spill module suffers from some serious
2236    performance problem.  But since I cannot decipher the old code fully,
2237    instead of patching the problems up, I'm reimplementing it
2238    with a different algorithm.  The new code is more modular,
2239    smaller when compiled, and substantially faster
2240    (O(n log n) time and O(n) space).  Timing of the new spill module:
2241    
2242    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2243          MLRISC ra                               0.25u  0.02s  0.15g
2244             (spills=1 reloads=3)
2245          TOTAL                                   7.74u  0.34s  1.62g
2246    
2247    Implementation details:
2248    
2249    As far as I can tell, the purpose of the CPS spill module is to make sure the
2250    number of live variables at any program point (the bandwidth)
2251    does not exceed a certain limit, which is determined by the
2252    size of the spill area.
2253    
2254    When the bandwidth is too large, we decrease the register pressure by
2255    packing live variables into spill records.  How we achieve this is
2256    completely different than what we did in the old code.
2257    
2258    First, there is something about the MLRiscGen code generator
2259    that we should be aware of:
2260    
2261    o MLRiscGen performs code motion!
2262    
2263       In particular, it will move floating point computations and
2264       address computations involving only the heap pointer to
2265       their use sites (if there is only a single use).
2266       What this means is that if we have a CPS record construction
2267       statement
2268    
2269           RECORD(k,vl,w,e)
2270    
2271       we should never count the new record address w as live if w
2272       has only one use (which is often the case).
2273    
2274       We should do something similar to floating point, but the transformation
2275       there is much more complex, so I won't deal with that.
2276    
2277    Secondly, there are now two new cps primops at our disposal:
2278    
2279     1. rawrecord of record_kind option
2280        This pure operator allocates some uninitialized storage from the heap.
2281        There are two forms:
2282    
2283         rawrecord NONE [INT n]  allocates a tagless record of length n
2284         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2285                                     and initializes the tag.
2286    
2287     2. rawupdate of cty
2288          rawupdate cty (v,i,x)
2289          Assigns to x to the ith component of record v.
2290          The storelist is not updated.
2291    
2292    We use these new primops for both spilling and increment record construction.
2293    
2294     1. Spilling.
2295    
2296        This is implemented with a linear scan algorithm (but generalized
2297        to trees).  The algorithm will create a single spill record at the
2298        beginning of the cps function and use rawupdate to spill to it,
2299        and SELECT or SELp to reload from it.  So both spills and reloads
2300        are fine-grain operations.  In contrast, in the old algorithm
2301        "spills" have to be bundled together in records.
2302    
2303        Ideally, we should sink the spill record construction to where
2304        it is needed.  We can even split the spill record into multiple ones
2305        at the places where they are needed.  But CPS is not a good
2306        representation for global code motion, so I'll keep it simple and
2307        am not attempting this.
2308    
2309     2. Incremental record construction (aka record splitting).
2310    
2311        Long records with many component values which are simulatenously live
2312        (recall that single use record addresses are not considered to
2313         be live) are constructed with rawrecord and rawupdate.
2314        We allocate space on the heap with rawrecord first, then gradually
2315        fill it in with rawupdate.  This is the technique suggested to me
2316        by Matthias.
2317    
2318        Some restrictions on when this is applicable:
2319        1. It is not a VECTOR record.  The code generator currently does not handle
2320           this case. VECTOR record uses double indirection like arrays.
2321        2. All the record component values are defined in the same "basic block"
2322           as the record constructor.  This is to prevent speculative
2323           record construction.
2324    
2325    ----------------------------------------------------------------------
2326    Name: Allen Leung
2327    Date: 2002/02/22 01:02:00 EST
2328    Tag: leunga-20020222-mlrisc-tools
2329    
2330    Minor bug fixes in the parser and rewriter
2331    
2332    ----------------------------------------------------------------------
2333    Name: Allen Leung
2334    Date: 2002/02/21 20:20:00 EST
2335    Tag: leunga-20020221-peephole
2336    
2337    Regenerated the peephole files.  Some contained typos in the specification
2338    and some didn't compile because of pretty printing bugs in the old version
2339    of 'nowhere'.
2340    
2341    ----------------------------------------------------------------------
2342    Name: Allen Leung
2343    Date: 2002/02/19 20:20:00 EST
2344    Tag: leunga-20020219-mlrisc-tools
2345    Description:
2346    
2347       Minor bug fixes to the mlrisc-tools library:
2348    
2349       1.  Fixed up parsing colon suffixed keywords
2350       2.  Added the ability to shut the error messages up
2351       3.  Reimplemented the pretty printer and fixed up/improved
2352           the pretty printing of handle and -> types.
2353       4.  Fixed up generation of literal symbols in the nowhere tool.
2354       5.  Added some SML keywords to to sml.sty
2355    
2356    ----------------------------------------------------------------------
2357    Name: Matthias Blume
2358    Date: 2002/02/19 16:20:00 EST
2359    Tag: blume-20020219-cmffi
2360    Description:
2361    
2362    A wild mix of changes, some minor, some major:
2363    
2364    * All C FFI-related libraries are now anchored under $c:
2365        $/c.cm      --> $c/c.cm
2366        $/c-int.cm  --> $c/internals/c-int.cm
2367        $/memory.cm --> $c/memory/memory.cm
2368    
2369    * "make" tool (in CM) now treats its argument pathname slightly
2370      differently:
2371        1. If the native expansion is an absolute name, then before invoking
2372           the "make" command on it, CM will apply OS.Path.mkRelative
2373           (with relativeTo = OS.FileSys.getDir()) to it.
2374        2. The argument will be passed through to subsequent phases of CM
2375           processing without "going native".  In particular, if the argument
2376           was an anchored path, then "make" will not lose track of that anchor.
2377    
2378    * Compiler backends now "know" their respective C calling conventions
2379      instead of having to be told about it by ml-nlffigen.  This relieves
2380      ml-nlffigen from one of its burdens.
2381    
2382    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
2383    
2384    * Export C_DEBUG and C_Debug from $c/c.cm.
2385    
2386    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
2387      subtyping relationship between incomplete pointers and their complete
2388      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
2389      with the convention of instantiating 'o with ('t, 'c) obj whenever
2390      the pointer target type is complete.  In the incomplete case, 'o
2391      will be instantiated with some "'c iobj" -- a type obtained by
2392      using one of the functors PointerToIncompleteType or PointerToCompleteType.
2393    
2394      Operations that work on both incomplete and complete pointer types are
2395      typed as taking an 'o ptr while operations that require the target to
2396      be known are typed as taking some ('t, 'c) obj ptr.
2397    
2398      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
2399      where void is an eqtype without any values.  This makes it possible
2400      to work on voidptr values using functions meant to operate on light
2401      incomplete pointers.
2402    
2403    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
2404      been vastly simplified.
2405    
2406    ----------------------------------------------------------------------
2407    Name: Matthias Blume
2408    Date: 2002/02/19 10:48:00 EST
2409    Tag: blume-20020219-pqfix
2410    Description:
2411    
2412    Applied Chris Okasaki's bug fix for priority queues.
2413    
2414    ----------------------------------------------------------------------
2415    Name: Matthias Blume
2416    Date: 2002/02/15 17:05:00
2417    Tag: Release_110_39
2418    Description:
2419    
2420    Last-minute retagging is becoming a tradition... :-(
2421    
2422    This is the working release 110.39.
2423    
2424    ----------------------------------------------------------------------
2425    Name: Matthias Blume
2426    Date: 2002/02/15 16:00:00 EST
2427    Tag: Release_110_39-orig
2428    Description:
2429    
2430    Working release 110.39.  New bootfiles.
2431    
2432    (Update: There was a small bug in the installer so it wouldn't work
2433    with all shells.  So I retagged. -Matthias)
2434    
2435    ----------------------------------------------------------------------
2436    Name: Matthias Blume
2437    Date: 2002/02/15 14:17:00 EST
2438    Tag: blume-20020215-showbindings
2439    Description:
2440    
2441    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
2442    the latter can be useful for exploring what bindings are available at
2443    the interactive prompt.  (The first function returns only the list
2444    of symbols that are really bound, the second prints those but also the
2445    ones that CM's autoloading mechanism knows about.)
2446    
2447    ----------------------------------------------------------------------
2448    Name: Matthias Blume
2449    Date: 2002/02/15 12:08:00 EST
2450    Tag: blume-20020215-iptrs
2451    Description:
2452    
2453    Two improvements to ml-nlffigen:
2454    
2455      1. Write files only if they do not exist or if their current contents
2456         do not coincide with what's being written.  (That is, avoid messing
2457         with the time stamps unless absolutely necessary.)
2458    
2459      2. Implement a "repository" mechanism for generated files related
2460         to "incomplete pointer types".   See the README file for details.
2461    
2462    ----------------------------------------------------------------------
2463    Name: Matthias Blume
2464    Date: 2002/02/14 11:50:00 EST
2465    Tag: blume-20020214-quote
2466    Description:
2467    
2468    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
2469    because of the new and improved tag generation scheme.  (Thanks to Allen
2470    Leung for pointing it out.)
2471    
2472    ----------------------------------------------------------------------
2473    Name: Lal George
2474    Date: 2002/02/14 09:55:27 EST 2002
2475    Tag: george-20020214-isabelle-bug
2476    Description:
2477    
2478    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
2479    of Isabelle on the x86.
2480    
2481    From Allen:
2482    -----------
2483     I've found the problem:
2484    
2485         in ra-core.sml, I use the counter "blocked" to keep track of the
2486         true number of elements in the freeze queue.  When the counter goes
2487         to zero, I skip examining the queue.  But I've messed up the
2488         bookkeeping in combine():
2489    
2490             else ();
2491             case !ucol of
2492               PSEUDO => (if !cntv > 0 then
2493                     (if !cntu > 0 then blocked := !blocked - 1 else ();
2494                                        ^^^^^^^^^^^^^^^^^^^^^^^
2495                      moveu := mergeMoveList(!movev, !moveu)
2496                     )
2497                  else ();
2498    
2499         combine() is called to coalesce two nodes u and v.
2500         I think I was thinking that if the move counts of u and v are both
2501         greater than zero then after they are coalesced then one node is
2502         removed from the freeze queue.  Apparently I was thinking that
2503         both u and v are of low degree, but that's clearly not necessarily true.
2504    
2505    
2506    02/12/2002:
2507        Here's the patch.  HOL now compiles.
2508    
2509        I don't know how this impact on performance (compile
2510        time or runtime).  This bug caused the RA (especially on the x86)
2511        to go thru the potential spill phase when there are still nodes on the
2512        freeze queue.
2513    
2514    
2515    
2516    
2517    ----------------------------------------------------------------------
2518    Name: Matthias Blume
2519    Date: 2002/02/13 22:40:00 EST
2520    Tag: blume-20020213-fptr-rtti
2521    Description:
2522    
2523    Fixed a bug in ml-nlffigen that was introduced with one of the previous
2524    updates.
2525    
2526    ----------------------------------------------------------------------
2527    Name: Matthias Blume
2528    Date: 2002/02/13 16:41:00 EST
2529    Tag: blume-20020213-cmlpq
2530    Description:
2531    
2532    Added new priority queue export symbols (which have just been added to
2533    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
2534    not compile and the installer would choke.)
2535    
2536    ----------------------------------------------------------------------
2537    Name: Matthias Blume
2538    Date: 2002/02/13 16:15:00 EST
2539    Tag: blume-20020213-various
2540    Description:
2541    
2542    1. More tweaks to ml-nlffigen:
2543    
2544       - better internal datastructures (resulting in slight speedup)
2545       - "-match" option requires exact match
2546       - "localized" gensym counters (untagged structs/unions nested within
2547         other structs/unions or within typedefs get a fresh counter; their
2548         tag will be prefixed by a concatenation of their parents' tags)
2549       - bug fixes (related to calculation of transitive closure of types
2550         to be included in the output)
2551    
2552    2. Minor Basis updates:
2553    
2554       - added implementations for List.collate and Option.app
2555    
2556    ----------------------------------------------------------------------
2557    Name: Matthias Blume
2558    Date: 2002/02/11 15:55:00 EST
2559    Tag: blume-20020211-gensym
2560    Description:
2561    
2562    Added a "-gensym" option to command line of ml-nlffigen.  This can be
2563    used to specify a "stem" -- a string that is inserted in all "gensym'd"
2564    names (ML structure names that correspond to unnamed C structs, unions,
2565    and enums), so that separate runs of ml-nlffigen do not clash.
2566    
2567    ----------------------------------------------------------------------
2568    Name: Matthias Blume
2569    Date: 2002/02/11 12:05:00 EST
2570    Tag: blume-20020211-gensml
2571    Description:
2572    
2573    A quick fix for a problem with GenSML (in the pgraph-util library):
2574    Make generation of toplevel "local" optional.  (Strictly speaking,
2575    signature definitions within "local" are not legal SML.)
2576    
2577    Other than that: updates to INSTALL and cm/TODO.
2578    
2579    ----------------------------------------------------------------------
2580    Name: Matthias Blume
2581    Date: 2002/02/08 15:00:00 EST
2582    Tag: blume-20020208-uniquepid
2583    Description:
2584    
2585    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
2586    
2587    1. The installer (config/install.sh) has gotten smarter:
2588    
2589         - Configuration options are a bit easier to specify now
2590           (in config/targets).
2591         - Bug in recognizing .tar.bz2 files fixed.
2592         - Installer automatically resolves dependencies between
2593           configuration options (e.g., if you ask for eXene, you will
2594           also get cml -- regardless whether you asked for it or not).
2595         - Installer can run in "quieter mode" by setting the environment
2596           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
2597           "completely silent", though.
2598         - Build HashCons library as part of smlnj-lib.
2599    
2600    2. A new scheme for assigning persistent identifiers to compilation
2601       units (and, by extension, to types etc.) has been put into place.
2602       This fixes a long-standing bug where types and even dynamic values
2603       can get internally confused, thereby compromising type safety
2604       (abstraction) and dynamic correctness.  See
2605    
2606         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
2607    
2608       for an example of how things could go wrong until now.
2609    
2610       The downside of the new scheme is that pids are not quite as
2611       persistent as they used to be: CM will generate a fresh pid
2612       for every compilation unit that it thinks it sees for the first
2613       time.  That means that if you compile starting from a clean, fresh
2614       source tree at two different times, you end up with different
2615       binaries.
2616    
2617       Cutoff recompilation, however, has not been compromised because
2618       CM keeps pid information in special caches between runs.
2619    
2620    ----------------------------------------------------------------------
2621    Name: Lal George
2622    Date: 2002/02/07 15:34:13 EST 2002
2623    Tag: <none>
2624    Description:
2625    
2626    Compilers that generate assembly code may produce  global labels
2627    whose value is resolved at link time. The various peephole optimization
2628    modules did not take this in account.
2629    
2630    TODO. The Labels.addrOf function should really return an option
2631    type so that clients are forced to deal with this issue, rather
2632    than an exception being raised.
2633    
2634    ----------------------------------------------------------------------
2635    Name: Lal George
2636    Date: 2002/02/06 13:55:02 EST
2637    Tag: george-20020206-ra-breakup
2638    Description:
2639    
2640    1. A bug fix from Allen.
2641    
2642        A typo causes extra fstp %st(0)'s to be generated at compensation
2643        edges, which might cause stack underflow traps at runtime.  This
2644        occurs in fft where there are extraneous fstps right before the 'into'
2645        trap instruction (in this case they are harmless since none of the
2646        integers overflow.)
2647    
2648    2. Pulled out various utility modules that were embedded in the modules
2649       of the register allocator. I need these modules for other purposes, but
2650       they are not complete enough to put into a library (just yet).
2651    ----------------------------------------------------------------------
2652    Name: Matthias Blume
2653    Date: 2002/01/31 16:05:00 EST
2654    Tag: blume-20020131-sparc-ccalls
2655    Description:
2656    
2657    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
2658       of extra stack space by mistake.  Fixed.
2659    
2660    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
2661    
2662    ----------------------------------------------------------------------
2663    Name: Allen Leung
2664    Date: 2002/01/30
2665    Tag: leunga-20020130-nowhere-bug-fix
2666    Description:
2667    
2668       MLRISC bug fixes:
2669       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
2670       2. MachineInt.fromString was negating its value.
2671    
2672    ----------------------------------------------------------------------
2673    Name: Matthias Blume
2674    Date: 2002/01/29
2675    Tag: blume-20020129-INSTALL
2676    Description:
2677    
2678    - Added somewhat detailed installation instructions (file INSTALL).
2679    - Fixed curl-detection bug in config/install.sh.
2680    - It is now possible to select the URL getter using the URLGETTER
2681      environment variable:
2682    
2683          not set / "unknown"      --> automatic detection (script tries wget,
2684                                       curl, and lynx)
2685          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
2686                                       how to properly invoke them)
2687          other                    --> use $URLGETTER directly, it must take
2688                                       precisely two command-line arguments
2689                                       (source URL and destination file name)
2690    
2691    ----------------------------------------------------------------------
2692    Name: Matthias Blume
2693    Date: 2002/01/28
2694    Tag: blume-20020128-sparc-ccalls
2695    Description:
2696    
2697    - Fixed problem with calculation of "used" registers in sparc-c-calls.
2698    - Make use of the allocParam argument in sparc-c-calls.
2699    
2700    ----------------------------------------------------------------------
2701    Name: Matthias Blume
2702    Date: 2002/01/28
2703    Tag: blume-20020128-allocParam
2704    Description:
2705    
2706    John Reppy:  Changes c-calls API to accept client-callback for
2707    allocating extra stack space.
2708    me: Corresponding changes to mlriscGen (using a dummy argument that
2709        does not change the current behavior).
2710    
2711    ----------------------------------------------------------------------
2712    Name: Matthias Blume
2713    Date: 2002/01/28 12:00:00
2714    Tag: Release_110_38
2715    Description:
2716    
2717    This time for real!!!
2718    
2719    ----------------------------------------------------------------------
2720    Name: Matthias Blume
2721    Date: 2002/01/28 10:56:00 EST
2722    Tag: blume-20020128-retraction
2723    Description:
2724    
2725    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
2726       with blume-Release_110_38-retracted.)
2727    
2728    1. Fixed a problem with incorrect rounding modes in real64.sml.
2729       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
2730    
2731    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
2732       and enums fixed.  The naming of corresponding ML identifiers should
2733       now be consistent again.
2734    
2735    ----------------------------------------------------------------------
2736    Name: Allen Leung
2737    Date: 2002/01/27
2738    Tag: leunga-20020127-nowhere
2739    Description:
2740    
2741       Added a target called nowhere in the configuration scripts.
2742       Enabling this will build the MLRISC 'nowhere' tool (for translating
2743       programs with where-clauses into legal SML code) during installation.
2744    
2745    ----------------------------------------------------------------------
2746    Name: Matthias Blume
2747    Date: 2002/01/25 21:27:00 EST
2748    Tag: blume-Release_110_38-retracted
2749    Description:
2750    
2751    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
2752    
2753    README will be added later.
2754    
2755    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
2756    (Release_110_38) removed.  Reason: Last-minute bug fixes.
2757    
2758    ----------------------------------------------------------------------
2759    Name: Matthias Blume
2760    Date: 2002/01/25
2761    Tag: blume-20020125-ffi
2762    Description:
2763    
2764    A large number of tweaks and improvements to ml-nlffi-lib and
2765    ml-nlffigen:
2766    
2767       - ML represenation types have been streamlined
2768       - getter and setter functions work with concrete values, not abstract
2769         ones where possible
2770       - ml-nlffigen command line more flexible (see README file there)
2771       - some bugs have been fixed (hopefully)
2772    
2773    ----------------------------------------------------------------------
2774    Name: Lal George
2775    Date: 2002/01/24
2776    Tag: george-20020124-risc-ra-interface
2777    Description:
2778    
2779       There is a dramatic simplification in the interface to the
2780       register allocator for RISC architectures as a result of making
2781       parallel copy instructions explicit.
2782    
2783    ----------------------------------------------------------------------
2784    Name: Matthias Blume
2785    Date: 2002/01/22
2786    Tag: blume-20020122-x86-ccalls
2787    Description:
2788    
2789    Bug fix for c-calls on x86 (having to do with how char- and
2790    short-arguments are being handled).
2791    
2792    ----------------------------------------------------------------------
2793    Name: Matthias Blume
2794    Date: 2002/01/21
2795    Tag: blume-20020121-ff
2796    Description:
2797    
2798    Another day of fiddling with the FFI...
2799    
2800    1. Bug fix/workaround:  CKIT does not complain about negative array
2801       dimensions, so ml-nlffigen has to guard itself against this possibility.
2802       (Otherwise a negative dimension would send it into an infinite loop.)
2803    
2804    2. Some of the abstract types (light objects, light pointers, most "base"
2805       types) in structure C are now eqtypes.
2806    
2807    3. Added constructors and test functions for NULL function pointers.
2808    
2809    ----------------------------------------------------------------------
2810    Name: Matthias Blume
2811    Date: 2002/01/18
2812    Tag: blume-20020118-ready-for-new-release
2813    Description:
2814    
2815    Made config/srcarchiveurl point to a new place.  (Will provide boot
2816    files shortly.)
2817    
2818    Maybe we christen this to be 110.38?
2819    
2820    ----------------------------------------------------------------------
2821    Name: Matthias Blume
2822    Date: 2002/01/18
2823    Tag: blume-20020118-more-ffifiddle
2824    Description:
2825    
2826    Today's FFI fiddling:
2827    
2828      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
2829        C.Set.* that use concrete (MLRep.*) arguments and results instead
2830        of abstract ones.
2831    
2832      - Provided word-style bit operations etc. for "int" representation
2833        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
2834        and Long.
2835    
2836    ----------------------------------------------------------------------
2837    Name: Matthias Blume
2838    Date: 2002/01/18
2839    Tag: blume-20020118-use-x86-fp
2840    Description:
2841    
2842    Now that x86-fast-fp seems to be working, I turned it back on again
2843    by default.  (Seems to work fine now, even with the FFI.)
2844    
2845    Other than that, I added some documentation about the FFI to
2846    src/ml-nlffigen/README and updated the FFI test examples in
2847    src/ml-nlffi-lib/Tests/*.
2848    
2849    ----------------------------------------------------------------------
2850    Name: Allen Leung
2851    Date: 2002/01/17
2852    Tag: leunga-20020117-x86-fast-fp-call
2853    Description:
2854    
2855       1. Fixed a problem with handling return fp values when x86's fast fp
2856          mode is turned on.
2857    
2858       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
2859          of %f32.
2860    
2861       3. Added a constructor INT32lit to the ast of MLRISC tools.
2862    
2863    ----------------------------------------------------------------------
2864    Name: Matthias Blume
2865    Date: 2002/01/16
2866    Tag: blume-20020116-ffifiddle
2867    Description:
2868    
2869    More fiddling with the FFI interface:
2870    
2871     - Make constness 'c instead of rw wherever possible.  This eliminates
2872       the need for certain explicit coercions.  (However, due to ML's
2873       value polymorphism, there will still be many cases where explicit
2874       coercions are necessary.  Phantom types are not the whole answer
2875       to modeling a subtyping relationship in ML.)
2876    
2877     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
2878    
2879     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
2880       clumsy workaround for varargs problem.)
2881    
2882    ----------------------------------------------------------------------
2883    Name: Lal George
2884    Date: 2002/01/15
2885    Tag: <none>
2886    Description:
2887    
2888    1. Since COPY instructions are no longer native to the architecture,
2889       a generic functor can be used to implement the expandCopies function.
2890    
2891    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
2892       TEXT segment.
2893    
2894    ----------------------------------------------------------------------
2895    Name: Matthias Blume
2896    Date: 2002/01/15
2897    Tag: blume-20020115-ffiupdates
2898    Description:
2899    
2900    1. Fix for bug resulting in single-precision float values being returned
2901       incorrectly from FFI calls.
2902    
2903    2. Small modifications to C FFI API:
2904    
2905        - memory-allocation routines return straight objects (no options)
2906          and raise an exception in out-of-memory situations
2907        - unsafe extensions to cast between function pointers and pointers
2908          from/to ints
2909        - added structure C_Debug as an alternative to structure C where
2910          pointer-dereferencing (|*| and |*!) always check for null-pointers
2911        - added open_lib' to DynLinkage;  open_lib' works like open_lib
2912          but also takes a (possibly empty) list of existing library handles
2913          that the current library depends on
2914    
2915    ----------------------------------------------------------------------
2916    Name: Matthias Blume
2917    Date: 2002/01/10
2918    Tag: blume-20020110-newffigen
2919    Description:
2920    
2921    1. Updates to portable graph code.
2922    
2923    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
2924       more scalable now so that even huge interfaces such as the one
2925       for GTK compile in finite time and space. :-)
2926       See src/ml-nlffigen/README for details on what's new.
2927    
2928    ----------------------------------------------------------------------
2929    Name: Lal George
2930    Date: 2001/01/09 14:31:35 EST 2002
2931    Tag: george-20011206-rm-native-copy
2932    Description:
2933    
2934            Removed the native COPY and FCOPY instructions
2935            from all the architectures and replaced it with the
2936            explicit COPY instruction from the previous commit.
2937    
2938            It is now possible to simplify many of the optimizations
2939            modules that manipulate copies. This has not been
2940            done in this change.
2941    
2942    ----------------------------------------------------------------------
2943    Name: Lal George
2944    Date: 2001/12/06 16:50:13 EST 2001
2945    Tag: george-20011206-mlrisc-instruction
2946    Description:
2947    
2948    Changed the representation of instructions from being fully abstract
2949    to being partially concrete. That is to say:
2950    
2951      from
2952            type instruction
2953    
2954      to
2955            type instr                              (* machine instruction *)
2956    
2957            datatype instruction =
2958                LIVE of {regs: C.cellset, spilled: C.cellset}
2959              | KILL of {regs: C.cellset, spilled: C.cellset}
2960              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
2961              | ANNOTATION of {i: instruction, a: Annotations.annotation}
2962              | INSTR of instr
2963    
2964    This makes the handling of certain special instructions that appear on
2965    all architectures easier and uniform.
2966    
2967    LIVE and KILL say that a list of registers are live or killed at the
2968    program point where they appear. No spill code is generated when an
2969    element of the 'regs' field is spilled, but the register is moved to
2970    the 'spilled' (which is present, more for debugging than anything else).
2971    
2972    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
2973    We used to generate:
2974    
2975            DEFFREG f1
2976            f1 := f2 + f3
2977            trapb
2978    
2979    but now generate:
2980    
2981            f1 := f2 + f3
2982            trapb
2983            LIVE {regs=[f1,f2,f3], spilled=[]}
2984    
2985    Furthermore, the DEFFREG (hack) required that all floating point instruction
2986    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
2987    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
2988    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
2989    this architecture).
2990    
2991    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
2992    all the architectures. This will result in further simplification of the
2993    register allocator that must be aware of them for coalescing purposes, and
2994    will also simplify certain aspects of the machine description that provides
2995    callbacks related to parallel copies.
2996    
2997    ANNOTATION should be obvious, and now INSTR represents the honest to God
2998    machine instruction set!
2999    
3000    The <arch>/instructions/<arch>Instr.sml files define certain utility
3001    functions for making porting easier -- essentially converting upper case
3002    to lower case. All machine instructions (of type instr) are in upper case,
3003    and the lower case form generates an MLRISC instruction. For example on
3004    the alpha we have:
3005    
3006      datatype instr =
3007         LDA of {r:cell, b:cell, d:operand}
3008       | ...
3009    
3010      val lda : {r:cell, b:cell, d:operand} -> instruction
3011        ...
3012    
3013    where lda is just (INSTR o LDA), etc.
3014    
3015    ----------------------------------------------------------------------
3016    Name: Matthias Blume
3017    Date: 2001/11/22 21:40:00 EST
3018    Tag: Release_110_37
3019    Description:
3020    
3021    Release 110.37.  This time for real.
3022    
3023    ----------------------------------------------------------------------
3024    Name: Matthias Blume
3025    Date: 2001/11/21 16:35:00 EST
3026    Tag: blume-20011121-foot-in-mouth
3027    Description:
3028    
3029    Removed the "Release_110_37" tag because of a serious bug.
3030    This will be re-tagged once the bug is fixed.
3031    
3032    ----------------------------------------------------------------------
3033    Name: Matthias Blume
3034    Date: 2001/11/21 16:14:00 EST
3035    Tag: blume-20011121-forgottenfile
3036    Description:
3037    
3038    Forgot to add a file.  (Just a .tex-file -- part of
3039    the CM manual source.)
3040    
3041    ----------------------------------------------------------------------
3042    Name: Matthias Blume
3043    Date: 2001/11/21 16:10:00 EST
3044    Tag: blume-20011121-invalid_110_37
3045    Description:
3046    
3047    Note: I removed the original tag "Release_110_37" from this commit
3048          because we found a serious bug in all non-x86 backends.
3049          - Matthias
3050    
3051    1. Modifications to the SML/NJ code generator and to the runtime system
3052       so that code object name strings are directly inserted into code
3053       objects at code generation time.  The only business the runtime system
3054       has with this is now to read the name strings on occasions.
3055       (The encoding of the name string has also changed somewhat.)
3056    
3057    2. CM now implements a simple "set calculus" for specifying export lists.
3058       In particular, it is now possible to refer to the export lists of
3059       other libraries/groups/sources and form unions as well as differences.
3060       See the latest CM manual for details.
3061    
3062    3. An separate notion of "proxy" libraries has again be eliminated from
3063       CM's model.  (Proxy libraries are now simply a special case of using
3064       the export list calculus.)
3065    
3066    4. Some of the existing libraries now take advantage of the new set
3067       calculus.
3068       (Notice that not all libraries have been converted because some
3069       of the existing .cm-files are supposed to be backward compatible
3070       with 110.0.x.)
3071    
3072    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3073       "General.exnMessage"!  The former relies on a certain hook to be
3074       initialized, and that often does not happen in the stand-alone case.)
3075    
3076    ----------------------------------------------------------------------
3077    Name: Lal George
3078    Date: 2001/11/21  13:56:18 EST
3079    Tag: george-2001121-pseudo-ops
3080    Description:
3081    
3082      Implemented a complete redesign of MLRISC pseudo-ops. Now there
3083      ought to never be any question of incompatabilities with
3084      pseudo-op syntax expected by host assemblers.
3085    
3086      For now, only modules supporting GAS syntax are implemented
3087      but more should follow, such as MASM, and vendor assembler
3088      syntax, e.g. IBM as, Sun as, etc.
3089    
3090    ----------------------------------------------------------------------
3091    Name: Matthias Blume
3092    Date: 2001/11/14 11:52:00 EST
3093    Tag: blume-20011114-srcname
3094    Description:
3095    
3096    1. Routed the name of the current source file to mlriscgen where it
3097       should be directly emitted into the code object.  (This last part
3098       is yet to be done.)
3099    
3100    2. Some cleanup of the pgraph code to make it match the proposal that
3101       I put out the other day.  (The proposal notwithstanding, things are
3102       still in flux here.)
3103    
3104    ----------------------------------------------------------------------
3105    Name: Lal George
3106    Date: 2001/11/14 09:44:04 EST
3107    Tag:
3108    Description:
3109    
3110      Fix for a backpatching bug reported by Allen.
3111    
3112      Because the boundary between short and long span-dependent
3113      instructions is +/- 128, there are an astounding number of
3114      span-dependent instructions whose size is over estimated.
3115    
3116      Allen came up with the idea of letting the size of span
3117      dependent instructions be non-monotonic, for a maxIter
3118      number of times, after which the size must be monotonically
3119      increasing.
3120    
3121      This table shows the number of span-dependent instructions
3122      whose size was over-estimated as a function of maxIter, for the
3123      file Parse/parse/ml.grm.sml:
3124    
3125         maxIter            # of instructions:
3126            10                      687
3127            20                      438
3128            30                      198
3129            40                        0
3130    
3131      In compiling the compiler, there is no significant difference in
3132      compilation speed between maxIter=10 and maxIter=40. Actually,
3133      my measurements showed that maxIter=40 was a tad faster than
3134      maxIter=10! Also 96% of the  files in the compiler reach a fix
3135      point within 13 iterations, so fixing maxIter at 40, while high,
3136      is okay.
3137    
3138    ----------------------------------------------------------------------
3139    Name: Matthias Blume
3140    Date: 2001/10/31 15:25:00 EST
3141    Tag: blume-20011031-pgraph
3142    Description:
3143    
3144    CKIT:
3145    * Changed the "Function" constructor of type Ast.ctype to carry optional
3146      argument identifiers.
3147    * Changed the return type of TypeUtil.getFunction accordingly.
3148    * Type equality ignores the argument names.
3149    * TypeUtil.composite tries to preserve argument names but gives up quickly
3150      if there is a mismatch.
3151    
3152    installation script:
3153    * attempts to use "curl" if available (unless "wget" is available as well)
3154    
3155    CM:
3156    * has an experimental implementation of "portable graphs" which I will
3157      soon propose as an implementation-independent library format
3158    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
3159    
3160    NLFFI-LIB:
3161    * some cleanup (all cosmetic)
3162    
3163    NLFFIGEN:
3164    * temporarily disabled the mechanism that suppresses ML output for
3165      C definitions whose identifiers start with an underscore character
3166    * generate val bindings for enum constants
3167    * user can request that only one style (light or heavy) is being used;
3168      default is to use both (command-line arguments: -heavy and -light)
3169    * fixed bug in handling of function types involving incomplete pointers
3170    * generate ML entry points that take record arguments (i.e., using
3171      named arguments) for C functions that have a prototype with named
3172      arguments
3173      (see changes to CKIT)
3174    
3175    ----------------------------------------------------------------------
3176    Name: Allen Leung
3177    Date: 2001/10/27 20:34:00 EDT
3178    Tag: leunga-20011027-x86-fast-fp-call
3179    Description:
3180    
3181       Fixed the bug described in blume-20010920-slowfp.
3182    
3183       The fix involves
3184          1. generating FCOPYs in FSTP in ia32-svid
3185          2. marking a CALL with the appropriate annotation
3186    
3187    ----------------------------------------------------------------------
3188    Name: Matthias Blume
3189    Date: 2001/10/16 11:32:00 EDT
3190    Tag: blume-20011016-netbsd
3191    Description:
3192    
3193    Underscore patch from Chris Richards (fixing problem with compiling
3194    runtime system under recent NetBSD).
3195    
3196    ----------------------------------------------------------------------
3197    Name: Allen Leung
3198    Date: 2001/10/12 17:18:32 EDT 2001
3199    Tag: leung-20011012-x86-printflowgraph
3200    Description:
3201    
3202    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
3203    
3204    ----------------------------------------------------------------------
3205    Name: Lal George
3206    Date: 2001/10/11 23:51:34 EDT
3207    Tag: george-20011011-too-many-instrs
3208    Description:
3209    
3210    The representation of a program point never expected to see more
3211    than 65536 instructions in a basic block!
3212    
3213    ----------------------------------------------------------------------
3214    Name: Lal George
3215    Date: 2001/10/09 09:41:37 EDT
3216    Tag: george-20011008-mlrisc-labels
3217    Description:
3218    
3219    Changed the machine description files to support printing of
3220    local and global labels in assembly code, based on host assembler
3221    conventions.
3222    
3223    ----------------------------------------------------------------------
3224    Name: Matthias Blume
3225    Date: 2001/09/25 15:25:00 EDT
3226    Tag: blume-20010925-exninfo
3227    Description:
3228    
3229    I provided a non-hook implementation of exnName (at the toplevel) and
3230    made the "dummy" implementation of exnMessage (at the toplevel) more
3231    useful: if nothing gets "hooked in", then at least you are going to
3232    see the exception name and a message indicating why you don't see more.
3233    
3234    [For the time being, programs that need exnMessage and want to use
3235    ml-build should either use General.exnMessage (strongly recommended) or
3236    refer to structure General at some other point so that CM sees a
3237    static dependency.]
3238    
3239    [Similar remarks go for "print" and "use":  If you want to use their
3240    functionality in stand-alone programs generated by ml-build, then use
3241    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
3242    
3243    ----------------------------------------------------------------------
3244    Name: Matthias Blume
3245    Date: 2001/09/20 17:28:00 EDT
3246    Tag: blume-20010920-slowfp
3247    Description:
3248    
3249    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
3250    
3251    ----------------------------------------------------------------------
3252    Name: Matthias Blume
3253    Date: 2001/09/20 17:20:00 EDT
3254    Tag: blume-20010920-canonicalpaths
3255    Description:
3256    
3257    0. Updated the BOOT file (something that I forgot to do earlier).
3258    
3259    1. Small internal change to CM so that it avoids "/../" in filenames
3260       as much as possible (but only where it is safe).
3261    
3262    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
3263       that arguments that contain delimiters are passed through correctly.
3264       This change also means that all "special" arguments of the form
3265       @SMLxxx... must come first.
3266    
3267    3. Changed install script to put relative anchor names for tool commands
3268       into pathconfig.
3269    
3270    ----------------------------------------------------------------------
3271    Name: Matthias Blume
3272    Date: 2001/09/18 15:35:00 EDT
3273    Tag: blume-20010918-readme11036
3274    Description:
3275    
3276    Added README files.
3277    
3278    ----------------------------------------------------------------------
3279    Name: Matthias Blume
3280    Date: 2001/09/18 11:45:00 EDT
3281    Tag: Release_110_36 (retag)
3282    Description:
3283    
3284    Fixed mistake in config/preloads. Retagged as 110.36.
3285    
3286    ----------------------------------------------------------------------
3287    Name: Matthias Blume
3288    Date: 2001/09/18 09:40:00 EDT
3289    Tag: Release_110_36_orig (tag changed)
3290    Description:
3291    
3292    New version (110.36).  New bootfiles.
3293    
3294    ----------------------------------------------------------------------
3295    Name: Matthias Blume
3296    Date: 2001/09/14 16:15:00 EDT
3297    Tag: blume-20010914-x86fastfp
3298    Description:
3299    
3300    John committed some changes that Allen made, in particular a (hopefully)
3301    correctly working version of the x86-fp module.
3302    
3303    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
3304    flag to "true".  Everything seems to compile to a fixpoint ok, and
3305    "mandelbrot" speeds up by about 15%.
3306    
3307    ----------------------------------------------------------------------
3308    Name: Matthias Blume
3309    Date: 2001/09/13 11:20:00 EDT
3310    Tag: blume-20010913-minimal
3311    Description:
3312    
3313    1. Stefan Monnier's patch to fix a miscompilation problem that
3314       was brought to light by John Reppy's work on Moby.
3315    
3316    2. Implemented a minimal "structure Compiler" that contains just
3317       "version" and "architecture".  The minimal version will be
3318       available when the full version is not.  This is for backward-
3319       compatibility with code that wants to test Compiler.version.
3320    
3321    ----------------------------------------------------------------------
3322    Name: Matthias Blume
3323    Date: 2001/08/28 14:03:00 EDT
3324    Tag: blume-20010828-ml-lex
3325    Description:
3326    
3327    Fix for bug 1581, received from Neophytos Michael.
3328    
3329    ----------------------------------------------------------------------
3330    Name: Matthias Blume
3331    Date: 2001/08/27 11:20:00 EDT
3332    Tag: blume-20010827-readme11035
3333    Description:
3334    
3335    Fleshed out the README file for 110.35.
3336    
3337    ----------------------------------------------------------------------
3338    Name: Matthias Blume
3339    Date: 2001/08/24 17:10:00 EDT
3340    Tag: Release_110_35
3341    Description:
3342    
3343    New version number (110.35).  New bootfiles.
3344    
3345    ----------------------------------------------------------------------
3346    Name: Lal George
3347    Date: 2001/08/24 13:47:18 EDT 2001
3348    Tag: george-20010824-MLRISC-graphs
3349    Description:
3350    
3351     removed clusters from MLRISC completely and replaced with graphs.
3352    
3353    ----------------------------------------------------------------------
3354    Name: Matthias Blume
3355    Date: 2001/08/23 17:50:00 EDT
3356    Tag: blume-20010823-toplevel
3357    Description:
3358    
3359    - some reorganization of the code that implements various kinds of
3360      environments in the compiler (static, dynamic, symbolic, combined)
3361    - re-implemented the EnvRef module so that evalStream works properly
3362      (if the stream contains references to "use", "CM.make", etc.)
3363    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
3364    
3365    ----------------------------------------------------------------------
3366    Name: Matthias Blume
3367    Date: 2001/08/20 15:50 EDT
3368    Tag: blume20010820-slipup
3369    Description:
3370    
3371    I forgot to commit a few files.  Here they are...
3372    
3373    ----------------------------------------------------------------------
3374    Name: Matthias Blume
3375    Date: 2001/08/20 15:35:00 EDT
3376    Tag: blume-20010820-debugprof
3377    Description:
3378    
3379    !!!! NEW BOOTFILES !!!!
3380    
3381    This is another round of reorganizing the compiler sources.  This
3382    time the main goal was to factor out all the "instrumentation"
3383    passes (for profiling and backtracing) into their own library.
3384    The difficulty was to do it in such a way that it does not depend
3385    on elaborate.cm but only on elabdata.cm.
3386    
3387    Therefore there have been further changes to both elaborate.cm and
3388    elabdata.cm -- more "generic" things have been moved from the former
3389    to the latter.  As a result, I was forced to split the assignment
3390    of numbers indicating "primtyc"s into two portions: SML-generic and
3391    SML/NJ-specific.  Since it would have been awkward to maintain,
3392    I bit the bullet and actually _changed_ the mapping between these
3393    numbers and primtycs.  The bottom line of this is that you need
3394    a new set of bin- and bootfiles.
3395    
3396    I have built new bootfiles for all architectures, so doing a fresh
3397    checkout and config/install.sh should be all you need.
3398    
3399    The newly created library's name is
3400    
3401        $smlnj/viscomp/debugprof.cm
3402    
3403    and its sources live under
3404    
3405        src/compiler/DebugProf
3406    
3407    ----------------------------------------------------------------------
3408    Name: Matthias Blume
3409    Date: 2001/08/15 17:15:00 EDT
3410    Tag: blume-20010815-compreorg
3411    Description:
3412    
3413    This is a first cut at reorganizing the CM libraries that make up the
3414    core of the compiler.  The idea is to separate out pieces that could
3415    be used independently by tools, e.g., the parser, the typechecker, etc.
3416    
3417    The current status is a step in this direction, but it is not quite
3418    satisfactory yet.  Expect more changes in the future.
3419    
3420    Here is the current (new) organization...
3421    
3422        What used to be $smlnj/viscomp/core.cm is now divided into
3423        six CM libraries:
3424    
3425             $smlnj/viscomp/basics.cm
3426                           /parser.cm
3427                           /elabdata.cm
3428                           /elaborate.cm
3429                           /execute.cm
3430                           /core.cm
3431    
3432        The CM files for these libraries live under src/system/smlnj/viscomp.
3433        All these libraries are proxy libraries that contain precisely
3434        one CM library component.  Here are the locations of the components
3435        (all within the src/compiler tree):
3436    
3437             Basics/basics.cm
3438             Parse/parser.cm
3439             ElabData/elabdata.cm
3440             Elaborator/elaborate.cm
3441             Execution/execute.cm
3442             core.cm
3443    
3444         [This organization is the same that has been used already
3445         for a while for the architecture-specific parts of the visible
3446         compiler and for the old version of core.cm.]
3447    
3448         As you will notice, many source files have been moved from their
3449         respective original locations to a new home in one of the above
3450         subtrees.
3451    
3452         The division of labor between the new libraries is the following:
3453    
3454             basics.cm:
3455                - Simple, basic definitions that pertain to many (or all) of
3456                  the other libraries.
3457             parser.cm:
3458                - The SML parser, producing output of type Ast.dec.
3459                - The type family for Ast is also defined and exported here.
3460             elabdata.cm:
3461                - The datatypes that describe input and output of the elaborator.
3462                  This includes types, absyn, and static environments.
3463             elaborator.cm:
3464                - The SML/NJ type checker and elaborator.
3465                  This maps an Ast.dec (with a given static environment) to
3466                  an Absyn.dec (with a new static environment).
3467                - This libraries implements certain modules that used to be
3468                  structures as functors (to remove dependencies on FLINT).
3469             execute.cm:
3470                - Everything having to do with executing binary code objects.
3471                - Dynamic environments.
3472             core.cm:
3473                - SML/NJ-specific instantiations of the elaborator and MLRISC.
3474                - Top-level modules.
3475                - FLINT (this should eventually become its own library)
3476    
3477    Notes:
3478    
3479    I am not 100% happy with the way I separated the elaborator (and its
3480    data structures) from FLINT.  Two instances of the same problem:
3481    
3482        1. Data structures contain certain fields that carry FLINT-specific
3483           information.  I hacked around this using exn and the property list
3484           module from smlnj-lib.  But the fact that there are middle-end
3485           specific fields around at all is a bit annoying.
3486    
3487        2. The elaborator calculates certain FLINT-related information.  I tried
3488           to make this as abstract as I could using functorization, but, again,
3489           the fact that the elaborator has to perform calculations on behalf
3490           of the middle-end at all is not nice.
3491    
3492        3. Having to used exn and property lists is unfortunate because it
3493           weakens type checking.  The other alternative (parameterizing
3494           nearly *everything*) is not appealing, though.
3495    
3496    I removed the "rebinding =" warning hack because due to the new organization
3497    it was awkward to maintain it.  As a result, the compiler now issues some of
3498    these warnings when compiling init.cmi during bootstrap compilation. On
3499    the plus side, you also get a warning when you do, for example:
3500       val op = = Int32.+
3501    which was not the case up to now.
3502    
3503    I placed "assign" and "deref" into the _Core structure so that the
3504    code that deals with the "lazy" keyword can find them there.  This
3505    removes the need for having access to the primitive environment
3506    during elaboration.
3507    
3508    ----------------------------------------------------------------------
3509    Name: Matthias Blume
3510    Date: 2001/08/13
3511    Tag: blume-20010813-closures
3512    Description:
3513    
3514    This fix was sent to us by Zhong Shao.  It is supposed to improve the
3515    performance of certain loops by avoiding needless closure allocation.
3516    
3517    ----------------------------------------------------------------------
3518    Name: Lal George
3519    Date: 2001/07/31 10:03:23 EDT 2001
3520    Tag: george-20010731-x86-fmalloc
3521    Description: Fixed bug in x86 calls
3522    
3523        There was a bug where call instructions would mysteriously
3524        vanish. The call instruction had to be one that returned
3525        a floating point value.
3526    
3527    ----------------------------------------------------------------------
3528    Name: Lal George
3529    Date: 2001/07/19 16:36:29 EDT 2001
3530    Tag: george-20010719-simple-cells
3531    Description:
3532    
3533    I have dramatically simplified the interface for CELLS in MLRISC.
3534    
3535    In summary, the cells interface is broken up into three parts:
3536    
3537      1. CellsBasis : CELLS_BASIS
3538    
3539            CellsBasis is a top level structure and common for all
3540            architectures.  it contains the definitions of basic datatypes
3541            and utility  functions over these types.
3542    
3543      2. functor Cells() : CELLS
3544    
3545            Cells generates an interface for CELLS that incorporates the
3546            specific resources on the target architecture, such as the
3547            presence of special register classes, their number and size,
3548            and various useful substructures.
3549    
3550      3. <ARCH>CELLS
3551    
3552            e.g. SparcCells: SPARCCELLS
3553    
3554            <ARCH>CELLS usually contains additional bindings for special
3555            registers  on the architecture, such as:
3556    
3557                    val r0 : cell           (* register zero *)
3558                    val y : cell            (* Y register *)
3559                    val psr : cell          (* processor status register *)
3560                    ...
3561    
3562            The structure returned by applying the Cells functor is opened
3563            in this interface.
3564    
3565    The main implication of all this is that the datatypes for cells is
3566    split between CellsBasis and CELLS -- a fairly simple change for user
3567    code.
3568    
3569    In the old scheme the CELLS interface had a definitional binding of
3570    the form:
3571    
3572            signature CELLS = sig
3573    
3574               structure CellsBasis = CellsBasis
3575    
3576               ...
3577    
3578            end
3579    
3580    With all the sharing constraints that goes on in MLRISC, this old
3581    design  quickly leads to errors such as:
3582    
3583            "structure definition spec inside of sharing ... "
3584    
3585    
3586    and appears to require an unacceptable amount of sharing and where
3587    constraint hackery.
3588    
3589    I think this error message (the interaction of definitional specs and
3590    sharing) requires more explanation on our web page.
3591    
3592    ----------------------------------------------------------------------
3593    Name: Matthias Blume
3594    Date: 2001/07/19 15:00:00 EDT
3595    Tag: blume-20010719-libreorg
3596    Description:
3597    
3598    This update puts together a fairly extensive but straightforward change
3599    to the way the libraries that implement the interactive system are
3600    organized:
3601    
3602       The biggest change is the elimination of structure Compiler.  As a
3603       replacement for this structure, there is now a CM library
3604       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
3605       that exports all the substructures of the original structure Compiler
3606       directly.  So instead of saying Compiler.Foo.bar one now simply
3607       says Foo.bar.  (The CM libraries actually export a collection of
3608       structures that is richer than the collection of substructures of
3609       structure Compiler.)
3610    
3611       To make the transition smooth, there is a separate library called
3612       $smlnj/compiler/compiler.cm which puts together and exports the
3613       original structure Compiler (or at least something very close to it).
3614    
3615       There are five members of the original structure Compiler
3616       that are not exported directly but which instead became members
3617       of a new structure Backend (described by signature BACKEND).  These are:
3618       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
3619       Interact (: INTERACT), structure Machine (: MACHINE), and val
3620       architecture (: string).
3621    
3622       Structure Compiler.Version has become structure CompilerVersion.
3623    
3624       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
3625       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
3626       or x86, respectively.
3627       Each of these exports the same frontend structures that
3628       $smlnj/compiler.cm exports.  But they do not have a structure Backend
3629       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
3630       Hppa, PPC, Sparc, or X86, respectively.
3631    
3632       Library $smlnj/compiler/all.cm exports the union of the exports of
3633       $smlnj/compiler/<arch>.cm
3634    
3635       There are no structures <Arch>Compiler anymore, use
3636       $smlnj/compiler/<arch>.cm instead.
3637    
3638       Library host-compiler-0.cm is gone.  Instead, the internal library
3639       that instantiates CM is now called cm0.cm.  Selection of the host
3640       compiler (backend) is no longer done here but. (Responsibility for it
3641       now lies with $smlnj/compiler/current.cm.  This seems to be more
3642       logical.)
3643    
3644       Many individual files have been moved or renamed.  Some files have
3645       been split into multiple files, and some "dead" files have been deleted.
3646    
3647    Aside from these changes to library organization, there are also changes
3648    to the way the code itself is organized:
3649    
3650       Structure Binfile has been re-implemented in such a way that it no
3651       longer needs any knowledge of the compiler.  It exclusively deals
3652       with the details of binfile layout.  It no longer invokes the
3653       compiler (for the purpose of creating new prospective binfile
3654       content), and it no longer has any knowledge of how to interpret
3655       pickles.
3656    
3657       Structure Compile (: COMPILE) has been stripped down to the bare
3658       essentials of compilation.  It no longer deals with linking/execution.
3659       The interface has been cleaned up considerably.
3660    
3661       Utility routines for dealing with linking and execution have been
3662       moved into their own substructures.
3663    
3664       (The ultimate goal of these changes is to provide a light-weight
3665       binfile loader/linker (at least for, e.g., stable libraries) that
3666       does not require CM or the compiler to be present.)
3667    
3668    CM documentation has been updated to reflect the changes to library
3669    organization.
3670    
3671    ----------------------------------------------------------------------
3672    Name: Matthias Blume
3673    Date: 2001/07/10 17:30:00 EDT
3674    Tag: Release_110_34
3675    Description:
3676    
3677    Minor tweak to 110.34 (re-tagged):
3678    
3679      - README.html file added to CVS repository
3680      - runtime compiles properly under FreeBSD 3.X and 4.X
3681    
3682    ----------------------------------------------------------------------
3683    Name: Matthias Blume
3684    Date: 2001/07/10 17:30:00 EDT
3685    Tag: Release_110_34
3686    Description:
3687    
3688    New version number (110.34). New bootfiles.
3689    
3690    ----------------------------------------------------------------------
3691    Name: Matthias Blume
3692    Date: 2001/07/09 16:00:00 EDT
3693    Tag: blume-20010709-more-varargs
3694    Description:
3695    
3696    I changed the handling of varargs in ml-nlffigen again:
3697    The ellipsis ... will now simply be ignored (with an accompanying warning).
3698    
3699    The immediate effect is that you can actually call a varargs function
3700    from ML -- but you can't actually supply any arguments beyond the ones
3701    specified explicitly.  (For example, you can call printf with its format
3702    string, but you cannot pass additional arguments.)
3703    
3704    This behavior is only marginally more useful than the one before, but
3705    it has the advantage that a function or, more importantly, a function
3706    type never gets dropped on the floor, thus avoiding follow-up problems with
3707    other types that refer to the offending one.
3708    
3709    ----------------------------------------------------------------------
3710    Name: Matthias Blume
3711    Date: 2001/07/09 11:25:00 EDT
3712    Tag: blume-20010709-varargs
3713    Description:
3714    
3715    1. ckit-lib.cm now exports structure Error
3716    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
3717       with a warning accompanied by a source location.  Moreover, it
3718       merely skips the offending function or type and proceeds with the
3719       rest of its work.u  As a result, one can safely feed C code containing
3720       "..." to ml-nlffigen.
3721    3. There are some internal improvements to CM, providing slightly
3722       more general string substitutions in the tools subsystem.
3723    
3724    ----------------------------------------------------------------------
3725    Name: Matthias Blume
3726    Date: 2001/06/27 15:10:00 EDT
3727    Tag: blume-20010627-concur
3728    Description:
3729    
3730    Fixed a small bug in CM's handling of parallel compilation.
3731    (You could observe the bug by Control-C-interrupting an ordinary
3732    CMB.make or CM.stabilize and then attaching some compile servers.
3733    The result was that all of a sudden the previously interrupted
3734    compilation would continue on its own.  This was because of
3735    an over-optimization: CM did not bother to clean out certain queues
3736    when no servers were attached "anyway", resulting in the contents
3737    of these queues to grab control when new servers did get attached.)
3738    
3739    There is also another minor update to the CM manual.
3740    
3741    ----------------------------------------------------------------------
3742    Name: Matthias Blume
3743    Date: 2001/06/26 16:15:00 EDT
3744    Tag: blume-20010626-cmdoc
3745    Description:
3746    
3747    Minor typo fixed in CM manual (syntax diagram for libraries).
3748    
3749    ----------------------------------------------------------------------
3750    Name: Matthias Blume
3751    Date: 2001/06/25 22:55:00 EDT
3752    Tag: blume-20010625-x86pc
3753    Description:
3754    
3755    Fixed a nasty bug in the X86 assembly code that caused signal
3756    handlers to fail (crash) randomly.
3757    
3758    ----------------------------------------------------------------------
3759    Name: Matthias Blume
3760    Date: 2001/06/25 12:05:00 EDT
3761    Tag: blume-20010625-nlffigen
3762    Description:
3763    
3764    This update fixes a number of minor bugs in ml-nlffigen as reported by
3765    Nick Carter <nbc@andrew.cmu.edu>.
3766    
3767      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
3768      2. Default names for generated files are now derived from the name of
3769         the C file *without its directory*.  In particular, this causes generated
3770         files to be placed locally even if the C file is in some system directory.
3771      3. Default names for generated signatures and structures are also derived
3772         from the C file name without its directory.  This avoids silly things
3773         like "structure GL/GL".
3774         (Other silly names are still possible because ml-nlffigen does not do
3775          a thorough check of whether generated names are legal ML identifiers.
3776          When in doubt, use command line arguments to force particular names.)
3777    
3778    ----------------------------------------------------------------------
3779    Name: Matthias Blume
3780    Date: 2001/06/21 12:25:00 EDT
3781    Tag: blume-20010621-eXene
3782    Description:
3783    
3784    eXene now compiles and (sort of) works again.
3785    
3786    The library name (for version > 110.33) is $/eXene.cm.
3787    
3788    I also added an new example in src/eXene/examples/nbody.  See the
3789    README file there for details.
3790    
3791    ----------------------------------------------------------------------
3792    Name: Matthias Blume
3793    Date: 2001/06/20 16:40:00 EDT
3794    Tag: blume-20010620-cml
3795    Description:
3796    
3797    CML now compiles and works again.
3798    
3799    Libraries (for version > 110.33):
3800    
3801      $cml/cml.cm            Main CML library.
3802      $cml/basis.cm          CML's version of $/basis.cm.
3803      $cml/cml-internal.cm   Internal helper library.
3804      $cml/core-cml.cm       Internal helper library.
3805      $cml-lib/trace-cml.cm  Tracing facility.
3806      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
3807    
3808    The installer (config/install.sh) has been taught how to properly
3809    install this stuff.
3810    
3811    ----------------------------------------------------------------------
3812    Name: Matthias Blume
3813    Date: 2001/06/19 17:55:00 EDT
3814    Tag: blume-20010619-instantiate
3815    Description:
3816    
3817    This un-breaks the fix for bug 1432.
3818    (The bug was originally fixed in 110.9 but I broke it again some
3819    time after that.)
3820    
3821    ----------------------------------------------------------------------
3822    Name: Matthias Blume
3823    Date: 2001/06/19 17:25:00 EDT
3824    Tag: blume-20010619-signals
3825    Description:
3826    
3827    This should (hopefully) fix the long-standing signal handling bug.
3828    (The runtime system was constructing a continuation record with an
3829    incorrect descriptor which would cause the GC to drop data on the floor...)
3830    
3831    ----------------------------------------------------------------------
3832    Name: Matthias Blume
3833    Date: 2001/06/15 15:05:00 EDT
3834    Tag: blume-20010615-moresparc
3835    Description:
3836    
3837    Here is a short late-hour update related to Sparc c-calls:
3838    
3839     -- made handling of double-word arguments a bit smarter
3840    
3841     -- instruction selection phase tries to collapse certain clumsily