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 701, Thu Aug 31 08:46:43 2000 UTC revision 1629, Thu Sep 23 22:38:06 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: 2004/09/23 17:30:00 CDT
18    Tag: blume-20040923-envvars
19    Description:
20    
21    1. Changed definition of type ControlRegistry.registry_tree to
22       include control_info (i.e., the name of the controlling environment
23       variable).
24    
25    2. Added command-line flags -e and -E to print the names of environment
26       variables that can be used to control internal settings.
27       (This uses the new API mentioned in 1.)
28    
29    ----------------------------------------------------------------------
30    Name: Matthias Blume (blume (at) tti - c (dot) org)
31    Date: 2004/09/13 16:50:00 CDT
32    Tag: Release_110_49
33    Description:
34    
35    New working version (110.49).  NEW BOOTFILES!
36    
37    ----------------------------------------------------------------------
38    Name: Matthias Blume
39    Date: 2004/09/13 16:20:00 CDT
40    Tag: blume-20040913-config-mlrisc
41    Description:
42    
43    Put target "mlrisc" back into the default list.
44    (There is no harm in having it, and some users have expressed their
45    wish to have "mlrisc" included by default.)
46    
47    ----------------------------------------------------------------------
48    Name: John Reppy
49    Date: 2004/09/13
50    Tag: jhr-20040913-signals
51    Description:
52    
53    Fixed the signal masking code to properly nest mask/unmask operations
54    on a per-signal basis.
55    
56    ----------------------------------------------------------------------
57    Name: Matthias Blume (blume (at) tti - c (dot) org)
58    Date: 2004/09/08 13:20:00 CDT
59    Tag: blume-20040908-heap-magic
60    Description:
61    
62    Bumped the heap macig number to 0x09082004 to account for the changed
63    layout of the ML frame under MacOS X.
64    
65    ----------------------------------------------------------------------
66    Name: Allen Leung (leunga (at) reservoir (dot) com)
67    Date: 2004/09/03 11:26:00 EST
68    Tag: leunga-20040903-cygwin-install
69    Description:
70    
71       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
72    The Cygwin runtime is turned on by setting the environment
73    variable SMLNJ_CYGWIN_RUNTIME to 1.
74    
75    ----------------------------------------------------------------------
76    Name: Matthias Blume (blume (at) tti - c (dot) org)
77    Date: 2004/08/31 17:15:00 CDT
78    Tag: blume-20040831-core
79    Description:
80    
81    Added some exports to src/compiler/core.cm upon request by J. Joemann.
82    
83    ----------------------------------------------------------------------
84    Name: Matthias Blume (blume (at) tti - c (dot) org)
85    Date: 2004/08/30 17:55:00 CDT
86    Tag: blume-20040830-installer
87    Description:
88    
89    Upon request by Johannes Joemann:
90    
91    - improved ML code of installer to fall back to coping when renaming
92      fails (i.e., when source and target are on different file systems);
93      the code compiles but has yet to be tested in anger
94    - removed mlrisc from list of default targets (config/targets)
95    
96    ----------------------------------------------------------------------
97    Name: Matthias Blume (blume (at) tti - c (dot) org)
98    Date: 2004/08/27 17:20:00 CDT
99    Tag: blume-20040827-ptreql
100    Description:
101    
102    Added ptreql primop to structure InlineT (upon request from Larry
103    Paulson).
104    
105    ----------------------------------------------------------------------
106    Name: Allen Leung (leunga (at) reservoir (dot) com)
107    Date: 2004/08/15 21:21:00 EST
108    Tag: leunga-110_48-udgraph
109    Description:
110    
111    Another bug fix from Carl Hauser:
112    
113    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
114    > 48c48
115    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
116    > e::L)
117    > ---
118    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
119    > Without this, any deletion of an edge in an undirected graph does severe
120    > violence to the graph.
121    
122    ----------------------------------------------------------------------
123    Name: Allen Leung (leunga (at) reservoir (dot) com)
124    Date: 2004/08/10 23:35:00 EST
125    Tag: leunga-110_48-ppc
126    Description:
127    
128       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
129    
130    ----------------------------------------------------------------------
131    Name: Matthias Blume (blume (at) tti - c (dot) org)
132    Date: 2004/08/10 12:00:00 CDT
133    Tag: Release_110_48
134    Description:
135    
136    New working version (110.48).  NEW BOOTFILES!
137    
138    ----------------------------------------------------------------------
139    Name: Allen Leung (leunga (at) reservoir (dot) com)
140    Date: 2004/08/09 12:21:00 EST
141    Tag: leunga-110_47-dijsktra
142    Description:
143    
144    Bug fix from Carl Hauser:
145    
146    single_source_shortest_paths in dijkstra.sml was observed to get wrong
147    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
148    
149    The problem is that following the expression A.update(dist,s,Num.zero)
150    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
151    
152    ----------------------------------------------------------------------
153    Name: Matthias Blume (blume (at) tti - c (dot) org)
154    Date: 2004/08/06 18:10:00 CDT
155    Tag: blume-20040806-cmdline
156    Description:
157    
158    Fiddled with handling of command-line options:
159    
160      * sml now quits after processing the command line
161        if -H, -S, -h<n>, or -s<n> appears as the last
162        command-line argument
163      * a new option -q terminates the session when encountered on
164        the command line; subsequent arguments will be ignored
165      * bug fixes: short (erroneous) arguments are no longer ignored
166        completely
167    
168    ----------------------------------------------------------------------
169    Name: Allen Leung (leunga (at) reservoir (dot) com)
170    Date: 2004/08/04 18:17:00 EST
171    Tag: leunga-110_47-ppc-ibm-asm
172    Description:
173    
174      - Added minimal IBM assembly syntax support for PowerPC.
175    
176      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
177        been altered in the runtime.  We need an automatic way to keep the file
178        in sync.
179    
180    ----------------------------------------------------------------------
181    Name: Matthias Blume (blume (at) tti - c (dot) org)
182    Date: 2004/08/04 14:00:00 CDT
183    Tag: Release_110_47
184    Description:
185    
186    New working version (110.47).  NEW BOOTFILES!
187    
188    ----------------------------------------------------------------------
189    Name: Matthias Blume (blume (at) tti - c (dot) org)
190    Date: 2004/08/03 14:25:00 CDT
191    Tag: blume-20040803-callingconv
192    Description:
193    
194    Added low-level support for choosing C calling conventions by
195    twiddling the type of rawccall.  (See
196    src/compiler/Semant/types/cproto.sml for details.)
197    
198    ----------------------------------------------------------------------
199    Name: Matthias Blume (blume (at) tti - c (dot) org)
200    Date: 2004/08/02 15:55:00 CDT
201    Tag: blume-20040802-backout
202    Description:
203    
204    Backed out of change to win32-filesys.c.  The earlier patch to
205    get_file_time caused CM to produce files with the wrong time stamp.
206    
207    ----------------------------------------------------------------------
208    Name: Matthias Blume (blume (at) tti - c (dot) org)
209    Date: 2004/08/02 14:45:00 CDT
210    Tag: blume-20040802-nlffi-win32
211    Description:
212    
213    Added NLFFI support for Win32, adapted from a patch provided by David
214    Hansel.  This is currently completely untested.  Also, the issue
215    concerning stdcall vs. ccall is still unresolved.
216    
217    ----------------------------------------------------------------------
218    Name: Matthias Blume (blume (at) tti - c (dot) org)
219    Date: 2004/07/30 17:55:00 CDT
220    Tag: blume-20040730-various
221    Description:
222    
223    Gearing up towards 110.47...
224    
225    - various minor bugfixes to ml-nlffigen
226    - a beginning of a manual for nlffi
227    
228    - eliminated 'export name=value' in config/install.sh as this does
229      not work with certain versions of /bin/sh
230      (Thanks to David King at Motorola for catching this.)
231    
232    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
233      - added a test for tm==NULL to gmtime.c and localtime.c
234      - applied patch for incorrect GetFileTime under win32
235      - toSeconds -> toMilliseconds in Win32/win32-process.sml
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume (blume (at) tti - c (dot) org)
239    Date: 2004/07/21 18:20:00 CDT
240    Tag: blume-20040721-nlffigen
241    Description:
242    
243    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
244      for a typedef to an incomplete type, but leave out the "typ" member.
245      (This is just for consistency.)
246    - Started to produce what is supposed to become better (i.e., comprehensive)
247      documentation of what ml-nlffigen does and produces.
248    
249    ----------------------------------------------------------------------
250    Name: Matthias Blume (blume (at) tti - c (dot) org)
251    Date: 2004/07/14 16:25:00 CDT
252    Tag: blume-20040714-union
253    Description:
254    
255    Added C_UNION to c-calls/c-types.sml and updated the machinery
256    (ml-nlffigen, cproto.sml) that conveys C function interface
257    information to the code generator.
258    
259    However, the actual architecture-specific implementation of function
260    arguments and results that are C unions is still not implemented.
261    
262    ----------------------------------------------------------------------
263    Name: Allen Leung (leunga (at) reservoir (dot) com)
264    Date: 2004/07/14 14:38:00 EST
265    Tag: leunga-110_46_1-ppc-lwzu
266    Description:
267    
268        Added these instructions to the PowerPC architecture:
269        LBZU(X), LHZU(X), LWZU(X),
270        STWU(X), STFDU, STFSU
271    
272        etc...
273    
274        Note: I haven't added their instruction encoding into the description.
275    
276    ----------------------------------------------------------------------
277    Name: Allen Leung (leunga (at) reservoir (dot) com)
278    Date: 2004/07/13 15:04:00 EST
279    Tag: leunga-110_46_1-ppc-lwarx
280    Description:
281    
282       Added the two instructions LWARX and STWCX to the PowerPC
283    instruction set.
284    
285       A (untested) rewrite of loop-structure.sml.  The old version
286    is completely broken.
287    
288    ----------------------------------------------------------------------
289    Name: Matthias Blume (blume (at) tti - c (dot) org)
290    Date: 2004/07/13 13:50:00 CDT
291    Tag: blume-20040713-nlffi
292    Description:
293    
294    - use paramAlloc to report c-calls with too many arguments
295      (for PPC version where parameter area is pre-allocated)
296    - added ccall_maxargspace to machspec (to implement the above)
297    - made "make" commend in CM's "make" tool configurable
298    - added option (default: on) for passing the name of the SML/NJ's "bin"
299      directory to "make"; the call looks like this:
300    
301         make <options> SMLNJ_BINDIR=<dir> <target>
302    
303      This can be used by the Makefile to, e.g., pick the "right" version
304      of ml-nlffigen.
305    - minor code tweaks
306    
307    ----------------------------------------------------------------------
308    Name: Matthias Blume (blume (at) tti - c (dot) org)
309    Date: 2004/07/12 22:50:00 CDT
310    Tag: blume-110_46_1-macosx-nlffi
311    Description:
312    
313    NLFFI under Mac OS X now working (sort of).  This is largely untested,
314    though.
315    
316    Note:  1. You have to make a new, clean build of the runtime system.
317           2. There are new BOOTFILES, you have to use them!
318              (Doing the bootstrap process yourself would be *very* painful!
319               If you absolutely have to do it, build the system under
320               a different architecture and then cross-compile.)
321    
322    Version bumped to 110.46.1 to account for runtime data format changes.
323    
324    ----------------------------------------------------------------------
325    Name: Matthias Blume (blume (at) tti - c (dot) org)
326    Date: 2004/06/18 14:30:00 CDT
327    Tag: blume-20040618-unix
328    Description:
329    
330    Changed the implementation of structure Unix so that the same stream
331    is returned every time one of the {text,bin}{In,Out}streamOf functions
332    is invoked on the same proc.  This is not what the spec currently
333    says -- although IMO it arguably should.  (See discussion below.)
334    
335    ----------------------------------------------------------------------
336    Name: Matthias Blume (blume (at) tti - c (dot) org)
337    Date: 2004/06/17 18:15:00 CDT
338    Tag: Release_110_46
339    Description:
340    
341    New working version (110.46).  NEW BOOTFILES!
342    
343    ----------------------------------------------------------------------
344    Name: Matthias Blume (blume (at) tti - c (dot) org)
345    Date: 2004/06/17 17:20:00 CDT
346    Tag: blume-20040617-timer-unix
347    Description:
348    
349    Changed the interface of structures Timer and Unix to match the most
350    recent Basis spec.
351    
352    In the case of Unix there still seems to be an open/weird issue:
353    
354       The {text,bin}{In,Out}streamOf functions are supposed to create
355       fresh streams whenever they are called -- as opposed to have them
356       return the same stream every time.  This design is supposed to
357       prevent space leaks caused by proc values hanging on to streams.
358    
359       The reap function, on the other hand, is supposed to close the
360       streams.  This cannot be done without having a handle on the
361       stream in proc after all...
362    
363    I took the liberty to implement the following stopgap solution:
364    
365      The proc value hangs on to the most recently created stream(s).
366      Reap closes those.  If either or both of the two streams hadn't
367      been created at all yet, then reap will close the corresponding
368      file descriptors directly.
369    
370    PS: I don't understand the original space leak argument anymore.  If
371    a proc hangs on to the imperative stream, then I/O operations on those
372    will advance the state of the cached stream and avoid the space leak.
373    
374    ----------------------------------------------------------------------
375    Name: Matthias Blume (blume (at) tti - c (dot) org)
376    Date: 2004/05/28 16:45:00 CDT
377    Tag: blume-20040528-basis
378    Description:
379    
380    Added signature PACK_REAL and exported functor PrimIO.
381    
382    ----------------------------------------------------------------------
383    Name: Matthias Blume (blume (at) tti - c (dot) org)
384    Date: 2004/05/25 16:00:00 CDT
385    Tag: blume-20040525-group-owner
386    Description:
387    
388    CM now ignores (but still accepts) the "owner" information in group
389    descriptions.  The owner of a group is its next enclosing
390    library. Each group must have a unique owner.  (There is a virtual
391    "toplevel" library that owns groups which are not nested within a real
392    library.)  Previously, each group had to explicitly declare its owner,
393    and CM would check that such a declaration is correct.  The new scheme
394    is to have CM check that for each group there is precisely one owning
395    library.
396    
397    The advantage of the new scheme is that the programmer no longer needs
398    to maintain the somewhat annoying owner information.  The downside is
399    that CM cannot enforce the ownership rule across multiple runs of
400    CM.make.  Fortunately, enclosing the same group in two different
401    libraries A and B which are not part of the same program does not
402    cause real problems.
403    
404    ----------------------------------------------------------------------
405    Name: Matthias Blume (blume (at) tti - c (dot) org)
406    Date: 2004/05/20 16:00:00 CDT
407    Tag: blume-20040520-win32
408    Description:
409    
410    Made the win32 version work again.  (Strangely, a misplaced comma had
411    slipped into win32-process.c which prevented the runtime from being
412    compiled correctly.)
413    
414    Also, included a minor addition to ml-build.bat analogous to what was
415    done in blume-20040519-ml-build.
416    
417    ----------------------------------------------------------------------
418    Name: Matthias Blume (blume (at) tti - c (dot) org)
419    Date: 2004/05/19 22:10:00 CDT
420    Tag: blume-20040519-ml-build
421    Description:
422    
423    Arranged for ml-build to clean up after itself a little bit better.
424    The script generates a temporary SML source file and compiles it using
425    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
426    gets rid of those at the end, so they don't accumulate under .cm.
427    
428    This required a minor change to install.sh because the name of the
429    metadata directory (default: .cm) is actually configurable at
430    installation time.
431    
432    ----------------------------------------------------------------------
433    Name: Matthias Blume (blume (at) tti - c (dot) org)
434    Date: 2004/05/18 15:50:00 CDT
435    Tag: blume-20040518-mkreader
436    Description:
437    
438    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
439    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
440    
441    ----------------------------------------------------------------------
442    Name: Matthias Blume (blume (at) tti - c (dot) org)
443    Date: 2004/05/11 14:35:00 CDT
444    Tag: blume-20040511-win32sock
445    Description:
446    
447    Added previously missing support for many socket-related functions
448    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
449    for the voluminous patch!
450    
451    (I have not tested this patch under win32 yet.)
452    
453    Here is David's e-mail:
454    
455    Hi,
456    
457    Attached to this email you find a diff against sml/nj 110.45
458    that will enable socket support under Windows.
459    
460    To apply the patch (using unix or cygwin)
461    1) gunzip runtime.diff.gz
462    2) "cd" into "src/runtime" in the source tree of a fresh
463       110.45 installation.
464    3) patch -p 1 < [your/path/to]runtime.diff
465    
466    The code compiles fine but has NOT yet been extensively tested.
467    I only ran a few tests for basic socket client functionality
468    (which worked fine).  Especially the functions that use ioctl
469    are not tested at all and might not work (see below).
470    
471    I implemented this since we want to move to a newer version of sml/nj
472    but need socket support in order to use it.  This is the first time I
473    even had a look at the sml/nj source,  so please review my changes
474    before making this part of the distribution!  Here are a few issues
475    that I think might be better for someone to solve who is more
476    familiar with the sml/nj source (and socket programming):
477    
478    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
479      exception since I could not figure out what the windows equivalent
480      of these functions is
481    
482    - In sockets-osdep.h there are a some #include statements that are
483      only used in a few files that include sockets-osdep.h
484    
485    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
486      does not process its return value since I don't know how to
487      report an error upwards.
488    
489    - It would probably be good to have a call to WSACleanup() when
490      the library is unloaded (if there is such a possibility).
491      Otherwise I think Windows will take care of this automatically
492      when the process finishes.
493    
494    - I used ioctlsocket() as a replacement for ioctl() but I have
495      no idea if that is actually the proper replacement on Windows.
496    
497    - All these issues are marked in the code by "FIXME" comments.
498    
499    We use sml/nj extensively in our products and are quite happy
500    with it.  I hope this contribution will help you.
501    
502    Keep up the good work!
503    
504    David
505    
506    ----------------------------------------------------------------------
507    Name: Matthias Blume (blume (at) tti - c (dot) org)
508    Date: 2004/05/11 14:20:00 CDT
509    Tag: blume-20040511-installml
510    Description:
511    
512    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
513    the patch.)
514    
515    ----------------------------------------------------------------------
516    Name: Matthias Blume (blume (at) tti - c (dot) org)
517    Date: 2004/05/11 14:05:00 CDT
518    Tag: blume-20040511-nlffi-netbsd
519    Description:
520    
521    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
522    the patch.)
523    
524    ----------------------------------------------------------------------
525    Name: Matthias Blume (blume (at) tti - c (dot) org)
526    Date: 2004/05/11 12:05:00 CDT
527    Tag: blume-20040511-exports
528    Description:
529    
530    As per request by Adam Chlipala <adam@hcoop.net>, extended various
531    export lists in compiler-related .cm-files.
532    
533    ----------------------------------------------------------------------
534    Name: Matthias Blume (blume (at) tti - c (dot) org)
535    Date: 2004/05/11 11:35:00 CDT
536    Tag: blume-20040511-allsource
537    Description:
538    
539    The installer now honors the "src-smlnj" target again, although its meaning
540    has changed from "all sources required for the compiler" to "all sources
541    the installer knows about".  In other words, if you enable "src-smlnj"
542    in the "targets" file, then the installer will pull in sources for
543    everything.  (Notice that this refers to source code only.  Compiled
544    code is still only installed for modules that were requested explicitly
545    or which are required for other modules that were requested explicitly.)
546    
547    ----------------------------------------------------------------------
548    Name: Matthias Blume (blume (at) tti - c (dot) org)
549    Date: 2004/04/23 17:40:00 CDT
550    Tag: blume-20040423-ieee-scan
551    Description:
552    
553    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
554    in the exponent calculation we get INF or ZERO (depending on the mantissa
555    and the sign of the exponent).
556    
557    ----------------------------------------------------------------------
558    Name: Matthias Blume (blume (at) tti - c (dot) org)
559    Date: 2004/04/23 10:40:00 CDT
560    Tag: blume-20040423-ml-build
561    Description:
562    
563    The ml-build script now terminates with a non-0 status when something
564    goes wrong.
565    
566    ----------------------------------------------------------------------
567    Name: Matthias Blume (blume (at) tti - c (dot) org)
568    Date: 2004/04/22 16:35:00 CDT
569    Tag: blume-20040422-Option
570    Description:
571    
572    Made exception Option to be the same as exception Option.Option
573    (as it should be).
574    
575    ----------------------------------------------------------------------
576    Name: Allen Leung (leunga (at) reservoir (dot) com)
577    Date: 2004/03/19 14:40:00 EST
578    Tag: leunga-20040319-cygwin-nlffi
579    Description:
580    
581    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
582    of SML/NJ.  The problem is that
583    
584       lib = dlopen(NULL, ...)
585       f   = dlsym(lib, "malloc");
586    
587    does not work on Windows unless we explicitly export symbols
588    such as 'malloc' during linking.  We fixed this by explicitly
589    exporting the required symbols with the magic gcc incantation:
590    
591        -Wl,--export-all cygwin.def
592    
593    where cygwin.def is a file containing all the symbols that
594    we wish to export.
595    
596    I suspect this is a Windows problem and we'll have to
597    do the same (somehow with windows compilers) when
598    we build the native win32 version with the system
599    calls LoadLibrary/GetProcAddress.
600    
601    ----------------------------------------------------------------------
602    Name: Matthias Blume (blume (at) tti - c (dot) org)
603    Date: 2004/03/04 16:35:00 CST
604    Tag: blume-20040304-intinf-fmt
605    Description:
606    
607    Fixed problem with IntInf.fmt (sign would show up on the right instead
608    of on the left for BIN, OCT, and HEX).
609    
610    ----------------------------------------------------------------------
611    Name: Matthias Blume (blume (at) tti - c (dot) org)
612    Date: 2004/03/04 11:25:00 CST
613    Tag: blume-20040304-symlinks
614    Description:
615    
616    Fixed problem with installer script (unix only) where bin/ml-yacc and
617    friends pointed (via symlinks) to absolute locations instead of just
618    .run-sml.  This was reported by Vesa A Norrman.
619    
620    ----------------------------------------------------------------------
621    Name: Matthias Blume (blume (at) tti - c (dot) org)
622    Date: 2004/02/13 14:50:00 CST
623    Tag: Release_110_45
624    Description:
625    
626    New working version (110.45). New bootfiles.
627    
628    ----------------------------------------------------------------------
629    Name: Matthias Blume (blume (at) tti - c (dot) org)
630    Date: 2004/01/26 15:15:15 CST
631    Tag: blume-20040126-toplevel
632    Description:
633    
634    Improved handling of exceptions at the interactive toplevel.
635    
636    ----------------------------------------------------------------------
637    Name: Matthias Blume (blume (at) tti - c (dot) org)
638    Date: 2004/01/26 11:25:00
639    Tag: blume-20040126-app
640    Description:
641    
642    Type of top-level "app" corrected.
643    Added code for setting vp_limitPtrMask to Win32-specific runtime.
644    
645    ----------------------------------------------------------------------
646    Name: Matthias Blume (blume (at) tti - c (dot) org)
647    Date: 2003/11/18 17:10 CST
648    Tag: blume-20031118-basis-fiddle
649    Description:
650    
651    - changed Timer interface to what might become the spec
652    - POSIX_FLAGS -> BIT_FLAGS according to spec
653    - some other minor discrepancies wrt. spec eliminated
654    
655    ----------------------------------------------------------------------
656    Name: Matthias Blume (blume (at) tti - c (dot) org)
657    Date: 2003/11/06 12:00:00 CST
658    Tag: Release_110_44
659    Description:
660    
661    New working version (110.44). New bootfiles.
662    
663    ----------------------------------------------------------------------
664    Name: Matthias Blume (blume (at) tti - c (dot) org)
665    Date: 2003/11/04 11:50:00 CST
666    Tag: blume-20031104-move-libraries
667    Description:
668    
669    Eliminated the "dont_move_libraries" directive in config/targets.
670    (The mechanism was broken and could not be fixed easily.  Moreover,
671    there does not seem to be any reason not to move all libraries into
672    lib during installation.  I originally implemented this directive as a
673    backward-compatibility feature when I first introduced the new CM.
674    Now that things have been stable for a long time and going back to the
675    old CM is not an option, there is no reason to keep it around.)
676    
677    ----------------------------------------------------------------------
678    Name: Matthias Blume (blume (at) tti - c (dot) org)
679    Date: 2003/11/03 16:00:00 CST
680    Tag: blume-20031103-installdir
681    Description:
682    
683    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
684    Richards for pointing out the problem and providing the solution.)
685    
686    ----------------------------------------------------------------------
687    Name: Matthias Blume (blume (at) tti - c (dot) org)
688    Date: 2003/10/01 17:05:00 CDT
689    Tag: blume-20031001-lal-mlrisc
690    Description:
691    
692    MLRISC bug fix from Lal.
693    
694    ----------------------------------------------------------------------
695    Name: Matthias Blume (blume (at) tti - c (dot) org)
696    Date: 2003/09/30 16:10:00 CDT
697    Tag: blume-20030930-primio-bat
698    Description:
699    
700    1. Added openVector, nullRd, and nullWr to PRIM_IO.
701    2. Improved .bat files (for Win32 port) to make things work under Win95.
702       (thanks to Aaron S. Hawley for this one)
703    
704    ----------------------------------------------------------------------
705    Name: Matthias Blume (blume (at) tti - c (dot) org)
706    Date: 2003/09/26 16:05:00 CDT
707    Tag: blume-20030926-wrappriv
708    Description:
709    
710    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
711    
712    ----------------------------------------------------------------------
713    Name: Matthias Blume (blume (at) tti - c (dot) org)
714    Date: 2003/09/26 15:00:00 CDT
715    Tag: blume-20030926-110_43_3
716    Description:
717    
718    - additional cleanup
719    - version number bump, NEW BOOTFILES
720    
721    ----------------------------------------------------------------------
722    Name: Matthias Blume (blume (at) tti - c (dot) org)
723    Date: 2003/09/26 12:00:00 CDT
724    Tag: blume-20030926-ppautoload
725    Description:
726    
727    I modified the read-eval-print loop so that the autoloader gets
728    invoked whenever the prettyprinter tries to look up a symbol that
729    is not currently defined in the toplevel environment but which
730    appears in CM's autoload registry.  As a result, we see far fewer of
731    those ?.Foo.Bar.xxx names in the prettyprinter's output.
732    
733    In addition to this I tried to clean up some pieces of the Basis
734    implementation (e.g., Socket, Word8Array) in order to prevent other
735    instances of these ?.Foo.Bar.xxx names from being printed.
736    
737    The mechanism that picks names for types still needs some work, though.
738    (Right now it seems that if there is a type A.t which is defined to
739    be B.u, but B is unavailable at toplevel, then A.t gets printed as
740    "?.B.u" although the perhaps more sensible solution would be to use
741    "A.t" in this case.  In other words, the prettyprinter should follow
742    a chain of DEFtycs not farther than there are corresponding toplevel
743    names in the current environment.)
744    
745    ----------------------------------------------------------------------
746    Name: Matthias Blume (blume (at) tti - c (dot) org)
747    Date: 2003/09/24 16:31:00 CDT
748    Tag: blume-20030924-installer
749    Description:
750    
751    Another installer tweak:  All the ML code for the installer is now
752    compiled during CMB.make and put into a little library called
753    $smlnj/installer.cm.  The installation then simply invokes
754    
755       sml -m $smlnj/installer.cm
756    
757    and everything happens automagically.
758    
759    Win32:  ML code senses value of environment variable SMLNJ_HOME.
760    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
761            and BINDIR.
762    
763    The new scheme guarantees that the ML code responsible for the installation
764    is in sync with the APIs of the main system.  Also, the installer is
765    somewhat faster because the installer script is precompiled.
766    
767    ----------------------------------------------------------------------
768    Name: Matthias Blume (blume (at) tti - c (dot) org)
769    Date: 2003/09/24 15:35:00 CDT
770    Tag: blume-20030924-synsock
771    Description:
772    
773    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
774    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
775    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
776    implementations of non-blocking operations from CML's Socket
777    structure.
778    
779    ----------------------------------------------------------------------
780    Name: Matthias Blume (blume (at) tti - c (dot) org)
781    Date: 2003/09/24 15:10:05 CDT
782    Tag: blume-20030924-sockets
783    Description:
784    
785    1. Fixed SOCKET API and implementation to match Basis spec.
786       This required changing the internal representation of sockets to one
787       that remembers (for each socket file descriptor) whether it is currently
788       blocking or non-blocking.  This state is maintained lazily (i.e., a system
789       call is made only if the state actually needs to change).
790    
791    2. OS-specific details of sockets were moved into separate files, thus
792       making it possible to unify the bulk of the socket implementations
793       between Unix and Win32.
794    
795    3. CML's socket API changed accordingly.
796       (Note that we need to remove non-blocking functions from this API
797       since they are redundant in the case of CML!)
798    
799    4. CML's socket implementation now makes use of non-blocking functions
800       provided by Basis, thus removing all OS-dependent code from this part
801       of CML.
802    
803    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
804    
805    ----------------------------------------------------------------------
806    Name: Matthias Blume (blume (at) tti - c (dot) org)
807    Date: 2003/09/22 12:10:00 CDT
808    Tag: blume-20030922-110_43_2
809    Description:
810    
811    Made a new interim version and bootfiles for developer's bootstrapping
812    convenience.
813    
814       110.43.2 -- NEW BOOTFILES
815    
816    ----------------------------------------------------------------------
817    Name: Matthias Blume (blume (at) tti - c (dot) org)
818    Date: 2003/09/19 15:55:00 CDT
819    Tag: blume-20030919-cmdir
820    Description:
821    
822    1. new-install.sh -> install.sh
823    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
824    3. tweaked installer so that another name instead of .cm can be chosen
825       at install time (by setting the CM_DIR_ARC environment variable
826       during installation); once installation is complete, the name is
827       fixed
828    
829    ----------------------------------------------------------------------
830    Name: Matthias Blume (blume (at) tti - c (dot) org)
831    Date: 2003/09/18 16:00:00 CDT
832    Tag: blume-20030918-110_43_1
833    Description:
834    
835    Made a new interim version and bootfiles for developer's bootstrapping
836    convenience.
837    
838       110.43.1 -- NEW BOOTFILES
839    
840    ----------------------------------------------------------------------
841    Name: Matthias Blume (blume (at) tti - c (dot) org)
842    Date: 2003/09/18 15:20:00 CDT
843    Tag: blume-20030918-misc
844    Description:
845    
846    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
847       this seems to be controversial at the moment) and used those in
848       Posix.ProcEnv.times.
849    
850    2. Added Time.{from,to}Nanoseconds to Time.
851    
852    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
853       For example, fromLargeInt never needs to look at more than 3 "big
854       digits" to get its 53 bits of precision.
855    
856    ----------------------------------------------------------------------
857    Name: Matthias Blume (blume (at) tti - c (dot) org)
858    Date: 2003/09/17 16:30:00 CDT
859    Tag: blume-20030917-real32-slices
860    Description:
861    
862    Added an entry to the primitive environment
863    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
864    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
865    
866    Removed some of the "magic" constants in real64.sml and replaced them
867    with code that generates these values from their corresponding
868    integer counterparts.
869    
870    Made all(?) the slice-related changes to the Basis and made everything
871    compile again...
872    
873    ----------------------------------------------------------------------
874    Name: Matthias Blume (blume (at) tti - c (dot) org)
875    Date: 2003/09/15 17:45:00 CDT
876    Tag: blume-20030915-rbase
877    Description:
878    
879    Fixed bug in Real.fromLargeInt.
880    
881    ----------------------------------------------------------------------
882    Name: Matthias Blume (blume (at) tti - c (dot) org)
883    Date: 2003/09/13 18:11:00 CDT
884    Tag: blume-20030913-libinstall
885    Description:
886    
887    Minor bugfix in config/libinstall (set anchor with path to
888    standalone tool after installing it, otherwise libraries that
889    need ml-lex or ml-yacc won't compile the first time the installer
890    runs).
891    
892    ----------------------------------------------------------------------
893    Name: Matthias Blume (blume (at) tti - c (dot) org)
894    Date: 2003/09/12 11:45:00 CDT
895    Tag: blume-20030912-various
896    Description:
897    
898     - fixed bug in Real.toLargeInt
899     - fixed bug in Posix.ProcEnv.times
900     - changed inputLine functions to return an option
901     - minor installer improvements / bugfixes
902     - changed default @SMLalloc parameter for x86/celeron to 64k
903    
904    ----------------------------------------------------------------------
905    Name: Matthias Blume (blume (at) tti - c (dot) org)
906    Date: 2003/09/09 22:00:00 CDT
907    Tag: Release_110_43
908    Description:
909    
910        New working release 110.43.  New bootfiles.
911    
912    ----------------------------------------------------------------------
913    Name: Matthias Blume (blume (at) tti - c (dot) org)
914    Date: 2003/09/09 19:20:00 CDT
915    Tag: blume-20030909-installer
916    Description:
917    
918    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
919    Modified config/install.bat to take advantage of it.  Also modified
920    config/install.sh (and called it config/new-install.sh) to take advantage
921    of it on Unix systems.  (The SML code is (supposed to be) platform-
922    independent.)
923    
924    The installer can now install everything under Win32
925    as well as under *nix as long as it compiles.
926    
927    Other changes:
928    
929      - made CML compile again under Win32
930      - made eXene compile under Win32 (by providing a fake structure UnixSock
931        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
932      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
933        same as type int;  under Win32 it isn't
934      - fixed some slice-related problems in the win32-specific parts of CML
935      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
936        (under Win32, the volume name is case-insensitive, and the
937        OS.Path code compares volume names for equality)
938    
939    ----------------------------------------------------------------------
940    Name: Matthias Blume (blume (at) tti - c (dot) org)
941    Date: 2003/09/08 11:55:00 CDT
942    Tag: blume-20030908-fullpath
943    Description:
944    
945    Made Win32 version of OS.FileSys.fullPath return current directory
946    when given an empty string.  This is what the spec says, and incidentally,
947    CM depends on it.  (CM otherwise goes into an infinite loop in certain
948    cases when presented with the name of a non-existing .cm file.)
949    
950    ----------------------------------------------------------------------
951    Name: Matthias Blume (blume (at) tti - c (dot) org)
952    Date: 2003/09/04 16:30:00 CDT
953    Tag: blume-20030905-slices-etc
954    Description:
955    
956      1. Changed interface to vectors and arrays in Basis to match
957         (draft) Basis spec.
958      2. Added signatures and implementations of slices according to
959         Basis spec.
960      3. Edited source code throughout the system to make it compile again
961         under 1. and 2.  (In some cases code had to be added to have it
962         match the new signatures.)
963      4. MLRISC should be backward-compatible: the copies of the originals
964         of files that needed to change under 3. were retained, the .cm files
965         check the compiler version number and use old versions when
966         appropriate.
967      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
968         dirstream -> string option (in accordance with Basis spec).
969      6. When generating code that counts lines, ml-lex used function
970         CharVector.foldli, taking advantage of its old interface.
971         This has been replaced with the corresponding code from
972         CharVectorSlice.  (html-lex must be re-lexed!)
973      7. BitArray in smlnj-lib/Util has been extended/modified to match the
974         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
975      8. Removed temporary additions (fromInternal, toInternal) from the
976         (now obsolete) IntInf in smlnj-lib/Util.
977      9. Cleaned up structure Byte.
978     10. Added localOffset, scan, and fromString to Date (according to spec).
979         Cleaned/corrected implementation of Date.
980         (Still need to check for correctness; implement better canonicalizeDate.)
981     11. Added "scan" to signature IEEE_REAL.
982     12. Some improvements to IntInf [in particular: efficiency-hack for
983         mod and rem when second operand is 2 (for parity checks).]
984     13. Changed representation of type Time.time, using a single IntInf.int
985         value counting microseconds.  This considerably simplified the
986         implementation of structure Time.  We now support negative time
987         values; scan and fromString handle signs.
988     14. Functor PrimIO now takes two additional arguments (VectorSlice and
989         ArraySlice).
990    
991    ----------------------------------------------------------------------
992    Name: Matthias Blume (blume (at) tti - c (dot) org)
993    Date: 2003/08/28 17:00:00 CDT
994    Tag: blume-20030828-intinf
995    Description:
996    
997    This is a major update which comes with a version number bump
998    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
999    and an implementation of IntInf in the Basis.
1000    
1001    There are a fairly large number of related changes and updates throughout
1002    the system:
1003    
1004    Basis:
1005            - Implemented IntInf.
1006            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1007            - Added some missing Real64 operations, most notably Real.toLargeInt.
1008            - Added FixedInt as a synonym for Int32.
1009    
1010    compiler:
1011            * Added support for a built-in intinf type.
1012              - literals
1013              - pattern matching
1014              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1015              - overloading on literals and operations
1016    
1017            This required adding a primitive type intinf, some additional
1018            primops, and implementations for several non-trivial intinf
1019            operations in Core.  (The intinf type is completely abstract
1020            to the compiler; all operations get delegated back to the Core.)
1021    
1022            * Intinf equality is handled by polyequal.  However, the compiler
1023            does not print its usual warning in this case (since polyequal
1024            is the right thing to do there).
1025    
1026            * Improved the organization of structure InlineT.
1027    
1028            * A word about conversion primops:
1029            If conversions involving intinf do not cancel out during
1030            CPS contract, then the compiler must insert calls to Core functions.
1031            Since all core access must be resolved already during the FLINT
1032            translate phase, it would be too late a the time of CPS contract
1033            to add new Core calls.  For this reason, conversion primops
1034            for intinf carry two arguments:  1. the numeric argument that
1035            they are supposed to convert, and 2. the Core function that
1036            can help with this conversion if necessary.  If CPS contract
1037            eliminates a primop, then the associated Core function becomes
1038            dead and goes away.  Intinf conversion primops that do not get
1039            eliminated by CPS contract get rewritten into calls of their
1040            core functions by a separate, new phase.
1041    
1042    interactive system:
1043            - Control.Print.intinfDepth controls max length of intinf constants
1044            being printed. (Analogous to Control.Print.stringDepth.)
1045            - Cleanup in printutil and pputil: got rid of unused stuff and
1046            duplicates; replaced some of the code with code that makes better
1047            use of library functionality.
1048    
1049    CM:
1050            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1051            are no longer silent.
1052    
1053    CKIT:
1054            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1055            in favor of LargeInt -- which is now the same as IntInf.
1056            CKIT-knowledgable people should check whether this is what's
1057            intended and otherwise change things back to using Int32 or
1058            FixedInt.
1059    
1060    Throughout the code:
1061            Started using IntInf.int literals and built-in operations
1062            (e.g., comparison with 0) where this seems appropriate.
1063    
1064    
1065    ----------------------------------------------------------------------
1066    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1067    Date: 2003/08/13 11:36:00 CDT
1068    Tag: dbm-20030813-mcz-merge1
1069    Description:
1070    
1071    Merging changes from the mcz-branch development branch into trunk.
1072    These changes involve replacement of the emulated old prettyprinter
1073    interface with direct use of the SML/NJ Lib PP library, and fixing
1074    of a couple of bugs (895, 1186) relating to error messages.  A new
1075    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1076    has been added.
1077    
1078    ----------------------------------------------------------------------
1079    Name: Matthias Blume (blume (at) tti - c (dot) org)
1080    Date: 2003/08/11 15:45:00 CDT
1081    Tag: blume-20030811-windows
1082    Description:
1083    
1084    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1085    
1086        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1087    
1088    This patch restores SML/NJ's ability to run under win32.  There are a
1089    number of changes, including fixes for several bugs that had gone
1090    unnoticed until now:
1091    
1092      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1093        trying to run the win32 version from within cygwin.)
1094    
1095      - There are a number of simple .bat scripts that substitute for their
1096        corresponding Unix shell-scripts.  (See below.)
1097    
1098      - The internals of ml-build have been modified slightly.  The main
1099        difference is that instead of calling ".link-sml" (or link-sml.bat)
1100        using OS.Process.system, the ML process delegates this task back
1101        to the script.  Otherwise problems arise in mixed environments such
1102        as Cygwin where scripts look and work like Unix scripts, but
1103        where OS.Process.system cannot run them.
1104    
1105      - In CM, the srcpath pickler used native pathname syntax -- which
1106        is incorrect in the case of cross-compilation.  The new pickle format
1107        is independent of platform-specific naming conventions.
1108    
1109      - Path configuration files (such as lib/pathconfig) can now choose
1110        between native and standard syntax.  Placing a line of the form
1111    
1112            standard!
1113    
1114        into the file causes all subsequent paths to be interpreted using
1115        CM standard pathname syntax (= Unix conventions); a line
1116    
1117            native!
1118    
1119        switches back to native style.  This was needed so that
1120        path config files can be written portably, see src/system/pathconfig.
1121    
1122      - Runtime system:
1123    
1124          - win32-filesys.c:  get_file_time and set_file_time now
1125            access modification time, not creation time.
1126    
1127          - I/O code made aware of new array representation.
1128    
1129          - Bug fixes in X86.prim.masm.
1130    
1131       - src/system/makeml made aware of win32. (For use under cygwin
1132            and other Unix-environments for windows.)
1133    
1134       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1135         which caused BinIO.inputAll to fail consistently.
1136    
1137    .bat scripts:
1138    
1139       Windows .bat scripts assume that SMLNJ_HOME is defined.
1140    
1141       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1142         applications (sml, ml-yacc, ml-lex).
1143       - ml-build.bat: analogous to ml-build.
1144       - config\install.bat:  Analogous to config/install.sh.  This requires
1145         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1146         (nmake etc. must be on the path, and vcvars32 must have been run.)
1147         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1148         and the bootfile hierarchy must have been unpacked under
1149         sml.boot.x86-win32.
1150         The script is very primitive and does a poor job at error checking.
1151         It only installs the base system, ml-lex, and ml-yacc.  No other
1152         libraries are being installed (i.e., you get only those that
1153         are part of the compiler.)
1154       - link-sml.bat:  analogous to .link-sml, but not currently used
1155    
1156    Unrelated bug fixes:
1157    
1158      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1159        types.
1160      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1161        otherwise.)
1162    
1163    ----------------------------------------------------------------------
1164    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1165    Date: 2003/06/17
1166    Tag: macqueen-20030617-bug895
1167    Description:
1168    
1169    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1170    Tag will be used for new development branch (mcz-branch) for
1171    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1172    
1173    ----------------------------------------------------------------------
1174    Name: Matthias Blume (blume (at) tti - c (dot) org)
1175    Date: 2003/05/27 16:55:00 CDT
1176    Tag: blume-20030527-polyeq
1177    Description:
1178    
1179    Tried to eliminated most cases of polymorphic equality.
1180    
1181    ----------------------------------------------------------------------
1182    Name: Matthias Blume (blume (at) tti - c (dot) org)
1183    Date: 2003/05/21 17:45:00 CDT
1184    Tag: blume-20030517-complete
1185    Description:
1186    
1187    Two changes:
1188    
1189      1. Added a flag for controlling whether non-exhaustive bindings will
1190         be treated as errors (default is false).
1191      2. Cleaned up the *entire* source tree so that CMB.make goes through
1192         without a single non-exhaustive match- or bind warning.
1193    
1194    ----------------------------------------------------------------------
1195    Name: Matthias Blume (blume (at) tti - c (dot) org)
1196    Date: 2003/05/17 10:20:00 CDT
1197    Tag: blume-20030517-absyn
1198    Description:
1199    
1200    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1201    
1202       This mainly affects the quality of error messages.  However, some
1203       of the code is now more straightforward than before. (Treatment of
1204       the above four constructs in translate.sml is much simpler than
1205       the "macro-expansion" that was going on before.  Plus, the mach-
1206       compiler no longer gets invoked just to be able to compile an
1207       if-expression.)
1208    
1209    2. The ErrorMsg.Error exception is now caught and absorbed by the
1210       interactive loop.
1211    
1212    ----------------------------------------------------------------------
1213    Name: Allen Leung
1214    Date: 2003/05/16 13:05:00 CDT
1215    Tag: leunga-20030516-cygwin-runtime
1216    Description:
1217    
1218        Ported the runtime system to cygwin, which uses the unix
1219    x86-unix bin files.  Missing/buggy features:
1220    
1221       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1222         the Cygwin library.
1223       o Ctrl-C handling may be flaky.
1224       o Windows system calls and Windows I/O are not supported.
1225    
1226       A new set of binfiles is located at:
1227    
1228         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1229    
1230       This is only needed for bootstrapping the cygwin version of smlnj.
1231       Other x86 versions can use the existing binfiles.
1232    
1233    ----------------------------------------------------------------------
1234    Name: Matthias Blume
1235    Date: 2003/04/08 15:42:00 CDT
1236    Tag: blume-20030408-listpair
1237    Description:
1238    
1239    1. Added a target 'mlrisc' to installer.
1240    
1241    2. Added missing elements to structure ListPair.
1242    
1243    ----------------------------------------------------------------------
1244    Name: Allen Leung
1245    Date: 2003/01/07 10:40:00 EST
1246    Tag: leunga-20030107-int-rem
1247    Description:
1248    
1249    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1250    The arguments to the SUBL instruction were swapped.
1251    
1252    ----------------------------------------------------------------------
1253    Name: Matthias Blume
1254    Date: 2002/12/12 16:25:00 EST
1255    Tag: blume-20021212-risc-ra
1256    Description:
1257    
1258    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1259    sent the RA into an infinite loop when floating point registers get
1260    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1261    110.37 and 110.38.)
1262    There was another set of potential problems related to the handling of
1263    MLRISC annotations (but those did not yet cause real problems, apparently).
1264    
1265    ----------------------------------------------------------------------
1266    Name: Matthias Blume
1267    Date: 2002/12/06 22:40:00 EST
1268    Tag: blume-20021206-cm-fileid
1269    Description:
1270    
1271    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1272    This fixes the problem of CM getting confused by files that suddenly
1273    change their identity (e.g., by getting unlinked and recreated by some
1274    text editor such as vi).  There might be a better/cheaper/cleaner way
1275    of doing this, but for now this will have to do.
1276    
1277    ----------------------------------------------------------------------
1278    Name: Matthias Blume
1279    Date: 2002/10/28 09:50:00 EST
1280    Tag: blume-20021028-typecheck
1281    Description:
1282    
1283    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1284    
1285    ----------------------------------------------------------------------
1286    Name: Matthias Blume
1287    Date: 2002/10/17 09:10:00 EDT
1288    Tag: Release_110_42
1289    Description:
1290    
1291    In good old tradition, there has been a slight hiccup so that we have
1292    to patch 110.42 after the fact.   The old release tag has been replaced
1293    (see below).
1294    
1295    The change solves a problem with two competing approaches the
1296    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1297    each other's way.
1298    
1299    This change only affects the runtime system code and the installer script.
1300    (No new bootfiles.)
1301    
1302    ----------------------------------------------------------------------
1303    Name: Matthias Blume
1304    Date: 2002/10/16 12:00:00 EDT
1305    Tag: Release_110_42_removed
1306    Description:
1307    
1308    New working release.  New bootfiles.
1309    
1310    ----------------------------------------------------------------------
1311    Name: Matthias Blume
1312    Date: 2002/10/10 13:10:00 EDT
1313    Tag: blume-20021010-ppc-divs
1314    Description:
1315    
1316    The mltree operator DIVS must be implemented with an overflow check on
1317    the PPC because the hardware indicates divide-by-zero using "overflow" as
1318    well.
1319    
1320    ----------------------------------------------------------------------
1321    Name: Matthias Blume
1322    Date: 2002/07/23 11:45:00
1323    Tag: blume-20020723-smlnj-home
1324    Description:
1325    
1326    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1327    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1328    CM_PATHCONFIG is also set), the path configuration file is assumed
1329    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1330    the entire tree to some other place and everything will "just work".
1331    
1332    (Companion commands such as ml-build and ml-makedepend also sense this
1333    variable.)
1334    
1335    ----------------------------------------------------------------------
1336    Name: Matthias Blume
1337    Date: 2002/07/12 21:19:00 EDT
1338    Tag: blume-20020712-liveness
1339    Description:
1340    
1341    Exported two useful "step" functions from liveness module (MLRISC).
1342    
1343    ----------------------------------------------------------------------
1344    Name: Matthias Blume
1345    Date: 2002/07/05 16:00 EDT
1346    Tag: Release_110_41
1347    Description:
1348    
1349    New working release.  New bootfiles.
1350    
1351    ----------------------------------------------------------------------
1352    Name: Matthias Blume
1353    Date: 2002/07/05 10:25:00 EDT
1354    Tag: blume-20020705-btimp
1355    Description:
1356    
1357    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1358    other clients can set up backtracing support.
1359    
1360    ----------------------------------------------------------------------
1361    Name: Matthias Blume
1362    Date: 2002/06/25 17:23:00 EDT
1363    Tag: blume-20020625-fpmax
1364    Description:
1365    
1366    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1367    case.  (The sense of the isNaN test was reversed -- which made min and
1368    max always return their first argument.)
1369    
1370    ----------------------------------------------------------------------
1371    Name: Matthias Blume
1372    Date: 2002/06/11
1373    Tag: blume-20020611-unixpath
1374    Description:
1375    
1376    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1377    
1378    ----------------------------------------------------------------------
1379    Name: Matthias Blume
1380    Date: 2002/06/10 16:35:00 EDT
1381    Tag: blume-20020610-ieeereal
1382    Description:
1383    
1384    I back-ported my implementation of IEEEReal.fromString from the
1385    idlbasis-devel branch so that we can test it.
1386    
1387    Another small change is that ppDec tries to give more information
1388    than just "<sig>" in the case of functors.  However, this code is
1389    broken in some mysterious way if the functor's body's signature
1390    has not been declared by ascription but gets inferred from the
1391    implementation.  This needs fixing...
1392    
1393    ----------------------------------------------------------------------
1394    Name: Matthias Blume
1395    Date: 2002/05/31
1396    Tag: blume-20020531-btrace-mode
1397    Description:
1398    
1399    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1400    the wayside when I switched over to using Controls everywhere.)
1401    
1402    ----------------------------------------------------------------------
1403    Name: Lal George
1404    Date: 2002/05/23 12:21:40 EDT
1405    Tag: george-20020523-visual-labels
1406    Description:
1407    
1408    Labels are now displayed in the graphical output to make
1409    the fall-through and target blocks obvious.
1410    
1411    ----------------------------------------------------------------------
1412    Name: Matthias Blume
1413    Date: 2002/05/22 11:03:00 EDT
1414    Tag: blume-20020522-shrink
1415    Description:
1416    
1417    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1418    situation that comes up when allocating huge arrays.
1419    
1420    ----------------------------------------------------------------------
1421    Name: Matthias Blume
1422    Date: 2002/05/21 16:00:00 EDT
1423    Tag: Release_110_40
1424    Description:
1425    
1426    New working release (110.40).  New bootfiles.
1427    
1428    [Also: John Reppy fixed GC bug 1131.]
1429    
1430    ----------------------------------------------------------------------
1431    Name: Matthias Blume
1432    Date: 2002/05/21 12:35:00 EDT
1433    Tag: blume-20020521-cmdoc
1434    Description:
1435    
1436    CM documentation update.
1437    
1438    ----------------------------------------------------------------------
1439    Name: Matthias Blume
1440    Date: 2002/05/21 10:55:00 EDT
1441    Tag: blume-20020521-misc
1442    Description:
1443    
1444    - John tweaked runtime to be silent on heap export (except when
1445      GC messages are on).
1446    - I added a few more things (cross-compiling versions of CMB) to
1447      config/preloads (as suggestions).
1448    
1449    ----------------------------------------------------------------------
1450    Name: Matthias Blume
1451    Date: 2002/05/20 22:25:00 EDT
1452    Tag: blume-20020520-controls
1453    Description:
1454    
1455    - Added ControlUtil structure to control-lib.cm.
1456    - Use it throughout.
1457    - Used Controls facility to define MLRISC controls (as opposed to
1458      registering MLRISC control ref cells with Controls after the fact)
1459    - Fixed messed-up controls priorities.
1460    
1461    * Removed again all the stuff from config/preloads that one wouldn't
1462      be able to preload at the time the initial heap image is built.
1463      (Many libraries, e.g., CML, do not exist yet at this time.  The
1464       only libraries that can be preloaded via config/preloads are those
1465       that come bundled with the bootfiles.)
1466    
1467    ----------------------------------------------------------------------
1468    Name: Matthias Blume
1469    Date: 2002/05/20 10:59:00 EDT
1470    Tag: blume-20020520-preloads
1471    Description:
1472    
1473    Added a lot of commented-out suggestions for things to be included
1474    in config/preloads.
1475    
1476    ----------------------------------------------------------------------
1477    Name: Allen Leung
1478    Date: 2002/05/18 14:20:00 EDT
1479    Tag: leunga-20020518-mdl
1480    Description:
1481    
1482    o Made the mdl tool stuff compile and run again.
1483    o I've disabled all the stuff that depends on RTL specifications; they
1484      are all badly broken anyway.
1485    
1486    ----------------------------------------------------------------------
1487    Name: Matthias Blume
1488    Date: 2002/05/17 16:49:00 EDT
1489    Tag: blume-20020517-controls
1490    Description:
1491    
1492    0. John Reppy made several modifications to the SML/NJ library.
1493       In particular, there is a shiny new controls-lib.cm.
1494    
1495    1. Pushed new controls interface through compiler so that everything
1496       compiles again.
1497    
1498    2. Added FormatComb and FORMAT_COMB to the CML version of the
1499       SML/NJ library (so that CML compiles again).
1500    
1501    3. Modified init scripts because XXX_DEFAULT environment variables
1502       are no longer with us.  (Boot-time initialization is now done
1503       using the same environment variables that are also used for
1504       startup-time initialization of controls.)
1505    
1506    ----------------------------------------------------------------------
1507    Name: Lal George
1508    Date: 2002/05/15 09:20:10 EDT
1509    Tag: george-20020515-pseudo-op-decls
1510    Description:
1511    
1512            All pseudo-ops emitted before the first segment declaration
1513    such as TEXT, DATA, and BSS directives are assumed to be global
1514    declarations and are emitted first in the assembly file. This is
1515    useful in a number of situations where one has pseudo-ops that are not
1516    specific to any segment, and also works around the constraint that one
1517    cannot have client pseudo-ops in the TEXT segment.
1518    
1519            Because no segment is associated with these declarations it is
1520    an error to allocate any space or objects before the first segment
1521    directive and an exception will be raised. However, we cannot make
1522    this check for client pseudo-ops.
1523    
1524            These top level declarations are a field in the CFG graph_info.
1525    In theory you can continue to add to this field after the CFG has been
1526    built -- provided you know what you are doing;-)
1527    
1528    ----------------------------------------------------------------------
1529    Name: Matthias Blume
1530    Date: 2002/05/13 16:40:00 EDT
1531    Tag: blume-20020513-pp-etc
1532    Description:
1533    
1534    A few minor bugfixes:
1535    
1536      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1537        (Bogus printouts for redefined bindings still occur.  Compiler
1538        bug should no longer occur now.  We need to redo the prettyprinter
1539        from scratch.)
1540    
1541      - CM pathname printer now also adds escape sequences for ( and )
1542    
1543      - commend and docu fixes for ml-nlffi
1544    
1545    ----------------------------------------------------------------------
1546    Name: Matthias Blume
1547    Date: 2002/05/10 16:40:00 EDT
1548    Tag: blume-20020510-erg-textio
1549    Description:
1550    
1551    Applied the following bugfix provided by Emden Gansner:
1552    
1553        Output is corrupted when outputSubstr is used rather than output.
1554    
1555        The problem occurs when a substring
1556    
1557            ss = (s, dataStart, dataLen)
1558    
1559        where dataStart > 0, fills a stream buffer with avail bytes left.
1560        avail bytes of s, starting at index dataStart, are copied into the
1561        buffer, the buffer is flushed, and then the remaining dataLen-avail
1562        bytes of ss are copied into the beginning of the buffer. Instead of
1563        starting this copy at index dataStart+avail in s, the current code
1564        starts the copy at index avail.
1565    
1566        Fix:
1567        In text-io-fn.sml, change line 695 from
1568             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1569        to
1570             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1571    
1572    ----------------------------------------------------------------------
1573    Name: Matthias Blume
1574    Date: 2002/04/12 13:55:00 EDT
1575    Tag: blume-20020412-assyntax
1576    Description:
1577    
1578    1. Grabbed newer assyntax.h from the XFree86 project.
1579    2. Fiddled with how to compile X86.prim.asm without warnings.
1580    3. (Very) Minor cleanup in CM.
1581    
1582    ----------------------------------------------------------------------
1583    Name: Matthias Blume
1584    Date: 2002/04/01 (no joke!) 17:07:00 EST
1585    Tag: blume-20020401-x86div
1586    Description:
1587    
1588    Added full support for div/mod/rem/quot on the x86, using the machine
1589    instruction's two results (without clumsily recomputing the remainder)
1590    directly where appropriate.
1591    
1592    Some more extensive power-of-two support was added to the x86 instruction
1593    selector (avoiding expensive divs, mods, and muls where they can be
1594    replaced with cheaper shifts and masks).  However, this sort of thing
1595    ought to be done earlier, e.g., within the CPS optimizer so that
1596    all architectures benefit from it.
1597    
1598    The compiler compiles to a fixed point, but changes might be somewhat
1599    fragile nevertheless.  Please, report any strange things that you might
1600    see wrt. div/mod/quot/rem...
1601    
1602    ----------------------------------------------------------------------
1603    Name: Matthias Blume
1604    Date: 2002/03/29 17:22:00
1605    Tag: blume-20020329-div
1606    Description:
1607    
1608    Fixed my broken div/mod logic.  Unfortunately, this means that the
1609    inline code for div/mod now has one more comparison than before.
1610    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
1611    The problem was with quotient = 0, because that alone does not tell
1612    us which way the rounding went.  One then has to look at whether
1613    remainder and divisor have the same sign...  :(
1614    
1615    Anyway, I replaced the bootfiles with fresh ones...
1616    
1617    ----------------------------------------------------------------------
1618    Name: Matthias Blume
1619    Date: 2002/03/29 14:10:00 EST
1620    Tag: blume-20020329-inlprims
1621    Description:
1622    
1623    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
1624    
1625    Primops have changed. This means that the bin/boot-file formats have
1626    changed as well.
1627    
1628    To make sure that there is no confusion, I made a new version.
1629    
1630    
1631    CHANGES:
1632    
1633    * removed REMT from mltree (remainder should never overflow).
1634    
1635    * added primops to deal with divisions of all flavors to the frontend
1636    
1637    * handled these primops all the way through so they map to their respective
1638      MLRISC support
1639    
1640    * used these primops in the implementation of Int, Int32, Word, Word32
1641    
1642    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
1643    
1644    * parameterized INLMIN, INLMAX, and INLABS by a numkind
1645    
1646    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
1647      floating point
1648    
1649    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
1650      and Real (but Real.abs maps to a separate floating-point-only primop)
1651    
1652    
1653    TODO items:
1654    
1655    * Hacked Alpha32 instruction selection, disabling the selection of REMx
1656      instructions because the machine instruction encoder cannot handle
1657      them.  (Hppa, PPC, and Sparc instruction selection did not handle
1658      REM in the first place, and REM is supported by the x86 machine coder.)
1659    
1660    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
1661      selection phase.  (The two can be streamlined because the hardware
1662      delivers both quotient and remainder at the same time anyway.)
1663    
1664    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
1665      (Currently the behavior is inconsistent both across architectures and
1666      wrt. the draft Basis spec.)
1667    
1668    * Word8 should eventually be handled natively, too.
1669    
1670    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
1671      as if there currently is no execution path that could trigger it in
1672      SML/NJ.  (The assumptions underlying functions arith and promotable do not
1673      hold for things like multiplication and division.)
1674    
1675    ----------------------------------------------------------------------
1676    Name: Matthias Blume
1677    Date: 2002/03/27 16:27:00 EST
1678    Tag: blume-20020327-mlrisc-divisions
1679    Description:
1680    
1681    Added support for all four division operations (ML's div, mod, quot,
1682    and rem) to MLRISC.  In the course of doing so, I also rationalized
1683    the naming (no more annoying switch-around of DIV and QUOT), by
1684    parameterizing the operation by div_rounding_mode (which can be either
1685    DIV_TO_ZERO or DIV_TO_NEGINF).
1686    
1687    The generic MLTreeGen functor takes care of compiling all four
1688    operations down to only round-to-zero div.
1689    
1690    Missing pieces:
1691    
1692      * Doing something smarter than relying on MLTreeGen on architectures
1693        like, e.g., the x86 where hardware division delivers both quotient and
1694        remainder at the same time.  With this, the implementation of the
1695        round-to-neginf operations could be further streamlined.
1696    
1697      * Remove inlining support for div/mod/rem from the frontend and replace it
1698        with primops that get carried through to the backend.  Do this for all
1699        int and word types.
1700    
1701    ----------------------------------------------------------------------
1702    Name: Matthias Blume
1703    Date: 2002/03/25 17:25:00 EST
1704    Tag: blume-20020325-divmod
1705    Description:
1706    
1707    I improved (hopefully without breaking them) the implementation of Int.div,
1708    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
1709    advantage of the following observations:
1710    
1711      Let  q = x quot y      r = x rem y
1712           d = x div  y      m = x mod y
1713    
1714    where "quot" is the round-to-zero version of integer division that
1715    hardware usually provides.  Then we have:
1716    
1717         r = x - q * y        where neither the * nor the - will overflow
1718         d = if q >= 0 orelse x = q * y then q else q - 1
1719                              where neither the * nor the - will overflow
1720         m = if q >= 0 orelse r = 0 then r else r + y
1721                              where the + will not overflow
1722    
1723    This results in substantial simplification of the generated code.
1724    The following table shows the number of CFG nodes and edges generated
1725    for
1726            fun f (x, y) = x OPER y
1727            (* with OPER \in div, mod, quot, rem *)
1728    
1729    
1730        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
1731        --------------------------------------------------------
1732         div |         24 |         39 |         12 |         16
1733         mod |         41 |         71 |         12 |         16
1734        quot |          8 |         10 |          8 |         10
1735         rem |         10 |         14 |          8 |         10
1736    
1737    
1738    ----------------------------------------------------------------------
1739    Name: Matthias Blume
1740    Date: 2002/03/25 22:06:00 EST
1741    Tag: blume-20020325-cprotobug
1742    Description:
1743    
1744    Fixed a bug in cproto (c prototype decoder).
1745    
1746    ----------------------------------------------------------------------
1747    Name: Matthias Blume
1748    Date: 2002/03/25 16:00:00 EST
1749    Tag: blume-20020325-raw-primops
1750    Description:
1751    
1752    I did some cleanup to Allen's new primop code and
1753    replaced yesterday's bootfiles with new ones.
1754    (But they are stored in the same place.)
1755    
1756    ----------------------------------------------------------------------
1757    Name: Matthias Blume
1758    Date: 2002/03/24 22:40:00 EST
1759    Tag: blume-20020324-bootfiles
1760    Description:
1761    
1762    Made the bootfiles that Allen asked for.
1763    
1764    ----------------------------------------------------------------------
1765    Name: Allen Leung
1766    Date: 2002/03/23 15:50:00 EST
1767    Tag: leunga-20020323-flint-cps-rcc-primops
1768    Description:
1769    
1770      1. Changes to FLINT primops:
1771    
1772        (* make a call to a C-function;
1773         * The primop carries C function prototype information and specifies
1774         * which of its (ML-) arguments are floating point. C prototype
1775         * information is for use by the backend, ML information is for
1776         * use by the CPS converter. *)
1777      | RAW_CCALL of { c_proto: CTypes.c_proto,
1778                       ml_args: ccall_type list,
1779                       ml_res_opt: ccall_type option,
1780                       reentrant : bool
1781                     } option
1782       (* Allocate uninitialized storage on the heap.
1783        * The record is meant to hold short-lived C objects, i.e., they
1784        * are not ML pointers.  With the tag, the representation is
1785        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1786        *)
1787      | RAW_RECORD of {tag:bool,sz:int}
1788      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1789    
1790      2.  These CPS primops are now overloaded:
1791    
1792           rawload of {kind:numkind}
1793           rawstore of {kind:numkind}
1794    
1795          The one argument form is:
1796    
1797             rawload {kind} address
1798    
1799          The two argument form is:
1800    
1801             rawload {kind} [ml object, byte-offset]
1802    
1803      3. RAW_CCALL/RCC now takes two extra arguments:
1804    
1805         a. The first is whether the C call is reentrant, i.e., whether
1806            ML state should be saved and restored.
1807         b. The second argument is a string argument specifying the name of
1808            library and the C function.
1809    
1810         These things are currently not handled in the code generator, yet.
1811    
1812      4. In CProto,
1813    
1814         An encoding type of "bool" means "ml object" and is mapped into
1815         C prototype of PTR.  Note that "bool" is different than "string",
1816         even though "string" is also mapped into PTR, because "bool"
1817         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
1818    
1819      5. Pickler/unpicker
1820    
1821         Changed to handle RAW_RECORD and newest RAW_CCALL
1822    
1823      6. MLRiscGen,
1824    
1825         1. Changed to handle the new rawload/rawstore/rawrecord operators.
1826         2. Code for handling C Calls has been moved to a new module CPSCCalls,
1827            in the file CodeGen/cpscompile/cps-c-calls.sml
1828    
1829      7. Added the conditional move operator
1830    
1831             condmove of branch
1832    
1833         to cps.  Generation of this is still buggy so it is currently
1834         disabled.
1835    
1836    ----------------------------------------------------------------------
1837    Name: Lal George
1838    Date: 2002/03/22 14:18:25 EST
1839    Tag: george-20020322-cps-branch-prob
1840    Description:
1841    
1842    Implemented the Ball-Larus branch prediction-heuristics, and
1843    incorporated graphical viewers for control flow graphs.
1844    
1845    Ball-Larus Heuristics:
1846    ---------------------
1847    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
1848    
1849    By design it uses the Dempster-Shafer theory for combining
1850    probabilities.  For example, in the function:
1851    
1852        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
1853    
1854    the ball-larus heuristics predicts that the n=0 is unlikely
1855    (OH-heuristic), and the 'then' branch is unlikely because of the
1856    RH-heuristic -- giving the 'then' branch an even lower combined
1857    probability using the Dempster-Shafer theory.
1858    
1859    Finally, John Reppy's loop analysis in MLRISC, further lowers the
1860    probability of the 'then' branch because of the loop in the else
1861    branch.
1862    
1863    
1864    Graphical Viewing:
1865    ------------------
1866    I merely plugged in Allen's graphical viewers into the compiler. The
1867    additional code is not much. At the top level, saying:
1868    
1869            Control.MLRISC.getFlag "cfg-graphical-view" := true;
1870    
1871    will display the graphical view of the control flow graph just before
1872    back-patching.  daVinci must be in your path for this to work. If
1873    daVinci is not available, then the default viewer can be changed
1874    using:
1875    
1876            Control.MLRISC.getString "viewer"
1877    
1878    which can be set to "dot" or "vcg" for the corresponding viewers. Of
1879    course, these viewers must be in your path.
1880    
1881    The above will display the compilation unit at the level of clusters,
1882    many of which are small, boring, and un-interesting. Also setting:
1883    
1884            Control.MLRISC.getInt "cfg-graphical-view_size"
1885    
1886    will display clusters that are larger than the value set by the above.
1887    
1888    
1889    ----------------------------------------------------------------------
1890    Name: Matthias Blume
1891    Date: 2002/03/21 22:20:00 EST
1892    Tag: blume-20020321-kmp-bugfix
1893    Description:
1894    
1895    Changed the interface to the KMP routine in PreString and fixed
1896    a minor bug in one place where it was used.
1897    
1898    ----------------------------------------------------------------------
1899    Name: Allen Leung
1900    Date: 2002/03/21 20:30:00 EST
1901    Tag: leunga-20020321-cfg
1902    Description:
1903    
1904      Fixed a potential problem in cfg edge splitting.
1905    
1906    ----------------------------------------------------------------------
1907    Name: Allen Leung
1908    Date: 2002/03/21 17:15:00 EST
1909    Tag: leunga-20020321-x86-fp-cfg
1910    Description:
1911    
1912      1. Recoded the buggy parts of x86-fp.
1913    
1914         a. All the block reordering code has been removed.
1915            We now depend on the block placement phases to do this work.
1916    
1917         b. Critical edge splitting code has been simplified and moved into the
1918            CFG modules, as where they belong.
1919    
1920         Both of these were quite buggy and complex.  The code is now much, much
1921         simpler.
1922    
1923      2. X86 backend.
1924    
1925         a. Added instructions for 64-bit support.  Instruction selection for
1926            64-bit has not been committed, however, since that
1927            requires changes to MLTREE which haven't been approved by
1928            Lal and John.
1929    
1930         b. Added support for FUCOMI and FUCOMIP when generating code for
1931            PentiumPro and above.  We only generate these instructions in
1932            the fast-fp mode.
1933    
1934         c. Added cases for JP and JNP in X86FreqProps.
1935    
1936      3. CFG
1937    
1938         CFG now has a bunch of methods for edge splitting and merging.
1939    
1940      4. Machine description.
1941    
1942         John's simplification of MLTREE_BASIS.fcond broke a few machine
1943         description things:
1944    
1945         rtl-build.{sig,sml} and hppa.mdl fixed.
1946    
1947         NOTE: the machine description stuff in the repository is still broken.
1948               Again, I can't put my fixes in because that involves
1949               changes to MLTREE.
1950    
1951    ----------------------------------------------------------------------
1952    Name: Matthias Blume
1953    Date: 2002/03/20 15:55:00 EST
1954    Tag: blume-20020320-kmp
1955    Description:
1956    
1957    Implemented Knuth-Morris-Pratt string matching in PreString and used
1958    it for String.isSubstring, Substring.isSubstring, and
1959    Substring.position.
1960    
1961    (Might need some stress-testing.  Simple examples worked fine.)
1962    
1963    ----------------------------------------------------------------------
1964    Name: Matthias Blume
1965    Date: 2002/03/19 16:37:00 EST
1966    Tag: blume-20020319-witnesses
1967    Description:
1968    
1969    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
1970    
1971    This implements a generic mechanism for changing constness qualifiers
1972    anywhere within big C types without resorting to outright "casts".
1973    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
1974    modify the constness at the outermost level.)
1975    The implementation of "convert" is based on the idea of "witness"
1976    values -- values that are not used by the operation but whose types
1977    "testify" to their applicability.  On the implementation side, "convert"
1978    is simply a projection (returning its second curried argument).  With
1979    cross-module inlining, it should not result in any machine code being
1980    generated.
1981    
1982    ----------------------------------------------------------------------
1983    Name: Matthias Blume
1984    Date: 2002/03/15 16:40:00 EST
1985    Tag: blume-20020315-basis
1986    Description:
1987    
1988    Provided (preliminary?) implementations for
1989    
1990      {String,Substring}.{concatWith,isSuffix,isSubstring}
1991    
1992    and
1993    
1994      Substring.full
1995    
1996    Those are in the Basis spec but they were missing in SML/NJ.
1997    
1998    ----------------------------------------------------------------------
1999    Name: Matthias Blume
2000    Date: 2002/03/14 21:30:00 EST
2001    Tag: blume-20020314-controls
2002    Description:
2003    
2004    Controls:
2005    ---------
2006    
2007    1. Factored out the recently-added Controls : CONTROLS stuff and put
2008       it into its own library $/controls-lib.cm.  The source tree for
2009       this is under src/smlnj-lib/Controls.
2010    
2011    2. Changed the names of types and functions in this interface, so they
2012       make a bit more "sense":
2013    
2014          module -> registry
2015          'a registry -> 'a group
2016    
2017    3. The interface now deals in ref cells only.  The getter/setter interface
2018       is (mostly) gone.
2019    
2020    4. Added a function that lets one register an already-existing ref cell.
2021    
2022    5. Made the corresponding modifications to the rest of the code so that
2023       everything compiles again.
2024    
2025    6. Changed the implementation of Controls.MLRISC back to something closer
2026       to the original.  In particular, this module (and therefore MLRISC)
2027       does not depend on Controls.  There now is some link-time code in
2028       int-sys.sml that registers the MLRISC controls with the Controls
2029       module.
2030    
2031    CM:
2032    ---
2033    
2034      * One can now specify the lambda-split aggressiveness in init.cmi.
2035    
2036    ----------------------------------------------------------------------
2037    Name: Allen Leung
2038    Date: 2002/03/13 17:30:00 EST
2039    Tag: leunga-20020313-x86-fp-unary
2040    Description:
2041    
2042    Bug fix for:
2043    
2044    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2045    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2046    > - fun f(x,(y,z)) = Real.~ y;
2047    > [autoloading]
2048    > [autoloading done]
2049    >       fchsl   (%eax), 184(%esp)
2050    > Error: MLRisc bug: X86MCEmitter.emitInstr
2051    >
2052    > uncaught exception Error
2053    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2054    
2055    The problem was that the code generator did not generate any fp registers
2056    in this case, and the ra didn't know that it needed to run the X86FP phase to
2057    translate the pseudo fp instruction.   This only happened with unary fp
2058    operators in certain situations.
2059    
2060    ----------------------------------------------------------------------
2061    Name: Matthias Blume
2062    Date: 2002/03/13 14:00:00 EST
2063    Tag: blume-20020313-overload-etc
2064    Description:
2065    
2066    1. Added _overload as a synonym for overload for backward compatibility.
2067       (Control.overloadKW must be true for either version to be accepted.)
2068    
2069    2. Fixed bug in install script that caused more things to be installed
2070       than what was requested in config/targets.
2071    
2072    3. Made CM aware of the (_)overload construct so that autoloading
2073       works.
2074    
2075    ----------------------------------------------------------------------
2076    Name: Matthias Blume
2077    Date: 2002/03/12 22:03:00 EST
2078    Tag: blume-20020312-url
2079    Description:
2080    
2081    Forgot to update BOOT and srcarchiveurl.
2082    
2083    ----------------------------------------------------------------------
2084    Name: Matthias Blume
2085    Date: 2002/03/12 17:30:00 EST
2086    Tag: blume-20020312-version110392
2087    Description:
2088    
2089    Yet another version number bump (because of small changes to the
2090    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2091    
2092    Changes:
2093    
2094      The new pid generation scheme described a few weeks ago was overly
2095      complicated.  I implemented a new mechanism that is simpler and
2096      provides a bit more "stability":  Once CM has seen a compilation
2097      unit, it keeps its identity constant (as long as you do not delete
2098      those crucial CM/GUID/* files).  This means that when you change
2099      an interface, compile, then go back to the old interface, and
2100      compile again, you arrive at the original pid.
2101    
2102      There now also is a mechanism that instructs CM to use the plain
2103      environment hash as a module's pid (effectively making its GUID
2104      the empty string).  For this, "noguid" must be specified as an
2105      option to the .sml file in question within its .cm file.
2106      This is most useful for code that is being generated by tools such
2107      as ml-nlffigen (because during development programmers tend to
2108      erase the tool's entire output directory tree including CM's cached
2109      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2110      revert to the old, broken behavior of SML/NJ, but in specific cases
2111      where there is no danger of interface confusion, its use is ok
2112      (I think).
2113    
2114      ml-nlffigen by default generates "noguid" annotations.  They can be
2115      turned off by specifying -guid in its command line.
2116    
2117    ----------------------------------------------------------------------
2118    Name: Lal George
2119    Date: 2002/03/12 12 14:42:36 EST
2120    Tag: george-20020312-frequency-computation
2121    Description:
2122    
2123    Integrated jump chaining and static block frequency into the
2124    compiler. More details and numbers later.
2125    
2126    ----------------------------------------------------------------------
2127    Name: Lal George
2128    Date: 2002/03/11 11 22:38:53 EST
2129    Tag: george-20020311-jump-chain-elim
2130    Description:
2131    
2132    Tested the jump chain elimination on all architectures (except the
2133    hppa).  This is on by default right now and is profitable for the
2134    alpha and x86, however, it may not be profitable for the sparc and ppc
2135    when compiling the compiler.
2136    
2137    The gc test will typically jump to a label at the end of the cluster,
2138    where there is another jump to an external cluster containing the actual
2139    code to invoke gc. This is to allow factoring of common gc invocation
2140    sequences. That is to say, we generate:
2141    
2142            f:
2143               testgc
2144               ja   L1      % jump if above to L1
2145    
2146            L1:
2147               jmp L2
2148    
2149    
2150    After jump chain elimination the 'ja L1' instructions is converted to
2151    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2152    up being implemented in their long form (if L2 is far away) using:
2153    
2154            jbe     L3      % jump if below or equal to L3
2155            jmp     L2
2156         L3:
2157            ...
2158    
2159    
2160    For large compilation units L2  may be far away.
2161    
2162    
2163    ----------------------------------------------------------------------
2164    Name: Matthias Blume
2165    Date: 2002/03/11 13:30:00 EST
2166    Tag: blume-20020311-mltreeeval
2167    Description:
2168    
2169    A functor parameter was missing.
2170    
2171    ----------------------------------------------------------------------
2172    Name: Allen Leung
2173    Date: 2002/03/11 10:30:00 EST
2174    Tag: leunga-20020311-runtime-string0
2175    Description:
2176    
2177       The representation of the empty string now points to a
2178    legal null terminated C string instead of unit.  It is now possible
2179    to convert an ML string into C string with InlineT.CharVector.getData.
2180    This compiles into one single machine instruction.
2181    
2182    ----------------------------------------------------------------------
2183    Name: Allen Leung
2184    Date: 2002/03/10 23:55:00 EST
2185    Tag: leunga-20020310-x86-call
2186    Description:
2187    
2188       Added machine generation for CALL instruction (relative displacement mode)
2189    
2190    ----------------------------------------------------------------------
2191    Name: Matthias Blume
2192    Date: 2002/03/08 16:05:00
2193    Tag: blume-20020308-entrypoints
2194    Description:
2195    
2196    Version number bumped to 110.39.1.  NEW BOOTFILES!
2197    
2198    Entrypoints: non-zero offset into a code object where execution should begin.
2199    
2200    - Added the notion of an entrypoint to CodeObj.
2201    - Added reading/writing of entrypoint info to Binfile.
2202    - Made runtime system bootloader aware of entrypoints.
2203    - Use the address of the label of the first function given to mlriscGen
2204      as the entrypoint.  This address is currently always 0, but it will
2205      not be 0 once we turn on block placement.
2206    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2207      with entry points) from mlriscGen.
2208    
2209    ----------------------------------------------------------------------
2210    Name: Allen Leung
2211    Date: 2002/03/07 20:45:00 EST
2212    Tag: leunga-20020307-x86-cmov
2213    Description:
2214    
2215       Bug fixes for CMOVcc on x86.
2216    
2217       1. Added machine code generation for CMOVcc
2218       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2219       3. CMOVcc cannot have an immediate operand as argument.
2220    
2221    ----------------------------------------------------------------------
2222    Name: Matthias Blume
2223    Date: 2002/03/07 16:15:00 EST
2224    Tag: blume-20020307-controls
2225    Description:
2226    
2227    This is a very large but mostly boring patch which makes (almost)
2228    every tuneable compiler knob (i.e., pretty much everything under
2229    Control.* plus a few other things) configurable via both the command
2230    line and environment variables in the style CM did its configuration
2231    until now.
2232    
2233    Try starting sml with '-h' (or, if you are brave, '-H')
2234    
2235    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2236    implements the underlying generic mechanism.
2237    
2238    The interface to some of the existing such facilities has changed somewhat.
2239    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2240    (The getFoo interface is still there for backward-compatibility, but its
2241    use is deprecated.)
2242    
2243    The ml-build script passes -Cxxx=yyy command-line arguments through so
2244    that one can now twiddle the compiler settings when using this "batch"
2245    compiler.
2246    
2247    TODO items:
2248    
2249    We should go through and throw out all controls that are no longer
2250    connected to anything.  Moreover, we should go through and provide
2251    meaningful (and correct!) documentation strings for those controls
2252    that still are connected.
2253    
2254    Currently, multiple calls to Controls.new are accepted (only the first
2255    has any effect).  Eventually we should make sure that every control
2256    is being made (via Controls.new) exactly once.  Future access can then
2257    be done using Controls.acc.
2258    
2259    Finally, it would probably be a good idea to use the getter-setter
2260    interface to controls rather than ref cells.  For the time being, both
2261    styles are provided by the Controls module, but getter-setter pairs are
2262    better if thread-safety is of any concern because they can be wrapped.
2263    
2264    *****************************************
2265    
2266    One bug fix: The function blockPlacement in three of the MLRISC
2267    backpatch files used to be hard-wired to one of two possibilities at
2268    link time (according to the value of the placementFlag).  But (I
2269    think) it should rather sense the flag every time.
2270    
2271    *****************************************
2272    
2273    Other assorted changes (by other people who did not supply a HISTORY entry):
2274    
2275    1. the cross-module inliner now works much better (Monnier)
2276    2. representation of weights, frequencies, and probabilities in MLRISC
2277       changed in preparation of using those for weighted block placement
2278       (Reppy, George)
2279    
2280    ----------------------------------------------------------------------
2281    Name: Lal George
2282    Date: 2002/03/07 14:44:24 EST 2002
2283    Tag: george-20020307-weighted-block-placement
2284    
2285    Tested the weighted block placement optimization on all architectures
2286    (except the hppa) using AMPL to generate the block and edge frequencies.
2287    Changes were required in the machine properties to correctly
2288    categorize trap instructions. There is an MLRISC flag
2289    "weighted-block-placement" that can be used to enable weighted block
2290    placement, but this will be ineffective without block/edge
2291    frequencies (coming soon).
2292    
2293    
2294    ----------------------------------------------------------------------
2295    Name: Lal George
2296    Date: 2002/03/05 17:24:48 EST
2297    Tag: george-20020305-linkage-cluster
2298    
2299    In order to support the block placement optimization, a new cluster
2300    is generated as the very first cluster (called the linkage cluster).
2301    It contains a single jump to the 'real' entry point for the compilation
2302    unit. Block placement has no effect on the linkage cluster itself, but
2303    all the other clusters  have full freedom in the manner in which they
2304    reorder blocks or functions.
2305    
2306    On the x86 the typical linkage code that is generated is:
2307       ----------------------
2308            .align 2
2309       L0:
2310            addl    $L1-L0, 72(%esp)
2311            jmp     L1
2312    
2313    
2314            .align  2
2315       L1:
2316       ----------------------
2317    
2318    72(%esp) is the memory location for the stdlink register. This
2319    must contain the address of the CPS function being called. In the
2320    above example, it contains the address of  L0; before
2321    calling L1 (the real entry point for the compilation unit), it
2322    must contain the address for L1, and hence
2323    
2324            addl $L1-L0, 72(%esp)
2325    
2326    I have tested this on all architectures except the hppa.The increase
2327    in code size is of course negligible
2328    
2329    ----------------------------------------------------------------------
2330    Name: Allen Leung
2331    Date: 2002/03/03 13:20:00 EST
2332    Tag: leunga-20020303-mlrisc-tools
2333    
2334      Added #[ ... ] expressions to mlrisc tools
2335    
2336    ----------------------------------------------------------------------
2337    Name: Matthias Blume
2338    Date: 2002/02/27 12:29:00 EST
2339    Tag: blume-20020227-cdebug
2340    Description:
2341    
2342    - made types in structure C and C_Debug to be equal
2343    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2344    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2345    
2346    ----------------------------------------------------------------------
2347    Name: Matthias Blume
2348    Date: 2002/02/26 12:00:00 EST
2349    Tag: blume-20020226-ffi
2350    Description:
2351    
2352    1. Fixed a minor bug in CM's "noweb" tool:
2353       If numbering is turned off, then truly don't number (i.e., do not
2354       supply the -L option to noweb).  The previous behavior was to supply
2355       -L'' -- which caused noweb to use the "default" line numbering scheme.
2356       Thanks to Chris Richards for pointing this out (and supplying the fix).
2357    
2358    2. Once again, I reworked some aspects of the FFI:
2359    
2360       A. The incomplete/complete type business:
2361    
2362       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2363         gone!
2364       - ML types representing an incomplete type are now *equal* to
2365         ML types representing their corresponding complete types (just like
2366         in C).  This is still safe because ml-nlffigen will not generate
2367         RTTI for incomplete types, nor will it generate functions that
2368         require access to such RTTI.   But when ML code generated from both
2369         incomplete and complete versions of the C type meet, the ML types
2370         are trivially interoperable.
2371    
2372         NOTE:  These changes restore the full generality of the translation
2373         (which was previously lost when I eliminated functorization)!
2374    
2375       B. Enum types:
2376    
2377       - Structure C now has a type constructor "enum" that is similar to
2378         how the "su" constructor works.  However, "enum" is not a phantom
2379         type because each "T enum" has values (and is isomorphic to
2380         MLRep.Signed.int).
2381       - There are generic access operations for enum objects (using
2382         MLRep.Signed.int).
2383       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2384         * The structure contains the definition of type "mlrep" (the ML-side
2385         representation type of the enum).  Normally, mlrep is the same
2386         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2387         then mlrep will be defined as a datatype -- thus facilitating
2388         pattern matching on mlrep values.
2389         ("-ec" will be suppressed if there are duplicate values in an
2390          enumeration.)
2391         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2392         will be generated for each C enum constant xxx.
2393         * Conversion functions m2i and i2m convert between mlrep and
2394         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2395         * Coversion functions c and ml convert between mlrep and "tag enum".
2396         * Access functions (get/set) fetch and store mlrep values.
2397       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2398         enumerations are merged into one single enumeration represented by
2399         structure E_'.
2400    
2401    ----------------------------------------------------------------------
2402    Name: Allen Leung
2403    Date: 2002/02/25 04:45:00 EST
2404    Tag: leunga-20020225-cps-spill
2405    
2406    This is a new implementation of the CPS spill phase.
2407    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2408    In case of problems, replace it with the old file spill.sml
2409    
2410    The current compiler runs into some serious performance problems when
2411    constructing a large record.  This can happen when we try to compile a
2412    structure with many items.  Even a very simple structure like the following
2413    makes the compiler slow down.
2414    
2415        structure Foo = struct
2416           val x_1 = 0w1 : Word32.int
2417           val x_2 = 0w2 : Word32.int
2418           val x_3 = 0w3 : Word32.int
2419           ...
2420           val x_N = 0wN : Word32.int
2421        end
2422    
2423    The following table shows the compile time, from N=1000 to N=4000,
2424    with the old compiler:
2425    
2426    N
2427    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2428           MLRISC ra                               0.06u  0.00s  0.05g
2429              (spills = 0 reloads = 0)
2430           TOTAL                                   0.63u  0.07s  0.21g
2431    
2432    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2433           MLRISC ra                               5.68u  0.59s  3.93g
2434              (spills = 0 reloads = 0)
2435           TOTAL                                   14.71u  0.99s  4.81g
2436    
2437    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2438           MLRISC ra                               5.54u  0.65s  3.91g
2439              (spills = 543 reloads = 1082)
2440           TOTAL                                   65.40u  3.13s  6.00g
2441    
2442    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2443           MLRISC ra                               0.80u  0.10s  0.55g
2444              (spills = 42 reloads = 84)
2445           TOTAL                                   129.42u  5.10s  4.13g
2446    
2447    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2448           MLRISC ra                               2.69u  0.27s  1.38g
2449              (spills = 62 reloads = 124)
2450           TOTAL                                   682.48u  19.61s  13.99g
2451    
2452    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2453           MLRISC ra                               4.96u  0.27s  2.72g
2454              (spills = 85 reloads = 170)
2455           TOTAL                                   2375.26u  57.21s  48.00g
2456    
2457    As you can see the old cps spill module suffers from some serious
2458    performance problem.  But since I cannot decipher the old code fully,
2459    instead of patching the problems up, I'm reimplementing it
2460    with a different algorithm.  The new code is more modular,
2461    smaller when compiled, and substantially faster
2462    (O(n log n) time and O(n) space).  Timing of the new spill module:
2463    
2464    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2465          MLRISC ra                               0.25u  0.02s  0.15g
2466             (spills=1 reloads=3)
2467          TOTAL                                   7.74u  0.34s  1.62g
2468    
2469    Implementation details:
2470    
2471    As far as I can tell, the purpose of the CPS spill module is to make sure the
2472    number of live variables at any program point (the bandwidth)
2473    does not exceed a certain limit, which is determined by the
2474    size of the spill area.
2475    
2476    When the bandwidth is too large, we decrease the register pressure by
2477    packing live variables into spill records.  How we achieve this is
2478    completely different than what we did in the old code.
2479    
2480    First, there is something about the MLRiscGen code generator
2481    that we should be aware of:
2482    
2483    o MLRiscGen performs code motion!
2484    
2485       In particular, it will move floating point computations and
2486       address computations involving only the heap pointer to
2487       their use sites (if there is only a single use).
2488       What this means is that if we have a CPS record construction
2489       statement
2490    
2491           RECORD(k,vl,w,e)
2492    
2493       we should never count the new record address w as live if w
2494       has only one use (which is often the case).
2495    
2496       We should do something similar to floating point, but the transformation
2497       there is much more complex, so I won't deal with that.
2498    
2499    Secondly, there are now two new cps primops at our disposal:
2500    
2501     1. rawrecord of record_kind option
2502        This pure operator allocates some uninitialized storage from the heap.
2503        There are two forms:
2504    
2505         rawrecord NONE [INT n]  allocates a tagless record of length n
2506         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2507                                     and initializes the tag.
2508    
2509     2. rawupdate of cty
2510          rawupdate cty (v,i,x)
2511          Assigns to x to the ith component of record v.
2512          The storelist is not updated.
2513    
2514    We use these new primops for both spilling and increment record construction.
2515    
2516     1. Spilling.
2517    
2518        This is implemented with a linear scan algorithm (but generalized
2519        to trees).  The algorithm will create a single spill record at the
2520        beginning of the cps function and use rawupdate to spill to it,
2521        and SELECT or SELp to reload from it.  So both spills and reloads
2522        are fine-grain operations.  In contrast, in the old algorithm
2523        "spills" have to be bundled together in records.
2524    
2525        Ideally, we should sink the spill record construction to where
2526        it is needed.  We can even split the spill record into multiple ones
2527        at the places where they are needed.  But CPS is not a good
2528        representation for global code motion, so I'll keep it simple and
2529        am not attempting this.
2530    
2531     2. Incremental record construction (aka record splitting).
2532    
2533        Long records with many component values which are simulatenously live
2534        (recall that single use record addresses are not considered to
2535         be live) are constructed with rawrecord and rawupdate.
2536        We allocate space on the heap with rawrecord first, then gradually
2537        fill it in with rawupdate.  This is the technique suggested to me
2538        by Matthias.
2539    
2540        Some restrictions on when this is applicable:
2541        1. It is not a VECTOR record.  The code generator currently does not handle
2542           this case. VECTOR record uses double indirection like arrays.
2543        2. All the record component values are defined in the same "basic block"
2544           as the record constructor.  This is to prevent speculative
2545           record construction.
2546    
2547    ----------------------------------------------------------------------
2548    Name: Allen Leung
2549    Date: 2002/02/22 01:02:00 EST
2550    Tag: leunga-20020222-mlrisc-tools
2551    
2552    Minor bug fixes in the parser and rewriter
2553    
2554    ----------------------------------------------------------------------
2555    Name: Allen Leung
2556    Date: 2002/02/21 20:20:00 EST
2557    Tag: leunga-20020221-peephole
2558    
2559    Regenerated the peephole files.  Some contained typos in the specification
2560    and some didn't compile because of pretty printing bugs in the old version
2561    of 'nowhere'.
2562    
2563    ----------------------------------------------------------------------
2564    Name: Allen Leung
2565    Date: 2002/02/19 20:20:00 EST
2566    Tag: leunga-20020219-mlrisc-tools
2567    Description:
2568    
2569       Minor bug fixes to the mlrisc-tools library:
2570    
2571       1.  Fixed up parsing colon suffixed keywords
2572       2.  Added the ability to shut the error messages up
2573       3.  Reimplemented the pretty printer and fixed up/improved
2574           the pretty printing of handle and -> types.
2575       4.  Fixed up generation of literal symbols in the nowhere tool.
2576       5.  Added some SML keywords to to sml.sty
2577    
2578    ----------------------------------------------------------------------
2579    Name: Matthias Blume
2580    Date: 2002/02/19 16:20:00 EST
2581    Tag: blume-20020219-cmffi
2582    Description:
2583    
2584    A wild mix of changes, some minor, some major:
2585    
2586    * All C FFI-related libraries are now anchored under $c:
2587        $/c.cm      --> $c/c.cm
2588        $/c-int.cm  --> $c/internals/c-int.cm
2589        $/memory.cm --> $c/memory/memory.cm
2590    
2591    * "make" tool (in CM) now treats its argument pathname slightly
2592      differently:
2593        1. If the native expansion is an absolute name, then before invoking
2594           the "make" command on it, CM will apply OS.Path.mkRelative
2595           (with relativeTo = OS.FileSys.getDir()) to it.
2596        2. The argument will be passed through to subsequent phases of CM
2597           processing without "going native".  In particular, if the argument
2598           was an anchored path, then "make" will not lose track of that anchor.
2599    
2600    * Compiler backends now "know" their respective C calling conventions
2601      instead of having to be told about it by ml-nlffigen.  This relieves
2602      ml-nlffigen from one of its burdens.
2603    
2604    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
2605    
2606    * Export C_DEBUG and C_Debug from $c/c.cm.
2607    
2608    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
2609      subtyping relationship between incomplete pointers and their complete
2610      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
2611      with the convention of instantiating 'o with ('t, 'c) obj whenever
2612      the pointer target type is complete.  In the incomplete case, 'o
2613      will be instantiated with some "'c iobj" -- a type obtained by
2614      using one of the functors PointerToIncompleteType or PointerToCompleteType.
2615    
2616      Operations that work on both incomplete and complete pointer types are
2617      typed as taking an 'o ptr while operations that require the target to
2618      be known are typed as taking some ('t, 'c) obj ptr.
2619    
2620      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
2621      where void is an eqtype without any values.  This makes it possible
2622      to work on voidptr values using functions meant to operate on light
2623      incomplete pointers.
2624    
2625    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
2626      been vastly simplified.
2627    
2628    ----------------------------------------------------------------------
2629    Name: Matthias Blume
2630    Date: 2002/02/19 10:48:00 EST
2631    Tag: blume-20020219-pqfix
2632    Description:
2633    
2634    Applied Chris Okasaki's bug fix for priority queues.
2635    
2636    ----------------------------------------------------------------------
2637    Name: Matthias Blume
2638    Date: 2002/02/15 17:05:00
2639    Tag: Release_110_39
2640    Description:
2641    
2642    Last-minute retagging is becoming a tradition... :-(
2643    
2644    This is the working release 110.39.
2645    
2646    ----------------------------------------------------------------------
2647    Name: Matthias Blume
2648    Date: 2002/02/15 16:00:00 EST
2649    Tag: Release_110_39-orig
2650    Description:
2651    
2652    Working release 110.39.  New bootfiles.
2653    
2654    (Update: There was a small bug in the installer so it wouldn't work
2655    with all shells.  So I retagged. -Matthias)
2656    
2657    ----------------------------------------------------------------------
2658    Name: Matthias Blume
2659    Date: 2002/02/15 14:17:00 EST
2660    Tag: blume-20020215-showbindings
2661    Description:
2662    
2663    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
2664    the latter can be useful for exploring what bindings are available at
2665    the interactive prompt.  (The first function returns only the list
2666    of symbols that are really bound, the second prints those but also the
2667    ones that CM's autoloading mechanism knows about.)
2668    
2669    ----------------------------------------------------------------------
2670    Name: Matthias Blume
2671    Date: 2002/02/15 12:08:00 EST
2672    Tag: blume-20020215-iptrs
2673    Description:
2674    
2675    Two improvements to ml-nlffigen:
2676    
2677      1. Write files only if they do not exist or if their current contents
2678         do not coincide with what's being written.  (That is, avoid messing
2679         with the time stamps unless absolutely necessary.)
2680    
2681      2. Implement a "repository" mechanism for generated files related
2682         to "incomplete pointer types".   See the README file for details.
2683    
2684    ----------------------------------------------------------------------
2685    Name: Matthias Blume
2686    Date: 2002/02/14 11:50:00 EST
2687    Tag: blume-20020214-quote
2688    Description:
2689    
2690    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
2691    because of the new and improved tag generation scheme.  (Thanks to Allen
2692    Leung for pointing it out.)
2693    
2694    ----------------------------------------------------------------------
2695    Name: Lal George
2696    Date: 2002/02/14 09:55:27 EST 2002
2697    Tag: george-20020214-isabelle-bug
2698    Description:
2699    
2700    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
2701    of Isabelle on the x86.
2702    
2703    From Allen:
2704    -----------
2705     I've found the problem:
2706    
2707         in ra-core.sml, I use the counter "blocked" to keep track of the
2708         true number of elements in the freeze queue.  When the counter goes
2709         to zero, I skip examining the queue.  But I've messed up the
2710         bookkeeping in combine():
2711    
2712             else ();
2713             case !ucol of
2714               PSEUDO => (if !cntv > 0 then
2715                     (if !cntu > 0 then blocked := !blocked - 1 else ();
2716                                        ^^^^^^^^^^^^^^^^^^^^^^^
2717                      moveu := mergeMoveList(!movev, !moveu)
2718                     )
2719                  else ();
2720    
2721         combine() is called to coalesce two nodes u and v.
2722         I think I was thinking that if the move counts of u and v are both
2723         greater than zero then after they are coalesced then one node is
2724         removed from the freeze queue.  Apparently I was thinking that
2725         both u and v are of low degree, but that's clearly not necessarily true.
2726    
2727    
2728    02/12/2002:
2729        Here's the patch.  HOL now compiles.
2730    
2731        I don't know how this impact on performance (compile
2732        time or runtime).  This bug caused the RA (especially on the x86)
2733        to go thru the potential spill phase when there are still nodes on the
2734        freeze queue.
2735    
2736    
2737    
2738    
2739    ----------------------------------------------------------------------
2740    Name: Matthias Blume
2741    Date: 2002/02/13 22:40:00 EST
2742    Tag: blume-20020213-fptr-rtti
2743    Description:
2744    
2745    Fixed a bug in ml-nlffigen that was introduced with one of the previous
2746    updates.
2747    
2748    ----------------------------------------------------------------------
2749    Name: Matthias Blume
2750    Date: 2002/02/13 16:41:00 EST
2751    Tag: blume-20020213-cmlpq
2752    Description:
2753    
2754    Added new priority queue export symbols (which have just been added to
2755    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
2756    not compile and the installer would choke.)
2757    
2758    ----------------------------------------------------------------------
2759    Name: Matthias Blume
2760    Date: 2002/02/13 16:15:00 EST
2761    Tag: blume-20020213-various
2762    Description:
2763    
2764    1. More tweaks to ml-nlffigen:
2765    
2766       - better internal datastructures (resulting in slight speedup)
2767       - "-match" option requires exact match
2768       - "localized" gensym counters (untagged structs/unions nested within
2769         other structs/unions or within typedefs get a fresh counter; their
2770         tag will be prefixed by a concatenation of their parents' tags)
2771       - bug fixes (related to calculation of transitive closure of types
2772         to be included in the output)
2773    
2774    2. Minor Basis updates:
2775    
2776       - added implementations for List.collate and Option.app
2777    
2778    ----------------------------------------------------------------------
2779    Name: Matthias Blume
2780    Date: 2002/02/11 15:55:00 EST
2781    Tag: blume-20020211-gensym
2782    Description:
2783    
2784    Added a "-gensym" option to command line of ml-nlffigen.  This can be
2785    used to specify a "stem" -- a string that is inserted in all "gensym'd"
2786    names (ML structure names that correspond to unnamed C structs, unions,
2787    and enums), so that separate runs of ml-nlffigen do not clash.
2788    
2789    ----------------------------------------------------------------------
2790    Name: Matthias Blume
2791    Date: 2002/02/11 12:05:00 EST
2792    Tag: blume-20020211-gensml
2793    Description:
2794    
2795    A quick fix for a problem with GenSML (in the pgraph-util library):
2796    Make generation of toplevel "local" optional.  (Strictly speaking,
2797    signature definitions within "local" are not legal SML.)
2798    
2799    Other than that: updates to INSTALL and cm/TODO.
2800    
2801    ----------------------------------------------------------------------
2802    Name: Matthias Blume
2803    Date: 2002/02/08 15:00:00 EST
2804    Tag: blume-20020208-uniquepid
2805    Description:
2806    
2807    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
2808    
2809    1. The installer (config/install.sh) has gotten smarter:
2810    
2811         - Configuration options are a bit easier to specify now
2812           (in config/targets).
2813         - Bug in recognizing .tar.bz2 files fixed.
2814         - Installer automatically resolves dependencies between
2815           configuration options (e.g., if you ask for eXene, you will
2816           also get cml -- regardless whether you asked for it or not).
2817         - Installer can run in "quieter mode" by setting the environment
2818           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
2819           "completely silent", though.
2820         - Build HashCons library as part of smlnj-lib.
2821    
2822    2. A new scheme for assigning persistent identifiers to compilation
2823       units (and, by extension, to types etc.) has been put into place.
2824       This fixes a long-standing bug where types and even dynamic values
2825       can get internally confused, thereby compromising type safety
2826       (abstraction) and dynamic correctness.  See
2827    
2828         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
2829    
2830       for an example of how things could go wrong until now.
2831    
2832       The downside of the new scheme is that pids are not quite as
2833       persistent as they used to be: CM will generate a fresh pid
2834       for every compilation unit that it thinks it sees for the first
2835       time.  That means that if you compile starting from a clean, fresh
2836       source tree at two different times, you end up with different
2837       binaries.
2838    
2839       Cutoff recompilation, however, has not been compromised because
2840       CM keeps pid information in special caches between runs.
2841    
2842    ----------------------------------------------------------------------
2843    Name: Lal George
2844    Date: 2002/02/07 15:34:13 EST 2002
2845    Tag: <none>
2846    Description:
2847    
2848    Compilers that generate assembly code may produce  global labels
2849    whose value is resolved at link time. The various peephole optimization
2850    modules did not take this in account.
2851    
2852    TODO. The Labels.addrOf function should really return an option
2853    type so that clients are forced to deal with this issue, rather
2854    than an exception being raised.
2855    
2856    ----------------------------------------------------------------------
2857    Name: Lal George
2858    Date: 2002/02/06 13:55:02 EST
2859    Tag: george-20020206-ra-breakup
2860    Description:
2861    
2862    1. A bug fix from Allen.
2863    
2864        A typo causes extra fstp %st(0)'s to be generated at compensation
2865        edges, which might cause stack underflow traps at runtime.  This
2866        occurs in fft where there are extraneous fstps right before the 'into'
2867        trap instruction (in this case they are harmless since none of the
2868        integers overflow.)
2869    
2870    2. Pulled out various utility modules that were embedded in the modules
2871       of the register allocator. I need these modules for other purposes, but
2872       they are not complete enough to put into a library (just yet).
2873    ----------------------------------------------------------------------
2874    Name: Matthias Blume
2875    Date: 2002/01/31 16:05:00 EST
2876    Tag: blume-20020131-sparc-ccalls
2877    Description:
2878    
2879    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
2880       of extra stack space by mistake.  Fixed.
2881    
2882    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
2883    
2884    ----------------------------------------------------------------------
2885    Name: Allen Leung
2886    Date: 2002/01/30
2887    Tag: leunga-20020130-nowhere-bug-fix
2888    Description:
2889    
2890       MLRISC bug fixes:
2891       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
2892       2. MachineInt.fromString was negating its value.
2893    
2894    ----------------------------------------------------------------------
2895    Name: Matthias Blume
2896    Date: 2002/01/29
2897    Tag: blume-20020129-INSTALL
2898    Description:
2899    
2900    - Added somewhat detailed installation instructions (file INSTALL).
2901    - Fixed curl-detection bug in config/install.sh.
2902    - It is now possible to select the URL getter using the URLGETTER
2903      environment variable:
2904    
2905          not set / "unknown"      --> automatic detection (script tries wget,
2906                                       curl, and lynx)
2907          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
2908                                       how to properly invoke them)
2909          other                    --> use $URLGETTER directly, it must take
2910                                       precisely two command-line arguments
2911                                       (source URL and destination file name)
2912    
2913    ----------------------------------------------------------------------
2914    Name: Matthias Blume
2915    Date: 2002/01/28
2916    Tag: blume-20020128-sparc-ccalls
2917    Description:
2918    
2919    - Fixed problem with calculation of "used" registers in sparc-c-calls.
2920    - Make use of the allocParam argument in sparc-c-calls.
2921    
2922    ----------------------------------------------------------------------
2923    Name: Matthias Blume
2924    Date: 2002/01/28
2925    Tag: blume-20020128-allocParam
2926    Description:
2927    
2928    John Reppy:  Changes c-calls API to accept client-callback for
2929    allocating extra stack space.
2930    me: Corresponding changes to mlriscGen (using a dummy argument that
2931        does not change the current behavior).
2932    
2933    ----------------------------------------------------------------------
2934    Name: Matthias Blume
2935    Date: 2002/01/28 12:00:00
2936    Tag: Release_110_38
2937    Description:
2938    
2939    This time for real!!!
2940    
2941    ----------------------------------------------------------------------
2942    Name: Matthias Blume
2943    Date: 2002/01/28 10:56:00 EST
2944    Tag: blume-20020128-retraction
2945    Description:
2946    
2947    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
2948       with blume-Release_110_38-retracted.)
2949    
2950    1. Fixed a problem with incorrect rounding modes in real64.sml.
2951       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
2952    
2953    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
2954       and enums fixed.  The naming of corresponding ML identifiers should
2955       now be consistent again.
2956    
2957    ----------------------------------------------------------------------
2958    Name: Allen Leung
2959    Date: 2002/01/27
2960    Tag: leunga-20020127-nowhere
2961    Description:
2962    
2963       Added a target called nowhere in the configuration scripts.
2964       Enabling this will build the MLRISC 'nowhere' tool (for translating
2965       programs with where-clauses into legal SML code) during installation.
2966    
2967    ----------------------------------------------------------------------
2968    Name: Matthias Blume
2969    Date: 2002/01/25 21:27:00 EST
2970    Tag: blume-Release_110_38-retracted
2971    Description:
2972    
2973    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
2974    
2975    README will be added later.
2976    
2977    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
2978    (Release_110_38) removed.  Reason: Last-minute bug fixes.
2979    
2980    ----------------------------------------------------------------------
2981    Name: Matthias Blume
2982    Date: 2002/01/25
2983    Tag: blume-20020125-ffi
2984    Description:
2985    
2986    A large number of tweaks and improvements to ml-nlffi-lib and
2987    ml-nlffigen:
2988    
2989       - ML represenation types have been streamlined
2990       - getter and setter functions work with concrete values, not abstract
2991         ones where possible
2992       - ml-nlffigen command line more flexible (see README file there)
2993       - some bugs have been fixed (hopefully)
2994    
2995    ----------------------------------------------------------------------
2996    Name: Lal George
2997    Date: 2002/01/24
2998    Tag: george-20020124-risc-ra-interface
2999    Description:
3000    
3001       There is a dramatic simplification in the interface to the
3002       register allocator for RISC architectures as a result of making
3003       parallel copy instructions explicit.
3004    
3005    ----------------------------------------------------------------------
3006    Name: Matthias Blume
3007    Date: 2002/01/22
3008    Tag: blume-20020122-x86-ccalls
3009    Description:
3010    
3011    Bug fix for c-calls on x86 (having to do with how char- and
3012    short-arguments are being handled).
3013    
3014    ----------------------------------------------------------------------
3015    Name: Matthias Blume
3016    Date: 2002/01/21
3017    Tag: blume-20020121-ff
3018    Description:
3019    
3020    Another day of fiddling with the FFI...
3021    
3022    1. Bug fix/workaround:  CKIT does not complain about negative array
3023       dimensions, so ml-nlffigen has to guard itself against this possibility.
3024       (Otherwise a negative dimension would send it into an infinite loop.)
3025    
3026    2. Some of the abstract types (light objects, light pointers, most "base"
3027       types) in structure C are now eqtypes.
3028    
3029    3. Added constructors and test functions for NULL function pointers.
3030    
3031    ----------------------------------------------------------------------
3032    Name: Matthias Blume
3033    Date: 2002/01/18
3034    Tag: blume-20020118-ready-for-new-release
3035    Description:
3036    
3037    Made config/srcarchiveurl point to a new place.  (Will provide boot
3038    files shortly.)
3039    
3040    Maybe we christen this to be 110.38?
3041    
3042    ----------------------------------------------------------------------
3043    Name: Matthias Blume
3044    Date: 2002/01/18
3045    Tag: blume-20020118-more-ffifiddle
3046    Description:
3047    
3048    Today's FFI fiddling:
3049    
3050      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3051        C.Set.* that use concrete (MLRep.*) arguments and results instead
3052        of abstract ones.
3053    
3054      - Provided word-style bit operations etc. for "int" representation
3055        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3056        and Long.
3057    
3058    ----------------------------------------------------------------------
3059    Name: Matthias Blume
3060    Date: 2002/01/18
3061    Tag: blume-20020118-use-x86-fp
3062    Description:
3063    
3064    Now that x86-fast-fp seems to be working, I turned it back on again
3065    by default.  (Seems to work fine now, even with the FFI.)
3066    
3067    Other than that, I added some documentation about the FFI to
3068    src/ml-nlffigen/README and updated the FFI test examples in
3069    src/ml-nlffi-lib/Tests/*.
3070    
3071    ----------------------------------------------------------------------
3072    Name: Allen Leung
3073    Date: 2002/01/17
3074    Tag: leunga-20020117-x86-fast-fp-call
3075    Description:
3076    
3077       1. Fixed a problem with handling return fp values when x86's fast fp
3078          mode is turned on.
3079    
3080       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3081          of %f32.
3082    
3083       3. Added a constructor INT32lit to the ast of MLRISC tools.
3084    
3085    ----------------------------------------------------------------------
3086    Name: Matthias Blume
3087    Date: 2002/01/16
3088    Tag: blume-20020116-ffifiddle
3089    Description:
3090    
3091    More fiddling with the FFI interface:
3092    
3093     - Make constness 'c instead of rw wherever possible.  This eliminates
3094       the need for certain explicit coercions.  (However, due to ML's
3095       value polymorphism, there will still be many cases where explicit
3096       coercions are necessary.  Phantom types are not the whole answer
3097       to modeling a subtyping relationship in ML.)
3098    
3099     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3100    
3101     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3102       clumsy workaround for varargs problem.)
3103    
3104    ----------------------------------------------------------------------
3105    Name: Lal George
3106    Date: 2002/01/15
3107    Tag: <none>
3108    Description:
3109    
3110    1. Since COPY instructions are no longer native to the architecture,
3111       a generic functor can be used to implement the expandCopies function.
3112    
3113    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3114       TEXT segment.
3115    
3116    ----------------------------------------------------------------------
3117    Name: Matthias Blume
3118    Date: 2002/01/15
3119    Tag: blume-20020115-ffiupdates
3120    Description:
3121    
3122    1. Fix for bug resulting in single-precision float values being returned
3123       incorrectly from FFI calls.
3124    
3125    2. Small modifications to C FFI API:
3126    
3127        - memory-allocation routines return straight objects (no options)
3128          and raise an exception in out-of-memory situations
3129        - unsafe extensions to cast between function pointers and pointers
3130          from/to ints
3131        - added structure C_Debug as an alternative to structure C where
3132          pointer-dereferencing (|*| and |*!) always check for null-pointers
3133        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3134          but also takes a (possibly empty) list of existing library handles
3135          that the current library depends on
3136    
3137    ----------------------------------------------------------------------
3138    Name: Matthias Blume
3139    Date: 2002/01/10
3140    Tag: blume-20020110-newffigen
3141    Description:
3142    
3143    1. Updates to portable graph code.
3144    
3145    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3146       more scalable now so that even huge interfaces such as the one
3147       for GTK compile in finite time and space. :-)
3148       See src/ml-nlffigen/README for details on what's new.
3149    
3150    ----------------------------------------------------------------------
3151    Name: Lal George
3152    Date: 2001/01/09 14:31:35 EST 2002
3153    Tag: george-20011206-rm-native-copy
3154    Description:
3155    
3156            Removed the native COPY and FCOPY instructions
3157            from all the architectures and replaced it with the
3158            explicit COPY instruction from the previous commit.
3159    
3160            It is now possible to simplify many of the optimizations
3161            modules that manipulate copies. This has not been
3162            done in this change.
3163    
3164    ----------------------------------------------------------------------
3165    Name: Lal George
3166    Date: 2001/12/06 16:50:13 EST 2001
3167    Tag: george-20011206-mlrisc-instruction
3168    Description:
3169    
3170    Changed the representation of instructions from being fully abstract
3171    to being partially concrete. That is to say:
3172    
3173      from
3174            type instruction
3175    
3176      to
3177            type instr                              (* machine instruction *)
3178    
3179            datatype instruction =
3180                LIVE of {regs: C.cellset, spilled: C.cellset}
3181              | KILL of {regs: C.cellset, spilled: C.cellset}
3182              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3183              | ANNOTATION of {i: instruction, a: Annotations.annotation}
3184              | INSTR of instr
3185    
3186    This makes the handling of certain special instructions that appear on
3187    all architectures easier and uniform.
3188    
3189    LIVE and KILL say that a list of registers are live or killed at the
3190    program point where they appear. No spill code is generated when an
3191    element of the 'regs' field is spilled, but the register is moved to
3192    the 'spilled' (which is present, more for debugging than anything else).
3193    
3194    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3195    We used to generate:
3196    
3197            DEFFREG f1
3198            f1 := f2 + f3
3199            trapb
3200    
3201    but now generate:
3202    
3203            f1 := f2 + f3
3204            trapb
3205            LIVE {regs=[f1,f2,f3], spilled=[]}
3206    
3207    Furthermore, the DEFFREG (hack) required that all floating point instruction
3208    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3209    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3210    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3211    this architecture).
3212    
3213    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3214    all the architectures. This will result in further simplification of the
3215    register allocator that must be aware of them for coalescing purposes, and
3216    will also simplify certain aspects of the machine description that provides
3217    callbacks related to parallel copies.
3218    
3219    ANNOTATION should be obvious, and now INSTR represents the honest to God
3220    machine instruction set!
3221    
3222    The <arch>/instructions/<arch>Instr.sml files define certain utility
3223    functions for making porting easier -- essentially converting upper case
3224    to lower case. All machine instructions (of type instr) are in upper case,
3225    and the lower case form generates an MLRISC instruction. For example on
3226    the alpha we have:
3227    
3228      datatype instr =
3229         LDA of {r:cell, b:cell, d:operand}
3230       | ...
3231    
3232      val lda : {r:cell, b:cell, d:operand} -> instruction
3233        ...
3234    
3235    where lda is just (INSTR o LDA), etc.
3236    
3237    ----------------------------------------------------------------------
3238    Name: Matthias Blume
3239    Date: 2001/11/22 21:40:00 EST
3240    Tag: Release_110_37
3241    Description:
3242    
3243    Release 110.37.  This time for real.
3244    
3245    ----------------------------------------------------------------------
3246    Name: Matthias Blume
3247    Date: 2001/11/21 16:35:00 EST
3248    Tag: blume-20011121-foot-in-mouth
3249    Description:
3250    
3251    Removed the "Release_110_37" tag because of a serious bug.
3252    This will be re-tagged once the bug is fixed.
3253    
3254    ----------------------------------------------------------------------
3255    Name: Matthias Blume
3256    Date: 2001/11/21 16:14:00 EST
3257    Tag: blume-20011121-forgottenfile
3258    Description:
3259    
3260    Forgot to add a file.  (Just a .tex-file -- part of
3261    the CM manual source.)
3262    
3263    ----------------------------------------------------------------------
3264    Name: Matthias Blume
3265    Date: 2001/11/21 16:10:00 EST
3266    Tag: blume-20011121-invalid_110_37
3267    Description:
3268    
3269    Note: I removed the original tag "Release_110_37" from this commit
3270          because we found a serious bug in all non-x86 backends.
3271          - Matthias
3272    
3273    1. Modifications to the SML/NJ code generator and to the runtime system
3274       so that code object name strings are directly inserted into code
3275       objects at code generation time.  The only business the runtime system
3276       has with this is now to read the name strings on occasions.
3277       (The encoding of the name string has also changed somewhat.)
3278    
3279    2. CM now implements a simple "set calculus" for specifying export lists.
3280       In particular, it is now possible to refer to the export lists of
3281       other libraries/groups/sources and form unions as well as differences.
3282       See the latest CM manual for details.
3283    
3284    3. An separate notion of "proxy" libraries has again be eliminated from
3285       CM's model.  (Proxy libraries are now simply a special case of using
3286       the export list calculus.)
3287    
3288    4. Some of the existing libraries now take advantage of the new set
3289       calculus.
3290       (Notice that not all libraries have been converted because some
3291       of the existing .cm-files are supposed to be backward compatible
3292       with 110.0.x.)
3293    
3294    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3295       "General.exnMessage"!  The former relies on a certain hook to be
3296       initialized, and that often does not happen in the stand-alone case.)
3297    
3298    ----------------------------------------------------------------------
3299    Name: Lal George
3300    Date: 2001/11/21  13:56:18 EST
3301    Tag: george-2001121-pseudo-ops
3302    Description:
3303    
3304      Implemented a complete redesign of MLRISC pseudo-ops. Now there
3305      ought to never be any question of incompatabilities with
3306      pseudo-op syntax expected by host assemblers.
3307    
3308      For now, only modules supporting GAS syntax are implemented
3309      but more should follow, such as MASM, and vendor assembler
3310      syntax, e.g. IBM as, Sun as, etc.
3311    
3312    ----------------------------------------------------------------------
3313    Name: Matthias Blume
3314    Date: 2001/11/14 11:52:00 EST
3315    Tag: blume-20011114-srcname
3316    Description:
3317    
3318    1. Routed the name of the current source file to mlriscgen where it
3319       should be directly emitted into the code object.  (This last part
3320       is yet to be done.)
3321    
3322    2. Some cleanup of the pgraph code to make it match the proposal that
3323       I put out the other day.  (The proposal notwithstanding, things are
3324       still in flux here.)
3325    
3326    ----------------------------------------------------------------------
3327    Name: Lal George
3328    Date: 2001/11/14 09:44:04 EST
3329    Tag:
3330    Description:
3331    
3332      Fix for a backpatching bug reported by Allen.
3333    
3334      Because the boundary between short and long span-dependent
3335      instructions is +/- 128, there are an astounding number of
3336      span-dependent instructions whose size is over estimated.
3337    
3338      Allen came up with the idea of letting the size of span
3339      dependent instructions be non-monotonic, for a maxIter
3340      number of times, after which the size must be monotonically
3341      increasing.
3342    
3343      This table shows the number of span-dependent instructions
3344      whose size was over-estimated as a function of maxIter, for the
3345      file Parse/parse/ml.grm.sml:
3346    
3347         maxIter            # of instructions:
3348            10                      687
3349            20                      438
3350            30                      198
3351            40                        0
3352    
3353      In compiling the compiler, there is no significant difference in
3354      compilation speed between maxIter=10 and maxIter=40. Actually,
3355      my measurements showed that maxIter=40 was a tad faster than
3356      maxIter=10! Also 96% of the  files in the compiler reach a fix
3357      point within 13 iterations, so fixing maxIter at 40, while high,
3358      is okay.
3359    
3360    ----------------------------------------------------------------------
3361    Name: Matthias Blume
3362    Date: 2001/10/31 15:25:00 EST
3363    Tag: blume-20011031-pgraph
3364    Description:
3365    
3366    CKIT:
3367    * Changed the "Function" constructor of type Ast.ctype to carry optional
3368      argument identifiers.
3369    * Changed the return type of TypeUtil.getFunction accordingly.
3370    * Type equality ignores the argument names.
3371    * TypeUtil.composite tries to preserve argument names but gives up quickly
3372      if there is a mismatch.
3373    
3374    installation script:
3375    * attempts to use "curl" if available (unless "wget" is available as well)
3376    
3377    CM:
3378    * has an experimental implementation of "portable graphs" which I will
3379      soon propose as an implementation-independent library format
3380    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
3381    
3382    NLFFI-LIB:
3383    * some cleanup (all cosmetic)
3384    
3385    NLFFIGEN:
3386    * temporarily disabled the mechanism that suppresses ML output for
3387      C definitions whose identifiers start with an underscore character
3388    * generate val bindings for enum constants
3389    * user can request that only one style (light or heavy) is being used;
3390      default is to use both (command-line arguments: -heavy and -light)
3391    * fixed bug in handling of function types involving incomplete pointers
3392    * generate ML entry points that take record arguments (i.e., using
3393      named arguments) for C functions that have a prototype with named
3394      arguments
3395      (see changes to CKIT)
3396    
3397    ----------------------------------------------------------------------
3398    Name: Allen Leung
3399    Date: 2001/10/27 20:34:00 EDT
3400    Tag: leunga-20011027-x86-fast-fp-call
3401    Description:
3402    
3403       Fixed the bug described in blume-20010920-slowfp.
3404    
3405       The fix involves
3406          1. generating FCOPYs in FSTP in ia32-svid
3407          2. marking a CALL with the appropriate annotation
3408    
3409    ----------------------------------------------------------------------
3410    Name: Matthias Blume
3411    Date: 2001/10/16 11:32:00 EDT
3412    Tag: blume-20011016-netbsd
3413    Description:
3414    
3415    Underscore patch from Chris Richards (fixing problem with compiling
3416    runtime system under recent NetBSD).
3417    
3418    ----------------------------------------------------------------------
3419    Name: Allen Leung
3420    Date: 2001/10/12 17:18:32 EDT 2001
3421    Tag: leung-20011012-x86-printflowgraph
3422    Description:
3423    
3424    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
3425    
3426    ----------------------------------------------------------------------
3427    Name: Lal George
3428    Date: 2001/10/11 23:51:34 EDT
3429    Tag: george-20011011-too-many-instrs
3430    Description:
3431    
3432    The representation of a program point never expected to see more
3433    than 65536 instructions in a basic block!
3434    
3435    ----------------------------------------------------------------------
3436    Name: Lal George
3437    Date: 2001/10/09 09:41:37 EDT
3438    Tag: george-20011008-mlrisc-labels
3439    Description:
3440    
3441    Changed the machine description files to support printing of
3442    local and global labels in assembly code, based on host assembler
3443    conventions.
3444    
3445    ----------------------------------------------------------------------
3446    Name: Matthias Blume
3447    Date: 2001/09/25 15:25:00 EDT
3448    Tag: blume-20010925-exninfo
3449    Description:
3450    
3451    I provided a non-hook implementation of exnName (at the toplevel) and
3452    made the "dummy" implementation of exnMessage (at the toplevel) more
3453    useful: if nothing gets "hooked in", then at least you are going to
3454    see the exception name and a message indicating why you don't see more.
3455    
3456    [For the time being, programs that need exnMessage and want to use
3457    ml-build should either use General.exnMessage (strongly recommended) or
3458    refer to structure General at some other point so that CM sees a
3459    static dependency.]
3460    
3461    [Similar remarks go for "print" and "use":  If you want to use their
3462    functionality in stand-alone programs generated by ml-build, then use
3463    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
3464    
3465    ----------------------------------------------------------------------
3466    Name: Matthias Blume
3467    Date: 2001/09/20 17:28:00 EDT
3468    Tag: blume-20010920-slowfp
3469    Description:
3470    
3471    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
3472    
3473    ----------------------------------------------------------------------
3474    Name: Matthias Blume
3475    Date: 2001/09/20 17:20:00 EDT
3476    Tag: blume-20010920-canonicalpaths
3477    Description:
3478    
3479    0. Updated the BOOT file (something that I forgot to do earlier).
3480    
3481    1. Small internal change to CM so that it avoids "/../" in filenames
3482       as much as possible (but only where it is safe).
3483    
3484    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
3485       that arguments that contain delimiters are passed through correctly.
3486       This change also means that all "special" arguments of the form
3487       @SMLxxx... must come first.
3488    
3489    3. Changed install script to put relative anchor names for tool commands
3490       into pathconfig.
3491    
3492    ----------------------------------------------------------------------
3493    Name: Matthias Blume
3494    Date: 2001/09/18 15:35:00 EDT
3495    Tag: blume-20010918-readme11036
3496    Description:
3497    
3498    Added README files.
3499    
3500    ----------------------------------------------------------------------
3501    Name: Matthias Blume
3502    Date: 2001/09/18 11:45:00 EDT
3503    Tag: Release_110_36 (retag)
3504    Description:
3505    
3506    Fixed mistake in config/preloads. Retagged as 110.36.
3507    
3508    ----------------------------------------------------------------------
3509    Name: Matthias Blume
3510    Date: 2001/09/18 09:40:00 EDT
3511    Tag: Release_110_36_orig (tag changed)
3512    Description:
3513    
3514    New version (110.36).  New bootfiles.
3515    
3516    ----------------------------------------------------------------------
3517    Name: Matthias Blume
3518    Date: 2001/09/14 16:15:00 EDT
3519    Tag: blume-20010914-x86fastfp
3520    Description:
3521    
3522    John committed some changes that Allen made, in particular a (hopefully)
3523    correctly working version of the x86-fp module.
3524    
3525    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
3526    flag to "true".  Everything seems to compile to a fixpoint ok, and
3527    "mandelbrot" speeds up by about 15%.
3528    
3529    ----------------------------------------------------------------------
3530    Name: Matthias Blume
3531    Date: 2001/09/13 11:20:00 EDT
3532    Tag: blume-20010913-minimal
3533    Description:
3534    
3535    1. Stefan Monnier's patch to fix a miscompilation problem that
3536       was brought to light by John Reppy's work on Moby.
3537    
3538    2. Implemented a minimal "structure Compiler" that contains just
3539       "version" and "architecture".  The minimal version will be
3540       available when the full version is not.  This is for backward-
3541       compatibility with code that wants to test Compiler.version.
3542    
3543    ----------------------------------------------------------------------
3544    Name: Matthias Blume
3545    Date: 2001/08/28 14:03:00 EDT
3546    Tag: blume-20010828-ml-lex
3547    Description:
3548    
3549    Fix for bug 1581, received from Neophytos Michael.
3550    
3551    ----------------------------------------------------------------------
3552    Name: Matthias Blume
3553    Date: 2001/08/27 11:20:00 EDT
3554    Tag: blume-20010827-readme11035
3555    Description:
3556    
3557    Fleshed out the README file for 110.35.
3558    
3559    ----------------------------------------------------------------------
3560    Name: Matthias Blume
3561    Date: 2001/08/24 17:10:00 EDT
3562    Tag: Release_110_35
3563    Description:
3564    
3565    New version number (110.35).  New bootfiles.
3566    
3567    ----------------------------------------------------------------------
3568    Name: Lal George
3569    Date: 2001/08/24 13:47:18 EDT 2001
3570    Tag: george-20010824-MLRISC-graphs
3571    Description:
3572    
3573     removed clusters from MLRISC completely and replaced with graphs.
3574    
3575    ----------------------------------------------------------------------
3576    Name: Matthias Blume
3577    Date: 2001/08/23 17:50:00 EDT
3578    Tag: blume-20010823-toplevel
3579    Description:
3580    
3581    - some reorganization of the code that implements various kinds of
3582      environments in the compiler (static, dynamic, symbolic, combined)
3583    - re-implemented the EnvRef module so that evalStream works properly
3584      (if the stream contains references to "use", "CM.make", etc.)
3585    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
3586    
3587    ----------------------------------------------------------------------
3588    Name: Matthias Blume
3589    Date: 2001/08/20 15:50 EDT
3590    Tag: blume20010820-slipup
3591    Description:
3592    
3593    I forgot to commit a few files.  Here they are...
3594    
3595    ----------------------------------------------------------------------
3596    Name: Matthias Blume
3597    Date: 2001/08/20 15:35:00 EDT
3598    Tag: blume-20010820-debugprof
3599    Description:
3600    
3601    !!!! NEW BOOTFILES !!!!
3602    
3603    This is another round of reorganizing the compiler sources.  This
3604    time the main goal was to factor out all the "instrumentation"
3605    passes (for profiling and backtracing) into their own library.
3606    The difficulty was to do it in such a way that it does not depend
3607    on elaborate.cm but only on elabdata.cm.
3608    
3609    Therefore there have been further changes to both elaborate.cm and
3610    elabdata.cm -- more "generic" things have been moved from the former
3611    to the latter.  As a result, I was forced to split the assignment
3612    of numbers indicating "primtyc"s into two portions: SML-generic and
3613    SML/NJ-specific.  Since it would have been awkward to maintain,
3614    I bit the bullet and actually _changed_ the mapping between these
3615    numbers and primtycs.  The bottom line of this is that you need
3616    a new set of bin- and bootfiles.
3617    
3618    I have built new bootfiles for all architectures, so doing a fresh
3619    checkout and config/install.sh should be all you need.
3620    
3621    The newly created library's name is
3622    
3623        $smlnj/viscomp/debugprof.cm
3624    
3625    and its sources live under
3626    
3627        src/compiler/DebugProf
3628    
3629    ----------------------------------------------------------------------
3630    Name: Matthias Blume
3631    Date: 2001/08/15 17:15:00 EDT
3632    Tag: blume-20010815-compreorg
3633    Description:
3634    
3635    This is a first cut at reorganizing the CM libraries that make up the
3636    core of the compiler.  The idea is to separate out pieces that could
3637    be used independently by tools, e.g., the parser, the typechecker, etc.
3638    
3639    The current status is a step in this direction, but it is not quite
3640    satisfactory yet.  Expect more changes in the future.
3641    
3642    Here is the current (new) organization...
3643    
3644        What used to be $smlnj/viscomp/core.cm is now divided into
3645        six CM libraries:
3646    
3647             $smlnj/viscomp/basics.cm
3648                           /parser.cm
3649                           /elabdata.cm
3650                           /elaborate.cm
3651                           /execute.cm
3652                           /core.cm
3653    
3654        The CM files for these libraries live under src/system/smlnj/viscomp.
3655        All these libraries are proxy libraries that contain precisely
3656        one CM library component.  Here are the locations of the components
3657        (all within the src/compiler tree):
3658    
3659             Basics/basics.cm
3660             Parse/parser.cm
3661             ElabData/elabdata.cm
3662             Elaborator/elaborate.cm
3663             Execution/execute.cm
3664             core.cm
3665    
3666         [This organization is the same that has been used already
3667         for a while for the architecture-specific parts of the visible
3668         compiler and for the old version of core.cm.]
3669    
3670         As you will notice, many source files have been moved from their
3671         respective original locations to a new home in one of the above
3672         subtrees.
3673    
3674         The division of labor between the new libraries is the following:
3675    
3676             basics.cm:
3677                - Simple, basic definitions that pertain to many (or all) of
3678                  the other libraries.
3679             parser.cm:
3680                - The SML parser, producing output of type Ast.dec.
3681                - The type family for Ast is also defined and exported here.
3682             elabdata.cm:
3683                - The datatypes that describe input and output of the elaborator.
3684                  This includes types, absyn, and static environments.
3685             elaborator.cm:
3686                - The SML/NJ type checker and elaborator.
3687                  This maps an Ast.dec (with a given static environment) to
3688                  an Absyn.dec (with a new static environment).
3689                - This libraries implements certain modules that used to be
3690                  structures as functors (to remove dependencies on FLINT).
3691             execute.cm:
3692                - Everything having to do with executing binary code objects.
3693                - Dynamic environments.
3694             core.cm:
3695                - SML/NJ-specific instantiations of the elaborator and MLRISC.
3696                - Top-level modules.
3697                - FLINT (this should eventually become its own library)
3698    
3699    Notes:
3700    
3701    I am not 100% happy with the way I separated the elaborator (and its
3702    data structures) from FLINT.  Two instances of the same problem:
3703    
3704        1. Data structures contain certain fields that carry FLINT-specific
3705           information.  I hacked around this using exn and the property list
3706           module from smlnj-lib.  But the fact that there are middle-end
3707           specific fields around at all is a bit annoying.
3708    
3709        2. The elaborator calculates certain FLINT-related information.  I tried
3710           to make this as abstract as I could using functorization, but, again,
3711           the fact that the elaborator has to perform calculations on behalf
3712           of the middle-end at all is not nice.
3713    
3714        3. Having to used exn and property lists is unfortunate because it
3715           weakens type checking.  The other alternative (parameterizing
3716           nearly *everything*) is not appealing, though.
3717    
3718    I removed the "rebinding =" warning hack because due to the new organization
3719    it was awkward to maintain it.  As a result, the compiler now issues some of
3720    these warnings when compiling init.cmi during bootstrap compilation. On
3721    the plus side, you also get a warning when you do, for example:
3722       val op = = Int32.+
3723    which was not the case up to now.
3724    
3725    I placed "assign" and "deref" into the _Core structure so that the
3726    code that deals with the "lazy" keyword can find them there.  This
3727    removes the need for having access to the primitive environment
3728    during elaboration.
3729    
3730    ----------------------------------------------------------------------
3731    Name: Matthias Blume
3732    Date: 2001/08/13
3733    Tag: blume-20010813-closures
3734    Description:
3735    
3736    This fix was sent to us by Zhong Shao.  It is supposed to improve the
3737    performance of certain loops by avoiding needless closure allocation.
3738    
3739    ----------------------------------------------------------------------
3740    Name: Lal George
3741    Date: 2001/07/31 10:03:23 EDT 2001
3742    Tag: george-20010731-x86-fmalloc
3743    Description: Fixed bug in x86 calls
3744    
3745        There was a bug where call instructions would mysteriously
3746        vanish. The call instruction had to be one that returned
3747        a floating point value.
3748    
3749    ----------------------------------------------------------------------
3750    Name: Lal George
3751    Date: 2001/07/19 16:36:29 EDT 2001
3752    Tag: george-20010719-simple-cells
3753    Description:
3754    
3755    I have dramatically simplified the interface for CELLS in MLRISC.
3756    
3757    In summary, the cells interface is broken up into three parts:
3758    
3759      1. CellsBasis : CELLS_BASIS
3760    
3761            CellsBasis is a top level structure and common for all
3762            architectures.  it contains the definitions of basic datatypes
3763            and utility  functions over these types.
3764    
3765      2. functor Cells() : CELLS
3766    
3767            Cells generates an interface for CELLS that incorporates the
3768            specific resources on the target architecture, such as the
3769            presence of special register classes, their number and size,
3770            and various useful substructures.
3771    
3772      3. <ARCH>CELLS
3773    
3774            e.g. SparcCells: SPARCCELLS
3775    
3776            <ARCH>CELLS usually contains additional bindings for special
3777            registers  on the architecture, such as:
3778    
3779                    val r0 : cell           (* register zero *)
3780                    val y : cell            (* Y register *)
3781                    val psr : cell          (* processor status register *)
3782                    ...
3783    
3784            The structure returned by applying the Cells functor is opened
3785            in this interface.
3786    
3787    The main implication of all this is that the datatypes for cells is
3788    split between CellsBasis and CELLS -- a fairly simple change for user
3789    code.
3790    
3791    In the old scheme the CELLS interface had a definitional binding of
3792    the form:
3793    
3794            signature CELLS = sig
3795    
3796               structure CellsBasis = CellsBasis
3797    
3798               ...
3799    
3800            end
3801    
3802    With all the sharing constraints that goes on in MLRISC, this old
3803    design  quickly leads to errors such as:
3804    
3805            "structure definition spec inside of sharing ... "
3806    
3807    
3808    and appears to require an unacceptable amount of sharing and where
3809    constraint hackery.
3810    
3811    I think this error message (the interaction of definitional specs and
3812    sharing) requires more explanation on our web page.
3813    
3814    ----------------------------------------------------------------------
3815    Name: Matthias Blume
3816    Date: 2001/07/19 15:00:00 EDT
3817    Tag: blume-20010719-libreorg
3818    Description:
3819    
3820    This update puts together a fairly extensive but straightforward change
3821    to the way the libraries that implement the interactive system are
3822    organized:
3823    
3824       The biggest change is the elimination of structure Compiler.  As a
3825       replacement for this structure, there is now a CM library
3826       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
3827       that exports all the substructures of the original structure Compiler
3828       directly.  So instead of saying Compiler.Foo.bar one now simply
3829       says Foo.bar.  (The CM libraries actually export a collection of
3830       structures that is richer than the collection of substructures of
3831       structure Compiler.)
3832    
3833       To make the transition smooth, there is a separate library called
3834       $smlnj/compiler/compiler.cm which puts together and exports the
3835       original structure Compiler (or at least something very close to it).
3836    
3837       There are five members of the original structure Compiler
3838       that are not exported directly but which instead became members
3839       of a new structure Backend (described by signature BACKEND).  These are:
3840       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
3841       Interact (: INTERACT), structure Machine (: MACHINE), and val
3842       architecture (: string).
3843    
3844       Structure Compiler.Version has become structure CompilerVersion.
3845    
3846