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 1066, Thu Feb 14 16:50:02 2002 UTC revision 1622, Mon Sep 13 21:50:40 2004 UTC
# Line 13  Line 13 
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/09/13 16:50:00 CDT
18    Tag: Release_110_49
19    Description:
20    
21    New working version (110.49).  NEW BOOTFILES!
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume
25    Date: 2004/09/13 16:20:00 CDT
26    Tag: blume-20040913-config-mlrisc
27    Description:
28    
29    Put target "mlrisc" back into the default list.
30    (There is no harm in having it, and some users have expressed their
31    wish to have "mlrisc" included by default.)
32    
33    ----------------------------------------------------------------------
34    Name: John Reppy
35    Date: 2004/09/13
36    Tag: jhr-20040913-signals
37    Description:
38    
39    Fixed the signal masking code to properly nest mask/unmask operations
40    on a per-signal basis.
41    
42    ----------------------------------------------------------------------
43    Name: Matthias Blume (blume (at) tti - c (dot) org)
44    Date: 2004/09/08 13:20:00 CDT
45    Tag: blume-20040908-heap-magic
46    Description:
47    
48    Bumped the heap macig number to 0x09082004 to account for the changed
49    layout of the ML frame under MacOS X.
50    
51    ----------------------------------------------------------------------
52    Name: Allen Leung (leunga (at) reservoir (dot) com)
53    Date: 2004/09/03 11:26:00 EST
54    Tag: leunga-20040903-cygwin-install
55    Description:
56    
57       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
58    The Cygwin runtime is turned on by setting the environment
59    variable SMLNJ_CYGWIN_RUNTIME to 1.
60    
61    ----------------------------------------------------------------------
62    Name: Matthias Blume (blume (at) tti - c (dot) org)
63    Date: 2004/08/31 17:15:00 CDT
64    Tag: blume-20040831-core
65    Description:
66    
67    Added some exports to src/compiler/core.cm upon request by J. Joemann.
68    
69    ----------------------------------------------------------------------
70    Name: Matthias Blume (blume (at) tti - c (dot) org)
71    Date: 2004/08/30 17:55:00 CDT
72    Tag: blume-20040830-installer
73    Description:
74    
75    Upon request by Johannes Joemann:
76    
77    - improved ML code of installer to fall back to coping when renaming
78      fails (i.e., when source and target are on different file systems);
79      the code compiles but has yet to be tested in anger
80    - removed mlrisc from list of default targets (config/targets)
81    
82    ----------------------------------------------------------------------
83    Name: Matthias Blume (blume (at) tti - c (dot) org)
84    Date: 2004/08/27 17:20:00 CDT
85    Tag: blume-20040827-ptreql
86    Description:
87    
88    Added ptreql primop to structure InlineT (upon request from Larry
89    Paulson).
90    
91    ----------------------------------------------------------------------
92    Name: Allen Leung (leunga (at) reservoir (dot) com)
93    Date: 2004/08/15 21:21:00 EST
94    Tag: leunga-110_48-udgraph
95    Description:
96    
97    Another bug fix from Carl Hauser:
98    
99    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
100    > 48c48
101    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
102    > e::L)
103    > ---
104    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
105    > Without this, any deletion of an edge in an undirected graph does severe
106    > violence to the graph.
107    
108    ----------------------------------------------------------------------
109    Name: Allen Leung (leunga (at) reservoir (dot) com)
110    Date: 2004/08/10 23:35:00 EST
111    Tag: leunga-110_48-ppc
112    Description:
113    
114       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
115    
116    ----------------------------------------------------------------------
117    Name: Matthias Blume (blume (at) tti - c (dot) org)
118    Date: 2004/08/10 12:00:00 CDT
119    Tag: Release_110_48
120    Description:
121    
122    New working version (110.48).  NEW BOOTFILES!
123    
124    ----------------------------------------------------------------------
125    Name: Allen Leung (leunga (at) reservoir (dot) com)
126    Date: 2004/08/09 12:21:00 EST
127    Tag: leunga-110_47-dijsktra
128    Description:
129    
130    Bug fix from Carl Hauser:
131    
132    single_source_shortest_paths in dijkstra.sml was observed to get wrong
133    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
134    
135    The problem is that following the expression A.update(dist,s,Num.zero)
136    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
137    
138    ----------------------------------------------------------------------
139    Name: Matthias Blume (blume (at) tti - c (dot) org)
140    Date: 2004/08/06 18:10:00 CDT
141    Tag: blume-20040806-cmdline
142    Description:
143    
144    Fiddled with handling of command-line options:
145    
146      * sml now quits after processing the command line
147        if -H, -S, -h<n>, or -s<n> appears as the last
148        command-line argument
149      * a new option -q terminates the session when encountered on
150        the command line; subsequent arguments will be ignored
151      * bug fixes: short (erroneous) arguments are no longer ignored
152        completely
153    
154    ----------------------------------------------------------------------
155    Name: Allen Leung (leunga (at) reservoir (dot) com)
156    Date: 2004/08/04 18:17:00 EST
157    Tag: leunga-110_47-ppc-ibm-asm
158    Description:
159    
160      - Added minimal IBM assembly syntax support for PowerPC.
161    
162      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
163        been altered in the runtime.  We need an automatic way to keep the file
164        in sync.
165    
166    ----------------------------------------------------------------------
167    Name: Matthias Blume (blume (at) tti - c (dot) org)
168    Date: 2004/08/04 14:00:00 CDT
169    Tag: Release_110_47
170    Description:
171    
172    New working version (110.47).  NEW BOOTFILES!
173    
174    ----------------------------------------------------------------------
175    Name: Matthias Blume (blume (at) tti - c (dot) org)
176    Date: 2004/08/03 14:25:00 CDT
177    Tag: blume-20040803-callingconv
178    Description:
179    
180    Added low-level support for choosing C calling conventions by
181    twiddling the type of rawccall.  (See
182    src/compiler/Semant/types/cproto.sml for details.)
183    
184    ----------------------------------------------------------------------
185    Name: Matthias Blume (blume (at) tti - c (dot) org)
186    Date: 2004/08/02 15:55:00 CDT
187    Tag: blume-20040802-backout
188    Description:
189    
190    Backed out of change to win32-filesys.c.  The earlier patch to
191    get_file_time caused CM to produce files with the wrong time stamp.
192    
193    ----------------------------------------------------------------------
194    Name: Matthias Blume (blume (at) tti - c (dot) org)
195    Date: 2004/08/02 14:45:00 CDT
196    Tag: blume-20040802-nlffi-win32
197    Description:
198    
199    Added NLFFI support for Win32, adapted from a patch provided by David
200    Hansel.  This is currently completely untested.  Also, the issue
201    concerning stdcall vs. ccall is still unresolved.
202    
203    ----------------------------------------------------------------------
204    Name: Matthias Blume (blume (at) tti - c (dot) org)
205    Date: 2004/07/30 17:55:00 CDT
206    Tag: blume-20040730-various
207    Description:
208    
209    Gearing up towards 110.47...
210    
211    - various minor bugfixes to ml-nlffigen
212    - a beginning of a manual for nlffi
213    
214    - eliminated 'export name=value' in config/install.sh as this does
215      not work with certain versions of /bin/sh
216      (Thanks to David King at Motorola for catching this.)
217    
218    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
219      - added a test for tm==NULL to gmtime.c and localtime.c
220      - applied patch for incorrect GetFileTime under win32
221      - toSeconds -> toMilliseconds in Win32/win32-process.sml
222    
223    ----------------------------------------------------------------------
224    Name: Matthias Blume (blume (at) tti - c (dot) org)
225    Date: 2004/07/21 18:20:00 CDT
226    Tag: blume-20040721-nlffigen
227    Description:
228    
229    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
230      for a typedef to an incomplete type, but leave out the "typ" member.
231      (This is just for consistency.)
232    - Started to produce what is supposed to become better (i.e., comprehensive)
233      documentation of what ml-nlffigen does and produces.
234    
235    ----------------------------------------------------------------------
236    Name: Matthias Blume (blume (at) tti - c (dot) org)
237    Date: 2004/07/14 16:25:00 CDT
238    Tag: blume-20040714-union
239    Description:
240    
241    Added C_UNION to c-calls/c-types.sml and updated the machinery
242    (ml-nlffigen, cproto.sml) that conveys C function interface
243    information to the code generator.
244    
245    However, the actual architecture-specific implementation of function
246    arguments and results that are C unions is still not implemented.
247    
248    ----------------------------------------------------------------------
249    Name: Allen Leung (leunga (at) reservoir (dot) com)
250    Date: 2004/07/14 14:38:00 EST
251    Tag: leunga-110_46_1-ppc-lwzu
252    Description:
253    
254        Added these instructions to the PowerPC architecture:
255        LBZU(X), LHZU(X), LWZU(X),
256        STWU(X), STFDU, STFSU
257    
258        etc...
259    
260        Note: I haven't added their instruction encoding into the description.
261    
262    ----------------------------------------------------------------------
263    Name: Allen Leung (leunga (at) reservoir (dot) com)
264    Date: 2004/07/13 15:04:00 EST
265    Tag: leunga-110_46_1-ppc-lwarx
266    Description:
267    
268       Added the two instructions LWARX and STWCX to the PowerPC
269    instruction set.
270    
271       A (untested) rewrite of loop-structure.sml.  The old version
272    is completely broken.
273    
274    ----------------------------------------------------------------------
275    Name: Matthias Blume (blume (at) tti - c (dot) org)
276    Date: 2004/07/13 13:50:00 CDT
277    Tag: blume-20040713-nlffi
278    Description:
279    
280    - use paramAlloc to report c-calls with too many arguments
281      (for PPC version where parameter area is pre-allocated)
282    - added ccall_maxargspace to machspec (to implement the above)
283    - made "make" commend in CM's "make" tool configurable
284    - added option (default: on) for passing the name of the SML/NJ's "bin"
285      directory to "make"; the call looks like this:
286    
287         make <options> SMLNJ_BINDIR=<dir> <target>
288    
289      This can be used by the Makefile to, e.g., pick the "right" version
290      of ml-nlffigen.
291    - minor code tweaks
292    
293    ----------------------------------------------------------------------
294    Name: Matthias Blume (blume (at) tti - c (dot) org)
295    Date: 2004/07/12 22:50:00 CDT
296    Tag: blume-110_46_1-macosx-nlffi
297    Description:
298    
299    NLFFI under Mac OS X now working (sort of).  This is largely untested,
300    though.
301    
302    Note:  1. You have to make a new, clean build of the runtime system.
303           2. There are new BOOTFILES, you have to use them!
304              (Doing the bootstrap process yourself would be *very* painful!
305               If you absolutely have to do it, build the system under
306               a different architecture and then cross-compile.)
307    
308    Version bumped to 110.46.1 to account for runtime data format changes.
309    
310    ----------------------------------------------------------------------
311    Name: Matthias Blume (blume (at) tti - c (dot) org)
312    Date: 2004/06/18 14:30:00 CDT
313    Tag: blume-20040618-unix
314    Description:
315    
316    Changed the implementation of structure Unix so that the same stream
317    is returned every time one of the {text,bin}{In,Out}streamOf functions
318    is invoked on the same proc.  This is not what the spec currently
319    says -- although IMO it arguably should.  (See discussion below.)
320    
321    ----------------------------------------------------------------------
322    Name: Matthias Blume (blume (at) tti - c (dot) org)
323    Date: 2004/06/17 18:15:00 CDT
324    Tag: Release_110_46
325    Description:
326    
327    New working version (110.46).  NEW BOOTFILES!
328    
329    ----------------------------------------------------------------------
330    Name: Matthias Blume (blume (at) tti - c (dot) org)
331    Date: 2004/06/17 17:20:00 CDT
332    Tag: blume-20040617-timer-unix
333    Description:
334    
335    Changed the interface of structures Timer and Unix to match the most
336    recent Basis spec.
337    
338    In the case of Unix there still seems to be an open/weird issue:
339    
340       The {text,bin}{In,Out}streamOf functions are supposed to create
341       fresh streams whenever they are called -- as opposed to have them
342       return the same stream every time.  This design is supposed to
343       prevent space leaks caused by proc values hanging on to streams.
344    
345       The reap function, on the other hand, is supposed to close the
346       streams.  This cannot be done without having a handle on the
347       stream in proc after all...
348    
349    I took the liberty to implement the following stopgap solution:
350    
351      The proc value hangs on to the most recently created stream(s).
352      Reap closes those.  If either or both of the two streams hadn't
353      been created at all yet, then reap will close the corresponding
354      file descriptors directly.
355    
356    PS: I don't understand the original space leak argument anymore.  If
357    a proc hangs on to the imperative stream, then I/O operations on those
358    will advance the state of the cached stream and avoid the space leak.
359    
360    ----------------------------------------------------------------------
361    Name: Matthias Blume (blume (at) tti - c (dot) org)
362    Date: 2004/05/28 16:45:00 CDT
363    Tag: blume-20040528-basis
364    Description:
365    
366    Added signature PACK_REAL and exported functor PrimIO.
367    
368    ----------------------------------------------------------------------
369    Name: Matthias Blume (blume (at) tti - c (dot) org)
370    Date: 2004/05/25 16:00:00 CDT
371    Tag: blume-20040525-group-owner
372    Description:
373    
374    CM now ignores (but still accepts) the "owner" information in group
375    descriptions.  The owner of a group is its next enclosing
376    library. Each group must have a unique owner.  (There is a virtual
377    "toplevel" library that owns groups which are not nested within a real
378    library.)  Previously, each group had to explicitly declare its owner,
379    and CM would check that such a declaration is correct.  The new scheme
380    is to have CM check that for each group there is precisely one owning
381    library.
382    
383    The advantage of the new scheme is that the programmer no longer needs
384    to maintain the somewhat annoying owner information.  The downside is
385    that CM cannot enforce the ownership rule across multiple runs of
386    CM.make.  Fortunately, enclosing the same group in two different
387    libraries A and B which are not part of the same program does not
388    cause real problems.
389    
390    ----------------------------------------------------------------------
391    Name: Matthias Blume (blume (at) tti - c (dot) org)
392    Date: 2004/05/20 16:00:00 CDT
393    Tag: blume-20040520-win32
394    Description:
395    
396    Made the win32 version work again.  (Strangely, a misplaced comma had
397    slipped into win32-process.c which prevented the runtime from being
398    compiled correctly.)
399    
400    Also, included a minor addition to ml-build.bat analogous to what was
401    done in blume-20040519-ml-build.
402    
403    ----------------------------------------------------------------------
404    Name: Matthias Blume (blume (at) tti - c (dot) org)
405    Date: 2004/05/19 22:10:00 CDT
406    Tag: blume-20040519-ml-build
407    Description:
408    
409    Arranged for ml-build to clean up after itself a little bit better.
410    The script generates a temporary SML source file and compiles it using
411    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
412    gets rid of those at the end, so they don't accumulate under .cm.
413    
414    This required a minor change to install.sh because the name of the
415    metadata directory (default: .cm) is actually configurable at
416    installation time.
417    
418    ----------------------------------------------------------------------
419    Name: Matthias Blume (blume (at) tti - c (dot) org)
420    Date: 2004/05/18 15:50:00 CDT
421    Tag: blume-20040518-mkreader
422    Description:
423    
424    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
425    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
426    
427    ----------------------------------------------------------------------
428    Name: Matthias Blume (blume (at) tti - c (dot) org)
429    Date: 2004/05/11 14:35:00 CDT
430    Tag: blume-20040511-win32sock
431    Description:
432    
433    Added previously missing support for many socket-related functions
434    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
435    for the voluminous patch!
436    
437    (I have not tested this patch under win32 yet.)
438    
439    Here is David's e-mail:
440    
441    Hi,
442    
443    Attached to this email you find a diff against sml/nj 110.45
444    that will enable socket support under Windows.
445    
446    To apply the patch (using unix or cygwin)
447    1) gunzip runtime.diff.gz
448    2) "cd" into "src/runtime" in the source tree of a fresh
449       110.45 installation.
450    3) patch -p 1 < [your/path/to]runtime.diff
451    
452    The code compiles fine but has NOT yet been extensively tested.
453    I only ran a few tests for basic socket client functionality
454    (which worked fine).  Especially the functions that use ioctl
455    are not tested at all and might not work (see below).
456    
457    I implemented this since we want to move to a newer version of sml/nj
458    but need socket support in order to use it.  This is the first time I
459    even had a look at the sml/nj source,  so please review my changes
460    before making this part of the distribution!  Here are a few issues
461    that I think might be better for someone to solve who is more
462    familiar with the sml/nj source (and socket programming):
463    
464    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
465      exception since I could not figure out what the windows equivalent
466      of these functions is
467    
468    - In sockets-osdep.h there are a some #include statements that are
469      only used in a few files that include sockets-osdep.h
470    
471    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
472      does not process its return value since I don't know how to
473      report an error upwards.
474    
475    - It would probably be good to have a call to WSACleanup() when
476      the library is unloaded (if there is such a possibility).
477      Otherwise I think Windows will take care of this automatically
478      when the process finishes.
479    
480    - I used ioctlsocket() as a replacement for ioctl() but I have
481      no idea if that is actually the proper replacement on Windows.
482    
483    - All these issues are marked in the code by "FIXME" comments.
484    
485    We use sml/nj extensively in our products and are quite happy
486    with it.  I hope this contribution will help you.
487    
488    Keep up the good work!
489    
490    David
491    
492    ----------------------------------------------------------------------
493    Name: Matthias Blume (blume (at) tti - c (dot) org)
494    Date: 2004/05/11 14:20:00 CDT
495    Tag: blume-20040511-installml
496    Description:
497    
498    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
499    the patch.)
500    
501    ----------------------------------------------------------------------
502    Name: Matthias Blume (blume (at) tti - c (dot) org)
503    Date: 2004/05/11 14:05:00 CDT
504    Tag: blume-20040511-nlffi-netbsd
505    Description:
506    
507    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
508    the patch.)
509    
510    ----------------------------------------------------------------------
511    Name: Matthias Blume (blume (at) tti - c (dot) org)
512    Date: 2004/05/11 12:05:00 CDT
513    Tag: blume-20040511-exports
514    Description:
515    
516    As per request by Adam Chlipala <adam@hcoop.net>, extended various
517    export lists in compiler-related .cm-files.
518    
519    ----------------------------------------------------------------------
520    Name: Matthias Blume (blume (at) tti - c (dot) org)
521    Date: 2004/05/11 11:35:00 CDT
522    Tag: blume-20040511-allsource
523    Description:
524    
525    The installer now honors the "src-smlnj" target again, although its meaning
526    has changed from "all sources required for the compiler" to "all sources
527    the installer knows about".  In other words, if you enable "src-smlnj"
528    in the "targets" file, then the installer will pull in sources for
529    everything.  (Notice that this refers to source code only.  Compiled
530    code is still only installed for modules that were requested explicitly
531    or which are required for other modules that were requested explicitly.)
532    
533    ----------------------------------------------------------------------
534    Name: Matthias Blume (blume (at) tti - c (dot) org)
535    Date: 2004/04/23 17:40:00 CDT
536    Tag: blume-20040423-ieee-scan
537    Description:
538    
539    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
540    in the exponent calculation we get INF or ZERO (depending on the mantissa
541    and the sign of the exponent).
542    
543    ----------------------------------------------------------------------
544    Name: Matthias Blume (blume (at) tti - c (dot) org)
545    Date: 2004/04/23 10:40:00 CDT
546    Tag: blume-20040423-ml-build
547    Description:
548    
549    The ml-build script now terminates with a non-0 status when something
550    goes wrong.
551    
552    ----------------------------------------------------------------------
553    Name: Matthias Blume (blume (at) tti - c (dot) org)
554    Date: 2004/04/22 16:35:00 CDT
555    Tag: blume-20040422-Option
556    Description:
557    
558    Made exception Option to be the same as exception Option.Option
559    (as it should be).
560    
561    ----------------------------------------------------------------------
562    Name: Allen Leung (leunga (at) reservoir (dot) com)
563    Date: 2004/03/19 14:40:00 EST
564    Tag: leunga-20040319-cygwin-nlffi
565    Description:
566    
567    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
568    of SML/NJ.  The problem is that
569    
570       lib = dlopen(NULL, ...)
571       f   = dlsym(lib, "malloc");
572    
573    does not work on Windows unless we explicitly export symbols
574    such as 'malloc' during linking.  We fixed this by explicitly
575    exporting the required symbols with the magic gcc incantation:
576    
577        -Wl,--export-all cygwin.def
578    
579    where cygwin.def is a file containing all the symbols that
580    we wish to export.
581    
582    I suspect this is a Windows problem and we'll have to
583    do the same (somehow with windows compilers) when
584    we build the native win32 version with the system
585    calls LoadLibrary/GetProcAddress.
586    
587    ----------------------------------------------------------------------
588    Name: Matthias Blume (blume (at) tti - c (dot) org)
589    Date: 2004/03/04 16:35:00 CST
590    Tag: blume-20040304-intinf-fmt
591    Description:
592    
593    Fixed problem with IntInf.fmt (sign would show up on the right instead
594    of on the left for BIN, OCT, and HEX).
595    
596    ----------------------------------------------------------------------
597    Name: Matthias Blume (blume (at) tti - c (dot) org)
598    Date: 2004/03/04 11:25:00 CST
599    Tag: blume-20040304-symlinks
600    Description:
601    
602    Fixed problem with installer script (unix only) where bin/ml-yacc and
603    friends pointed (via symlinks) to absolute locations instead of just
604    .run-sml.  This was reported by Vesa A Norrman.
605    
606    ----------------------------------------------------------------------
607    Name: Matthias Blume (blume (at) tti - c (dot) org)
608    Date: 2004/02/13 14:50:00 CST
609    Tag: Release_110_45
610    Description:
611    
612    New working version (110.45). New bootfiles.
613    
614    ----------------------------------------------------------------------
615    Name: Matthias Blume (blume (at) tti - c (dot) org)
616    Date: 2004/01/26 15:15:15 CST
617    Tag: blume-20040126-toplevel
618    Description:
619    
620    Improved handling of exceptions at the interactive toplevel.
621    
622    ----------------------------------------------------------------------
623    Name: Matthias Blume (blume (at) tti - c (dot) org)
624    Date: 2004/01/26 11:25:00
625    Tag: blume-20040126-app
626    Description:
627    
628    Type of top-level "app" corrected.
629    Added code for setting vp_limitPtrMask to Win32-specific runtime.
630    
631    ----------------------------------------------------------------------
632    Name: Matthias Blume (blume (at) tti - c (dot) org)
633    Date: 2003/11/18 17:10 CST
634    Tag: blume-20031118-basis-fiddle
635    Description:
636    
637    - changed Timer interface to what might become the spec
638    - POSIX_FLAGS -> BIT_FLAGS according to spec
639    - some other minor discrepancies wrt. spec eliminated
640    
641    ----------------------------------------------------------------------
642    Name: Matthias Blume (blume (at) tti - c (dot) org)
643    Date: 2003/11/06 12:00:00 CST
644    Tag: Release_110_44
645    Description:
646    
647    New working version (110.44). New bootfiles.
648    
649    ----------------------------------------------------------------------
650    Name: Matthias Blume (blume (at) tti - c (dot) org)
651    Date: 2003/11/04 11:50:00 CST
652    Tag: blume-20031104-move-libraries
653    Description:
654    
655    Eliminated the "dont_move_libraries" directive in config/targets.
656    (The mechanism was broken and could not be fixed easily.  Moreover,
657    there does not seem to be any reason not to move all libraries into
658    lib during installation.  I originally implemented this directive as a
659    backward-compatibility feature when I first introduced the new CM.
660    Now that things have been stable for a long time and going back to the
661    old CM is not an option, there is no reason to keep it around.)
662    
663    ----------------------------------------------------------------------
664    Name: Matthias Blume (blume (at) tti - c (dot) org)
665    Date: 2003/11/03 16:00:00 CST
666    Tag: blume-20031103-installdir
667    Description:
668    
669    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
670    Richards for pointing out the problem and providing the solution.)
671    
672    ----------------------------------------------------------------------
673    Name: Matthias Blume (blume (at) tti - c (dot) org)
674    Date: 2003/10/01 17:05:00 CDT
675    Tag: blume-20031001-lal-mlrisc
676    Description:
677    
678    MLRISC bug fix from Lal.
679    
680    ----------------------------------------------------------------------
681    Name: Matthias Blume (blume (at) tti - c (dot) org)
682    Date: 2003/09/30 16:10:00 CDT
683    Tag: blume-20030930-primio-bat
684    Description:
685    
686    1. Added openVector, nullRd, and nullWr to PRIM_IO.
687    2. Improved .bat files (for Win32 port) to make things work under Win95.
688       (thanks to Aaron S. Hawley for this one)
689    
690    ----------------------------------------------------------------------
691    Name: Matthias Blume (blume (at) tti - c (dot) org)
692    Date: 2003/09/26 16:05:00 CDT
693    Tag: blume-20030926-wrappriv
694    Description:
695    
696    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
697    
698    ----------------------------------------------------------------------
699    Name: Matthias Blume (blume (at) tti - c (dot) org)
700    Date: 2003/09/26 15:00:00 CDT
701    Tag: blume-20030926-110_43_3
702    Description:
703    
704    - additional cleanup
705    - version number bump, NEW BOOTFILES
706    
707    ----------------------------------------------------------------------
708    Name: Matthias Blume (blume (at) tti - c (dot) org)
709    Date: 2003/09/26 12:00:00 CDT
710    Tag: blume-20030926-ppautoload
711    Description:
712    
713    I modified the read-eval-print loop so that the autoloader gets
714    invoked whenever the prettyprinter tries to look up a symbol that
715    is not currently defined in the toplevel environment but which
716    appears in CM's autoload registry.  As a result, we see far fewer of
717    those ?.Foo.Bar.xxx names in the prettyprinter's output.
718    
719    In addition to this I tried to clean up some pieces of the Basis
720    implementation (e.g., Socket, Word8Array) in order to prevent other
721    instances of these ?.Foo.Bar.xxx names from being printed.
722    
723    The mechanism that picks names for types still needs some work, though.
724    (Right now it seems that if there is a type A.t which is defined to
725    be B.u, but B is unavailable at toplevel, then A.t gets printed as
726    "?.B.u" although the perhaps more sensible solution would be to use
727    "A.t" in this case.  In other words, the prettyprinter should follow
728    a chain of DEFtycs not farther than there are corresponding toplevel
729    names in the current environment.)
730    
731    ----------------------------------------------------------------------
732    Name: Matthias Blume (blume (at) tti - c (dot) org)
733    Date: 2003/09/24 16:31:00 CDT
734    Tag: blume-20030924-installer
735    Description:
736    
737    Another installer tweak:  All the ML code for the installer is now
738    compiled during CMB.make and put into a little library called
739    $smlnj/installer.cm.  The installation then simply invokes
740    
741       sml -m $smlnj/installer.cm
742    
743    and everything happens automagically.
744    
745    Win32:  ML code senses value of environment variable SMLNJ_HOME.
746    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
747            and BINDIR.
748    
749    The new scheme guarantees that the ML code responsible for the installation
750    is in sync with the APIs of the main system.  Also, the installer is
751    somewhat faster because the installer script is precompiled.
752    
753    ----------------------------------------------------------------------
754    Name: Matthias Blume (blume (at) tti - c (dot) org)
755    Date: 2003/09/24 15:35:00 CDT
756    Tag: blume-20030924-synsock
757    Description:
758    
759    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
760    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
761    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
762    implementations of non-blocking operations from CML's Socket
763    structure.
764    
765    ----------------------------------------------------------------------
766    Name: Matthias Blume (blume (at) tti - c (dot) org)
767    Date: 2003/09/24 15:10:05 CDT
768    Tag: blume-20030924-sockets
769    Description:
770    
771    1. Fixed SOCKET API and implementation to match Basis spec.
772       This required changing the internal representation of sockets to one
773       that remembers (for each socket file descriptor) whether it is currently
774       blocking or non-blocking.  This state is maintained lazily (i.e., a system
775       call is made only if the state actually needs to change).
776    
777    2. OS-specific details of sockets were moved into separate files, thus
778       making it possible to unify the bulk of the socket implementations
779       between Unix and Win32.
780    
781    3. CML's socket API changed accordingly.
782       (Note that we need to remove non-blocking functions from this API
783       since they are redundant in the case of CML!)
784    
785    4. CML's socket implementation now makes use of non-blocking functions
786       provided by Basis, thus removing all OS-dependent code from this part
787       of CML.
788    
789    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
790    
791    ----------------------------------------------------------------------
792    Name: Matthias Blume (blume (at) tti - c (dot) org)
793    Date: 2003/09/22 12:10:00 CDT
794    Tag: blume-20030922-110_43_2
795    Description:
796    
797    Made a new interim version and bootfiles for developer's bootstrapping
798    convenience.
799    
800       110.43.2 -- NEW BOOTFILES
801    
802    ----------------------------------------------------------------------
803    Name: Matthias Blume (blume (at) tti - c (dot) org)
804    Date: 2003/09/19 15:55:00 CDT
805    Tag: blume-20030919-cmdir
806    Description:
807    
808    1. new-install.sh -> install.sh
809    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
810    3. tweaked installer so that another name instead of .cm can be chosen
811       at install time (by setting the CM_DIR_ARC environment variable
812       during installation); once installation is complete, the name is
813       fixed
814    
815    ----------------------------------------------------------------------
816    Name: Matthias Blume (blume (at) tti - c (dot) org)
817    Date: 2003/09/18 16:00:00 CDT
818    Tag: blume-20030918-110_43_1
819    Description:
820    
821    Made a new interim version and bootfiles for developer's bootstrapping
822    convenience.
823    
824       110.43.1 -- NEW BOOTFILES
825    
826    ----------------------------------------------------------------------
827    Name: Matthias Blume (blume (at) tti - c (dot) org)
828    Date: 2003/09/18 15:20:00 CDT
829    Tag: blume-20030918-misc
830    Description:
831    
832    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
833       this seems to be controversial at the moment) and used those in
834       Posix.ProcEnv.times.
835    
836    2. Added Time.{from,to}Nanoseconds to Time.
837    
838    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
839       For example, fromLargeInt never needs to look at more than 3 "big
840       digits" to get its 53 bits of precision.
841    
842    ----------------------------------------------------------------------
843    Name: Matthias Blume (blume (at) tti - c (dot) org)
844    Date: 2003/09/17 16:30:00 CDT
845    Tag: blume-20030917-real32-slices
846    Description:
847    
848    Added an entry to the primitive environment
849    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
850    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
851    
852    Removed some of the "magic" constants in real64.sml and replaced them
853    with code that generates these values from their corresponding
854    integer counterparts.
855    
856    Made all(?) the slice-related changes to the Basis and made everything
857    compile again...
858    
859    ----------------------------------------------------------------------
860    Name: Matthias Blume (blume (at) tti - c (dot) org)
861    Date: 2003/09/15 17:45:00 CDT
862    Tag: blume-20030915-rbase
863    Description:
864    
865    Fixed bug in Real.fromLargeInt.
866    
867    ----------------------------------------------------------------------
868    Name: Matthias Blume (blume (at) tti - c (dot) org)
869    Date: 2003/09/13 18:11:00 CDT
870    Tag: blume-20030913-libinstall
871    Description:
872    
873    Minor bugfix in config/libinstall (set anchor with path to
874    standalone tool after installing it, otherwise libraries that
875    need ml-lex or ml-yacc won't compile the first time the installer
876    runs).
877    
878    ----------------------------------------------------------------------
879    Name: Matthias Blume (blume (at) tti - c (dot) org)
880    Date: 2003/09/12 11:45:00 CDT
881    Tag: blume-20030912-various
882    Description:
883    
884     - fixed bug in Real.toLargeInt
885     - fixed bug in Posix.ProcEnv.times
886     - changed inputLine functions to return an option
887     - minor installer improvements / bugfixes
888     - changed default @SMLalloc parameter for x86/celeron to 64k
889    
890    ----------------------------------------------------------------------
891    Name: Matthias Blume (blume (at) tti - c (dot) org)
892    Date: 2003/09/09 22:00:00 CDT
893    Tag: Release_110_43
894    Description:
895    
896        New working release 110.43.  New bootfiles.
897    
898    ----------------------------------------------------------------------
899    Name: Matthias Blume (blume (at) tti - c (dot) org)
900    Date: 2003/09/09 19:20:00 CDT
901    Tag: blume-20030909-installer
902    Description:
903    
904    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
905    Modified config/install.bat to take advantage of it.  Also modified
906    config/install.sh (and called it config/new-install.sh) to take advantage
907    of it on Unix systems.  (The SML code is (supposed to be) platform-
908    independent.)
909    
910    The installer can now install everything under Win32
911    as well as under *nix as long as it compiles.
912    
913    Other changes:
914    
915      - made CML compile again under Win32
916      - made eXene compile under Win32 (by providing a fake structure UnixSock
917        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
918      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
919        same as type int;  under Win32 it isn't
920      - fixed some slice-related problems in the win32-specific parts of CML
921      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
922        (under Win32, the volume name is case-insensitive, and the
923        OS.Path code compares volume names for equality)
924    
925    ----------------------------------------------------------------------
926    Name: Matthias Blume (blume (at) tti - c (dot) org)
927    Date: 2003/09/08 11:55:00 CDT
928    Tag: blume-20030908-fullpath
929    Description:
930    
931    Made Win32 version of OS.FileSys.fullPath return current directory
932    when given an empty string.  This is what the spec says, and incidentally,
933    CM depends on it.  (CM otherwise goes into an infinite loop in certain
934    cases when presented with the name of a non-existing .cm file.)
935    
936    ----------------------------------------------------------------------
937    Name: Matthias Blume (blume (at) tti - c (dot) org)
938    Date: 2003/09/04 16:30:00 CDT
939    Tag: blume-20030905-slices-etc
940    Description:
941    
942      1. Changed interface to vectors and arrays in Basis to match
943         (draft) Basis spec.
944      2. Added signatures and implementations of slices according to
945         Basis spec.
946      3. Edited source code throughout the system to make it compile again
947         under 1. and 2.  (In some cases code had to be added to have it
948         match the new signatures.)
949      4. MLRISC should be backward-compatible: the copies of the originals
950         of files that needed to change under 3. were retained, the .cm files
951         check the compiler version number and use old versions when
952         appropriate.
953      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
954         dirstream -> string option (in accordance with Basis spec).
955      6. When generating code that counts lines, ml-lex used function
956         CharVector.foldli, taking advantage of its old interface.
957         This has been replaced with the corresponding code from
958         CharVectorSlice.  (html-lex must be re-lexed!)
959      7. BitArray in smlnj-lib/Util has been extended/modified to match the
960         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
961      8. Removed temporary additions (fromInternal, toInternal) from the
962         (now obsolete) IntInf in smlnj-lib/Util.
963      9. Cleaned up structure Byte.
964     10. Added localOffset, scan, and fromString to Date (according to spec).
965         Cleaned/corrected implementation of Date.
966         (Still need to check for correctness; implement better canonicalizeDate.)
967     11. Added "scan" to signature IEEE_REAL.
968     12. Some improvements to IntInf [in particular: efficiency-hack for
969         mod and rem when second operand is 2 (for parity checks).]
970     13. Changed representation of type Time.time, using a single IntInf.int
971         value counting microseconds.  This considerably simplified the
972         implementation of structure Time.  We now support negative time
973         values; scan and fromString handle signs.
974     14. Functor PrimIO now takes two additional arguments (VectorSlice and
975         ArraySlice).
976    
977    ----------------------------------------------------------------------
978    Name: Matthias Blume (blume (at) tti - c (dot) org)
979    Date: 2003/08/28 17:00:00 CDT
980    Tag: blume-20030828-intinf
981    Description:
982    
983    This is a major update which comes with a version number bump
984    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
985    and an implementation of IntInf in the Basis.
986    
987    There are a fairly large number of related changes and updates throughout
988    the system:
989    
990    Basis:
991            - Implemented IntInf.
992            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
993            - Added some missing Real64 operations, most notably Real.toLargeInt.
994            - Added FixedInt as a synonym for Int32.
995    
996    compiler:
997            * Added support for a built-in intinf type.
998              - literals
999              - pattern matching
1000              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1001              - overloading on literals and operations
1002    
1003            This required adding a primitive type intinf, some additional
1004            primops, and implementations for several non-trivial intinf
1005            operations in Core.  (The intinf type is completely abstract
1006            to the compiler; all operations get delegated back to the Core.)
1007    
1008            * Intinf equality is handled by polyequal.  However, the compiler
1009            does not print its usual warning in this case (since polyequal
1010            is the right thing to do there).
1011    
1012            * Improved the organization of structure InlineT.
1013    
1014            * A word about conversion primops:
1015            If conversions involving intinf do not cancel out during
1016            CPS contract, then the compiler must insert calls to Core functions.
1017            Since all core access must be resolved already during the FLINT
1018            translate phase, it would be too late a the time of CPS contract
1019            to add new Core calls.  For this reason, conversion primops
1020            for intinf carry two arguments:  1. the numeric argument that
1021            they are supposed to convert, and 2. the Core function that
1022            can help with this conversion if necessary.  If CPS contract
1023            eliminates a primop, then the associated Core function becomes
1024            dead and goes away.  Intinf conversion primops that do not get
1025            eliminated by CPS contract get rewritten into calls of their
1026            core functions by a separate, new phase.
1027    
1028    interactive system:
1029            - Control.Print.intinfDepth controls max length of intinf constants
1030            being printed. (Analogous to Control.Print.stringDepth.)
1031            - Cleanup in printutil and pputil: got rid of unused stuff and
1032            duplicates; replaced some of the code with code that makes better
1033            use of library functionality.
1034    
1035    CM:
1036            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1037            are no longer silent.
1038    
1039    CKIT:
1040            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1041            in favor of LargeInt -- which is now the same as IntInf.
1042            CKIT-knowledgable people should check whether this is what's
1043            intended and otherwise change things back to using Int32 or
1044            FixedInt.
1045    
1046    Throughout the code:
1047            Started using IntInf.int literals and built-in operations
1048            (e.g., comparison with 0) where this seems appropriate.
1049    
1050    
1051    ----------------------------------------------------------------------
1052    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1053    Date: 2003/08/13 11:36:00 CDT
1054    Tag: dbm-20030813-mcz-merge1
1055    Description:
1056    
1057    Merging changes from the mcz-branch development branch into trunk.
1058    These changes involve replacement of the emulated old prettyprinter
1059    interface with direct use of the SML/NJ Lib PP library, and fixing
1060    of a couple of bugs (895, 1186) relating to error messages.  A new
1061    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1062    has been added.
1063    
1064    ----------------------------------------------------------------------
1065    Name: Matthias Blume (blume (at) tti - c (dot) org)
1066    Date: 2003/08/11 15:45:00 CDT
1067    Tag: blume-20030811-windows
1068    Description:
1069    
1070    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1071    
1072        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1073    
1074    This patch restores SML/NJ's ability to run under win32.  There are a
1075    number of changes, including fixes for several bugs that had gone
1076    unnoticed until now:
1077    
1078      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1079        trying to run the win32 version from within cygwin.)
1080    
1081      - There are a number of simple .bat scripts that substitute for their
1082        corresponding Unix shell-scripts.  (See below.)
1083    
1084      - The internals of ml-build have been modified slightly.  The main
1085        difference is that instead of calling ".link-sml" (or link-sml.bat)
1086        using OS.Process.system, the ML process delegates this task back
1087        to the script.  Otherwise problems arise in mixed environments such
1088        as Cygwin where scripts look and work like Unix scripts, but
1089        where OS.Process.system cannot run them.
1090    
1091      - In CM, the srcpath pickler used native pathname syntax -- which
1092        is incorrect in the case of cross-compilation.  The new pickle format
1093        is independent of platform-specific naming conventions.
1094    
1095      - Path configuration files (such as lib/pathconfig) can now choose
1096        between native and standard syntax.  Placing a line of the form
1097    
1098            standard!
1099    
1100        into the file causes all subsequent paths to be interpreted using
1101        CM standard pathname syntax (= Unix conventions); a line
1102    
1103            native!
1104    
1105        switches back to native style.  This was needed so that
1106        path config files can be written portably, see src/system/pathconfig.
1107    
1108      - Runtime system:
1109    
1110          - win32-filesys.c:  get_file_time and set_file_time now
1111            access modification time, not creation time.
1112    
1113          - I/O code made aware of new array representation.
1114    
1115          - Bug fixes in X86.prim.masm.
1116    
1117       - src/system/makeml made aware of win32. (For use under cygwin
1118            and other Unix-environments for windows.)
1119    
1120       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1121         which caused BinIO.inputAll to fail consistently.
1122    
1123    .bat scripts:
1124    
1125       Windows .bat scripts assume that SMLNJ_HOME is defined.
1126    
1127       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1128         applications (sml, ml-yacc, ml-lex).
1129       - ml-build.bat: analogous to ml-build.
1130       - config\install.bat:  Analogous to config/install.sh.  This requires
1131         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1132         (nmake etc. must be on the path, and vcvars32 must have been run.)
1133         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1134         and the bootfile hierarchy must have been unpacked under
1135         sml.boot.x86-win32.
1136         The script is very primitive and does a poor job at error checking.
1137         It only installs the base system, ml-lex, and ml-yacc.  No other
1138         libraries are being installed (i.e., you get only those that
1139         are part of the compiler.)
1140       - link-sml.bat:  analogous to .link-sml, but not currently used
1141    
1142    Unrelated bug fixes:
1143    
1144      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1145        types.
1146      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1147        otherwise.)
1148    
1149    ----------------------------------------------------------------------
1150    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1151    Date: 2003/06/17
1152    Tag: macqueen-20030617-bug895
1153    Description:
1154    
1155    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1156    Tag will be used for new development branch (mcz-branch) for
1157    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1158    
1159    ----------------------------------------------------------------------
1160    Name: Matthias Blume (blume (at) tti - c (dot) org)
1161    Date: 2003/05/27 16:55:00 CDT
1162    Tag: blume-20030527-polyeq
1163    Description:
1164    
1165    Tried to eliminated most cases of polymorphic equality.
1166    
1167    ----------------------------------------------------------------------
1168    Name: Matthias Blume (blume (at) tti - c (dot) org)
1169    Date: 2003/05/21 17:45:00 CDT
1170    Tag: blume-20030517-complete
1171    Description:
1172    
1173    Two changes:
1174    
1175      1. Added a flag for controlling whether non-exhaustive bindings will
1176         be treated as errors (default is false).
1177      2. Cleaned up the *entire* source tree so that CMB.make goes through
1178         without a single non-exhaustive match- or bind warning.
1179    
1180    ----------------------------------------------------------------------
1181    Name: Matthias Blume (blume (at) tti - c (dot) org)
1182    Date: 2003/05/17 10:20:00 CDT
1183    Tag: blume-20030517-absyn
1184    Description:
1185    
1186    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1187    
1188       This mainly affects the quality of error messages.  However, some
1189       of the code is now more straightforward than before. (Treatment of
1190       the above four constructs in translate.sml is much simpler than
1191       the "macro-expansion" that was going on before.  Plus, the mach-
1192       compiler no longer gets invoked just to be able to compile an
1193       if-expression.)
1194    
1195    2. The ErrorMsg.Error exception is now caught and absorbed by the
1196       interactive loop.
1197    
1198    ----------------------------------------------------------------------
1199    Name: Allen Leung
1200    Date: 2003/05/16 13:05:00 CDT
1201    Tag: leunga-20030516-cygwin-runtime
1202    Description:
1203    
1204        Ported the runtime system to cygwin, which uses the unix
1205    x86-unix bin files.  Missing/buggy features:
1206    
1207       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1208         the Cygwin library.
1209       o Ctrl-C handling may be flaky.
1210       o Windows system calls and Windows I/O are not supported.
1211    
1212       A new set of binfiles is located at:
1213    
1214         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1215    
1216       This is only needed for bootstrapping the cygwin version of smlnj.
1217       Other x86 versions can use the existing binfiles.
1218    
1219    ----------------------------------------------------------------------
1220    Name: Matthias Blume
1221    Date: 2003/04/08 15:42:00 CDT
1222    Tag: blume-20030408-listpair
1223    Description:
1224    
1225    1. Added a target 'mlrisc' to installer.
1226    
1227    2. Added missing elements to structure ListPair.
1228    
1229    ----------------------------------------------------------------------
1230    Name: Allen Leung
1231    Date: 2003/01/07 10:40:00 EST
1232    Tag: leunga-20030107-int-rem
1233    Description:
1234    
1235    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1236    The arguments to the SUBL instruction were swapped.
1237    
1238    ----------------------------------------------------------------------
1239    Name: Matthias Blume
1240    Date: 2002/12/12 16:25:00 EST
1241    Tag: blume-20021212-risc-ra
1242    Description:
1243    
1244    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1245    sent the RA into an infinite loop when floating point registers get
1246    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1247    110.37 and 110.38.)
1248    There was another set of potential problems related to the handling of
1249    MLRISC annotations (but those did not yet cause real problems, apparently).
1250    
1251    ----------------------------------------------------------------------
1252    Name: Matthias Blume
1253    Date: 2002/12/06 22:40:00 EST
1254    Tag: blume-20021206-cm-fileid
1255    Description:
1256    
1257    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1258    This fixes the problem of CM getting confused by files that suddenly
1259    change their identity (e.g., by getting unlinked and recreated by some
1260    text editor such as vi).  There might be a better/cheaper/cleaner way
1261    of doing this, but for now this will have to do.
1262    
1263    ----------------------------------------------------------------------
1264    Name: Matthias Blume
1265    Date: 2002/10/28 09:50:00 EST
1266    Tag: blume-20021028-typecheck
1267    Description:
1268    
1269    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1270    
1271    ----------------------------------------------------------------------
1272    Name: Matthias Blume
1273    Date: 2002/10/17 09:10:00 EDT
1274    Tag: Release_110_42
1275    Description:
1276    
1277    In good old tradition, there has been a slight hiccup so that we have
1278    to patch 110.42 after the fact.   The old release tag has been replaced
1279    (see below).
1280    
1281    The change solves a problem with two competing approaches the
1282    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1283    each other's way.
1284    
1285    This change only affects the runtime system code and the installer script.
1286    (No new bootfiles.)
1287    
1288    ----------------------------------------------------------------------
1289    Name: Matthias Blume
1290    Date: 2002/10/16 12:00:00 EDT
1291    Tag: Release_110_42_removed
1292    Description:
1293    
1294    New working release.  New bootfiles.
1295    
1296    ----------------------------------------------------------------------
1297    Name: Matthias Blume
1298    Date: 2002/10/10 13:10:00 EDT
1299    Tag: blume-20021010-ppc-divs
1300    Description:
1301    
1302    The mltree operator DIVS must be implemented with an overflow check on
1303    the PPC because the hardware indicates divide-by-zero using "overflow" as
1304    well.
1305    
1306    ----------------------------------------------------------------------
1307    Name: Matthias Blume
1308    Date: 2002/07/23 11:45:00
1309    Tag: blume-20020723-smlnj-home
1310    Description:
1311    
1312    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1313    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1314    CM_PATHCONFIG is also set), the path configuration file is assumed
1315    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1316    the entire tree to some other place and everything will "just work".
1317    
1318    (Companion commands such as ml-build and ml-makedepend also sense this
1319    variable.)
1320    
1321    ----------------------------------------------------------------------
1322    Name: Matthias Blume
1323    Date: 2002/07/12 21:19:00 EDT
1324    Tag: blume-20020712-liveness
1325    Description:
1326    
1327    Exported two useful "step" functions from liveness module (MLRISC).
1328    
1329    ----------------------------------------------------------------------
1330    Name: Matthias Blume
1331    Date: 2002/07/05 16:00 EDT
1332    Tag: Release_110_41
1333    Description:
1334    
1335    New working release.  New bootfiles.
1336    
1337    ----------------------------------------------------------------------
1338    Name: Matthias Blume
1339    Date: 2002/07/05 10:25:00 EDT
1340    Tag: blume-20020705-btimp
1341    Description:
1342    
1343    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1344    other clients can set up backtracing support.
1345    
1346    ----------------------------------------------------------------------
1347    Name: Matthias Blume
1348    Date: 2002/06/25 17:23:00 EDT
1349    Tag: blume-20020625-fpmax
1350    Description:
1351    
1352    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1353    case.  (The sense of the isNaN test was reversed -- which made min and
1354    max always return their first argument.)
1355    
1356    ----------------------------------------------------------------------
1357    Name: Matthias Blume
1358    Date: 2002/06/11
1359    Tag: blume-20020611-unixpath
1360    Description:
1361    
1362    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1363    
1364    ----------------------------------------------------------------------
1365    Name: Matthias Blume
1366    Date: 2002/06/10 16:35:00 EDT
1367    Tag: blume-20020610-ieeereal
1368    Description:
1369    
1370    I back-ported my implementation of IEEEReal.fromString from the
1371    idlbasis-devel branch so that we can test it.
1372    
1373    Another small change is that ppDec tries to give more information
1374    than just "<sig>" in the case of functors.  However, this code is
1375    broken in some mysterious way if the functor's body's signature
1376    has not been declared by ascription but gets inferred from the
1377    implementation.  This needs fixing...
1378    
1379    ----------------------------------------------------------------------
1380    Name: Matthias Blume
1381    Date: 2002/05/31
1382    Tag: blume-20020531-btrace-mode
1383    Description:
1384    
1385    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1386    the wayside when I switched over to using Controls everywhere.)
1387    
1388    ----------------------------------------------------------------------
1389    Name: Lal George
1390    Date: 2002/05/23 12:21:40 EDT
1391    Tag: george-20020523-visual-labels
1392    Description:
1393    
1394    Labels are now displayed in the graphical output to make
1395    the fall-through and target blocks obvious.
1396    
1397    ----------------------------------------------------------------------
1398    Name: Matthias Blume
1399    Date: 2002/05/22 11:03:00 EDT
1400    Tag: blume-20020522-shrink
1401    Description:
1402    
1403    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1404    situation that comes up when allocating huge arrays.
1405    
1406    ----------------------------------------------------------------------
1407    Name: Matthias Blume
1408    Date: 2002/05/21 16:00:00 EDT
1409    Tag: Release_110_40
1410    Description:
1411    
1412    New working release (110.40).  New bootfiles.
1413    
1414    [Also: John Reppy fixed GC bug 1131.]
1415    
1416    ----------------------------------------------------------------------
1417    Name: Matthias Blume
1418    Date: 2002/05/21 12:35:00 EDT
1419    Tag: blume-20020521-cmdoc
1420    Description:
1421    
1422    CM documentation update.
1423    
1424    ----------------------------------------------------------------------
1425    Name: Matthias Blume
1426    Date: 2002/05/21 10:55:00 EDT
1427    Tag: blume-20020521-misc
1428    Description:
1429    
1430    - John tweaked runtime to be silent on heap export (except when
1431      GC messages are on).
1432    - I added a few more things (cross-compiling versions of CMB) to
1433      config/preloads (as suggestions).
1434    
1435    ----------------------------------------------------------------------
1436    Name: Matthias Blume
1437    Date: 2002/05/20 22:25:00 EDT
1438    Tag: blume-20020520-controls
1439    Description:
1440    
1441    - Added ControlUtil structure to control-lib.cm.
1442    - Use it throughout.
1443    - Used Controls facility to define MLRISC controls (as opposed to
1444      registering MLRISC control ref cells with Controls after the fact)
1445    - Fixed messed-up controls priorities.
1446    
1447    * Removed again all the stuff from config/preloads that one wouldn't
1448      be able to preload at the time the initial heap image is built.
1449      (Many libraries, e.g., CML, do not exist yet at this time.  The
1450       only libraries that can be preloaded via config/preloads are those
1451       that come bundled with the bootfiles.)
1452    
1453    ----------------------------------------------------------------------
1454    Name: Matthias Blume
1455    Date: 2002/05/20 10:59:00 EDT
1456    Tag: blume-20020520-preloads
1457    Description:
1458    
1459    Added a lot of commented-out suggestions for things to be included
1460    in config/preloads.
1461    
1462    ----------------------------------------------------------------------
1463    Name: Allen Leung
1464    Date: 2002/05/18 14:20:00 EDT
1465    Tag: leunga-20020518-mdl
1466    Description:
1467    
1468    o Made the mdl tool stuff compile and run again.
1469    o I've disabled all the stuff that depends on RTL specifications; they
1470      are all badly broken anyway.
1471    
1472    ----------------------------------------------------------------------
1473    Name: Matthias Blume
1474    Date: 2002/05/17 16:49:00 EDT
1475    Tag: blume-20020517-controls
1476    Description:
1477    
1478    0. John Reppy made several modifications to the SML/NJ library.
1479       In particular, there is a shiny new controls-lib.cm.
1480    
1481    1. Pushed new controls interface through compiler so that everything
1482       compiles again.
1483    
1484    2. Added FormatComb and FORMAT_COMB to the CML version of the
1485       SML/NJ library (so that CML compiles again).
1486    
1487    3. Modified init scripts because XXX_DEFAULT environment variables
1488       are no longer with us.  (Boot-time initialization is now done
1489       using the same environment variables that are also used for
1490       startup-time initialization of controls.)
1491    
1492    ----------------------------------------------------------------------
1493    Name: Lal George
1494    Date: 2002/05/15 09:20:10 EDT
1495    Tag: george-20020515-pseudo-op-decls
1496    Description:
1497    
1498            All pseudo-ops emitted before the first segment declaration
1499    such as TEXT, DATA, and BSS directives are assumed to be global
1500    declarations and are emitted first in the assembly file. This is
1501    useful in a number of situations where one has pseudo-ops that are not
1502    specific to any segment, and also works around the constraint that one
1503    cannot have client pseudo-ops in the TEXT segment.
1504    
1505            Because no segment is associated with these declarations it is
1506    an error to allocate any space or objects before the first segment
1507    directive and an exception will be raised. However, we cannot make
1508    this check for client pseudo-ops.
1509    
1510            These top level declarations are a field in the CFG graph_info.
1511    In theory you can continue to add to this field after the CFG has been
1512    built -- provided you know what you are doing;-)
1513    
1514    ----------------------------------------------------------------------
1515    Name: Matthias Blume
1516    Date: 2002/05/13 16:40:00 EDT
1517    Tag: blume-20020513-pp-etc
1518    Description:
1519    
1520    A few minor bugfixes:
1521    
1522      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1523        (Bogus printouts for redefined bindings still occur.  Compiler
1524        bug should no longer occur now.  We need to redo the prettyprinter
1525        from scratch.)
1526    
1527      - CM pathname printer now also adds escape sequences for ( and )
1528    
1529      - commend and docu fixes for ml-nlffi
1530    
1531    ----------------------------------------------------------------------
1532    Name: Matthias Blume
1533    Date: 2002/05/10 16:40:00 EDT
1534    Tag: blume-20020510-erg-textio
1535    Description:
1536    
1537    Applied the following bugfix provided by Emden Gansner:
1538    
1539        Output is corrupted when outputSubstr is used rather than output.
1540    
1541        The problem occurs when a substring
1542    
1543            ss = (s, dataStart, dataLen)
1544    
1545        where dataStart > 0, fills a stream buffer with avail bytes left.
1546        avail bytes of s, starting at index dataStart, are copied into the
1547        buffer, the buffer is flushed, and then the remaining dataLen-avail
1548        bytes of ss are copied into the beginning of the buffer. Instead of
1549        starting this copy at index dataStart+avail in s, the current code
1550        starts the copy at index avail.
1551    
1552        Fix:
1553        In text-io-fn.sml, change line 695 from
1554             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1555        to
1556             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1557    
1558    ----------------------------------------------------------------------
1559    Name: Matthias Blume
1560    Date: 2002/04/12 13:55:00 EDT
1561    Tag: blume-20020412-assyntax
1562    Description:
1563    
1564    1. Grabbed newer assyntax.h from the XFree86 project.
1565    2. Fiddled with how to compile X86.prim.asm without warnings.
1566    3. (Very) Minor cleanup in CM.
1567    
1568    ----------------------------------------------------------------------
1569    Name: Matthias Blume
1570    Date: 2002/04/01 (no joke!) 17:07:00 EST
1571    Tag: blume-20020401-x86div
1572    Description:
1573    
1574    Added full support for div/mod/rem/quot on the x86, using the machine
1575    instruction's two results (without clumsily recomputing the remainder)
1576    directly where appropriate.
1577    
1578    Some more extensive power-of-two support was added to the x86 instruction
1579    selector (avoiding expensive divs, mods, and muls where they can be
1580    replaced with cheaper shifts and masks).  However, this sort of thing
1581    ought to be done earlier, e.g., within the CPS optimizer so that
1582    all architectures benefit from it.
1583    
1584    The compiler compiles to a fixed point, but changes might be somewhat
1585    fragile nevertheless.  Please, report any strange things that you might
1586    see wrt. div/mod/quot/rem...
1587    
1588    ----------------------------------------------------------------------
1589    Name: Matthias Blume
1590    Date: 2002/03/29 17:22:00
1591    Tag: blume-20020329-div
1592    Description:
1593    
1594    Fixed my broken div/mod logic.  Unfortunately, this means that the
1595    inline code for div/mod now has one more comparison than before.
1596    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
1597    The problem was with quotient = 0, because that alone does not tell
1598    us which way the rounding went.  One then has to look at whether
1599    remainder and divisor have the same sign...  :(
1600    
1601    Anyway, I replaced the bootfiles with fresh ones...
1602    
1603    ----------------------------------------------------------------------
1604    Name: Matthias Blume
1605    Date: 2002/03/29 14:10:00 EST
1606    Tag: blume-20020329-inlprims
1607    Description:
1608    
1609    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
1610    
1611    Primops have changed. This means that the bin/boot-file formats have
1612    changed as well.
1613    
1614    To make sure that there is no confusion, I made a new version.
1615    
1616    
1617    CHANGES:
1618    
1619    * removed REMT from mltree (remainder should never overflow).
1620    
1621    * added primops to deal with divisions of all flavors to the frontend
1622    
1623    * handled these primops all the way through so they map to their respective
1624      MLRISC support
1625    
1626    * used these primops in the implementation of Int, Int32, Word, Word32
1627    
1628    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
1629    
1630    * parameterized INLMIN, INLMAX, and INLABS by a numkind
1631    
1632    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
1633      floating point
1634    
1635    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
1636      and Real (but Real.abs maps to a separate floating-point-only primop)
1637    
1638    
1639    TODO items:
1640    
1641    * Hacked Alpha32 instruction selection, disabling the selection of REMx
1642      instructions because the machine instruction encoder cannot handle
1643      them.  (Hppa, PPC, and Sparc instruction selection did not handle
1644      REM in the first place, and REM is supported by the x86 machine coder.)
1645    
1646    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
1647      selection phase.  (The two can be streamlined because the hardware
1648      delivers both quotient and remainder at the same time anyway.)
1649    
1650    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
1651      (Currently the behavior is inconsistent both across architectures and
1652      wrt. the draft Basis spec.)
1653    
1654    * Word8 should eventually be handled natively, too.
1655    
1656    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
1657      as if there currently is no execution path that could trigger it in
1658      SML/NJ.  (The assumptions underlying functions arith and promotable do not
1659      hold for things like multiplication and division.)
1660    
1661    ----------------------------------------------------------------------
1662    Name: Matthias Blume
1663    Date: 2002/03/27 16:27:00 EST
1664    Tag: blume-20020327-mlrisc-divisions
1665    Description:
1666    
1667    Added support for all four division operations (ML's div, mod, quot,
1668    and rem) to MLRISC.  In the course of doing so, I also rationalized
1669    the naming (no more annoying switch-around of DIV and QUOT), by
1670    parameterizing the operation by div_rounding_mode (which can be either
1671    DIV_TO_ZERO or DIV_TO_NEGINF).
1672    
1673    The generic MLTreeGen functor takes care of compiling all four
1674    operations down to only round-to-zero div.
1675    
1676    Missing pieces:
1677    
1678      * Doing something smarter than relying on MLTreeGen on architectures
1679        like, e.g., the x86 where hardware division delivers both quotient and
1680        remainder at the same time.  With this, the implementation of the
1681        round-to-neginf operations could be further streamlined.
1682    
1683      * Remove inlining support for div/mod/rem from the frontend and replace it
1684        with primops that get carried through to the backend.  Do this for all
1685        int and word types.
1686    
1687    ----------------------------------------------------------------------
1688    Name: Matthias Blume
1689    Date: 2002/03/25 17:25:00 EST
1690    Tag: blume-20020325-divmod
1691    Description:
1692    
1693    I improved (hopefully without breaking them) the implementation of Int.div,
1694    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
1695    advantage of the following observations:
1696    
1697      Let  q = x quot y      r = x rem y
1698           d = x div  y      m = x mod y
1699    
1700    where "quot" is the round-to-zero version of integer division that
1701    hardware usually provides.  Then we have:
1702    
1703         r = x - q * y        where neither the * nor the - will overflow
1704         d = if q >= 0 orelse x = q * y then q else q - 1
1705                              where neither the * nor the - will overflow
1706         m = if q >= 0 orelse r = 0 then r else r + y
1707                              where the + will not overflow
1708    
1709    This results in substantial simplification of the generated code.
1710    The following table shows the number of CFG nodes and edges generated
1711    for
1712            fun f (x, y) = x OPER y
1713            (* with OPER \in div, mod, quot, rem *)
1714    
1715    
1716        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
1717        --------------------------------------------------------
1718         div |         24 |         39 |         12 |         16
1719         mod |         41 |         71 |         12 |         16
1720        quot |          8 |         10 |          8 |         10
1721         rem |         10 |         14 |          8 |         10
1722    
1723    
1724    ----------------------------------------------------------------------
1725    Name: Matthias Blume
1726    Date: 2002/03/25 22:06:00 EST
1727    Tag: blume-20020325-cprotobug
1728    Description:
1729    
1730    Fixed a bug in cproto (c prototype decoder).
1731    
1732    ----------------------------------------------------------------------
1733    Name: Matthias Blume
1734    Date: 2002/03/25 16:00:00 EST
1735    Tag: blume-20020325-raw-primops
1736    Description:
1737    
1738    I did some cleanup to Allen's new primop code and
1739    replaced yesterday's bootfiles with new ones.
1740    (But they are stored in the same place.)
1741    
1742    ----------------------------------------------------------------------
1743    Name: Matthias Blume
1744    Date: 2002/03/24 22:40:00 EST
1745    Tag: blume-20020324-bootfiles
1746    Description:
1747    
1748    Made the bootfiles that Allen asked for.
1749    
1750    ----------------------------------------------------------------------
1751    Name: Allen Leung
1752    Date: 2002/03/23 15:50:00 EST
1753    Tag: leunga-20020323-flint-cps-rcc-primops
1754    Description:
1755    
1756      1. Changes to FLINT primops:
1757    
1758        (* make a call to a C-function;
1759         * The primop carries C function prototype information and specifies
1760         * which of its (ML-) arguments are floating point. C prototype
1761         * information is for use by the backend, ML information is for
1762         * use by the CPS converter. *)
1763      | RAW_CCALL of { c_proto: CTypes.c_proto,
1764                       ml_args: ccall_type list,
1765                       ml_res_opt: ccall_type option,
1766                       reentrant : bool
1767                     } option
1768       (* Allocate uninitialized storage on the heap.
1769        * The record is meant to hold short-lived C objects, i.e., they
1770        * are not ML pointers.  With the tag, the representation is
1771        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1772        *)
1773      | RAW_RECORD of {tag:bool,sz:int}
1774      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1775    
1776      2.  These CPS primops are now overloaded:
1777    
1778           rawload of {kind:numkind}
1779           rawstore of {kind:numkind}
1780    
1781          The one argument form is:
1782    
1783             rawload {kind} address
1784    
1785          The two argument form is:
1786    
1787             rawload {kind} [ml object, byte-offset]
1788    
1789      3. RAW_CCALL/RCC now takes two extra arguments:
1790    
1791         a. The first is whether the C call is reentrant, i.e., whether
1792            ML state should be saved and restored.
1793         b. The second argument is a string argument specifying the name of
1794            library and the C function.
1795    
1796         These things are currently not handled in the code generator, yet.
1797    
1798      4. In CProto,
1799    
1800         An encoding type of "bool" means "ml object" and is mapped into
1801         C prototype of PTR.  Note that "bool" is different than "string",
1802         even though "string" is also mapped into PTR, because "bool"
1803         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
1804    
1805      5. Pickler/unpicker
1806    
1807         Changed to handle RAW_RECORD and newest RAW_CCALL
1808    
1809      6. MLRiscGen,
1810    
1811         1. Changed to handle the new rawload/rawstore/rawrecord operators.
1812         2. Code for handling C Calls has been moved to a new module CPSCCalls,
1813            in the file CodeGen/cpscompile/cps-c-calls.sml
1814    
1815      7. Added the conditional move operator
1816    
1817             condmove of branch
1818    
1819         to cps.  Generation of this is still buggy so it is currently
1820         disabled.
1821    
1822    ----------------------------------------------------------------------
1823    Name: Lal George
1824    Date: 2002/03/22 14:18:25 EST
1825    Tag: george-20020322-cps-branch-prob
1826    Description:
1827    
1828    Implemented the Ball-Larus branch prediction-heuristics, and
1829    incorporated graphical viewers for control flow graphs.
1830    
1831    Ball-Larus Heuristics:
1832    ---------------------
1833    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
1834    
1835    By design it uses the Dempster-Shafer theory for combining
1836    probabilities.  For example, in the function:
1837    
1838        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
1839    
1840    the ball-larus heuristics predicts that the n=0 is unlikely
1841    (OH-heuristic), and the 'then' branch is unlikely because of the
1842    RH-heuristic -- giving the 'then' branch an even lower combined
1843    probability using the Dempster-Shafer theory.
1844    
1845    Finally, John Reppy's loop analysis in MLRISC, further lowers the
1846    probability of the 'then' branch because of the loop in the else
1847    branch.
1848    
1849    
1850    Graphical Viewing:
1851    ------------------
1852    I merely plugged in Allen's graphical viewers into the compiler. The
1853    additional code is not much. At the top level, saying:
1854    
1855            Control.MLRISC.getFlag "cfg-graphical-view" := true;
1856    
1857    will display the graphical view of the control flow graph just before
1858    back-patching.  daVinci must be in your path for this to work. If
1859    daVinci is not available, then the default viewer can be changed
1860    using:
1861    
1862            Control.MLRISC.getString "viewer"
1863    
1864    which can be set to "dot" or "vcg" for the corresponding viewers. Of
1865    course, these viewers must be in your path.
1866    
1867    The above will display the compilation unit at the level of clusters,
1868    many of which are small, boring, and un-interesting. Also setting:
1869    
1870            Control.MLRISC.getInt "cfg-graphical-view_size"
1871    
1872    will display clusters that are larger than the value set by the above.
1873    
1874    
1875    ----------------------------------------------------------------------
1876    Name: Matthias Blume
1877    Date: 2002/03/21 22:20:00 EST
1878    Tag: blume-20020321-kmp-bugfix
1879    Description:
1880    
1881    Changed the interface to the KMP routine in PreString and fixed
1882    a minor bug in one place where it was used.
1883    
1884    ----------------------------------------------------------------------
1885    Name: Allen Leung
1886    Date: 2002/03/21 20:30:00 EST
1887    Tag: leunga-20020321-cfg
1888    Description:
1889    
1890      Fixed a potential problem in cfg edge splitting.
1891    
1892    ----------------------------------------------------------------------
1893    Name: Allen Leung
1894    Date: 2002/03/21 17:15:00 EST
1895    Tag: leunga-20020321-x86-fp-cfg
1896    Description:
1897    
1898      1. Recoded the buggy parts of x86-fp.
1899    
1900         a. All the block reordering code has been removed.
1901            We now depend on the block placement phases to do this work.
1902    
1903         b. Critical edge splitting code has been simplified and moved into the
1904            CFG modules, as where they belong.
1905    
1906         Both of these were quite buggy and complex.  The code is now much, much
1907         simpler.
1908    
1909      2. X86 backend.
1910    
1911         a. Added instructions for 64-bit support.  Instruction selection for
1912            64-bit has not been committed, however, since that
1913            requires changes to MLTREE which haven't been approved by
1914            Lal and John.
1915    
1916         b. Added support for FUCOMI and FUCOMIP when generating code for
1917            PentiumPro and above.  We only generate these instructions in
1918            the fast-fp mode.
1919    
1920         c. Added cases for JP and JNP in X86FreqProps.
1921    
1922      3. CFG
1923    
1924         CFG now has a bunch of methods for edge splitting and merging.
1925    
1926      4. Machine description.
1927    
1928         John's simplification of MLTREE_BASIS.fcond broke a few machine
1929         description things:
1930    
1931         rtl-build.{sig,sml} and hppa.mdl fixed.
1932    
1933         NOTE: the machine description stuff in the repository is still broken.
1934               Again, I can't put my fixes in because that involves
1935               changes to MLTREE.
1936    
1937    ----------------------------------------------------------------------
1938    Name: Matthias Blume
1939    Date: 2002/03/20 15:55:00 EST
1940    Tag: blume-20020320-kmp
1941    Description:
1942    
1943    Implemented Knuth-Morris-Pratt string matching in PreString and used
1944    it for String.isSubstring, Substring.isSubstring, and
1945    Substring.position.
1946    
1947    (Might need some stress-testing.  Simple examples worked fine.)
1948    
1949    ----------------------------------------------------------------------
1950    Name: Matthias Blume
1951    Date: 2002/03/19 16:37:00 EST
1952    Tag: blume-20020319-witnesses
1953    Description:
1954    
1955    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
1956    
1957    This implements a generic mechanism for changing constness qualifiers
1958    anywhere within big C types without resorting to outright "casts".
1959    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
1960    modify the constness at the outermost level.)
1961    The implementation of "convert" is based on the idea of "witness"
1962    values -- values that are not used by the operation but whose types
1963    "testify" to their applicability.  On the implementation side, "convert"
1964    is simply a projection (returning its second curried argument).  With
1965    cross-module inlining, it should not result in any machine code being
1966    generated.
1967    
1968    ----------------------------------------------------------------------
1969    Name: Matthias Blume
1970    Date: 2002/03/15 16:40:00 EST
1971    Tag: blume-20020315-basis
1972    Description:
1973    
1974    Provided (preliminary?) implementations for
1975    
1976      {String,Substring}.{concatWith,isSuffix,isSubstring}
1977    
1978    and
1979    
1980      Substring.full
1981    
1982    Those are in the Basis spec but they were missing in SML/NJ.
1983    
1984    ----------------------------------------------------------------------
1985    Name: Matthias Blume
1986    Date: 2002/03/14 21:30:00 EST
1987    Tag: blume-20020314-controls
1988    Description:
1989    
1990    Controls:
1991    ---------
1992    
1993    1. Factored out the recently-added Controls : CONTROLS stuff and put
1994       it into its own library $/controls-lib.cm.  The source tree for
1995       this is under src/smlnj-lib/Controls.
1996    
1997    2. Changed the names of types and functions in this interface, so they
1998       make a bit more "sense":
1999    
2000          module -> registry
2001          'a registry -> 'a group
2002    
2003    3. The interface now deals in ref cells only.  The getter/setter interface
2004       is (mostly) gone.
2005    
2006    4. Added a function that lets one register an already-existing ref cell.
2007    
2008    5. Made the corresponding modifications to the rest of the code so that
2009       everything compiles again.
2010    
2011    6. Changed the implementation of Controls.MLRISC back to something closer
2012       to the original.  In particular, this module (and therefore MLRISC)
2013       does not depend on Controls.  There now is some link-time code in
2014       int-sys.sml that registers the MLRISC controls with the Controls
2015       module.
2016    
2017    CM:
2018    ---
2019    
2020      * One can now specify the lambda-split aggressiveness in init.cmi.
2021    
2022    ----------------------------------------------------------------------
2023    Name: Allen Leung
2024    Date: 2002/03/13 17:30:00 EST
2025    Tag: leunga-20020313-x86-fp-unary
2026    Description:
2027    
2028    Bug fix for:
2029    
2030    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2031    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2032    > - fun f(x,(y,z)) = Real.~ y;
2033    > [autoloading]
2034    > [autoloading done]
2035    >       fchsl   (%eax), 184(%esp)
2036    > Error: MLRisc bug: X86MCEmitter.emitInstr
2037    >
2038    > uncaught exception Error
2039    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2040    
2041    The problem was that the code generator did not generate any fp registers
2042    in this case, and the ra didn't know that it needed to run the X86FP phase to
2043    translate the pseudo fp instruction.   This only happened with unary fp
2044    operators in certain situations.
2045    
2046    ----------------------------------------------------------------------
2047    Name: Matthias Blume
2048    Date: 2002/03/13 14:00:00 EST
2049    Tag: blume-20020313-overload-etc
2050    Description:
2051    
2052    1. Added _overload as a synonym for overload for backward compatibility.
2053       (Control.overloadKW must be true for either version to be accepted.)
2054    
2055    2. Fixed bug in install script that caused more things to be installed
2056       than what was requested in config/targets.
2057    
2058    3. Made CM aware of the (_)overload construct so that autoloading
2059       works.
2060    
2061    ----------------------------------------------------------------------
2062    Name: Matthias Blume
2063    Date: 2002/03/12 22:03:00 EST
2064    Tag: blume-20020312-url
2065    Description:
2066    
2067    Forgot to update BOOT and srcarchiveurl.
2068    
2069    ----------------------------------------------------------------------
2070    Name: Matthias Blume
2071    Date: 2002/03/12 17:30:00 EST
2072    Tag: blume-20020312-version110392
2073    Description:
2074    
2075    Yet another version number bump (because of small changes to the
2076    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2077    
2078    Changes:
2079    
2080      The new pid generation scheme described a few weeks ago was overly
2081      complicated.  I implemented a new mechanism that is simpler and
2082      provides a bit more "stability":  Once CM has seen a compilation
2083      unit, it keeps its identity constant (as long as you do not delete
2084      those crucial CM/GUID/* files).  This means that when you change
2085      an interface, compile, then go back to the old interface, and
2086      compile again, you arrive at the original pid.
2087    
2088      There now also is a mechanism that instructs CM to use the plain
2089      environment hash as a module's pid (effectively making its GUID
2090      the empty string).  For this, "noguid" must be specified as an
2091      option to the .sml file in question within its .cm file.
2092      This is most useful for code that is being generated by tools such
2093      as ml-nlffigen (because during development programmers tend to
2094      erase the tool's entire output directory tree including CM's cached
2095      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2096      revert to the old, broken behavior of SML/NJ, but in specific cases
2097      where there is no danger of interface confusion, its use is ok
2098      (I think).
2099    
2100      ml-nlffigen by default generates "noguid" annotations.  They can be
2101      turned off by specifying -guid in its command line.
2102    
2103    ----------------------------------------------------------------------
2104    Name: Lal George
2105    Date: 2002/03/12 12 14:42:36 EST
2106    Tag: george-20020312-frequency-computation
2107    Description:
2108    
2109    Integrated jump chaining and static block frequency into the
2110    compiler. More details and numbers later.
2111    
2112    ----------------------------------------------------------------------
2113    Name: Lal George
2114    Date: 2002/03/11 11 22:38:53 EST
2115    Tag: george-20020311-jump-chain-elim
2116    Description:
2117    
2118    Tested the jump chain elimination on all architectures (except the
2119    hppa).  This is on by default right now and is profitable for the
2120    alpha and x86, however, it may not be profitable for the sparc and ppc
2121    when compiling the compiler.
2122    
2123    The gc test will typically jump to a label at the end of the cluster,
2124    where there is another jump to an external cluster containing the actual
2125    code to invoke gc. This is to allow factoring of common gc invocation
2126    sequences. That is to say, we generate:
2127    
2128            f:
2129               testgc
2130               ja   L1      % jump if above to L1
2131    
2132            L1:
2133               jmp L2
2134    
2135    
2136    After jump chain elimination the 'ja L1' instructions is converted to
2137    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2138    up being implemented in their long form (if L2 is far away) using:
2139    
2140            jbe     L3      % jump if below or equal to L3
2141            jmp     L2
2142         L3:
2143            ...
2144    
2145    
2146    For large compilation units L2  may be far away.
2147    
2148    
2149    ----------------------------------------------------------------------
2150    Name: Matthias Blume
2151    Date: 2002/03/11 13:30:00 EST
2152    Tag: blume-20020311-mltreeeval
2153    Description:
2154    
2155    A functor parameter was missing.
2156    
2157    ----------------------------------------------------------------------
2158    Name: Allen Leung
2159    Date: 2002/03/11 10:30:00 EST
2160    Tag: leunga-20020311-runtime-string0
2161    Description:
2162    
2163       The representation of the empty string now points to a
2164    legal null terminated C string instead of unit.  It is now possible
2165    to convert an ML string into C string with InlineT.CharVector.getData.
2166    This compiles into one single machine instruction.
2167    
2168    ----------------------------------------------------------------------
2169    Name: Allen Leung
2170    Date: 2002/03/10 23:55:00 EST
2171    Tag: leunga-20020310-x86-call
2172    Description:
2173    
2174       Added machine generation for CALL instruction (relative displacement mode)
2175    
2176    ----------------------------------------------------------------------
2177    Name: Matthias Blume
2178    Date: 2002/03/08 16:05:00
2179    Tag: blume-20020308-entrypoints
2180    Description:
2181    
2182    Version number bumped to 110.39.1.  NEW BOOTFILES!
2183    
2184    Entrypoints: non-zero offset into a code object where execution should begin.
2185    
2186    - Added the notion of an entrypoint to CodeObj.
2187    - Added reading/writing of entrypoint info to Binfile.
2188    - Made runtime system bootloader aware of entrypoints.
2189    - Use the address of the label of the first function given to mlriscGen
2190      as the entrypoint.  This address is currently always 0, but it will
2191      not be 0 once we turn on block placement.
2192    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2193      with entry points) from mlriscGen.
2194    
2195    ----------------------------------------------------------------------
2196    Name: Allen Leung
2197    Date: 2002/03/07 20:45:00 EST
2198    Tag: leunga-20020307-x86-cmov
2199    Description:
2200    
2201       Bug fixes for CMOVcc on x86.
2202    
2203       1. Added machine code generation for CMOVcc
2204       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2205       3. CMOVcc cannot have an immediate operand as argument.
2206    
2207    ----------------------------------------------------------------------
2208    Name: Matthias Blume
2209    Date: 2002/03/07 16:15:00 EST
2210    Tag: blume-20020307-controls
2211    Description:
2212    
2213    This is a very large but mostly boring patch which makes (almost)
2214    every tuneable compiler knob (i.e., pretty much everything under
2215    Control.* plus a few other things) configurable via both the command
2216    line and environment variables in the style CM did its configuration
2217    until now.
2218    
2219    Try starting sml with '-h' (or, if you are brave, '-H')
2220    
2221    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2222    implements the underlying generic mechanism.
2223    
2224    The interface to some of the existing such facilities has changed somewhat.
2225    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2226    (The getFoo interface is still there for backward-compatibility, but its
2227    use is deprecated.)
2228    
2229    The ml-build script passes -Cxxx=yyy command-line arguments through so
2230    that one can now twiddle the compiler settings when using this "batch"
2231    compiler.
2232    
2233    TODO items:
2234    
2235    We should go through and throw out all controls that are no longer
2236    connected to anything.  Moreover, we should go through and provide
2237    meaningful (and correct!) documentation strings for those controls
2238    that still are connected.
2239    
2240    Currently, multiple calls to Controls.new are accepted (only the first
2241    has any effect).  Eventually we should make sure that every control
2242    is being made (via Controls.new) exactly once.  Future access can then
2243    be done using Controls.acc.
2244    
2245    Finally, it would probably be a good idea to use the getter-setter
2246    interface to controls rather than ref cells.  For the time being, both
2247    styles are provided by the Controls module, but getter-setter pairs are
2248    better if thread-safety is of any concern because they can be wrapped.
2249    
2250    *****************************************
2251    
2252    One bug fix: The function blockPlacement in three of the MLRISC
2253    backpatch files used to be hard-wired to one of two possibilities at
2254    link time (according to the value of the placementFlag).  But (I
2255    think) it should rather sense the flag every time.
2256    
2257    *****************************************
2258    
2259    Other assorted changes (by other people who did not supply a HISTORY entry):
2260    
2261    1. the cross-module inliner now works much better (Monnier)
2262    2. representation of weights, frequencies, and probabilities in MLRISC
2263       changed in preparation of using those for weighted block placement
2264       (Reppy, George)
2265    
2266    ----------------------------------------------------------------------
2267    Name: Lal George
2268    Date: 2002/03/07 14:44:24 EST 2002
2269    Tag: george-20020307-weighted-block-placement
2270    
2271    Tested the weighted block placement optimization on all architectures
2272    (except the hppa) using AMPL to generate the block and edge frequencies.
2273    Changes were required in the machine properties to correctly
2274    categorize trap instructions. There is an MLRISC flag
2275    "weighted-block-placement" that can be used to enable weighted block
2276    placement, but this will be ineffective without block/edge
2277    frequencies (coming soon).
2278    
2279    
2280    ----------------------------------------------------------------------
2281    Name: Lal George
2282    Date: 2002/03/05 17:24:48 EST
2283    Tag: george-20020305-linkage-cluster
2284    
2285    In order to support the block placement optimization, a new cluster
2286    is generated as the very first cluster (called the linkage cluster).
2287    It contains a single jump to the 'real' entry point for the compilation
2288    unit. Block placement has no effect on the linkage cluster itself, but
2289    all the other clusters  have full freedom in the manner in which they
2290    reorder blocks or functions.
2291    
2292    On the x86 the typical linkage code that is generated is:
2293       ----------------------
2294            .align 2
2295       L0:
2296            addl    $L1-L0, 72(%esp)
2297            jmp     L1
2298    
2299    
2300            .align  2
2301       L1:
2302       ----------------------
2303    
2304    72(%esp) is the memory location for the stdlink register. This
2305    must contain the address of the CPS function being called. In the
2306    above example, it contains the address of  L0; before
2307    calling L1 (the real entry point for the compilation unit), it
2308    must contain the address for L1, and hence
2309    
2310            addl $L1-L0, 72(%esp)
2311    
2312    I have tested this on all architectures except the hppa.The increase
2313    in code size is of course negligible
2314    
2315    ----------------------------------------------------------------------
2316    Name: Allen Leung
2317    Date: 2002/03/03 13:20:00 EST
2318    Tag: leunga-20020303-mlrisc-tools
2319    
2320      Added #[ ... ] expressions to mlrisc tools
2321    
2322    ----------------------------------------------------------------------
2323    Name: Matthias Blume
2324    Date: 2002/02/27 12:29:00 EST
2325    Tag: blume-20020227-cdebug
2326    Description:
2327    
2328    - made types in structure C and C_Debug to be equal
2329    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2330    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2331    
2332    ----------------------------------------------------------------------
2333    Name: Matthias Blume
2334    Date: 2002/02/26 12:00:00 EST
2335    Tag: blume-20020226-ffi
2336    Description:
2337    
2338    1. Fixed a minor bug in CM's "noweb" tool:
2339       If numbering is turned off, then truly don't number (i.e., do not
2340       supply the -L option to noweb).  The previous behavior was to supply
2341       -L'' -- which caused noweb to use the "default" line numbering scheme.
2342       Thanks to Chris Richards for pointing this out (and supplying the fix).
2343    
2344    2. Once again, I reworked some aspects of the FFI:
2345    
2346       A. The incomplete/complete type business:
2347    
2348       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2349         gone!
2350       - ML types representing an incomplete type are now *equal* to
2351         ML types representing their corresponding complete types (just like
2352         in C).  This is still safe because ml-nlffigen will not generate
2353         RTTI for incomplete types, nor will it generate functions that
2354         require access to such RTTI.   But when ML code generated from both
2355         incomplete and complete versions of the C type meet, the ML types
2356         are trivially interoperable.
2357    
2358         NOTE:  These changes restore the full generality of the translation
2359         (which was previously lost when I eliminated functorization)!
2360    
2361       B. Enum types:
2362    
2363       - Structure C now has a type constructor "enum" that is similar to
2364         how the "su" constructor works.  However, "enum" is not a phantom
2365         type because each "T enum" has values (and is isomorphic to
2366         MLRep.Signed.int).
2367       - There are generic access operations for enum objects (using
2368         MLRep.Signed.int).
2369       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2370         * The structure contains the definition of type "mlrep" (the ML-side
2371         representation type of the enum).  Normally, mlrep is the same
2372         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2373         then mlrep will be defined as a datatype -- thus facilitating
2374         pattern matching on mlrep values.
2375         ("-ec" will be suppressed if there are duplicate values in an
2376          enumeration.)
2377         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2378         will be generated for each C enum constant xxx.
2379         * Conversion functions m2i and i2m convert between mlrep and
2380         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2381         * Coversion functions c and ml convert between mlrep and "tag enum".
2382         * Access functions (get/set) fetch and store mlrep values.
2383       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2384         enumerations are merged into one single enumeration represented by
2385         structure E_'.
2386    
2387    ----------------------------------------------------------------------
2388    Name: Allen Leung
2389    Date: 2002/02/25 04:45:00 EST
2390    Tag: leunga-20020225-cps-spill
2391    
2392    This is a new implementation of the CPS spill phase.
2393    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2394    In case of problems, replace it with the old file spill.sml
2395    
2396    The current compiler runs into some serious performance problems when
2397    constructing a large record.  This can happen when we try to compile a
2398    structure with many items.  Even a very simple structure like the following
2399    makes the compiler slow down.
2400    
2401        structure Foo = struct
2402           val x_1 = 0w1 : Word32.int
2403           val x_2 = 0w2 : Word32.int
2404           val x_3 = 0w3 : Word32.int
2405           ...
2406           val x_N = 0wN : Word32.int
2407        end
2408    
2409    The following table shows the compile time, from N=1000 to N=4000,
2410    with the old compiler:
2411    
2412    N
2413    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2414           MLRISC ra                               0.06u  0.00s  0.05g
2415              (spills = 0 reloads = 0)
2416           TOTAL                                   0.63u  0.07s  0.21g
2417    
2418    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2419           MLRISC ra                               5.68u  0.59s  3.93g
2420              (spills = 0 reloads = 0)
2421           TOTAL                                   14.71u  0.99s  4.81g
2422    
2423    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2424           MLRISC ra                               5.54u  0.65s  3.91g
2425              (spills = 543 reloads = 1082)
2426           TOTAL                                   65.40u  3.13s  6.00g
2427    
2428    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2429           MLRISC ra                               0.80u  0.10s  0.55g
2430              (spills = 42 reloads = 84)
2431           TOTAL                                   129.42u  5.10s  4.13g
2432    
2433    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2434           MLRISC ra                               2.69u  0.27s  1.38g
2435              (spills = 62 reloads = 124)
2436           TOTAL                                   682.48u  19.61s  13.99g
2437    
2438    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2439           MLRISC ra                               4.96u  0.27s  2.72g
2440              (spills = 85 reloads = 170)
2441           TOTAL                                   2375.26u  57.21s  48.00g
2442    
2443    As you can see the old cps spill module suffers from some serious
2444    performance problem.  But since I cannot decipher the old code fully,
2445    instead of patching the problems up, I'm reimplementing it
2446    with a different algorithm.  The new code is more modular,
2447    smaller when compiled, and substantially faster
2448    (O(n log n) time and O(n) space).  Timing of the new spill module:
2449    
2450    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2451          MLRISC ra                               0.25u  0.02s  0.15g
2452             (spills=1 reloads=3)
2453          TOTAL                                   7.74u  0.34s  1.62g
2454    
2455    Implementation details:
2456    
2457    As far as I can tell, the purpose of the CPS spill module is to make sure the
2458    number of live variables at any program point (the bandwidth)
2459    does not exceed a certain limit, which is determined by the
2460    size of the spill area.
2461    
2462    When the bandwidth is too large, we decrease the register pressure by
2463    packing live variables into spill records.  How we achieve this is
2464    completely different than what we did in the old code.
2465    
2466    First, there is something about the MLRiscGen code generator
2467    that we should be aware of:
2468    
2469    o MLRiscGen performs code motion!
2470    
2471       In particular, it will move floating point computations and
2472       address computations involving only the heap pointer to
2473       their use sites (if there is only a single use).
2474       What this means is that if we have a CPS record construction
2475       statement
2476    
2477           RECORD(k,vl,w,e)
2478    
2479       we should never count the new record address w as live if w
2480       has only one use (which is often the case).
2481    
2482       We should do something similar to floating point, but the transformation
2483       there is much more complex, so I won't deal with that.
2484    
2485    Secondly, there are now two new cps primops at our disposal:
2486    
2487     1. rawrecord of record_kind option
2488        This pure operator allocates some uninitialized storage from the heap.
2489        There are two forms:
2490    
2491         rawrecord NONE [INT n]  allocates a tagless record of length n
2492         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2493                                     and initializes the tag.
2494    
2495     2. rawupdate of cty
2496          rawupdate cty (v,i,x)
2497          Assigns to x to the ith component of record v.
2498          The storelist is not updated.
2499    
2500    We use these new primops for both spilling and increment record construction.
2501    
2502     1. Spilling.
2503    
2504        This is implemented with a linear scan algorithm (but generalized
2505        to trees).  The algorithm will create a single spill record at the
2506        beginning of the cps function and use rawupdate to spill to it,
2507        and SELECT or SELp to reload from it.  So both spills and reloads
2508        are fine-grain operations.  In contrast, in the old algorithm
2509        "spills" have to be bundled together in records.
2510    
2511        Ideally, we should sink the spill record construction to where
2512        it is needed.  We can even split the spill record into multiple ones
2513        at the places where they are needed.  But CPS is not a good
2514        representation for global code motion, so I'll keep it simple and
2515        am not attempting this.
2516    
2517     2. Incremental record construction (aka record splitting).
2518    
2519        Long records with many component values which are simulatenously live
2520        (recall that single use record addresses are not considered to
2521         be live) are constructed with rawrecord and rawupdate.
2522        We allocate space on the heap with rawrecord first, then gradually
2523        fill it in with rawupdate.  This is the technique suggested to me
2524        by Matthias.
2525    
2526        Some restrictions on when this is applicable:
2527        1. It is not a VECTOR record.  The code generator currently does not handle
2528           this case. VECTOR record uses double indirection like arrays.
2529        2. All the record component values are defined in the same "basic block"
2530           as the record constructor.  This is to prevent speculative
2531           record construction.
2532    
2533    ----------------------------------------------------------------------
2534    Name: Allen Leung
2535    Date: 2002/02/22 01:02:00 EST
2536    Tag: leunga-20020222-mlrisc-tools
2537    
2538    Minor bug fixes in the parser and rewriter
2539    
2540    ----------------------------------------------------------------------
2541    Name: Allen Leung
2542    Date: 2002/02/21 20:20:00 EST
2543    Tag: leunga-20020221-peephole
2544    
2545    Regenerated the peephole files.  Some contained typos in the specification
2546    and some didn't compile because of pretty printing bugs in the old version
2547    of 'nowhere'.
2548    
2549    ----------------------------------------------------------------------
2550    Name: Allen Leung
2551    Date: 2002/02/19 20:20:00 EST
2552    Tag: leunga-20020219-mlrisc-tools
2553    Description:
2554    
2555       Minor bug fixes to the mlrisc-tools library:
2556    
2557       1.  Fixed up parsing colon suffixed keywords
2558       2.  Added the ability to shut the error messages up
2559       3.  Reimplemented the pretty printer and fixed up/improved
2560           the pretty printing of handle and -> types.
2561       4.  Fixed up generation of literal symbols in the nowhere tool.
2562       5.  Added some SML keywords to to sml.sty
2563    
2564    ----------------------------------------------------------------------
2565    Name: Matthias Blume
2566    Date: 2002/02/19 16:20:00 EST
2567    Tag: blume-20020219-cmffi
2568    Description:
2569    
2570    A wild mix of changes, some minor, some major:
2571    
2572    * All C FFI-related libraries are now anchored under $c:
2573        $/c.cm      --> $c/c.cm
2574        $/c-int.cm  --> $c/internals/c-int.cm
2575        $/memory.cm --> $c/memory/memory.cm
2576    
2577    * "make" tool (in CM) now treats its argument pathname slightly
2578      differently:
2579        1. If the native expansion is an absolute name, then before invoking
2580           the "make" command on it, CM will apply OS.Path.mkRelative
2581           (with relativeTo = OS.FileSys.getDir()) to it.
2582        2. The argument will be passed through to subsequent phases of CM
2583           processing without "going native".  In particular, if the argument
2584           was an anchored path, then "make" will not lose track of that anchor.
2585    
2586    * Compiler backends now "know" their respective C calling conventions
2587      instead of having to be told about it by ml-nlffigen.  This relieves
2588      ml-nlffigen from one of its burdens.
2589    
2590    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
2591    
2592    * Export C_DEBUG and C_Debug from $c/c.cm.
2593    
2594    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
2595      subtyping relationship between incomplete pointers and their complete
2596      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
2597      with the convention of instantiating 'o with ('t, 'c) obj whenever
2598      the pointer target type is complete.  In the incomplete case, 'o
2599      will be instantiated with some "'c iobj" -- a type obtained by
2600      using one of the functors PointerToIncompleteType or PointerToCompleteType.
2601    
2602      Operations that work on both incomplete and complete pointer types are
2603      typed as taking an 'o ptr while operations that require the target to
2604      be known are typed as taking some ('t, 'c) obj ptr.
2605    
2606      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
2607      where void is an eqtype without any values.  This makes it possible
2608      to work on voidptr values using functions meant to operate on light
2609      incomplete pointers.
2610    
2611    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
2612      been vastly simplified.
2613    
2614    ----------------------------------------------------------------------
2615    Name: Matthias Blume
2616    Date: 2002/02/19 10:48:00 EST
2617    Tag: blume-20020219-pqfix
2618    Description:
2619    
2620    Applied Chris Okasaki's bug fix for priority queues.
2621    
2622    ----------------------------------------------------------------------
2623    Name: Matthias Blume
2624    Date: 2002/02/15 17:05:00
2625    Tag: Release_110_39
2626    Description:
2627    
2628    Last-minute retagging is becoming a tradition... :-(
2629    
2630    This is the working release 110.39.
2631    
2632    ----------------------------------------------------------------------
2633    Name: Matthias Blume
2634    Date: 2002/02/15 16:00:00 EST
2635    Tag: Release_110_39-orig
2636    Description:
2637    
2638    Working release 110.39.  New bootfiles.
2639    
2640    (Update: There was a small bug in the installer so it wouldn't work
2641    with all shells.  So I retagged. -Matthias)
2642    
2643    ----------------------------------------------------------------------
2644    Name: Matthias Blume
2645    Date: 2002/02/15 14:17:00 EST
2646    Tag: blume-20020215-showbindings
2647    Description:
2648    
2649    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
2650    the latter can be useful for exploring what bindings are available at
2651    the interactive prompt.  (The first function returns only the list
2652    of symbols that are really bound, the second prints those but also the
2653    ones that CM's autoloading mechanism knows about.)
2654    
2655    ----------------------------------------------------------------------
2656    Name: Matthias Blume
2657    Date: 2002/02/15 12:08:00 EST
2658    Tag: blume-20020215-iptrs
2659    Description:
2660    
2661    Two improvements to ml-nlffigen:
2662    
2663      1. Write files only if they do not exist or if their current contents
2664         do not coincide with what's being written.  (That is, avoid messing
2665         with the time stamps unless absolutely necessary.)
2666    
2667      2. Implement a "repository" mechanism for generated files related
2668         to "incomplete pointer types".   See the README file for details.
2669    
2670    ----------------------------------------------------------------------
2671  Name: Matthias Blume  Name: Matthias Blume
2672  Date: 2002/02/14 11:50:00 EST  Date: 2002/02/14 11:50:00 EST
2673  Tag: blume-20020214-quote  Tag: blume-20020214-quote

Legend:
Removed from v.1066  
changed lines
  Added in v.1622

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