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 984, Wed Nov 21 19:00:08 2001 UTC revision 1642, Thu Oct 7 21:00:04 2004 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2004/10/07 16:00:00 CDT
18    Tag: blume-20041007-cleanup
19    Description:
20    
21    Assorted cleanup work:
22    
23      - got rid of intstrmap in favor of using the library's
24        hash table implementation
25      - threw out most of the pathnames stuff, as it was not used anyway
26      - simplified tokentable implementation
27      - fixed some minor spelling errors
28    
29    ----------------------------------------------------------------------
30    Name: Matthias Blume (blume (at) tti - c (dot) org)
31    Date: 2004/10/06 15:15:15 CDT
32    Tag: blume-20041006-handler
33    Description:
34    
35    Cleaned up the absyn to reflect the invariant that HANDLE always
36    carries a FNexp as part of the type definition.  This eliminates some
37    superfluous sanity checks at runtime down the road.
38    
39    Some minor cleanup of the btrace code.
40    
41    ----------------------------------------------------------------------
42    Name: Matthias Blume (blume (at) tti - c (dot) org)
43    Date: 2004/10/01 10:20:30 CDT
44    Tag: blume-20041001-slave
45    Description:
46    
47    Added hack to make slave mode work in the presence of the version
48    tool.  (Still, since the master does two passes over the code for
49    CMB.make, the release number gets bumped twice when slaves are
50    attached.  I don't know if this is worth fixing...)
51    
52    ----------------------------------------------------------------------
53    Name: Matthias Blume (blume (at) tti - c (dot) org)
54    Date: 2004/09/30 10:55:00 CDT
55    Tag: blume-20040930-version
56    Description:
57    
58    * Moved the "version" magic into its own little library under
59    src/system/smlnj/internal.  This avoids expensive reconstruction
60    of a stable src/compiler/core.cm.
61    
62    * At the same time, structure CompilerVersion is now known as structure
63    SMLNJVersion.
64    
65    * Arranged for the version tool to NOT kick in when rebuilding the system
66      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
67      Also, loading the versiontool does not work when rebuilding the system
68      because CM is not properly initialized at that time.
69    
70    ----------------------------------------------------------------------
71    Name: Matthias Blume (blume (at) tti - c (dot) org)
72    Date: 2004/09/29 14:00:00 CDT
73    Tag: blume-20040929-autoversion
74    Description:
75    
76    Implemented some CM magic to have
77        file src/compiler/TopLevel/main/version.sml
78    generated automagically.
79    The version is taken from two files: config/version and config/release.
80    The first is expected to contain a two-part version number such as 110.49.
81    The second should contain a single number, but it may be missing.
82    
83    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
84    time the version tool is loaded (which is the first time you say CMB.make),
85    then the tool will increment the value stored in config/release every
86    time CMB.make is invoked.
87    
88    The binfile format is now insensitive to anything beyond the first
89    two components of a version number, so bumping the release does not render
90    binfiles incompatible.  Auto-bumping can be used to keep track of versions
91    during development without invalidating existing binfiles.
92    
93    In any case, every CMB.make updates the date information in version.sml.
94    (This is the date that is printed in the banner.)
95    
96    ----------------------------------------------------------------------
97    Name: Matthias Blume (blume (at) tti - c (dot) org)
98    Date: 2004/09/28 10:53:00 CDT
99    Tag: blume-20040928-controls
100    Description:
101    
102    Some cleanup of the controls code.
103    
104    ----------------------------------------------------------------------
105    Name: Matthias Blume (blume (at) tti - c (dot) org)
106    Date: 2004/09/27 22:08:00 CDT
107    Tag: blume-20040927-controls
108    Description:
109    
110    Added two pieces of functionality to the Controls interface:
111    
112    1.   val save'restore: 'a control -> unit -> unit
113    
114       grabs the current value of the control in stage 1 and restores it
115       in stage 2.
116    
117    2.   val set' : 'a control * 'a -> unit -> unit
118    
119       stores the given value into the control in stage 2 (i.e., delayed)
120       but does all error checking in stage 1.
121       (This is for string controls that need to do parse their argument --
122        something that might fail.  In some cases, notably in CM, one
123        already knows the intended argument but wants to delay the actual
124        assignment until a time when error recovery would be more difficult.)
125    
126    Changed the handling of controls in tool arguments to classes "sml" and
127    "lazysml":
128       - use Controls.save'restore as a more robust way of restoring the
129         old value (in particular: without having to re-parse the string)
130       - use controls to handle the "overload" keyword in the init group
131         (I believe this change actually fixes a long-standing obscure bug.)
132    
133    ----------------------------------------------------------------------
134    Name: Matthias Blume (blume (at) tti - c (dot) org)
135    Date: 2004/09/27 17:00:00 CDT
136    Tag: blume-20040927-lazysml
137    Description:
138    
139    Added a new tool class called "lazysml" to CM's tool chest.  The only
140    difference to "sml" is that compilation is done with Control.lazysml
141    set to true.  A source of class "lazysml" is automatically recognized
142    by a file name suffix of ".lml".
143    
144    In addition to the above feature, the original class "sml" now also
145    supports a tool argument "lazy" which has the same effect.  As a
146    result, the following three lines are equivalent:
147    
148        foo.sml : lazysml
149        foo.sml : sml (lazy)
150        foo.sml (lazy)
151    
152    The setting goes into effect both during parsing and during
153    compilation.  The original setting is restored right after parsing and
154    after compilation, respectively.
155    
156    In addition to all the above, there is also a general mechanism to set
157    ANY of the "controls" that are available at the command line via
158    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
159    apply as well. (In fact, "lazy" is implemented as a special case of
160    the general mechanism.)
161    
162    The .cm file syntax uses a new keyword tool argument called "with".
163    There are several ways of indicating the desired settings:
164    
165        foo.sml (with:parser.quotations=true)
166        foo.sml (with:(name:parser.quotations value:true))
167        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
168        foo.sml (with:(name1=value1 name2=value2 ...))
169        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
170    
171        etc.
172    
173    Another possible abbreviation is to leave out the =v or value:v part
174    if the name refers to a boolean control (in which case the value is
175    taken to be true).  Thus, one could get lazy sml also by saying:
176    
177        foo.sml (with:parser.lazy-keyword=true)
178        foo.sml (with:parser.lazy-keyword)
179        foo.sml (with:(name:parser.lazy-keyword value:true))
180        foo.sml (with:(name:parser.lazy-keyword))
181    
182    ----------------------------------------------------------------------
183    Name: Matthias Blume (blume (at) tti - c (dot) org)
184    Date: 2004/09/24 16:22:00 CDT
185    Tag: blume-20040924-ppc-long-branch
186    Description:
187    
188    Turned message about "emiting long form of branch" off by default.
189    Added a control flag to turn it back on when desired.
190    
191    ----------------------------------------------------------------------
192    Name: Matthias Blume (blume (at) tti - c (dot) org)
193    Date: 2004/09/24 16:05:00 CDT
194    Tag: blume-20040924-rounding
195    Description:
196    
197    Applied patch for setting rounding modes under Mac OS X.  Thanks to
198    Melissa O'Neill for providing the code!
199    
200    ----------------------------------------------------------------------
201    Name: Matthias Blume (blume (at) tti - c (dot) org)
202    Date: 2004/09/23 17:30:00 CDT
203    Tag: blume-20040923-envvars
204    Description:
205    
206    1. Changed definition of type ControlRegistry.registry_tree to
207       include control_info (i.e., the name of the controlling environment
208       variable).
209    
210    2. Added command-line flags -e and -E to print the names of environment
211       variables that can be used to control internal settings.
212       (This uses the new API mentioned in 1.)
213    
214    ----------------------------------------------------------------------
215    Name: Matthias Blume (blume (at) tti - c (dot) org)
216    Date: 2004/09/13 16:50:00 CDT
217    Tag: Release_110_49
218    Description:
219    
220    New working version (110.49).  NEW BOOTFILES!
221    
222    ----------------------------------------------------------------------
223    Name: Matthias Blume
224    Date: 2004/09/13 16:20:00 CDT
225    Tag: blume-20040913-config-mlrisc
226    Description:
227    
228    Put target "mlrisc" back into the default list.
229    (There is no harm in having it, and some users have expressed their
230    wish to have "mlrisc" included by default.)
231    
232    ----------------------------------------------------------------------
233    Name: John Reppy
234    Date: 2004/09/13
235    Tag: jhr-20040913-signals
236    Description:
237    
238    Fixed the signal masking code to properly nest mask/unmask operations
239    on a per-signal basis.
240    
241    ----------------------------------------------------------------------
242    Name: Matthias Blume (blume (at) tti - c (dot) org)
243    Date: 2004/09/08 13:20:00 CDT
244    Tag: blume-20040908-heap-magic
245    Description:
246    
247    Bumped the heap macig number to 0x09082004 to account for the changed
248    layout of the ML frame under MacOS X.
249    
250    ----------------------------------------------------------------------
251    Name: Allen Leung (leunga (at) reservoir (dot) com)
252    Date: 2004/09/03 11:26:00 EST
253    Tag: leunga-20040903-cygwin-install
254    Description:
255    
256       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
257    The Cygwin runtime is turned on by setting the environment
258    variable SMLNJ_CYGWIN_RUNTIME to 1.
259    
260    ----------------------------------------------------------------------
261    Name: Matthias Blume (blume (at) tti - c (dot) org)
262    Date: 2004/08/31 17:15:00 CDT
263    Tag: blume-20040831-core
264    Description:
265    
266    Added some exports to src/compiler/core.cm upon request by J. Joemann.
267    
268    ----------------------------------------------------------------------
269    Name: Matthias Blume (blume (at) tti - c (dot) org)
270    Date: 2004/08/30 17:55:00 CDT
271    Tag: blume-20040830-installer
272    Description:
273    
274    Upon request by Johannes Joemann:
275    
276    - improved ML code of installer to fall back to coping when renaming
277      fails (i.e., when source and target are on different file systems);
278      the code compiles but has yet to be tested in anger
279    - removed mlrisc from list of default targets (config/targets)
280    
281    ----------------------------------------------------------------------
282    Name: Matthias Blume (blume (at) tti - c (dot) org)
283    Date: 2004/08/27 17:20:00 CDT
284    Tag: blume-20040827-ptreql
285    Description:
286    
287    Added ptreql primop to structure InlineT (upon request from Larry
288    Paulson).
289    
290    ----------------------------------------------------------------------
291    Name: Allen Leung (leunga (at) reservoir (dot) com)
292    Date: 2004/08/15 21:21:00 EST
293    Tag: leunga-110_48-udgraph
294    Description:
295    
296    Another bug fix from Carl Hauser:
297    
298    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
299    > 48c48
300    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
301    > e::L)
302    > ---
303    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
304    > Without this, any deletion of an edge in an undirected graph does severe
305    > violence to the graph.
306    
307    ----------------------------------------------------------------------
308    Name: Allen Leung (leunga (at) reservoir (dot) com)
309    Date: 2004/08/10 23:35:00 EST
310    Tag: leunga-110_48-ppc
311    Description:
312    
313       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
314    
315    ----------------------------------------------------------------------
316    Name: Matthias Blume (blume (at) tti - c (dot) org)
317    Date: 2004/08/10 12:00:00 CDT
318    Tag: Release_110_48
319    Description:
320    
321    New working version (110.48).  NEW BOOTFILES!
322    
323    ----------------------------------------------------------------------
324    Name: Allen Leung (leunga (at) reservoir (dot) com)
325    Date: 2004/08/09 12:21:00 EST
326    Tag: leunga-110_47-dijsktra
327    Description:
328    
329    Bug fix from Carl Hauser:
330    
331    single_source_shortest_paths in dijkstra.sml was observed to get wrong
332    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
333    
334    The problem is that following the expression A.update(dist,s,Num.zero)
335    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
336    
337    ----------------------------------------------------------------------
338    Name: Matthias Blume (blume (at) tti - c (dot) org)
339    Date: 2004/08/06 18:10:00 CDT
340    Tag: blume-20040806-cmdline
341    Description:
342    
343    Fiddled with handling of command-line options:
344    
345      * sml now quits after processing the command line
346        if -H, -S, -h<n>, or -s<n> appears as the last
347        command-line argument
348      * a new option -q terminates the session when encountered on
349        the command line; subsequent arguments will be ignored
350      * bug fixes: short (erroneous) arguments are no longer ignored
351        completely
352    
353    ----------------------------------------------------------------------
354    Name: Allen Leung (leunga (at) reservoir (dot) com)
355    Date: 2004/08/04 18:17:00 EST
356    Tag: leunga-110_47-ppc-ibm-asm
357    Description:
358    
359      - Added minimal IBM assembly syntax support for PowerPC.
360    
361      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
362        been altered in the runtime.  We need an automatic way to keep the file
363        in sync.
364    
365    ----------------------------------------------------------------------
366    Name: Matthias Blume (blume (at) tti - c (dot) org)
367    Date: 2004/08/04 14:00:00 CDT
368    Tag: Release_110_47
369    Description:
370    
371    New working version (110.47).  NEW BOOTFILES!
372    
373    ----------------------------------------------------------------------
374    Name: Matthias Blume (blume (at) tti - c (dot) org)
375    Date: 2004/08/03 14:25:00 CDT
376    Tag: blume-20040803-callingconv
377    Description:
378    
379    Added low-level support for choosing C calling conventions by
380    twiddling the type of rawccall.  (See
381    src/compiler/Semant/types/cproto.sml for details.)
382    
383    ----------------------------------------------------------------------
384    Name: Matthias Blume (blume (at) tti - c (dot) org)
385    Date: 2004/08/02 15:55:00 CDT
386    Tag: blume-20040802-backout
387    Description:
388    
389    Backed out of change to win32-filesys.c.  The earlier patch to
390    get_file_time caused CM to produce files with the wrong time stamp.
391    
392    ----------------------------------------------------------------------
393    Name: Matthias Blume (blume (at) tti - c (dot) org)
394    Date: 2004/08/02 14:45:00 CDT
395    Tag: blume-20040802-nlffi-win32
396    Description:
397    
398    Added NLFFI support for Win32, adapted from a patch provided by David
399    Hansel.  This is currently completely untested.  Also, the issue
400    concerning stdcall vs. ccall is still unresolved.
401    
402    ----------------------------------------------------------------------
403    Name: Matthias Blume (blume (at) tti - c (dot) org)
404    Date: 2004/07/30 17:55:00 CDT
405    Tag: blume-20040730-various
406    Description:
407    
408    Gearing up towards 110.47...
409    
410    - various minor bugfixes to ml-nlffigen
411    - a beginning of a manual for nlffi
412    
413    - eliminated 'export name=value' in config/install.sh as this does
414      not work with certain versions of /bin/sh
415      (Thanks to David King at Motorola for catching this.)
416    
417    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
418      - added a test for tm==NULL to gmtime.c and localtime.c
419      - applied patch for incorrect GetFileTime under win32
420      - toSeconds -> toMilliseconds in Win32/win32-process.sml
421    
422    ----------------------------------------------------------------------
423    Name: Matthias Blume (blume (at) tti - c (dot) org)
424    Date: 2004/07/21 18:20:00 CDT
425    Tag: blume-20040721-nlffigen
426    Description:
427    
428    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
429      for a typedef to an incomplete type, but leave out the "typ" member.
430      (This is just for consistency.)
431    - Started to produce what is supposed to become better (i.e., comprehensive)
432      documentation of what ml-nlffigen does and produces.
433    
434    ----------------------------------------------------------------------
435    Name: Matthias Blume (blume (at) tti - c (dot) org)
436    Date: 2004/07/14 16:25:00 CDT
437    Tag: blume-20040714-union
438    Description:
439    
440    Added C_UNION to c-calls/c-types.sml and updated the machinery
441    (ml-nlffigen, cproto.sml) that conveys C function interface
442    information to the code generator.
443    
444    However, the actual architecture-specific implementation of function
445    arguments and results that are C unions is still not implemented.
446    
447    ----------------------------------------------------------------------
448    Name: Allen Leung (leunga (at) reservoir (dot) com)
449    Date: 2004/07/14 14:38:00 EST
450    Tag: leunga-110_46_1-ppc-lwzu
451    Description:
452    
453        Added these instructions to the PowerPC architecture:
454        LBZU(X), LHZU(X), LWZU(X),
455        STWU(X), STFDU, STFSU
456    
457        etc...
458    
459        Note: I haven't added their instruction encoding into the description.
460    
461    ----------------------------------------------------------------------
462    Name: Allen Leung (leunga (at) reservoir (dot) com)
463    Date: 2004/07/13 15:04:00 EST
464    Tag: leunga-110_46_1-ppc-lwarx
465    Description:
466    
467       Added the two instructions LWARX and STWCX to the PowerPC
468    instruction set.
469    
470       A (untested) rewrite of loop-structure.sml.  The old version
471    is completely broken.
472    
473    ----------------------------------------------------------------------
474    Name: Matthias Blume (blume (at) tti - c (dot) org)
475    Date: 2004/07/13 13:50:00 CDT
476    Tag: blume-20040713-nlffi
477    Description:
478    
479    - use paramAlloc to report c-calls with too many arguments
480      (for PPC version where parameter area is pre-allocated)
481    - added ccall_maxargspace to machspec (to implement the above)
482    - made "make" commend in CM's "make" tool configurable
483    - added option (default: on) for passing the name of the SML/NJ's "bin"
484      directory to "make"; the call looks like this:
485    
486         make <options> SMLNJ_BINDIR=<dir> <target>
487    
488      This can be used by the Makefile to, e.g., pick the "right" version
489      of ml-nlffigen.
490    - minor code tweaks
491    
492    ----------------------------------------------------------------------
493    Name: Matthias Blume (blume (at) tti - c (dot) org)
494    Date: 2004/07/12 22:50:00 CDT
495    Tag: blume-110_46_1-macosx-nlffi
496    Description:
497    
498    NLFFI under Mac OS X now working (sort of).  This is largely untested,
499    though.
500    
501    Note:  1. You have to make a new, clean build of the runtime system.
502           2. There are new BOOTFILES, you have to use them!
503              (Doing the bootstrap process yourself would be *very* painful!
504               If you absolutely have to do it, build the system under
505               a different architecture and then cross-compile.)
506    
507    Version bumped to 110.46.1 to account for runtime data format changes.
508    
509    ----------------------------------------------------------------------
510    Name: Matthias Blume (blume (at) tti - c (dot) org)
511    Date: 2004/06/18 14:30:00 CDT
512    Tag: blume-20040618-unix
513    Description:
514    
515    Changed the implementation of structure Unix so that the same stream
516    is returned every time one of the {text,bin}{In,Out}streamOf functions
517    is invoked on the same proc.  This is not what the spec currently
518    says -- although IMO it arguably should.  (See discussion below.)
519    
520    ----------------------------------------------------------------------
521    Name: Matthias Blume (blume (at) tti - c (dot) org)
522    Date: 2004/06/17 18:15:00 CDT
523    Tag: Release_110_46
524    Description:
525    
526    New working version (110.46).  NEW BOOTFILES!
527    
528    ----------------------------------------------------------------------
529    Name: Matthias Blume (blume (at) tti - c (dot) org)
530    Date: 2004/06/17 17:20:00 CDT
531    Tag: blume-20040617-timer-unix
532    Description:
533    
534    Changed the interface of structures Timer and Unix to match the most
535    recent Basis spec.
536    
537    In the case of Unix there still seems to be an open/weird issue:
538    
539       The {text,bin}{In,Out}streamOf functions are supposed to create
540       fresh streams whenever they are called -- as opposed to have them
541       return the same stream every time.  This design is supposed to
542       prevent space leaks caused by proc values hanging on to streams.
543    
544       The reap function, on the other hand, is supposed to close the
545       streams.  This cannot be done without having a handle on the
546       stream in proc after all...
547    
548    I took the liberty to implement the following stopgap solution:
549    
550      The proc value hangs on to the most recently created stream(s).
551      Reap closes those.  If either or both of the two streams hadn't
552      been created at all yet, then reap will close the corresponding
553      file descriptors directly.
554    
555    PS: I don't understand the original space leak argument anymore.  If
556    a proc hangs on to the imperative stream, then I/O operations on those
557    will advance the state of the cached stream and avoid the space leak.
558    
559    ----------------------------------------------------------------------
560    Name: Matthias Blume (blume (at) tti - c (dot) org)
561    Date: 2004/05/28 16:45:00 CDT
562    Tag: blume-20040528-basis
563    Description:
564    
565    Added signature PACK_REAL and exported functor PrimIO.
566    
567    ----------------------------------------------------------------------
568    Name: Matthias Blume (blume (at) tti - c (dot) org)
569    Date: 2004/05/25 16:00:00 CDT
570    Tag: blume-20040525-group-owner
571    Description:
572    
573    CM now ignores (but still accepts) the "owner" information in group
574    descriptions.  The owner of a group is its next enclosing
575    library. Each group must have a unique owner.  (There is a virtual
576    "toplevel" library that owns groups which are not nested within a real
577    library.)  Previously, each group had to explicitly declare its owner,
578    and CM would check that such a declaration is correct.  The new scheme
579    is to have CM check that for each group there is precisely one owning
580    library.
581    
582    The advantage of the new scheme is that the programmer no longer needs
583    to maintain the somewhat annoying owner information.  The downside is
584    that CM cannot enforce the ownership rule across multiple runs of
585    CM.make.  Fortunately, enclosing the same group in two different
586    libraries A and B which are not part of the same program does not
587    cause real problems.
588    
589    ----------------------------------------------------------------------
590    Name: Matthias Blume (blume (at) tti - c (dot) org)
591    Date: 2004/05/20 16:00:00 CDT
592    Tag: blume-20040520-win32
593    Description:
594    
595    Made the win32 version work again.  (Strangely, a misplaced comma had
596    slipped into win32-process.c which prevented the runtime from being
597    compiled correctly.)
598    
599    Also, included a minor addition to ml-build.bat analogous to what was
600    done in blume-20040519-ml-build.
601    
602    ----------------------------------------------------------------------
603    Name: Matthias Blume (blume (at) tti - c (dot) org)
604    Date: 2004/05/19 22:10:00 CDT
605    Tag: blume-20040519-ml-build
606    Description:
607    
608    Arranged for ml-build to clean up after itself a little bit better.
609    The script generates a temporary SML source file and compiles it using
610    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
611    gets rid of those at the end, so they don't accumulate under .cm.
612    
613    This required a minor change to install.sh because the name of the
614    metadata directory (default: .cm) is actually configurable at
615    installation time.
616    
617    ----------------------------------------------------------------------
618    Name: Matthias Blume (blume (at) tti - c (dot) org)
619    Date: 2004/05/18 15:50:00 CDT
620    Tag: blume-20040518-mkreader
621    Description:
622    
623    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
624    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
625    
626    ----------------------------------------------------------------------
627    Name: Matthias Blume (blume (at) tti - c (dot) org)
628    Date: 2004/05/11 14:35:00 CDT
629    Tag: blume-20040511-win32sock
630    Description:
631    
632    Added previously missing support for many socket-related functions
633    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
634    for the voluminous patch!
635    
636    (I have not tested this patch under win32 yet.)
637    
638    Here is David's e-mail:
639    
640    Hi,
641    
642    Attached to this email you find a diff against sml/nj 110.45
643    that will enable socket support under Windows.
644    
645    To apply the patch (using unix or cygwin)
646    1) gunzip runtime.diff.gz
647    2) "cd" into "src/runtime" in the source tree of a fresh
648       110.45 installation.
649    3) patch -p 1 < [your/path/to]runtime.diff
650    
651    The code compiles fine but has NOT yet been extensively tested.
652    I only ran a few tests for basic socket client functionality
653    (which worked fine).  Especially the functions that use ioctl
654    are not tested at all and might not work (see below).
655    
656    I implemented this since we want to move to a newer version of sml/nj
657    but need socket support in order to use it.  This is the first time I
658    even had a look at the sml/nj source,  so please review my changes
659    before making this part of the distribution!  Here are a few issues
660    that I think might be better for someone to solve who is more
661    familiar with the sml/nj source (and socket programming):
662    
663    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
664      exception since I could not figure out what the windows equivalent
665      of these functions is
666    
667    - In sockets-osdep.h there are a some #include statements that are
668      only used in a few files that include sockets-osdep.h
669    
670    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
671      does not process its return value since I don't know how to
672      report an error upwards.
673    
674    - It would probably be good to have a call to WSACleanup() when
675      the library is unloaded (if there is such a possibility).
676      Otherwise I think Windows will take care of this automatically
677      when the process finishes.
678    
679    - I used ioctlsocket() as a replacement for ioctl() but I have
680      no idea if that is actually the proper replacement on Windows.
681    
682    - All these issues are marked in the code by "FIXME" comments.
683    
684    We use sml/nj extensively in our products and are quite happy
685    with it.  I hope this contribution will help you.
686    
687    Keep up the good work!
688    
689    David
690    
691    ----------------------------------------------------------------------
692    Name: Matthias Blume (blume (at) tti - c (dot) org)
693    Date: 2004/05/11 14:20:00 CDT
694    Tag: blume-20040511-installml
695    Description:
696    
697    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
698    the patch.)
699    
700    ----------------------------------------------------------------------
701    Name: Matthias Blume (blume (at) tti - c (dot) org)
702    Date: 2004/05/11 14:05:00 CDT
703    Tag: blume-20040511-nlffi-netbsd
704    Description:
705    
706    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
707    the patch.)
708    
709    ----------------------------------------------------------------------
710    Name: Matthias Blume (blume (at) tti - c (dot) org)
711    Date: 2004/05/11 12:05:00 CDT
712    Tag: blume-20040511-exports
713    Description:
714    
715    As per request by Adam Chlipala <adam@hcoop.net>, extended various
716    export lists in compiler-related .cm-files.
717    
718    ----------------------------------------------------------------------
719    Name: Matthias Blume (blume (at) tti - c (dot) org)
720    Date: 2004/05/11 11:35:00 CDT
721    Tag: blume-20040511-allsource
722    Description:
723    
724    The installer now honors the "src-smlnj" target again, although its meaning
725    has changed from "all sources required for the compiler" to "all sources
726    the installer knows about".  In other words, if you enable "src-smlnj"
727    in the "targets" file, then the installer will pull in sources for
728    everything.  (Notice that this refers to source code only.  Compiled
729    code is still only installed for modules that were requested explicitly
730    or which are required for other modules that were requested explicitly.)
731    
732    ----------------------------------------------------------------------
733    Name: Matthias Blume (blume (at) tti - c (dot) org)
734    Date: 2004/04/23 17:40:00 CDT
735    Tag: blume-20040423-ieee-scan
736    Description:
737    
738    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
739    in the exponent calculation we get INF or ZERO (depending on the mantissa
740    and the sign of the exponent).
741    
742    ----------------------------------------------------------------------
743    Name: Matthias Blume (blume (at) tti - c (dot) org)
744    Date: 2004/04/23 10:40:00 CDT
745    Tag: blume-20040423-ml-build
746    Description:
747    
748    The ml-build script now terminates with a non-0 status when something
749    goes wrong.
750    
751    ----------------------------------------------------------------------
752    Name: Matthias Blume (blume (at) tti - c (dot) org)
753    Date: 2004/04/22 16:35:00 CDT
754    Tag: blume-20040422-Option
755    Description:
756    
757    Made exception Option to be the same as exception Option.Option
758    (as it should be).
759    
760    ----------------------------------------------------------------------
761    Name: Allen Leung (leunga (at) reservoir (dot) com)
762    Date: 2004/03/19 14:40:00 EST
763    Tag: leunga-20040319-cygwin-nlffi
764    Description:
765    
766    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
767    of SML/NJ.  The problem is that
768    
769       lib = dlopen(NULL, ...)
770       f   = dlsym(lib, "malloc");
771    
772    does not work on Windows unless we explicitly export symbols
773    such as 'malloc' during linking.  We fixed this by explicitly
774    exporting the required symbols with the magic gcc incantation:
775    
776        -Wl,--export-all cygwin.def
777    
778    where cygwin.def is a file containing all the symbols that
779    we wish to export.
780    
781    I suspect this is a Windows problem and we'll have to
782    do the same (somehow with windows compilers) when
783    we build the native win32 version with the system
784    calls LoadLibrary/GetProcAddress.
785    
786    ----------------------------------------------------------------------
787    Name: Matthias Blume (blume (at) tti - c (dot) org)
788    Date: 2004/03/04 16:35:00 CST
789    Tag: blume-20040304-intinf-fmt
790    Description:
791    
792    Fixed problem with IntInf.fmt (sign would show up on the right instead
793    of on the left for BIN, OCT, and HEX).
794    
795    ----------------------------------------------------------------------
796    Name: Matthias Blume (blume (at) tti - c (dot) org)
797    Date: 2004/03/04 11:25:00 CST
798    Tag: blume-20040304-symlinks
799    Description:
800    
801    Fixed problem with installer script (unix only) where bin/ml-yacc and
802    friends pointed (via symlinks) to absolute locations instead of just
803    .run-sml.  This was reported by Vesa A Norrman.
804    
805    ----------------------------------------------------------------------
806    Name: Matthias Blume (blume (at) tti - c (dot) org)
807    Date: 2004/02/13 14:50:00 CST
808    Tag: Release_110_45
809    Description:
810    
811    New working version (110.45). New bootfiles.
812    
813    ----------------------------------------------------------------------
814    Name: Matthias Blume (blume (at) tti - c (dot) org)
815    Date: 2004/01/26 15:15:15 CST
816    Tag: blume-20040126-toplevel
817    Description:
818    
819    Improved handling of exceptions at the interactive toplevel.
820    
821    ----------------------------------------------------------------------
822    Name: Matthias Blume (blume (at) tti - c (dot) org)
823    Date: 2004/01/26 11:25:00
824    Tag: blume-20040126-app
825    Description:
826    
827    Type of top-level "app" corrected.
828    Added code for setting vp_limitPtrMask to Win32-specific runtime.
829    
830    ----------------------------------------------------------------------
831    Name: Matthias Blume (blume (at) tti - c (dot) org)
832    Date: 2003/11/18 17:10 CST
833    Tag: blume-20031118-basis-fiddle
834    Description:
835    
836    - changed Timer interface to what might become the spec
837    - POSIX_FLAGS -> BIT_FLAGS according to spec
838    - some other minor discrepancies wrt. spec eliminated
839    
840    ----------------------------------------------------------------------
841    Name: Matthias Blume (blume (at) tti - c (dot) org)
842    Date: 2003/11/06 12:00:00 CST
843    Tag: Release_110_44
844    Description:
845    
846    New working version (110.44). New bootfiles.
847    
848    ----------------------------------------------------------------------
849    Name: Matthias Blume (blume (at) tti - c (dot) org)
850    Date: 2003/11/04 11:50:00 CST
851    Tag: blume-20031104-move-libraries
852    Description:
853    
854    Eliminated the "dont_move_libraries" directive in config/targets.
855    (The mechanism was broken and could not be fixed easily.  Moreover,
856    there does not seem to be any reason not to move all libraries into
857    lib during installation.  I originally implemented this directive as a
858    backward-compatibility feature when I first introduced the new CM.
859    Now that things have been stable for a long time and going back to the
860    old CM is not an option, there is no reason to keep it around.)
861    
862    ----------------------------------------------------------------------
863    Name: Matthias Blume (blume (at) tti - c (dot) org)
864    Date: 2003/11/03 16:00:00 CST
865    Tag: blume-20031103-installdir
866    Description:
867    
868    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
869    Richards for pointing out the problem and providing the solution.)
870    
871    ----------------------------------------------------------------------
872    Name: Matthias Blume (blume (at) tti - c (dot) org)
873    Date: 2003/10/01 17:05:00 CDT
874    Tag: blume-20031001-lal-mlrisc
875    Description:
876    
877    MLRISC bug fix from Lal.
878    
879    ----------------------------------------------------------------------
880    Name: Matthias Blume (blume (at) tti - c (dot) org)
881    Date: 2003/09/30 16:10:00 CDT
882    Tag: blume-20030930-primio-bat
883    Description:
884    
885    1. Added openVector, nullRd, and nullWr to PRIM_IO.
886    2. Improved .bat files (for Win32 port) to make things work under Win95.
887       (thanks to Aaron S. Hawley for this one)
888    
889    ----------------------------------------------------------------------
890    Name: Matthias Blume (blume (at) tti - c (dot) org)
891    Date: 2003/09/26 16:05:00 CDT
892    Tag: blume-20030926-wrappriv
893    Description:
894    
895    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
896    
897    ----------------------------------------------------------------------
898    Name: Matthias Blume (blume (at) tti - c (dot) org)
899    Date: 2003/09/26 15:00:00 CDT
900    Tag: blume-20030926-110_43_3
901    Description:
902    
903    - additional cleanup
904    - version number bump, NEW BOOTFILES
905    
906    ----------------------------------------------------------------------
907    Name: Matthias Blume (blume (at) tti - c (dot) org)
908    Date: 2003/09/26 12:00:00 CDT
909    Tag: blume-20030926-ppautoload
910    Description:
911    
912    I modified the read-eval-print loop so that the autoloader gets
913    invoked whenever the prettyprinter tries to look up a symbol that
914    is not currently defined in the toplevel environment but which
915    appears in CM's autoload registry.  As a result, we see far fewer of
916    those ?.Foo.Bar.xxx names in the prettyprinter's output.
917    
918    In addition to this I tried to clean up some pieces of the Basis
919    implementation (e.g., Socket, Word8Array) in order to prevent other
920    instances of these ?.Foo.Bar.xxx names from being printed.
921    
922    The mechanism that picks names for types still needs some work, though.
923    (Right now it seems that if there is a type A.t which is defined to
924    be B.u, but B is unavailable at toplevel, then A.t gets printed as
925    "?.B.u" although the perhaps more sensible solution would be to use
926    "A.t" in this case.  In other words, the prettyprinter should follow
927    a chain of DEFtycs not farther than there are corresponding toplevel
928    names in the current environment.)
929    
930    ----------------------------------------------------------------------
931    Name: Matthias Blume (blume (at) tti - c (dot) org)
932    Date: 2003/09/24 16:31:00 CDT
933    Tag: blume-20030924-installer
934    Description:
935    
936    Another installer tweak:  All the ML code for the installer is now
937    compiled during CMB.make and put into a little library called
938    $smlnj/installer.cm.  The installation then simply invokes
939    
940       sml -m $smlnj/installer.cm
941    
942    and everything happens automagically.
943    
944    Win32:  ML code senses value of environment variable SMLNJ_HOME.
945    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
946            and BINDIR.
947    
948    The new scheme guarantees that the ML code responsible for the installation
949    is in sync with the APIs of the main system.  Also, the installer is
950    somewhat faster because the installer script is precompiled.
951    
952    ----------------------------------------------------------------------
953    Name: Matthias Blume (blume (at) tti - c (dot) org)
954    Date: 2003/09/24 15:35:00 CDT
955    Tag: blume-20030924-synsock
956    Description:
957    
958    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
959    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
960    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
961    implementations of non-blocking operations from CML's Socket
962    structure.
963    
964    ----------------------------------------------------------------------
965    Name: Matthias Blume (blume (at) tti - c (dot) org)
966    Date: 2003/09/24 15:10:05 CDT
967    Tag: blume-20030924-sockets
968    Description:
969    
970    1. Fixed SOCKET API and implementation to match Basis spec.
971       This required changing the internal representation of sockets to one
972       that remembers (for each socket file descriptor) whether it is currently
973       blocking or non-blocking.  This state is maintained lazily (i.e., a system
974       call is made only if the state actually needs to change).
975    
976    2. OS-specific details of sockets were moved into separate files, thus
977       making it possible to unify the bulk of the socket implementations
978       between Unix and Win32.
979    
980    3. CML's socket API changed accordingly.
981       (Note that we need to remove non-blocking functions from this API
982       since they are redundant in the case of CML!)
983    
984    4. CML's socket implementation now makes use of non-blocking functions
985       provided by Basis, thus removing all OS-dependent code from this part
986       of CML.
987    
988    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
989    
990    ----------------------------------------------------------------------
991    Name: Matthias Blume (blume (at) tti - c (dot) org)
992    Date: 2003/09/22 12:10:00 CDT
993    Tag: blume-20030922-110_43_2
994    Description:
995    
996    Made a new interim version and bootfiles for developer's bootstrapping
997    convenience.
998    
999       110.43.2 -- NEW BOOTFILES
1000    
1001    ----------------------------------------------------------------------
1002    Name: Matthias Blume (blume (at) tti - c (dot) org)
1003    Date: 2003/09/19 15:55:00 CDT
1004    Tag: blume-20030919-cmdir
1005    Description:
1006    
1007    1. new-install.sh -> install.sh
1008    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1009    3. tweaked installer so that another name instead of .cm can be chosen
1010       at install time (by setting the CM_DIR_ARC environment variable
1011       during installation); once installation is complete, the name is
1012       fixed
1013    
1014    ----------------------------------------------------------------------
1015    Name: Matthias Blume (blume (at) tti - c (dot) org)
1016    Date: 2003/09/18 16:00:00 CDT
1017    Tag: blume-20030918-110_43_1
1018    Description:
1019    
1020    Made a new interim version and bootfiles for developer's bootstrapping
1021    convenience.
1022    
1023       110.43.1 -- NEW BOOTFILES
1024    
1025    ----------------------------------------------------------------------
1026    Name: Matthias Blume (blume (at) tti - c (dot) org)
1027    Date: 2003/09/18 15:20:00 CDT
1028    Tag: blume-20030918-misc
1029    Description:
1030    
1031    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1032       this seems to be controversial at the moment) and used those in
1033       Posix.ProcEnv.times.
1034    
1035    2. Added Time.{from,to}Nanoseconds to Time.
1036    
1037    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1038       For example, fromLargeInt never needs to look at more than 3 "big
1039       digits" to get its 53 bits of precision.
1040    
1041    ----------------------------------------------------------------------
1042    Name: Matthias Blume (blume (at) tti - c (dot) org)
1043    Date: 2003/09/17 16:30:00 CDT
1044    Tag: blume-20030917-real32-slices
1045    Description:
1046    
1047    Added an entry to the primitive environment
1048    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1049    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1050    
1051    Removed some of the "magic" constants in real64.sml and replaced them
1052    with code that generates these values from their corresponding
1053    integer counterparts.
1054    
1055    Made all(?) the slice-related changes to the Basis and made everything
1056    compile again...
1057    
1058    ----------------------------------------------------------------------
1059    Name: Matthias Blume (blume (at) tti - c (dot) org)
1060    Date: 2003/09/15 17:45:00 CDT
1061    Tag: blume-20030915-rbase
1062    Description:
1063    
1064    Fixed bug in Real.fromLargeInt.
1065    
1066    ----------------------------------------------------------------------
1067    Name: Matthias Blume (blume (at) tti - c (dot) org)
1068    Date: 2003/09/13 18:11:00 CDT
1069    Tag: blume-20030913-libinstall
1070    Description:
1071    
1072    Minor bugfix in config/libinstall (set anchor with path to
1073    standalone tool after installing it, otherwise libraries that
1074    need ml-lex or ml-yacc won't compile the first time the installer
1075    runs).
1076    
1077    ----------------------------------------------------------------------
1078    Name: Matthias Blume (blume (at) tti - c (dot) org)
1079    Date: 2003/09/12 11:45:00 CDT
1080    Tag: blume-20030912-various
1081    Description:
1082    
1083     - fixed bug in Real.toLargeInt
1084     - fixed bug in Posix.ProcEnv.times
1085     - changed inputLine functions to return an option
1086     - minor installer improvements / bugfixes
1087     - changed default @SMLalloc parameter for x86/celeron to 64k
1088    
1089    ----------------------------------------------------------------------
1090    Name: Matthias Blume (blume (at) tti - c (dot) org)
1091    Date: 2003/09/09 22:00:00 CDT
1092    Tag: Release_110_43
1093    Description:
1094    
1095        New working release 110.43.  New bootfiles.
1096    
1097    ----------------------------------------------------------------------
1098    Name: Matthias Blume (blume (at) tti - c (dot) org)
1099    Date: 2003/09/09 19:20:00 CDT
1100    Tag: blume-20030909-installer
1101    Description:
1102    
1103    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1104    Modified config/install.bat to take advantage of it.  Also modified
1105    config/install.sh (and called it config/new-install.sh) to take advantage
1106    of it on Unix systems.  (The SML code is (supposed to be) platform-
1107    independent.)
1108    
1109    The installer can now install everything under Win32
1110    as well as under *nix as long as it compiles.
1111    
1112    Other changes:
1113    
1114      - made CML compile again under Win32
1115      - made eXene compile under Win32 (by providing a fake structure UnixSock
1116        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1117      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1118        same as type int;  under Win32 it isn't
1119      - fixed some slice-related problems in the win32-specific parts of CML
1120      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1121        (under Win32, the volume name is case-insensitive, and the
1122        OS.Path code compares volume names for equality)
1123    
1124    ----------------------------------------------------------------------
1125    Name: Matthias Blume (blume (at) tti - c (dot) org)
1126    Date: 2003/09/08 11:55:00 CDT
1127    Tag: blume-20030908-fullpath
1128    Description:
1129    
1130    Made Win32 version of OS.FileSys.fullPath return current directory
1131    when given an empty string.  This is what the spec says, and incidentally,
1132    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1133    cases when presented with the name of a non-existing .cm file.)
1134    
1135    ----------------------------------------------------------------------
1136    Name: Matthias Blume (blume (at) tti - c (dot) org)
1137    Date: 2003/09/04 16:30:00 CDT
1138    Tag: blume-20030905-slices-etc
1139    Description:
1140    
1141      1. Changed interface to vectors and arrays in Basis to match
1142         (draft) Basis spec.
1143      2. Added signatures and implementations of slices according to
1144         Basis spec.
1145      3. Edited source code throughout the system to make it compile again
1146         under 1. and 2.  (In some cases code had to be added to have it
1147         match the new signatures.)
1148      4. MLRISC should be backward-compatible: the copies of the originals
1149         of files that needed to change under 3. were retained, the .cm files
1150         check the compiler version number and use old versions when
1151         appropriate.
1152      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1153         dirstream -> string option (in accordance with Basis spec).
1154      6. When generating code that counts lines, ml-lex used function
1155         CharVector.foldli, taking advantage of its old interface.
1156         This has been replaced with the corresponding code from
1157         CharVectorSlice.  (html-lex must be re-lexed!)
1158      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1159         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1160      8. Removed temporary additions (fromInternal, toInternal) from the
1161         (now obsolete) IntInf in smlnj-lib/Util.
1162      9. Cleaned up structure Byte.
1163     10. Added localOffset, scan, and fromString to Date (according to spec).
1164         Cleaned/corrected implementation of Date.
1165         (Still need to check for correctness; implement better canonicalizeDate.)
1166     11. Added "scan" to signature IEEE_REAL.
1167     12. Some improvements to IntInf [in particular: efficiency-hack for
1168         mod and rem when second operand is 2 (for parity checks).]
1169     13. Changed representation of type Time.time, using a single IntInf.int
1170         value counting microseconds.  This considerably simplified the
1171         implementation of structure Time.  We now support negative time
1172         values; scan and fromString handle signs.
1173     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1174         ArraySlice).
1175    
1176    ----------------------------------------------------------------------
1177    Name: Matthias Blume (blume (at) tti - c (dot) org)
1178    Date: 2003/08/28 17:00:00 CDT
1179    Tag: blume-20030828-intinf
1180    Description:
1181    
1182    This is a major update which comes with a version number bump
1183    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1184    and an implementation of IntInf in the Basis.
1185    
1186    There are a fairly large number of related changes and updates throughout
1187    the system:
1188    
1189    Basis:
1190            - Implemented IntInf.
1191            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1192            - Added some missing Real64 operations, most notably Real.toLargeInt.
1193            - Added FixedInt as a synonym for Int32.
1194    
1195    compiler:
1196            * Added support for a built-in intinf type.
1197              - literals
1198              - pattern matching
1199              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1200              - overloading on literals and operations
1201    
1202            This required adding a primitive type intinf, some additional
1203            primops, and implementations for several non-trivial intinf
1204            operations in Core.  (The intinf type is completely abstract
1205            to the compiler; all operations get delegated back to the Core.)
1206    
1207            * Intinf equality is handled by polyequal.  However, the compiler
1208            does not print its usual warning in this case (since polyequal
1209            is the right thing to do there).
1210    
1211            * Improved the organization of structure InlineT.
1212    
1213            * A word about conversion primops:
1214            If conversions involving intinf do not cancel out during
1215            CPS contract, then the compiler must insert calls to Core functions.
1216            Since all core access must be resolved already during the FLINT
1217            translate phase, it would be too late a the time of CPS contract
1218            to add new Core calls.  For this reason, conversion primops
1219            for intinf carry two arguments:  1. the numeric argument that
1220            they are supposed to convert, and 2. the Core function that
1221            can help with this conversion if necessary.  If CPS contract
1222            eliminates a primop, then the associated Core function becomes
1223            dead and goes away.  Intinf conversion primops that do not get
1224            eliminated by CPS contract get rewritten into calls of their
1225            core functions by a separate, new phase.
1226    
1227    interactive system:
1228            - Control.Print.intinfDepth controls max length of intinf constants
1229            being printed. (Analogous to Control.Print.stringDepth.)
1230            - Cleanup in printutil and pputil: got rid of unused stuff and
1231            duplicates; replaced some of the code with code that makes better
1232            use of library functionality.
1233    
1234    CM:
1235            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1236            are no longer silent.
1237    
1238    CKIT:
1239            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1240            in favor of LargeInt -- which is now the same as IntInf.
1241            CKIT-knowledgable people should check whether this is what's
1242            intended and otherwise change things back to using Int32 or
1243            FixedInt.
1244    
1245    Throughout the code:
1246            Started using IntInf.int literals and built-in operations
1247            (e.g., comparison with 0) where this seems appropriate.
1248    
1249    
1250    ----------------------------------------------------------------------
1251    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1252    Date: 2003/08/13 11:36:00 CDT
1253    Tag: dbm-20030813-mcz-merge1
1254    Description:
1255    
1256    Merging changes from the mcz-branch development branch into trunk.
1257    These changes involve replacement of the emulated old prettyprinter
1258    interface with direct use of the SML/NJ Lib PP library, and fixing
1259    of a couple of bugs (895, 1186) relating to error messages.  A new
1260    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1261    has been added.
1262    
1263    ----------------------------------------------------------------------
1264    Name: Matthias Blume (blume (at) tti - c (dot) org)
1265    Date: 2003/08/11 15:45:00 CDT
1266    Tag: blume-20030811-windows
1267    Description:
1268    
1269    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1270    
1271        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1272    
1273    This patch restores SML/NJ's ability to run under win32.  There are a
1274    number of changes, including fixes for several bugs that had gone
1275    unnoticed until now:
1276    
1277      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1278        trying to run the win32 version from within cygwin.)
1279    
1280      - There are a number of simple .bat scripts that substitute for their
1281        corresponding Unix shell-scripts.  (See below.)
1282    
1283      - The internals of ml-build have been modified slightly.  The main
1284        difference is that instead of calling ".link-sml" (or link-sml.bat)
1285        using OS.Process.system, the ML process delegates this task back
1286        to the script.  Otherwise problems arise in mixed environments such
1287        as Cygwin where scripts look and work like Unix scripts, but
1288        where OS.Process.system cannot run them.
1289    
1290      - In CM, the srcpath pickler used native pathname syntax -- which
1291        is incorrect in the case of cross-compilation.  The new pickle format
1292        is independent of platform-specific naming conventions.
1293    
1294      - Path configuration files (such as lib/pathconfig) can now choose
1295        between native and standard syntax.  Placing a line of the form
1296    
1297            standard!
1298    
1299        into the file causes all subsequent paths to be interpreted using
1300        CM standard pathname syntax (= Unix conventions); a line
1301    
1302            native!
1303    
1304        switches back to native style.  This was needed so that
1305        path config files can be written portably, see src/system/pathconfig.
1306    
1307      - Runtime system:
1308    
1309          - win32-filesys.c:  get_file_time and set_file_time now
1310            access modification time, not creation time.
1311    
1312          - I/O code made aware of new array representation.
1313    
1314          - Bug fixes in X86.prim.masm.
1315    
1316       - src/system/makeml made aware of win32. (For use under cygwin
1317            and other Unix-environments for windows.)
1318    
1319       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1320         which caused BinIO.inputAll to fail consistently.
1321    
1322    .bat scripts:
1323    
1324       Windows .bat scripts assume that SMLNJ_HOME is defined.
1325    
1326       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1327         applications (sml, ml-yacc, ml-lex).
1328       - ml-build.bat: analogous to ml-build.
1329       - config\install.bat:  Analogous to config/install.sh.  This requires
1330         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1331         (nmake etc. must be on the path, and vcvars32 must have been run.)
1332         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1333         and the bootfile hierarchy must have been unpacked under
1334         sml.boot.x86-win32.
1335         The script is very primitive and does a poor job at error checking.
1336         It only installs the base system, ml-lex, and ml-yacc.  No other
1337         libraries are being installed (i.e., you get only those that
1338         are part of the compiler.)
1339       - link-sml.bat:  analogous to .link-sml, but not currently used
1340    
1341    Unrelated bug fixes:
1342    
1343      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1344        types.
1345      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1346        otherwise.)
1347    
1348    ----------------------------------------------------------------------
1349    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1350    Date: 2003/06/17
1351    Tag: macqueen-20030617-bug895
1352    Description:
1353    
1354    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1355    Tag will be used for new development branch (mcz-branch) for
1356    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1357    
1358    ----------------------------------------------------------------------
1359    Name: Matthias Blume (blume (at) tti - c (dot) org)
1360    Date: 2003/05/27 16:55:00 CDT
1361    Tag: blume-20030527-polyeq
1362    Description:
1363    
1364    Tried to eliminated most cases of polymorphic equality.
1365    
1366    ----------------------------------------------------------------------
1367    Name: Matthias Blume (blume (at) tti - c (dot) org)
1368    Date: 2003/05/21 17:45:00 CDT
1369    Tag: blume-20030517-complete
1370    Description:
1371    
1372    Two changes:
1373    
1374      1. Added a flag for controlling whether non-exhaustive bindings will
1375         be treated as errors (default is false).
1376      2. Cleaned up the *entire* source tree so that CMB.make goes through
1377         without a single non-exhaustive match- or bind warning.
1378    
1379    ----------------------------------------------------------------------
1380    Name: Matthias Blume (blume (at) tti - c (dot) org)
1381    Date: 2003/05/17 10:20:00 CDT
1382    Tag: blume-20030517-absyn
1383    Description:
1384    
1385    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1386    
1387       This mainly affects the quality of error messages.  However, some
1388       of the code is now more straightforward than before. (Treatment of
1389       the above four constructs in translate.sml is much simpler than
1390       the "macro-expansion" that was going on before.  Plus, the mach-
1391       compiler no longer gets invoked just to be able to compile an
1392       if-expression.)
1393    
1394    2. The ErrorMsg.Error exception is now caught and absorbed by the
1395       interactive loop.
1396    
1397    ----------------------------------------------------------------------
1398    Name: Allen Leung
1399    Date: 2003/05/16 13:05:00 CDT
1400    Tag: leunga-20030516-cygwin-runtime
1401    Description:
1402    
1403        Ported the runtime system to cygwin, which uses the unix
1404    x86-unix bin files.  Missing/buggy features:
1405    
1406       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1407         the Cygwin library.
1408       o Ctrl-C handling may be flaky.
1409       o Windows system calls and Windows I/O are not supported.
1410    
1411       A new set of binfiles is located at:
1412    
1413         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1414    
1415       This is only needed for bootstrapping the cygwin version of smlnj.
1416       Other x86 versions can use the existing binfiles.
1417    
1418    ----------------------------------------------------------------------
1419    Name: Matthias Blume
1420    Date: 2003/04/08 15:42:00 CDT
1421    Tag: blume-20030408-listpair
1422    Description:
1423    
1424    1. Added a target 'mlrisc' to installer.
1425    
1426    2. Added missing elements to structure ListPair.
1427    
1428    ----------------------------------------------------------------------
1429    Name: Allen Leung
1430    Date: 2003/01/07 10:40:00 EST
1431    Tag: leunga-20030107-int-rem
1432    Description:
1433    
1434    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1435    The arguments to the SUBL instruction were swapped.
1436    
1437    ----------------------------------------------------------------------
1438    Name: Matthias Blume
1439    Date: 2002/12/12 16:25:00 EST
1440    Tag: blume-20021212-risc-ra
1441    Description:
1442    
1443    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1444    sent the RA into an infinite loop when floating point registers get
1445    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1446    110.37 and 110.38.)
1447    There was another set of potential problems related to the handling of
1448    MLRISC annotations (but those did not yet cause real problems, apparently).
1449    
1450    ----------------------------------------------------------------------
1451    Name: Matthias Blume
1452    Date: 2002/12/06 22:40:00 EST
1453    Tag: blume-20021206-cm-fileid
1454    Description:
1455    
1456    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1457    This fixes the problem of CM getting confused by files that suddenly
1458    change their identity (e.g., by getting unlinked and recreated by some
1459    text editor such as vi).  There might be a better/cheaper/cleaner way
1460    of doing this, but for now this will have to do.
1461    
1462    ----------------------------------------------------------------------
1463    Name: Matthias Blume
1464    Date: 2002/10/28 09:50:00 EST
1465    Tag: blume-20021028-typecheck
1466    Description:
1467    
1468    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1469    
1470    ----------------------------------------------------------------------
1471    Name: Matthias Blume
1472    Date: 2002/10/17 09:10:00 EDT
1473    Tag: Release_110_42
1474    Description:
1475    
1476    In good old tradition, there has been a slight hiccup so that we have
1477    to patch 110.42 after the fact.   The old release tag has been replaced
1478    (see below).
1479    
1480    The change solves a problem with two competing approaches the
1481    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
1482    each other's way.
1483    
1484    This change only affects the runtime system code and the installer script.
1485    (No new bootfiles.)
1486    
1487    ----------------------------------------------------------------------
1488    Name: Matthias Blume
1489    Date: 2002/10/16 12:00:00 EDT
1490    Tag: Release_110_42_removed
1491    Description:
1492    
1493    New working release.  New bootfiles.
1494    
1495    ----------------------------------------------------------------------
1496    Name: Matthias Blume
1497    Date: 2002/10/10 13:10:00 EDT
1498    Tag: blume-20021010-ppc-divs
1499    Description:
1500    
1501    The mltree operator DIVS must be implemented with an overflow check on
1502    the PPC because the hardware indicates divide-by-zero using "overflow" as
1503    well.
1504    
1505    ----------------------------------------------------------------------
1506    Name: Matthias Blume
1507    Date: 2002/07/23 11:45:00
1508    Tag: blume-20020723-smlnj-home
1509    Description:
1510    
1511    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
1512    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
1513    CM_PATHCONFIG is also set), the path configuration file is assumed
1514    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
1515    the entire tree to some other place and everything will "just work".
1516    
1517    (Companion commands such as ml-build and ml-makedepend also sense this
1518    variable.)
1519    
1520    ----------------------------------------------------------------------
1521    Name: Matthias Blume
1522    Date: 2002/07/12 21:19:00 EDT
1523    Tag: blume-20020712-liveness
1524    Description:
1525    
1526    Exported two useful "step" functions from liveness module (MLRISC).
1527    
1528    ----------------------------------------------------------------------
1529    Name: Matthias Blume
1530    Date: 2002/07/05 16:00 EDT
1531    Tag: Release_110_41
1532    Description:
1533    
1534    New working release.  New bootfiles.
1535    
1536    ----------------------------------------------------------------------
1537    Name: Matthias Blume
1538    Date: 2002/07/05 10:25:00 EDT
1539    Tag: blume-20020705-btimp
1540    Description:
1541    
1542    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
1543    other clients can set up backtracing support.
1544    
1545    ----------------------------------------------------------------------
1546    Name: Matthias Blume
1547    Date: 2002/06/25 17:23:00 EDT
1548    Tag: blume-20020625-fpmax
1549    Description:
1550    
1551    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
1552    case.  (The sense of the isNaN test was reversed -- which made min and
1553    max always return their first argument.)
1554    
1555    ----------------------------------------------------------------------
1556    Name: Matthias Blume
1557    Date: 2002/06/11
1558    Tag: blume-20020611-unixpath
1559    Description:
1560    
1561    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
1562    
1563    ----------------------------------------------------------------------
1564    Name: Matthias Blume
1565    Date: 2002/06/10 16:35:00 EDT
1566    Tag: blume-20020610-ieeereal
1567    Description:
1568    
1569    I back-ported my implementation of IEEEReal.fromString from the
1570    idlbasis-devel branch so that we can test it.
1571    
1572    Another small change is that ppDec tries to give more information
1573    than just "<sig>" in the case of functors.  However, this code is
1574    broken in some mysterious way if the functor's body's signature
1575    has not been declared by ascription but gets inferred from the
1576    implementation.  This needs fixing...
1577    
1578    ----------------------------------------------------------------------
1579    Name: Matthias Blume
1580    Date: 2002/05/31
1581    Tag: blume-20020531-btrace-mode
1582    Description:
1583    
1584    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
1585    the wayside when I switched over to using Controls everywhere.)
1586    
1587    ----------------------------------------------------------------------
1588    Name: Lal George
1589    Date: 2002/05/23 12:21:40 EDT
1590    Tag: george-20020523-visual-labels
1591    Description:
1592    
1593    Labels are now displayed in the graphical output to make
1594    the fall-through and target blocks obvious.
1595    
1596    ----------------------------------------------------------------------
1597    Name: Matthias Blume
1598    Date: 2002/05/22 11:03:00 EDT
1599    Tag: blume-20020522-shrink
1600    Description:
1601    
1602    John tweaked yesterday's fix for 1131 to handle an out-of-memory
1603    situation that comes up when allocating huge arrays.
1604    
1605    ----------------------------------------------------------------------
1606    Name: Matthias Blume
1607    Date: 2002/05/21 16:00:00 EDT
1608    Tag: Release_110_40
1609    Description:
1610    
1611    New working release (110.40).  New bootfiles.
1612    
1613    [Also: John Reppy fixed GC bug 1131.]
1614    
1615    ----------------------------------------------------------------------
1616    Name: Matthias Blume
1617    Date: 2002/05/21 12:35:00 EDT
1618    Tag: blume-20020521-cmdoc
1619    Description:
1620    
1621    CM documentation update.
1622    
1623    ----------------------------------------------------------------------
1624    Name: Matthias Blume
1625    Date: 2002/05/21 10:55:00 EDT
1626    Tag: blume-20020521-misc
1627    Description:
1628    
1629    - John tweaked runtime to be silent on heap export (except when
1630      GC messages are on).
1631    - I added a few more things (cross-compiling versions of CMB) to
1632      config/preloads (as suggestions).
1633    
1634    ----------------------------------------------------------------------
1635    Name: Matthias Blume
1636    Date: 2002/05/20 22:25:00 EDT
1637    Tag: blume-20020520-controls
1638    Description:
1639    
1640    - Added ControlUtil structure to control-lib.cm.
1641    - Use it throughout.
1642    - Used Controls facility to define MLRISC controls (as opposed to
1643      registering MLRISC control ref cells with Controls after the fact)
1644    - Fixed messed-up controls priorities.
1645    
1646    * Removed again all the stuff from config/preloads that one wouldn't
1647      be able to preload at the time the initial heap image is built.
1648      (Many libraries, e.g., CML, do not exist yet at this time.  The
1649       only libraries that can be preloaded via config/preloads are those
1650       that come bundled with the bootfiles.)
1651    
1652    ----------------------------------------------------------------------
1653    Name: Matthias Blume
1654    Date: 2002/05/20 10:59:00 EDT
1655    Tag: blume-20020520-preloads
1656    Description:
1657    
1658    Added a lot of commented-out suggestions for things to be included
1659    in config/preloads.
1660    
1661    ----------------------------------------------------------------------
1662    Name: Allen Leung
1663    Date: 2002/05/18 14:20:00 EDT
1664    Tag: leunga-20020518-mdl
1665    Description:
1666    
1667    o Made the mdl tool stuff compile and run again.
1668    o I've disabled all the stuff that depends on RTL specifications; they
1669      are all badly broken anyway.
1670    
1671    ----------------------------------------------------------------------
1672    Name: Matthias Blume
1673    Date: 2002/05/17 16:49:00 EDT
1674    Tag: blume-20020517-controls
1675    Description:
1676    
1677    0. John Reppy made several modifications to the SML/NJ library.
1678       In particular, there is a shiny new controls-lib.cm.
1679    
1680    1. Pushed new controls interface through compiler so that everything
1681       compiles again.
1682    
1683    2. Added FormatComb and FORMAT_COMB to the CML version of the
1684       SML/NJ library (so that CML compiles again).
1685    
1686    3. Modified init scripts because XXX_DEFAULT environment variables
1687       are no longer with us.  (Boot-time initialization is now done
1688       using the same environment variables that are also used for
1689       startup-time initialization of controls.)
1690    
1691    ----------------------------------------------------------------------
1692    Name: Lal George
1693    Date: 2002/05/15 09:20:10 EDT
1694    Tag: george-20020515-pseudo-op-decls
1695    Description:
1696    
1697            All pseudo-ops emitted before the first segment declaration
1698    such as TEXT, DATA, and BSS directives are assumed to be global
1699    declarations and are emitted first in the assembly file. This is
1700    useful in a number of situations where one has pseudo-ops that are not
1701    specific to any segment, and also works around the constraint that one
1702    cannot have client pseudo-ops in the TEXT segment.
1703    
1704            Because no segment is associated with these declarations it is
1705    an error to allocate any space or objects before the first segment
1706    directive and an exception will be raised. However, we cannot make
1707    this check for client pseudo-ops.
1708    
1709            These top level declarations are a field in the CFG graph_info.
1710    In theory you can continue to add to this field after the CFG has been
1711    built -- provided you know what you are doing;-)
1712    
1713    ----------------------------------------------------------------------
1714    Name: Matthias Blume
1715    Date: 2002/05/13 16:40:00 EDT
1716    Tag: blume-20020513-pp-etc
1717    Description:
1718    
1719    A few minor bugfixes:
1720    
1721      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
1722        (Bogus printouts for redefined bindings still occur.  Compiler
1723        bug should no longer occur now.  We need to redo the prettyprinter
1724        from scratch.)
1725    
1726      - CM pathname printer now also adds escape sequences for ( and )
1727    
1728      - commend and docu fixes for ml-nlffi
1729    
1730    ----------------------------------------------------------------------
1731    Name: Matthias Blume
1732    Date: 2002/05/10 16:40:00 EDT
1733    Tag: blume-20020510-erg-textio
1734    Description:
1735    
1736    Applied the following bugfix provided by Emden Gansner:
1737    
1738        Output is corrupted when outputSubstr is used rather than output.
1739    
1740        The problem occurs when a substring
1741    
1742            ss = (s, dataStart, dataLen)
1743    
1744        where dataStart > 0, fills a stream buffer with avail bytes left.
1745        avail bytes of s, starting at index dataStart, are copied into the
1746        buffer, the buffer is flushed, and then the remaining dataLen-avail
1747        bytes of ss are copied into the beginning of the buffer. Instead of
1748        starting this copy at index dataStart+avail in s, the current code
1749        starts the copy at index avail.
1750    
1751        Fix:
1752        In text-io-fn.sml, change line 695 from
1753             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
1754        to
1755             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
1756    
1757    ----------------------------------------------------------------------
1758    Name: Matthias Blume
1759    Date: 2002/04/12 13:55:00 EDT
1760    Tag: blume-20020412-assyntax
1761    Description:
1762    
1763    1. Grabbed newer assyntax.h from the XFree86 project.
1764    2. Fiddled with how to compile X86.prim.asm without warnings.
1765    3. (Very) Minor cleanup in CM.
1766    
1767    ----------------------------------------------------------------------
1768    Name: Matthias Blume
1769    Date: 2002/04/01 (no joke!) 17:07:00 EST
1770    Tag: blume-20020401-x86div
1771    Description:
1772    
1773    Added full support for div/mod/rem/quot on the x86, using the machine
1774    instruction's two results (without clumsily recomputing the remainder)
1775    directly where appropriate.
1776    
1777    Some more extensive power-of-two support was added to the x86 instruction
1778    selector (avoiding expensive divs, mods, and muls where they can be
1779    replaced with cheaper shifts and masks).  However, this sort of thing
1780    ought to be done earlier, e.g., within the CPS optimizer so that
1781    all architectures benefit from it.
1782    
1783    The compiler compiles to a fixed point, but changes might be somewhat
1784    fragile nevertheless.  Please, report any strange things that you might
1785    see wrt. div/mod/quot/rem...
1786    
1787    ----------------------------------------------------------------------
1788    Name: Matthias Blume
1789    Date: 2002/03/29 17:22:00
1790    Tag: blume-20020329-div
1791    Description:
1792    
1793    Fixed my broken div/mod logic.  Unfortunately, this means that the
1794    inline code for div/mod now has one more comparison than before.
1795    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
1796    The problem was with quotient = 0, because that alone does not tell
1797    us which way the rounding went.  One then has to look at whether
1798    remainder and divisor have the same sign...  :(
1799    
1800    Anyway, I replaced the bootfiles with fresh ones...
1801    
1802    ----------------------------------------------------------------------
1803    Name: Matthias Blume
1804    Date: 2002/03/29 14:10:00 EST
1805    Tag: blume-20020329-inlprims
1806    Description:
1807    
1808    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
1809    
1810    Primops have changed. This means that the bin/boot-file formats have
1811    changed as well.
1812    
1813    To make sure that there is no confusion, I made a new version.
1814    
1815    
1816    CHANGES:
1817    
1818    * removed REMT from mltree (remainder should never overflow).
1819    
1820    * added primops to deal with divisions of all flavors to the frontend
1821    
1822    * handled these primops all the way through so they map to their respective
1823      MLRISC support
1824    
1825    * used these primops in the implementation of Int, Int32, Word, Word32
1826    
1827    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
1828    
1829    * parameterized INLMIN, INLMAX, and INLABS by a numkind
1830    
1831    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
1832      floating point
1833    
1834    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
1835      and Real (but Real.abs maps to a separate floating-point-only primop)
1836    
1837    
1838    TODO items:
1839    
1840    * Hacked Alpha32 instruction selection, disabling the selection of REMx
1841      instructions because the machine instruction encoder cannot handle
1842      them.  (Hppa, PPC, and Sparc instruction selection did not handle
1843      REM in the first place, and REM is supported by the x86 machine coder.)
1844    
1845    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
1846      selection phase.  (The two can be streamlined because the hardware
1847      delivers both quotient and remainder at the same time anyway.)
1848    
1849    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
1850      (Currently the behavior is inconsistent both across architectures and
1851      wrt. the draft Basis spec.)
1852    
1853    * Word8 should eventually be handled natively, too.
1854    
1855    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
1856      as if there currently is no execution path that could trigger it in
1857      SML/NJ.  (The assumptions underlying functions arith and promotable do not
1858      hold for things like multiplication and division.)
1859    
1860    ----------------------------------------------------------------------
1861    Name: Matthias Blume
1862    Date: 2002/03/27 16:27:00 EST
1863    Tag: blume-20020327-mlrisc-divisions
1864    Description:
1865    
1866    Added support for all four division operations (ML's div, mod, quot,
1867    and rem) to MLRISC.  In the course of doing so, I also rationalized
1868    the naming (no more annoying switch-around of DIV and QUOT), by
1869    parameterizing the operation by div_rounding_mode (which can be either
1870    DIV_TO_ZERO or DIV_TO_NEGINF).
1871    
1872    The generic MLTreeGen functor takes care of compiling all four
1873    operations down to only round-to-zero div.
1874    
1875    Missing pieces:
1876    
1877      * Doing something smarter than relying on MLTreeGen on architectures
1878        like, e.g., the x86 where hardware division delivers both quotient and
1879        remainder at the same time.  With this, the implementation of the
1880        round-to-neginf operations could be further streamlined.
1881    
1882      * Remove inlining support for div/mod/rem from the frontend and replace it
1883        with primops that get carried through to the backend.  Do this for all
1884        int and word types.
1885    
1886    ----------------------------------------------------------------------
1887    Name: Matthias Blume
1888    Date: 2002/03/25 17:25:00 EST
1889    Tag: blume-20020325-divmod
1890    Description:
1891    
1892    I improved (hopefully without breaking them) the implementation of Int.div,
1893    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
1894    advantage of the following observations:
1895    
1896      Let  q = x quot y      r = x rem y
1897           d = x div  y      m = x mod y
1898    
1899    where "quot" is the round-to-zero version of integer division that
1900    hardware usually provides.  Then we have:
1901    
1902         r = x - q * y        where neither the * nor the - will overflow
1903         d = if q >= 0 orelse x = q * y then q else q - 1
1904                              where neither the * nor the - will overflow
1905         m = if q >= 0 orelse r = 0 then r else r + y
1906                              where the + will not overflow
1907    
1908    This results in substantial simplification of the generated code.
1909    The following table shows the number of CFG nodes and edges generated
1910    for
1911            fun f (x, y) = x OPER y
1912            (* with OPER \in div, mod, quot, rem *)
1913    
1914    
1915        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
1916        --------------------------------------------------------
1917         div |         24 |         39 |         12 |         16
1918         mod |         41 |         71 |         12 |         16
1919        quot |          8 |         10 |          8 |         10
1920         rem |         10 |         14 |          8 |         10
1921    
1922    
1923    ----------------------------------------------------------------------
1924    Name: Matthias Blume
1925    Date: 2002/03/25 22:06:00 EST
1926    Tag: blume-20020325-cprotobug
1927    Description:
1928    
1929    Fixed a bug in cproto (c prototype decoder).
1930    
1931    ----------------------------------------------------------------------
1932    Name: Matthias Blume
1933    Date: 2002/03/25 16:00:00 EST
1934    Tag: blume-20020325-raw-primops
1935    Description:
1936    
1937    I did some cleanup to Allen's new primop code and
1938    replaced yesterday's bootfiles with new ones.
1939    (But they are stored in the same place.)
1940    
1941    ----------------------------------------------------------------------
1942    Name: Matthias Blume
1943    Date: 2002/03/24 22:40:00 EST
1944    Tag: blume-20020324-bootfiles
1945    Description:
1946    
1947    Made the bootfiles that Allen asked for.
1948    
1949    ----------------------------------------------------------------------
1950    Name: Allen Leung
1951    Date: 2002/03/23 15:50:00 EST
1952    Tag: leunga-20020323-flint-cps-rcc-primops
1953    Description:
1954    
1955      1. Changes to FLINT primops:
1956    
1957        (* make a call to a C-function;
1958         * The primop carries C function prototype information and specifies
1959         * which of its (ML-) arguments are floating point. C prototype
1960         * information is for use by the backend, ML information is for
1961         * use by the CPS converter. *)
1962      | RAW_CCALL of { c_proto: CTypes.c_proto,
1963                       ml_args: ccall_type list,
1964                       ml_res_opt: ccall_type option,
1965                       reentrant : bool
1966                     } option
1967       (* Allocate uninitialized storage on the heap.
1968        * The record is meant to hold short-lived C objects, i.e., they
1969        * are not ML pointers.  With the tag, the representation is
1970        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
1971        *)
1972      | RAW_RECORD of {tag:bool,sz:int}
1973      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
1974    
1975      2.  These CPS primops are now overloaded:
1976    
1977           rawload of {kind:numkind}
1978           rawstore of {kind:numkind}
1979    
1980          The one argument form is:
1981    
1982             rawload {kind} address
1983    
1984          The two argument form is:
1985    
1986             rawload {kind} [ml object, byte-offset]
1987    
1988      3. RAW_CCALL/RCC now takes two extra arguments:
1989    
1990         a. The first is whether the C call is reentrant, i.e., whether
1991            ML state should be saved and restored.
1992         b. The second argument is a string argument specifying the name of
1993            library and the C function.
1994    
1995         These things are currently not handled in the code generator, yet.
1996    
1997      4. In CProto,
1998    
1999         An encoding type of "bool" means "ml object" and is mapped into
2000         C prototype of PTR.  Note that "bool" is different than "string",
2001         even though "string" is also mapped into PTR, because "bool"
2002         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2003    
2004      5. Pickler/unpicker
2005    
2006         Changed to handle RAW_RECORD and newest RAW_CCALL
2007    
2008      6. MLRiscGen,
2009    
2010         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2011         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2012            in the file CodeGen/cpscompile/cps-c-calls.sml
2013    
2014      7. Added the conditional move operator
2015    
2016             condmove of branch
2017    
2018         to cps.  Generation of this is still buggy so it is currently
2019         disabled.
2020    
2021    ----------------------------------------------------------------------
2022    Name: Lal George
2023    Date: 2002/03/22 14:18:25 EST
2024    Tag: george-20020322-cps-branch-prob
2025    Description:
2026    
2027    Implemented the Ball-Larus branch prediction-heuristics, and
2028    incorporated graphical viewers for control flow graphs.
2029    
2030    Ball-Larus Heuristics:
2031    ---------------------
2032    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2033    
2034    By design it uses the Dempster-Shafer theory for combining
2035    probabilities.  For example, in the function:
2036    
2037        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2038    
2039    the ball-larus heuristics predicts that the n=0 is unlikely
2040    (OH-heuristic), and the 'then' branch is unlikely because of the
2041    RH-heuristic -- giving the 'then' branch an even lower combined
2042    probability using the Dempster-Shafer theory.
2043    
2044    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2045    probability of the 'then' branch because of the loop in the else
2046    branch.
2047    
2048    
2049    Graphical Viewing:
2050    ------------------
2051    I merely plugged in Allen's graphical viewers into the compiler. The
2052    additional code is not much. At the top level, saying:
2053    
2054            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2055    
2056    will display the graphical view of the control flow graph just before
2057    back-patching.  daVinci must be in your path for this to work. If
2058    daVinci is not available, then the default viewer can be changed
2059    using:
2060    
2061            Control.MLRISC.getString "viewer"
2062    
2063    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2064    course, these viewers must be in your path.
2065    
2066    The above will display the compilation unit at the level of clusters,
2067    many of which are small, boring, and un-interesting. Also setting:
2068    
2069            Control.MLRISC.getInt "cfg-graphical-view_size"
2070    
2071    will display clusters that are larger than the value set by the above.
2072    
2073    
2074    ----------------------------------------------------------------------
2075    Name: Matthias Blume
2076    Date: 2002/03/21 22:20:00 EST
2077    Tag: blume-20020321-kmp-bugfix
2078    Description:
2079    
2080    Changed the interface to the KMP routine in PreString and fixed
2081    a minor bug in one place where it was used.
2082    
2083    ----------------------------------------------------------------------
2084    Name: Allen Leung
2085    Date: 2002/03/21 20:30:00 EST
2086    Tag: leunga-20020321-cfg
2087    Description:
2088    
2089      Fixed a potential problem in cfg edge splitting.
2090    
2091    ----------------------------------------------------------------------
2092    Name: Allen Leung
2093    Date: 2002/03/21 17:15:00 EST
2094    Tag: leunga-20020321-x86-fp-cfg
2095    Description:
2096    
2097      1. Recoded the buggy parts of x86-fp.
2098    
2099         a. All the block reordering code has been removed.
2100            We now depend on the block placement phases to do this work.
2101    
2102         b. Critical edge splitting code has been simplified and moved into the
2103            CFG modules, as where they belong.
2104    
2105         Both of these were quite buggy and complex.  The code is now much, much
2106         simpler.
2107    
2108      2. X86 backend.
2109    
2110         a. Added instructions for 64-bit support.  Instruction selection for
2111            64-bit has not been committed, however, since that
2112            requires changes to MLTREE which haven't been approved by
2113            Lal and John.
2114    
2115         b. Added support for FUCOMI and FUCOMIP when generating code for
2116            PentiumPro and above.  We only generate these instructions in
2117            the fast-fp mode.
2118    
2119         c. Added cases for JP and JNP in X86FreqProps.
2120    
2121      3. CFG
2122    
2123         CFG now has a bunch of methods for edge splitting and merging.
2124    
2125      4. Machine description.
2126    
2127         John's simplification of MLTREE_BASIS.fcond broke a few machine
2128         description things:
2129    
2130         rtl-build.{sig,sml} and hppa.mdl fixed.
2131    
2132         NOTE: the machine description stuff in the repository is still broken.
2133               Again, I can't put my fixes in because that involves
2134               changes to MLTREE.
2135    
2136    ----------------------------------------------------------------------
2137    Name: Matthias Blume
2138    Date: 2002/03/20 15:55:00 EST
2139    Tag: blume-20020320-kmp
2140    Description:
2141    
2142    Implemented Knuth-Morris-Pratt string matching in PreString and used
2143    it for String.isSubstring, Substring.isSubstring, and
2144    Substring.position.
2145    
2146    (Might need some stress-testing.  Simple examples worked fine.)
2147    
2148    ----------------------------------------------------------------------
2149    Name: Matthias Blume
2150    Date: 2002/03/19 16:37:00 EST
2151    Tag: blume-20020319-witnesses
2152    Description:
2153    
2154    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2155    
2156    This implements a generic mechanism for changing constness qualifiers
2157    anywhere within big C types without resorting to outright "casts".
2158    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2159    modify the constness at the outermost level.)
2160    The implementation of "convert" is based on the idea of "witness"
2161    values -- values that are not used by the operation but whose types
2162    "testify" to their applicability.  On the implementation side, "convert"
2163    is simply a projection (returning its second curried argument).  With
2164    cross-module inlining, it should not result in any machine code being
2165    generated.
2166    
2167    ----------------------------------------------------------------------
2168    Name: Matthias Blume
2169    Date: 2002/03/15 16:40:00 EST
2170    Tag: blume-20020315-basis
2171    Description:
2172    
2173    Provided (preliminary?) implementations for
2174    
2175      {String,Substring}.{concatWith,isSuffix,isSubstring}
2176    
2177    and
2178    
2179      Substring.full
2180    
2181    Those are in the Basis spec but they were missing in SML/NJ.
2182    
2183    ----------------------------------------------------------------------
2184    Name: Matthias Blume
2185    Date: 2002/03/14 21:30:00 EST
2186    Tag: blume-20020314-controls
2187    Description:
2188    
2189    Controls:
2190    ---------
2191    
2192    1. Factored out the recently-added Controls : CONTROLS stuff and put
2193       it into its own library $/controls-lib.cm.  The source tree for
2194       this is under src/smlnj-lib/Controls.
2195    
2196    2. Changed the names of types and functions in this interface, so they
2197       make a bit more "sense":
2198    
2199          module -> registry
2200          'a registry -> 'a group
2201    
2202    3. The interface now deals in ref cells only.  The getter/setter interface
2203       is (mostly) gone.
2204    
2205    4. Added a function that lets one register an already-existing ref cell.
2206    
2207    5. Made the corresponding modifications to the rest of the code so that
2208       everything compiles again.
2209    
2210    6. Changed the implementation of Controls.MLRISC back to something closer
2211       to the original.  In particular, this module (and therefore MLRISC)
2212       does not depend on Controls.  There now is some link-time code in
2213       int-sys.sml that registers the MLRISC controls with the Controls
2214       module.
2215    
2216    CM:
2217    ---
2218    
2219      * One can now specify the lambda-split aggressiveness in init.cmi.
2220    
2221    ----------------------------------------------------------------------
2222    Name: Allen Leung
2223    Date: 2002/03/13 17:30:00 EST
2224    Tag: leunga-20020313-x86-fp-unary
2225    Description:
2226    
2227    Bug fix for:
2228    
2229    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2230    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2231    > - fun f(x,(y,z)) = Real.~ y;
2232    > [autoloading]
2233    > [autoloading done]
2234    >       fchsl   (%eax), 184(%esp)
2235    > Error: MLRisc bug: X86MCEmitter.emitInstr
2236    >
2237    > uncaught exception Error
2238    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2239    
2240    The problem was that the code generator did not generate any fp registers
2241    in this case, and the ra didn't know that it needed to run the X86FP phase to
2242    translate the pseudo fp instruction.   This only happened with unary fp
2243    operators in certain situations.
2244    
2245    ----------------------------------------------------------------------
2246    Name: Matthias Blume
2247    Date: 2002/03/13 14:00:00 EST
2248    Tag: blume-20020313-overload-etc
2249    Description:
2250    
2251    1. Added _overload as a synonym for overload for backward compatibility.
2252       (Control.overloadKW must be true for either version to be accepted.)
2253    
2254    2. Fixed bug in install script that caused more things to be installed
2255       than what was requested in config/targets.
2256    
2257    3. Made CM aware of the (_)overload construct so that autoloading
2258       works.
2259    
2260    ----------------------------------------------------------------------
2261    Name: Matthias Blume
2262    Date: 2002/03/12 22:03:00 EST
2263    Tag: blume-20020312-url
2264    Description:
2265    
2266    Forgot to update BOOT and srcarchiveurl.
2267    
2268    ----------------------------------------------------------------------
2269    Name: Matthias Blume
2270    Date: 2002/03/12 17:30:00 EST
2271    Tag: blume-20020312-version110392
2272    Description:
2273    
2274    Yet another version number bump (because of small changes to the
2275    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2276    
2277    Changes:
2278    
2279      The new pid generation scheme described a few weeks ago was overly
2280      complicated.  I implemented a new mechanism that is simpler and
2281      provides a bit more "stability":  Once CM has seen a compilation
2282      unit, it keeps its identity constant (as long as you do not delete
2283      those crucial CM/GUID/* files).  This means that when you change
2284      an interface, compile, then go back to the old interface, and
2285      compile again, you arrive at the original pid.
2286    
2287      There now also is a mechanism that instructs CM to use the plain
2288      environment hash as a module's pid (effectively making its GUID
2289      the empty string).  For this, "noguid" must be specified as an
2290      option to the .sml file in question within its .cm file.
2291      This is most useful for code that is being generated by tools such
2292      as ml-nlffigen (because during development programmers tend to
2293      erase the tool's entire output directory tree including CM's cached
2294      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2295      revert to the old, broken behavior of SML/NJ, but in specific cases
2296      where there is no danger of interface confusion, its use is ok
2297      (I think).
2298    
2299      ml-nlffigen by default generates "noguid" annotations.  They can be
2300      turned off by specifying -guid in its command line.
2301    
2302    ----------------------------------------------------------------------
2303    Name: Lal George
2304    Date: 2002/03/12 12 14:42:36 EST
2305    Tag: george-20020312-frequency-computation
2306    Description:
2307    
2308    Integrated jump chaining and static block frequency into the
2309    compiler. More details and numbers later.
2310    
2311    ----------------------------------------------------------------------
2312    Name: Lal George
2313    Date: 2002/03/11 11 22:38:53 EST
2314    Tag: george-20020311-jump-chain-elim
2315    Description:
2316    
2317    Tested the jump chain elimination on all architectures (except the
2318    hppa).  This is on by default right now and is profitable for the
2319    alpha and x86, however, it may not be profitable for the sparc and ppc
2320    when compiling the compiler.
2321    
2322    The gc test will typically jump to a label at the end of the cluster,
2323    where there is another jump to an external cluster containing the actual
2324    code to invoke gc. This is to allow factoring of common gc invocation
2325    sequences. That is to say, we generate:
2326    
2327            f:
2328               testgc
2329               ja   L1      % jump if above to L1
2330    
2331            L1:
2332               jmp L2
2333    
2334    
2335    After jump chain elimination the 'ja L1' instructions is converted to
2336    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2337    up being implemented in their long form (if L2 is far away) using:
2338    
2339            jbe     L3      % jump if below or equal to L3
2340            jmp     L2
2341         L3:
2342            ...
2343    
2344    
2345    For large compilation units L2  may be far away.
2346    
2347    
2348    ----------------------------------------------------------------------
2349    Name: Matthias Blume
2350    Date: 2002/03/11 13:30:00 EST
2351    Tag: blume-20020311-mltreeeval
2352    Description:
2353    
2354    A functor parameter was missing.
2355    
2356    ----------------------------------------------------------------------
2357    Name: Allen Leung
2358    Date: 2002/03/11 10:30:00 EST
2359    Tag: leunga-20020311-runtime-string0
2360    Description:
2361    
2362       The representation of the empty string now points to a
2363    legal null terminated C string instead of unit.  It is now possible
2364    to convert an ML string into C string with InlineT.CharVector.getData.
2365    This compiles into one single machine instruction.
2366    
2367    ----------------------------------------------------------------------
2368    Name: Allen Leung
2369    Date: 2002/03/10 23:55:00 EST
2370    Tag: leunga-20020310-x86-call
2371    Description:
2372    
2373       Added machine generation for CALL instruction (relative displacement mode)
2374    
2375    ----------------------------------------------------------------------
2376    Name: Matthias Blume
2377    Date: 2002/03/08 16:05:00
2378    Tag: blume-20020308-entrypoints
2379    Description:
2380    
2381    Version number bumped to 110.39.1.  NEW BOOTFILES!
2382    
2383    Entrypoints: non-zero offset into a code object where execution should begin.
2384    
2385    - Added the notion of an entrypoint to CodeObj.
2386    - Added reading/writing of entrypoint info to Binfile.
2387    - Made runtime system bootloader aware of entrypoints.
2388    - Use the address of the label of the first function given to mlriscGen
2389      as the entrypoint.  This address is currently always 0, but it will
2390      not be 0 once we turn on block placement.
2391    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2392      with entry points) from mlriscGen.
2393    
2394    ----------------------------------------------------------------------
2395    Name: Allen Leung
2396    Date: 2002/03/07 20:45:00 EST
2397    Tag: leunga-20020307-x86-cmov
2398    Description:
2399    
2400       Bug fixes for CMOVcc on x86.
2401    
2402       1. Added machine code generation for CMOVcc
2403       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2404       3. CMOVcc cannot have an immediate operand as argument.
2405    
2406    ----------------------------------------------------------------------
2407    Name: Matthias Blume
2408    Date: 2002/03/07 16:15:00 EST
2409    Tag: blume-20020307-controls
2410    Description:
2411    
2412    This is a very large but mostly boring patch which makes (almost)
2413    every tuneable compiler knob (i.e., pretty much everything under
2414    Control.* plus a few other things) configurable via both the command
2415    line and environment variables in the style CM did its configuration
2416    until now.
2417    
2418    Try starting sml with '-h' (or, if you are brave, '-H')
2419    
2420    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2421    implements the underlying generic mechanism.
2422    
2423    The interface to some of the existing such facilities has changed somewhat.
2424    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2425    (The getFoo interface is still there for backward-compatibility, but its
2426    use is deprecated.)
2427    
2428    The ml-build script passes -Cxxx=yyy command-line arguments through so
2429    that one can now twiddle the compiler settings when using this "batch"
2430    compiler.
2431    
2432    TODO items:
2433    
2434    We should go through and throw out all controls that are no longer
2435    connected to anything.  Moreover, we should go through and provide
2436    meaningful (and correct!) documentation strings for those controls
2437    that still are connected.
2438    
2439    Currently, multiple calls to Controls.new are accepted (only the first
2440    has any effect).  Eventually we should make sure that every control
2441    is being made (via Controls.new) exactly once.  Future access can then
2442    be done using Controls.acc.
2443    
2444    Finally, it would probably be a good idea to use the getter-setter
2445    interface to controls rather than ref cells.  For the time being, both
2446    styles are provided by the Controls module, but getter-setter pairs are
2447    better if thread-safety is of any concern because they can be wrapped.
2448    
2449    *****************************************
2450    
2451    One bug fix: The function blockPlacement in three of the MLRISC
2452    backpatch files used to be hard-wired to one of two possibilities at
2453    link time (according to the value of the placementFlag).  But (I
2454    think) it should rather sense the flag every time.
2455    
2456    *****************************************
2457    
2458    Other assorted changes (by other people who did not supply a HISTORY entry):
2459    
2460    1. the cross-module inliner now works much better (Monnier)
2461    2. representation of weights, frequencies, and probabilities in MLRISC
2462       changed in preparation of using those for weighted block placement
2463       (Reppy, George)
2464    
2465    ----------------------------------------------------------------------
2466    Name: Lal George
2467    Date: 2002/03/07 14:44:24 EST 2002
2468    Tag: george-20020307-weighted-block-placement
2469    
2470    Tested the weighted block placement optimization on all architectures
2471    (except the hppa) using AMPL to generate the block and edge frequencies.
2472    Changes were required in the machine properties to correctly
2473    categorize trap instructions. There is an MLRISC flag
2474    "weighted-block-placement" that can be used to enable weighted block
2475    placement, but this will be ineffective without block/edge
2476    frequencies (coming soon).
2477    
2478    
2479    ----------------------------------------------------------------------
2480    Name: Lal George
2481    Date: 2002/03/05 17:24:48 EST
2482    Tag: george-20020305-linkage-cluster
2483    
2484    In order to support the block placement optimization, a new cluster
2485    is generated as the very first cluster (called the linkage cluster).
2486    It contains a single jump to the 'real' entry point for the compilation
2487    unit. Block placement has no effect on the linkage cluster itself, but
2488    all the other clusters  have full freedom in the manner in which they
2489    reorder blocks or functions.
2490    
2491    On the x86 the typical linkage code that is generated is:
2492       ----------------------
2493            .align 2
2494       L0:
2495            addl    $L1-L0, 72(%esp)
2496            jmp     L1
2497    
2498    
2499            .align  2
2500       L1:
2501       ----------------------
2502    
2503    72(%esp) is the memory location for the stdlink register. This
2504    must contain the address of the CPS function being called. In the
2505    above example, it contains the address of  L0; before
2506    calling L1 (the real entry point for the compilation unit), it
2507    must contain the address for L1, and hence
2508    
2509            addl $L1-L0, 72(%esp)
2510    
2511    I have tested this on all architectures except the hppa.The increase
2512    in code size is of course negligible
2513    
2514    ----------------------------------------------------------------------
2515    Name: Allen Leung
2516    Date: 2002/03/03 13:20:00 EST
2517    Tag: leunga-20020303-mlrisc-tools
2518    
2519      Added #[ ... ] expressions to mlrisc tools
2520    
2521    ----------------------------------------------------------------------
2522    Name: Matthias Blume
2523    Date: 2002/02/27 12:29:00 EST
2524    Tag: blume-20020227-cdebug
2525    Description:
2526    
2527    - made types in structure C and C_Debug to be equal
2528    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
2529    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
2530    
2531    ----------------------------------------------------------------------
2532    Name: Matthias Blume
2533    Date: 2002/02/26 12:00:00 EST
2534    Tag: blume-20020226-ffi
2535    Description:
2536    
2537    1. Fixed a minor bug in CM's "noweb" tool:
2538       If numbering is turned off, then truly don't number (i.e., do not
2539       supply the -L option to noweb).  The previous behavior was to supply
2540       -L'' -- which caused noweb to use the "default" line numbering scheme.
2541       Thanks to Chris Richards for pointing this out (and supplying the fix).
2542    
2543    2. Once again, I reworked some aspects of the FFI:
2544    
2545       A. The incomplete/complete type business:
2546    
2547       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
2548         gone!
2549       - ML types representing an incomplete type are now *equal* to
2550         ML types representing their corresponding complete types (just like
2551         in C).  This is still safe because ml-nlffigen will not generate
2552         RTTI for incomplete types, nor will it generate functions that
2553         require access to such RTTI.   But when ML code generated from both
2554         incomplete and complete versions of the C type meet, the ML types
2555         are trivially interoperable.
2556    
2557         NOTE:  These changes restore the full generality of the translation
2558         (which was previously lost when I eliminated functorization)!
2559    
2560       B. Enum types:
2561    
2562       - Structure C now has a type constructor "enum" that is similar to
2563         how the "su" constructor works.  However, "enum" is not a phantom
2564         type because each "T enum" has values (and is isomorphic to
2565         MLRep.Signed.int).
2566       - There are generic access operations for enum objects (using
2567         MLRep.Signed.int).
2568       - ml-nlffigen will generate a structure E_foo for each "enum foo".
2569         * The structure contains the definition of type "mlrep" (the ML-side
2570         representation type of the enum).  Normally, mlrep is the same
2571         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
2572         then mlrep will be defined as a datatype -- thus facilitating
2573         pattern matching on mlrep values.
2574         ("-ec" will be suppressed if there are duplicate values in an
2575          enumeration.)
2576         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
2577         will be generated for each C enum constant xxx.
2578         * Conversion functions m2i and i2m convert between mlrep and
2579         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
2580         * Coversion functions c and ml convert between mlrep and "tag enum".
2581         * Access functions (get/set) fetch and store mlrep values.
2582       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
2583         enumerations are merged into one single enumeration represented by
2584         structure E_'.
2585    
2586    ----------------------------------------------------------------------
2587    Name: Allen Leung
2588    Date: 2002/02/25 04:45:00 EST
2589    Tag: leunga-20020225-cps-spill
2590    
2591    This is a new implementation of the CPS spill phase.
2592    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
2593    In case of problems, replace it with the old file spill.sml
2594    
2595    The current compiler runs into some serious performance problems when
2596    constructing a large record.  This can happen when we try to compile a
2597    structure with many items.  Even a very simple structure like the following
2598    makes the compiler slow down.
2599    
2600        structure Foo = struct
2601           val x_1 = 0w1 : Word32.int
2602           val x_2 = 0w2 : Word32.int
2603           val x_3 = 0w3 : Word32.int
2604           ...
2605           val x_N = 0wN : Word32.int
2606        end
2607    
2608    The following table shows the compile time, from N=1000 to N=4000,
2609    with the old compiler:
2610    
2611    N
2612    1000   CPS 100 spill                           0.04u  0.00s  0.00g
2613           MLRISC ra                               0.06u  0.00s  0.05g
2614              (spills = 0 reloads = 0)
2615           TOTAL                                   0.63u  0.07s  0.21g
2616    
2617    1100   CPS 100 spill                           8.25u  0.32s  0.64g
2618           MLRISC ra                               5.68u  0.59s  3.93g
2619              (spills = 0 reloads = 0)
2620           TOTAL                                   14.71u  0.99s  4.81g
2621    
2622    1500   CPS 100 spill                           58.55u  2.34s  1.74g
2623           MLRISC ra                               5.54u  0.65s  3.91g
2624              (spills = 543 reloads = 1082)
2625           TOTAL                                   65.40u  3.13s  6.00g
2626    
2627    2000   CPS 100 spill                           126.69u  4.84s  3.08g
2628           MLRISC ra                               0.80u  0.10s  0.55g
2629              (spills = 42 reloads = 84)
2630           TOTAL                                   129.42u  5.10s  4.13g
2631    
2632    3000   CPS 100 spill                           675.59u  19.03s  11.64g
2633           MLRISC ra                               2.69u  0.27s  1.38g
2634              (spills = 62 reloads = 124)
2635           TOTAL                                   682.48u  19.61s  13.99g
2636    
2637    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
2638           MLRISC ra                               4.96u  0.27s  2.72g
2639              (spills = 85 reloads = 170)
2640           TOTAL                                   2375.26u  57.21s  48.00g
2641    
2642    As you can see the old cps spill module suffers from some serious
2643    performance problem.  But since I cannot decipher the old code fully,
2644    instead of patching the problems up, I'm reimplementing it
2645    with a different algorithm.  The new code is more modular,
2646    smaller when compiled, and substantially faster
2647    (O(n log n) time and O(n) space).  Timing of the new spill module:
2648    
2649    4000  CPS 100 spill                           0.02u  0.00s  0.00g
2650          MLRISC ra                               0.25u  0.02s  0.15g
2651             (spills=1 reloads=3)
2652          TOTAL                                   7.74u  0.34s  1.62g
2653    
2654    Implementation details:
2655    
2656    As far as I can tell, the purpose of the CPS spill module is to make sure the
2657    number of live variables at any program point (the bandwidth)
2658    does not exceed a certain limit, which is determined by the
2659    size of the spill area.
2660    
2661    When the bandwidth is too large, we decrease the register pressure by
2662    packing live variables into spill records.  How we achieve this is
2663    completely different than what we did in the old code.
2664    
2665    First, there is something about the MLRiscGen code generator
2666    that we should be aware of:
2667    
2668    o MLRiscGen performs code motion!
2669    
2670       In particular, it will move floating point computations and
2671       address computations involving only the heap pointer to
2672       their use sites (if there is only a single use).
2673       What this means is that if we have a CPS record construction
2674       statement
2675    
2676           RECORD(k,vl,w,e)
2677    
2678       we should never count the new record address w as live if w
2679       has only one use (which is often the case).
2680    
2681       We should do something similar to floating point, but the transformation
2682       there is much more complex, so I won't deal with that.
2683    
2684    Secondly, there are now two new cps primops at our disposal:
2685    
2686     1. rawrecord of record_kind option
2687        This pure operator allocates some uninitialized storage from the heap.
2688        There are two forms:
2689    
2690         rawrecord NONE [INT n]  allocates a tagless record of length n
2691         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
2692                                     and initializes the tag.
2693    
2694     2. rawupdate of cty
2695          rawupdate cty (v,i,x)
2696          Assigns to x to the ith component of record v.
2697          The storelist is not updated.
2698    
2699    We use these new primops for both spilling and increment record construction.
2700    
2701     1. Spilling.
2702    
2703        This is implemented with a linear scan algorithm (but generalized
2704        to trees).  The algorithm will create a single spill record at the
2705        beginning of the cps function and use rawupdate to spill to it,
2706        and SELECT or SELp to reload from it.  So both spills and reloads
2707        are fine-grain operations.  In contrast, in the old algorithm
2708        "spills" have to be bundled together in records.
2709    
2710        Ideally, we should sink the spill record construction to where
2711        it is needed.  We can even split the spill record into multiple ones
2712        at the places where they are needed.  But CPS is not a good
2713        representation for global code motion, so I'll keep it simple and
2714        am not attempting this.
2715    
2716     2. Incremental record construction (aka record splitting).
2717    
2718        Long records with many component values which are simulatenously live
2719        (recall that single use record addresses are not considered to
2720         be live) are constructed with rawrecord and rawupdate.
2721        We allocate space on the heap with rawrecord first, then gradually
2722        fill it in with rawupdate.  This is the technique suggested to me
2723        by Matthias.
2724    
2725        Some restrictions on when this is applicable:
2726        1. It is not a VECTOR record.  The code generator currently does not handle
2727           this case. VECTOR record uses double indirection like arrays.
2728        2. All the record component values are defined in the same "basic block"
2729           as the record constructor.  This is to prevent speculative
2730           record construction.
2731    
2732    ----------------------------------------------------------------------
2733    Name: Allen Leung
2734    Date: 2002/02/22 01:02:00 EST
2735    Tag: leunga-20020222-mlrisc-tools
2736    
2737    Minor bug fixes in the parser and rewriter
2738    
2739    ----------------------------------------------------------------------
2740    Name: Allen Leung
2741    Date: 2002/02/21 20:20:00 EST
2742    Tag: leunga-20020221-peephole
2743    
2744    Regenerated the peephole files.  Some contained typos in the specification
2745    and some didn't compile because of pretty printing bugs in the old version
2746    of 'nowhere'.
2747    
2748    ----------------------------------------------------------------------
2749    Name: Allen Leung
2750    Date: 2002/02/19 20:20:00 EST
2751    Tag: leunga-20020219-mlrisc-tools
2752    Description:
2753    
2754       Minor bug fixes to the mlrisc-tools library:
2755    
2756       1.  Fixed up parsing colon suffixed keywords
2757       2.  Added the ability to shut the error messages up
2758       3.  Reimplemented the pretty printer and fixed up/improved
2759           the pretty printing of handle and -> types.
2760       4.  Fixed up generation of literal symbols in the nowhere tool.
2761       5.  Added some SML keywords to to sml.sty
2762    
2763    ----------------------------------------------------------------------
2764    Name: Matthias Blume
2765    Date: 2002/02/19 16:20:00 EST
2766    Tag: blume-20020219-cmffi
2767    Description:
2768    
2769    A wild mix of changes, some minor, some major:
2770    
2771    * All C FFI-related libraries are now anchored under $c:
2772        $/c.cm      --> $c/c.cm
2773        $/c-int.cm  --> $c/internals/c-int.cm
2774        $/memory.cm --> $c/memory/memory.cm
2775    
2776    * "make" tool (in CM) now treats its argument pathname slightly
2777      differently:
2778        1. If the native expansion is an absolute name, then before invoking
2779           the "make" command on it, CM will apply OS.Path.mkRelative
2780           (with relativeTo = OS.FileSys.getDir()) to it.
2781        2. The argument will be passed through to subsequent phases of CM
2782           processing without "going native".  In particular, if the argument
2783           was an anchored path, then "make" will not lose track of that anchor.
2784    
2785    * Compiler backends now "know" their respective C calling conventions
2786      instead of having to be told about it by ml-nlffigen.  This relieves
2787      ml-nlffigen from one of its burdens.
2788    
2789    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
2790    
2791    * Export C_DEBUG and C_Debug from $c/c.cm.
2792    
2793    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
2794      subtyping relationship between incomplete pointers and their complete
2795      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
2796      with the convention of instantiating 'o with ('t, 'c) obj whenever
2797      the pointer target type is complete.  In the incomplete case, 'o
2798      will be instantiated with some "'c iobj" -- a type obtained by
2799      using one of the functors PointerToIncompleteType or PointerToCompleteType.
2800    
2801      Operations that work on both incomplete and complete pointer types are
2802      typed as taking an 'o ptr while operations that require the target to
2803      be known are typed as taking some ('t, 'c) obj ptr.
2804    
2805      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
2806      where void is an eqtype without any values.  This makes it possible
2807      to work on voidptr values using functions meant to operate on light
2808      incomplete pointers.
2809    
2810    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
2811      been vastly simplified.
2812    
2813    ----------------------------------------------------------------------
2814    Name: Matthias Blume
2815    Date: 2002/02/19 10:48:00 EST
2816    Tag: blume-20020219-pqfix
2817    Description:
2818    
2819    Applied Chris Okasaki's bug fix for priority queues.
2820    
2821    ----------------------------------------------------------------------
2822    Name: Matthias Blume
2823    Date: 2002/02/15 17:05:00
2824    Tag: Release_110_39
2825    Description:
2826    
2827    Last-minute retagging is becoming a tradition... :-(
2828    
2829    This is the working release 110.39.
2830    
2831    ----------------------------------------------------------------------
2832    Name: Matthias Blume
2833    Date: 2002/02/15 16:00:00 EST
2834    Tag: Release_110_39-orig
2835    Description:
2836    
2837    Working release 110.39.  New bootfiles.
2838    
2839    (Update: There was a small bug in the installer so it wouldn't work
2840    with all shells.  So I retagged. -Matthias)
2841    
2842    ----------------------------------------------------------------------
2843    Name: Matthias Blume
2844    Date: 2002/02/15 14:17:00 EST
2845    Tag: blume-20020215-showbindings
2846    Description:
2847    
2848    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
2849    the latter can be useful for exploring what bindings are available at
2850    the interactive prompt.  (The first function returns only the list
2851    of symbols that are really bound, the second prints those but also the
2852    ones that CM's autoloading mechanism knows about.)
2853    
2854    ----------------------------------------------------------------------
2855    Name: Matthias Blume
2856    Date: 2002/02/15 12:08:00 EST
2857    Tag: blume-20020215-iptrs
2858    Description:
2859    
2860    Two improvements to ml-nlffigen:
2861    
2862      1. Write files only if they do not exist or if their current contents
2863         do not coincide with what's being written.  (That is, avoid messing
2864         with the time stamps unless absolutely necessary.)
2865    
2866      2. Implement a "repository" mechanism for generated files related
2867         to "incomplete pointer types".   See the README file for details.
2868    
2869    ----------------------------------------------------------------------
2870    Name: Matthias Blume
2871    Date: 2002/02/14 11:50:00 EST
2872    Tag: blume-20020214-quote
2873    Description:
2874    
2875    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
2876    because of the new and improved tag generation scheme.  (Thanks to Allen
2877    Leung for pointing it out.)
2878    
2879    ----------------------------------------------------------------------
2880    Name: Lal George
2881    Date: 2002/02/14 09:55:27 EST 2002
2882    Tag: george-20020214-isabelle-bug
2883    Description:
2884    
2885    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
2886    of Isabelle on the x86.
2887    
2888    From Allen:
2889    -----------
2890     I've found the problem:
2891    
2892         in ra-core.sml, I use the counter "blocked" to keep track of the
2893         true number of elements in the freeze queue.  When the counter goes
2894         to zero, I skip examining the queue.  But I've messed up the
2895         bookkeeping in combine():
2896    
2897             else ();
2898             case !ucol of
2899               PSEUDO => (if !cntv > 0 then
2900                     (if !cntu > 0 then blocked := !blocked - 1 else ();
2901                                        ^^^^^^^^^^^^^^^^^^^^^^^
2902                      moveu := mergeMoveList(!movev, !moveu)
2903                     )
2904                  else ();
2905    
2906         combine() is called to coalesce two nodes u and v.
2907         I think I was thinking that if the move counts of u and v are both
2908         greater than zero then after they are coalesced then one node is
2909         removed from the freeze queue.  Apparently I was thinking that
2910         both u and v are of low degree, but that's clearly not necessarily true.
2911    
2912    
2913    02/12/2002:
2914        Here's the patch.  HOL now compiles.
2915    
2916        I don't know how this impact on performance (compile
2917        time or runtime).  This bug caused the RA (especially on the x86)
2918        to go thru the potential spill phase when there are still nodes on the
2919        freeze queue.
2920    
2921    
2922    
2923    
2924    ----------------------------------------------------------------------
2925    Name: Matthias Blume
2926    Date: 2002/02/13 22:40:00 EST
2927    Tag: blume-20020213-fptr-rtti
2928    Description:
2929    
2930    Fixed a bug in ml-nlffigen that was introduced with one of the previous
2931    updates.
2932    
2933    ----------------------------------------------------------------------
2934    Name: Matthias Blume
2935    Date: 2002/02/13 16:41:00 EST
2936    Tag: blume-20020213-cmlpq
2937    Description:
2938    
2939    Added new priority queue export symbols (which have just been added to
2940    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
2941    not compile and the installer would choke.)
2942    
2943    ----------------------------------------------------------------------
2944    Name: Matthias Blume
2945    Date: 2002/02/13 16:15:00 EST
2946    Tag: blume-20020213-various
2947    Description:
2948    
2949    1. More tweaks to ml-nlffigen:
2950    
2951       - better internal datastructures (resulting in slight speedup)
2952       - "-match" option requires exact match
2953       - "localized" gensym counters (untagged structs/unions nested within
2954         other structs/unions or within typedefs get a fresh counter; their
2955         tag will be prefixed by a concatenation of their parents' tags)
2956       - bug fixes (related to calculation of transitive closure of types
2957         to be included in the output)
2958    
2959    2. Minor Basis updates:
2960    
2961       - added implementations for List.collate and Option.app
2962    
2963    ----------------------------------------------------------------------
2964    Name: Matthias Blume
2965    Date: 2002/02/11 15:55:00 EST
2966    Tag: blume-20020211-gensym
2967    Description:
2968    
2969    Added a "-gensym" option to command line of ml-nlffigen.  This can be
2970    used to specify a "stem" -- a string that is inserted in all "gensym'd"
2971    names (ML structure names that correspond to unnamed C structs, unions,
2972    and enums), so that separate runs of ml-nlffigen do not clash.
2973    
2974    ----------------------------------------------------------------------
2975    Name: Matthias Blume
2976    Date: 2002/02/11 12:05:00 EST
2977    Tag: blume-20020211-gensml
2978    Description:
2979    
2980    A quick fix for a problem with GenSML (in the pgraph-util library):
2981    Make generation of toplevel "local" optional.  (Strictly speaking,
2982    signature definitions within "local" are not legal SML.)
2983    
2984    Other than that: updates to INSTALL and cm/TODO.
2985    
2986    ----------------------------------------------------------------------
2987    Name: Matthias Blume
2988    Date: 2002/02/08 15:00:00 EST
2989    Tag: blume-20020208-uniquepid
2990    Description:
2991    
2992    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
2993    
2994    1. The installer (config/install.sh) has gotten smarter:
2995    
2996         - Configuration options are a bit easier to specify now
2997           (in config/targets).
2998         - Bug in recognizing .tar.bz2 files fixed.
2999         - Installer automatically resolves dependencies between
3000           configuration options (e.g., if you ask for eXene, you will
3001           also get cml -- regardless whether you asked for it or not).
3002         - Installer can run in "quieter mode" by setting the environment
3003           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3004           "completely silent", though.
3005         - Build HashCons library as part of smlnj-lib.
3006    
3007    2. A new scheme for assigning persistent identifiers to compilation
3008       units (and, by extension, to types etc.) has been put into place.
3009       This fixes a long-standing bug where types and even dynamic values
3010       can get internally confused, thereby compromising type safety
3011       (abstraction) and dynamic correctness.  See
3012    
3013         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3014    
3015       for an example of how things could go wrong until now.
3016    
3017       The downside of the new scheme is that pids are not quite as
3018       persistent as they used to be: CM will generate a fresh pid
3019       for every compilation unit that it thinks it sees for the first
3020       time.  That means that if you compile starting from a clean, fresh
3021       source tree at two different times, you end up with different
3022       binaries.
3023    
3024       Cutoff recompilation, however, has not been compromised because
3025       CM keeps pid information in special caches between runs.
3026    
3027    ----------------------------------------------------------------------
3028    Name: Lal George
3029    Date: 2002/02/07 15:34:13 EST 2002
3030    Tag: <none>
3031    Description:
3032    
3033    Compilers that generate assembly code may produce  global labels
3034    whose value is resolved at link time. The various peephole optimization
3035    modules did not take this in account.
3036    
3037    TODO. The Labels.addrOf function should really return an option
3038    type so that clients are forced to deal with this issue, rather
3039    than an exception being raised.
3040    
3041    ----------------------------------------------------------------------
3042    Name: Lal George
3043    Date: 2002/02/06 13:55:02 EST
3044    Tag: george-20020206-ra-breakup
3045    Description:
3046    
3047    1. A bug fix from Allen.
3048    
3049        A typo causes extra fstp %st(0)'s to be generated at compensation
3050        edges, which might cause stack underflow traps at runtime.  This
3051        occurs in fft where there are extraneous fstps right before the 'into'
3052        trap instruction (in this case they are harmless since none of the
3053        integers overflow.)
3054    
3055    2. Pulled out various utility modules that were embedded in the modules
3056       of the register allocator. I need these modules for other purposes, but
3057       they are not complete enough to put into a library (just yet).
3058    ----------------------------------------------------------------------
3059    Name: Matthias Blume
3060    Date: 2002/01/31 16:05:00 EST
3061    Tag: blume-20020131-sparc-ccalls
3062    Description:
3063    
3064    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3065       of extra stack space by mistake.  Fixed.
3066    
3067    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3068    
3069    ----------------------------------------------------------------------
3070    Name: Allen Leung
3071    Date: 2002/01/30
3072    Tag: leunga-20020130-nowhere-bug-fix
3073    Description:
3074    
3075       MLRISC bug fixes:
3076       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3077       2. MachineInt.fromString was negating its value.
3078    
3079    ----------------------------------------------------------------------
3080    Name: Matthias Blume
3081    Date: 2002/01/29
3082    Tag: blume-20020129-INSTALL
3083    Description:
3084    
3085    - Added somewhat detailed installation instructions (file INSTALL).
3086    - Fixed curl-detection bug in config/install.sh.
3087    - It is now possible to select the URL getter using the URLGETTER
3088      environment variable:
3089    
3090          not set / "unknown"      --> automatic detection (script tries wget,
3091                                       curl, and lynx)
3092          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3093                                       how to properly invoke them)
3094          other                    --> use $URLGETTER directly, it must take
3095                                       precisely two command-line arguments
3096                                       (source URL and destination file name)
3097    
3098    ----------------------------------------------------------------------
3099    Name: Matthias Blume
3100    Date: 2002/01/28
3101    Tag: blume-20020128-sparc-ccalls
3102    Description:
3103    
3104    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3105    - Make use of the allocParam argument in sparc-c-calls.
3106    
3107    ----------------------------------------------------------------------
3108    Name: Matthias Blume
3109    Date: 2002/01/28
3110    Tag: blume-20020128-allocParam
3111    Description:
3112    
3113    John Reppy:  Changes c-calls API to accept client-callback for
3114    allocating extra stack space.
3115    me: Corresponding changes to mlriscGen (using a dummy argument that
3116        does not change the current behavior).
3117    
3118    ----------------------------------------------------------------------
3119    Name: Matthias Blume
3120    Date: 2002/01/28 12:00:00
3121    Tag: Release_110_38
3122    Description:
3123    
3124    This time for real!!!
3125    
3126    ----------------------------------------------------------------------
3127    Name: Matthias Blume
3128    Date: 2002/01/28 10:56:00 EST
3129    Tag: blume-20020128-retraction
3130    Description:
3131    
3132    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3133       with blume-Release_110_38-retracted.)
3134    
3135    1. Fixed a problem with incorrect rounding modes in real64.sml.
3136       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3137    
3138    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3139       and enums fixed.  The naming of corresponding ML identifiers should
3140       now be consistent again.
3141    
3142    ----------------------------------------------------------------------
3143    Name: Allen Leung
3144    Date: 2002/01/27
3145    Tag: leunga-20020127-nowhere
3146    Description:
3147    
3148       Added a target called nowhere in the configuration scripts.
3149       Enabling this will build the MLRISC 'nowhere' tool (for translating
3150       programs with where-clauses into legal SML code) during installation.
3151    
3152    ----------------------------------------------------------------------
3153    Name: Matthias Blume
3154    Date: 2002/01/25 21:27:00 EST
3155    Tag: blume-Release_110_38-retracted
3156    Description:
3157    
3158    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3159    
3160    README will be added later.
3161    
3162    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3163    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3164    
3165    ----------------------------------------------------------------------
3166    Name: Matthias Blume
3167    Date: 2002/01/25
3168    Tag: blume-20020125-ffi
3169    Description:
3170    
3171    A large number of tweaks and improvements to ml-nlffi-lib and
3172    ml-nlffigen:
3173    
3174       - ML represenation types have been streamlined
3175       - getter and setter functions work with concrete values, not abstract
3176         ones where possible
3177       - ml-nlffigen command line more flexible (see README file there)
3178       - some bugs have been fixed (hopefully)
3179    
3180    ----------------------------------------------------------------------
3181    Name: Lal George
3182    Date: 2002/01/24
3183    Tag: george-20020124-risc-ra-interface
3184    Description:
3185    
3186       There is a dramatic simplification in the interface to the
3187       register allocator for RISC architectures as a result of making
3188       parallel copy instructions explicit.
3189    
3190    ----------------------------------------------------------------------
3191    Name: Matthias Blume
3192    Date: 2002/01/22
3193    Tag: blume-20020122-x86-ccalls
3194    Description:
3195    
3196    Bug fix for c-calls on x86 (having to do with how char- and
3197    short-arguments are being handled).
3198    
3199    ----------------------------------------------------------------------
3200    Name: Matthias Blume
3201    Date: 2002/01/21
3202    Tag: blume-20020121-ff
3203    Description:
3204    
3205    Another day of fiddling with the FFI...
3206    
3207    1. Bug fix/workaround:  CKIT does not complain about negative array
3208       dimensions, so ml-nlffigen has to guard itself against this possibility.
3209       (Otherwise a negative dimension would send it into an infinite loop.)
3210    
3211    2. Some of the abstract types (light objects, light pointers, most "base"
3212       types) in structure C are now eqtypes.
3213    
3214    3. Added constructors and test functions for NULL function pointers.
3215    
3216    ----------------------------------------------------------------------
3217    Name: Matthias Blume
3218    Date: 2002/01/18
3219    Tag: blume-20020118-ready-for-new-release
3220    Description:
3221    
3222    Made config/srcarchiveurl point to a new place.  (Will provide boot
3223    files shortly.)
3224    
3225    Maybe we christen this to be 110.38?
3226    
3227    ----------------------------------------------------------------------
3228    Name: Matthias Blume
3229    Date: 2002/01/18
3230    Tag: blume-20020118-more-ffifiddle
3231    Description:
3232    
3233    Today's FFI fiddling:
3234    
3235      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3236        C.Set.* that use concrete (MLRep.*) arguments and results instead
3237        of abstract ones.
3238    
3239      - Provided word-style bit operations etc. for "int" representation
3240        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3241        and Long.
3242    
3243    ----------------------------------------------------------------------
3244    Name: Matthias Blume
3245    Date: 2002/01/18
3246    Tag: blume-20020118-use-x86-fp
3247    Description:
3248    
3249    Now that x86-fast-fp seems to be working, I turned it back on again
3250    by default.  (Seems to work fine now, even with the FFI.)
3251    
3252    Other than that, I added some documentation about the FFI to
3253    src/ml-nlffigen/README and updated the FFI test examples in
3254    src/ml-nlffi-lib/Tests/*.
3255    
3256    ----------------------------------------------------------------------
3257    Name: Allen Leung
3258    Date: 2002/01/17
3259    Tag: leunga-20020117-x86-fast-fp-call
3260    Description:
3261    
3262       1. Fixed a problem with handling return fp values when x86's fast fp
3263          mode is turned on.
3264    
3265       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3266          of %f32.
3267    
3268       3. Added a constructor INT32lit to the ast of MLRISC tools.
3269    
3270    ----------------------------------------------------------------------
3271    Name: Matthias Blume
3272    Date: 2002/01/16
3273    Tag: blume-20020116-ffifiddle
3274    Description:
3275    
3276    More fiddling with the FFI interface:
3277    
3278     - Make constness 'c instead of rw wherever possible.  This eliminates
3279       the need for certain explicit coercions.  (However, due to ML's
3280       value polymorphism, there will still be many cases where explicit
3281       coercions are necessary.  Phantom types are not the whole answer
3282       to modeling a subtyping relationship in ML.)
3283    
3284     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3285    
3286     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3287       clumsy workaround for varargs problem.)
3288    
3289    ----------------------------------------------------------------------
3290    Name: Lal George
3291    Date: 2002/01/15
3292    Tag: <none>
3293    Description:
3294    
3295    1. Since COPY instructions are no longer native to the architecture,
3296       a generic functor can be used to implement the expandCopies function.
3297    
3298    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3299       TEXT segment.
3300    
3301    ----------------------------------------------------------------------
3302    Name: Matthias Blume
3303    Date: 2002/01/15
3304    Tag: blume-20020115-ffiupdates
3305    Description:
3306    
3307    1. Fix for bug resulting in single-precision float values being returned
3308       incorrectly from FFI calls.
3309    
3310    2. Small modifications to C FFI API:
3311    
3312        - memory-allocation routines return straight objects (no options)
3313          and raise an exception in out-of-memory situations
3314        - unsafe extensions to cast between function pointers and pointers
3315          from/to ints
3316        - added structure C_Debug as an alternative to structure C where
3317          pointer-dereferencing (|*| and |*!) always check for null-pointers
3318        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3319          but also takes a (possibly empty) list of existing library handles
3320          that the current library depends on
3321    
3322    ----------------------------------------------------------------------
3323    Name: Matthias Blume
3324    Date: 2002/01/10
3325    Tag: blume-20020110-newffigen
3326    Description:
3327    
3328    1. Updates to portable graph code.
3329    
3330    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
3331       more scalable now so that even huge interfaces such as the one
3332       for GTK compile in finite time and space. :-)
3333       See src/ml-nlffigen/README for details on what's new.
3334    
3335    ----------------------------------------------------------------------
3336    Name: Lal George
3337    Date: 2001/01/09 14:31:35 EST 2002
3338    Tag: george-20011206-rm-native-copy
3339    Description:
3340    
3341            Removed the native COPY and FCOPY instructions
3342            from all the architectures and replaced it with the
3343            explicit COPY instruction from the previous commit.
3344    
3345            It is now possible to simplify many of the optimizations
3346            modules that manipulate copies. This has not been
3347            done in this change.
3348    
3349    ----------------------------------------------------------------------
3350    Name: Lal George
3351    Date: 2001/12/06 16:50:13 EST 2001
3352    Tag: george-20011206-mlrisc-instruction
3353    Description:
3354    
3355    Changed the representation of instructions from being fully abstract
3356    to being partially concrete. That is to say:
3357    
3358      from
3359            type instruction
3360    
3361      to
3362            type instr                              (* machine instruction *)
3363    
3364            datatype instruction =
3365                LIVE of {regs: C.cellset, spilled: C.cellset}
3366              | KILL of {regs: C.cellset, spilled: C.cellset}
3367              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
3368              | ANNOTATION of {i: instruction, a: Annotations.annotation}
3369              | INSTR of instr
3370    
3371    This makes the handling of certain special instructions that appear on
3372    all architectures easier and uniform.
3373    
3374    LIVE and KILL say that a list of registers are live or killed at the
3375    program point where they appear. No spill code is generated when an
3376    element of the 'regs' field is spilled, but the register is moved to
3377    the 'spilled' (which is present, more for debugging than anything else).
3378    
3379    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
3380    We used to generate:
3381    
3382            DEFFREG f1
3383            f1 := f2 + f3
3384            trapb
3385    
3386    but now generate:
3387    
3388            f1 := f2 + f3
3389            trapb
3390            LIVE {regs=[f1,f2,f3], spilled=[]}
3391    
3392    Furthermore, the DEFFREG (hack) required that all floating point instruction
3393    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
3394    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
3395    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
3396    this architecture).
3397    
3398    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
3399    all the architectures. This will result in further simplification of the
3400    register allocator that must be aware of them for coalescing purposes, and
3401    will also simplify certain aspects of the machine description that provides
3402    callbacks related to parallel copies.
3403    
3404    ANNOTATION should be obvious, and now INSTR represents the honest to God
3405    machine instruction set!
3406    
3407    The <arch>/instructions/<arch>Instr.sml files define certain utility
3408    functions for making porting easier -- essentially converting upper case
3409    to lower case. All machine instructions (of type instr) are in upper case,
3410    and the lower case form generates an MLRISC instruction. For example on
3411    the alpha we have:
3412    
3413      datatype instr =
3414         LDA of {r:cell, b:cell, d:operand}
3415       | ...
3416    
3417      val lda : {r:cell, b:cell, d:operand} -> instruction
3418        ...
3419    
3420    where lda is just (INSTR o LDA), etc.
3421    
3422    ----------------------------------------------------------------------
3423    Name: Matthias Blume
3424    Date: 2001/11/22 21:40:00 EST
3425    Tag: Release_110_37
3426    Description:
3427    
3428    Release 110.37.  This time for real.
3429    
3430    ----------------------------------------------------------------------
3431    Name: Matthias Blume
3432    Date: 2001/11/21 16:35:00 EST
3433    Tag: blume-20011121-foot-in-mouth
3434    Description:
3435    
3436    Removed the "Release_110_37" tag because of a serious bug.
3437    This will be re-tagged once the bug is fixed.
3438    
3439    ----------------------------------------------------------------------
3440    Name: Matthias Blume
3441    Date: 2001/11/21 16:14:00 EST
3442    Tag: blume-20011121-forgottenfile
3443    Description:
3444    
3445    Forgot to add a file.  (Just a .tex-file -- part of
3446    the CM manual source.)
3447    
3448    ----------------------------------------------------------------------
3449    Name: Matthias Blume
3450    Date: 2001/11/21 16:10:00 EST
3451    Tag: blume-20011121-invalid_110_37
3452    Description:
3453    
3454    Note: I removed the original tag "Release_110_37" from this commit
3455          because we found a serious bug in all non-x86 backends.
3456          - Matthias
3457    
3458    1. Modifications to the SML/NJ code generator and to the runtime system
3459       so that code object name strings are directly inserted into code
3460       objects at code generation time.  The only business the runtime system
3461       has with this is now to read the name strings on occasions.
3462       (The encoding of the name string has also changed somewhat.)
3463    
3464    2. CM now implements a simple "set calculus" for specifying export lists.
3465       In particular, it is now possible to refer to the export lists of
3466       other libraries/groups/sources and form unions as well as differences.
3467       See the latest CM manual for details.
3468    
3469    3. An separate notion of "proxy" libraries has again be eliminated from
3470       CM's model.  (Proxy libraries are now simply a special case of using
3471       the export list calculus.)
3472    
3473    4. Some of the existing libraries now take advantage of the new set
3474       calculus.
3475       (Notice that not all libraries have been converted because some
3476       of the existing .cm-files are supposed to be backward compatible
3477       with 110.0.x.)
3478    
3479    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
3480       "General.exnMessage"!  The former relies on a certain hook to be
3481       initialized, and that often does not happen in the stand-alone case.)
3482    
3483    ----------------------------------------------------------------------
3484  Name: Lal George  Name: Lal George
3485  Date: 2001/11/21  13:56:18 EST  Date: 2001/11/21  13:56:18 EST
3486  Tag: george-2001121-pseudo-ops  Tag: george-2001121-pseudo-ops
# Line 208  Line 3676 
3676    
3677  ----------------------------------------------------------------------  ----------------------------------------------------------------------
3678  Name: Matthias Blume  Name: Matthias Blume
 >>>>>>> 1.169  
3679  Date: 2001/09/18 15:35:00 EDT  Date: 2001/09/18 15:35:00 EDT
3680  Tag: blume-20010918-readme11036  Tag: blume-20010918-readme11036
3681  Description:  Description:

Legend:
Removed from v.984  
changed lines
  Added in v.1642

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