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/branches/primop-branch-3/NOTES/HISTORY
ViewVC logotype

Diff of /sml/branches/primop-branch-3/NOTES/HISTORY

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 759, Fri Dec 22 12:29:26 2000 UTC revision 890, Thu Jul 19 20:38:56 2001 UTC
# Line 11  Line 11 
11  Date:  Date:
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15    ----------------------------------------------------------------------
16    Name: Lal George
17    Date: 2001/07/19 16:36:29 EDT 2001
18    Tag: george-20010719-simple-cells
19    Description:
20    
21    I have dramatically simplified the interface for CELLS in MLRISC.
22    
23    In summary, the cells interface is broken up into three parts:
24    
25      1. CellsBasis : CELLS_BASIS
26    
27            CellsBasis is a top level structure and common for all
28            architectures.  it contains the definitions of basic datatypes
29            and utility  functions over these types.
30    
31      2. functor Cells() : CELLS
32    
33            Cells generates an interface for CELLS that incorporates the
34            specific resources on the target architecture, such as the
35            presence of special register classes, their number and size,
36            and various useful substructures.
37    
38      3. <ARCH>CELLS
39    
40            e.g. SparcCells: SPARCCELLS
41    
42            <ARCH>CELLS usually contains additional bindings for special
43            registers  on the architecture, such as:
44    
45                    val r0 : cell           (* register zero *)
46                    val y : cell            (* Y register *)
47                    val psr : cell          (* processor status register *)
48                    ...
49    
50            The structure returned by applying the Cells functor is opened
51            in this interface.
52    
53    The main implication of all this is that the datatypes for cells is
54    split between CellsBasis and CELLS -- a fairly simple change for user
55    code.
56    
57    In the old scheme the CELLS interface had a definitional binding of
58    the form:
59    
60            signature CELLS = sig
61    
62               structure CellsBasis = CellsBasis
63    
64               ...
65    
66            end
67    
68    With all the sharing constraints that goes on in MLRISC, this old
69    design  quickly leads to errors such as:
70    
71            "structure definition spec inside of sharing ... "
72    
73    
74    and appears to require an unacceptable amount of sharing and where
75    constraint hackery.
76    
77    I think this error message (the interaction of definitional specs and
78    sharing) requires more explanation on our web page.
79    
80    ----------------------------------------------------------------------
81    Name: Matthias Blume
82    Date: 2001/07/19 15:00:00 EDT
83    Tag: blume-20010719-libreorg
84    Description:
85    
86    This update puts together a fairly extensive but straightforward change
87    to the way the libraries that implement the interactive system are
88    organized:
89    
90       The biggest change is the elimination of structure Compiler.  As a
91       replacement for this structure, there is now a CM library
92       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
93       that exports all the substructures of the original structure Compiler
94       directly.  So instead of saying Compiler.Foo.bar one now simply
95       says Foo.bar.  (The CM libraries actually export a collection of
96       structures that is richer than the collection of substructures of
97       structure Compiler.)
98    
99       To make the transition smooth, there is a separate library called
100       $smlnj/compiler/compiler.cm which puts together and exports the
101       original structure Compiler (or at least something very close to it).
102    
103       There are five members of the original structure Compiler
104       that are not exported directly but which instead became members
105       of a new structure Backend (described by signature BACKEND).  These are:
106       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
107       Interact (: INTERACT), structure Machine (: MACHINE), and val
108       architecture (: string).
109    
110       Structure Compiler.Version has become structure CompilerVersion.
111    
112       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
113       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
114       or x86, respectively.
115       Each of these exports the same frontend structures that
116       $smlnj/compiler.cm exports.  But they do not have a structure Backend
117       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
118       Hppa, PPC, Sparc, or X86, respectively.
119    
120       Library $smlnj/compiler/all.cm exports the union of the exports of
121       $smlnj/compiler/<arch>.cm
122    
123       There are no structures <Arch>Compiler anymore, use
124       $smlnj/compiler/<arch>.cm instead.
125    
126       Library host-compiler-0.cm is gone.  Instead, the internal library
127       that instantiates CM is now called cm0.cm.  Selection of the host
128       compiler (backend) is no longer done here but. (Responsibility for it
129       now lies with $smlnj/compiler/current.cm.  This seems to be more
130       logical.)
131    
132       Many individual files have been moved or renamed.  Some files have
133       been split into multiple files, and some "dead" files have been deleted.
134    
135    Aside from these changes to library organization, there are also changes
136    to the way the code itself is organized:
137    
138       Structure Binfile has been re-implemented in such a way that it no
139       longer needs any knowledge of the compiler.  It exclusively deals
140       with the details of binfile layout.  It no longer invokes the
141       compiler (for the purpose of creating new prospective binfile
142       content), and it no longer has any knowledge of how to interpret
143       pickles.
144    
145       Structure Compile (: COMPILE) has been stripped down to the bare
146       essentials of compilation.  It no longer deals with linking/execution.
147       The interface has been cleaned up considerably.
148    
149       Utility routines for dealing with linking and execution have been
150       moved into their own substructures.
151    
152       (The ultimate goal of these changes is to provide a light-weight
153       binfile loader/linker (at least for, e.g., stable libraries) that
154       does not require CM or the compiler to be present.)
155    
156    CM documentation has been updated to reflect the changes to library
157    organization.
158    
159    ----------------------------------------------------------------------
160    Name: Matthias Blume
161    Date: 2001/07/10 17:30:00 EDT
162    Tag: Release_110_34
163    Description:
164    
165    Minor tweak to 110.34 (re-tagged):
166    
167      - README.html file added to CVS repository
168      - runtime compiles properly under FreeBSD 3.X and 4.X
169    
170    ----------------------------------------------------------------------
171    Name: Matthias Blume
172    Date: 2001/07/10 17:30:00 EDT
173    Tag: Release_110_34
174    Description:
175    
176    New version number (110.34). New bootfiles.
177    
178    ----------------------------------------------------------------------
179    Name: Matthias Blume
180    Date: 2001/07/09 16:00:00 EDT
181    Tag: blume-20010709-more-varargs
182    Description:
183    
184    I changed the handling of varargs in ml-nlffigen again:
185    The ellipsis ... will now simply be ignored (with an accompanying warning).
186    
187    The immediate effect is that you can actually call a varargs function
188    from ML -- but you can't actually supply any arguments beyond the ones
189    specified explicitly.  (For example, you can call printf with its format
190    string, but you cannot pass additional arguments.)
191    
192    This behavior is only marginally more useful than the one before, but
193    it has the advantage that a function or, more importantly, a function
194    type never gets dropped on the floor, thus avoiding follow-up problems with
195    other types that refer to the offending one.
196    
197    ----------------------------------------------------------------------
198    Name: Matthias Blume
199    Date: 2001/07/09 11:25:00 EDT
200    Tag: blume-20010709-varargs
201    Description:
202    
203    1. ckit-lib.cm now exports structure Error
204    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
205       with a warning accompanied by a source location.  Moreover, it
206       merely skips the offending function or type and proceeds with the
207       rest of its work.u  As a result, one can safely feed C code containing
208       "..." to ml-nlffigen.
209    3. There are some internal improvements to CM, providing slightly
210       more general string substitutions in the tools subsystem.
211    
212    ----------------------------------------------------------------------
213    Name: Matthias Blume
214    Date: 2001/06/27 15:10:00 EDT
215    Tag: blume-20010627-concur
216    Description:
217    
218    Fixed a small bug in CM's handling of parallel compilation.
219    (You could observe the bug by Control-C-interrupting an ordinary
220    CMB.make or CM.stabilize and then attaching some compile servers.
221    The result was that all of a sudden the previously interrupted
222    compilation would continue on its own.  This was because of
223    an over-optimization: CM did not bother to clean out certain queues
224    when no servers were attached "anyway", resulting in the contents
225    of these queues to grab control when new servers did get attached.)
226    
227    There is also another minor update to the CM manual.
228    
229    ----------------------------------------------------------------------
230    Name: Matthias Blume
231    Date: 2001/06/26 16:15:00 EDT
232    Tag: blume-20010626-cmdoc
233    Description:
234    
235    Minor typo fixed in CM manual (syntax diagram for libraries).
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume
239    Date: 2001/06/25 22:55:00 EDT
240    Tag: blume-20010625-x86pc
241    Description:
242    
243    Fixed a nasty bug in the X86 assembly code that caused signal
244    handlers to fail (crash) randomly.
245    
246    ----------------------------------------------------------------------
247    Name: Matthias Blume
248    Date: 2001/06/25 12:05:00 EDT
249    Tag: blume-20010625-nlffigen
250    Description:
251    
252    This update fixes a number of minor bugs in ml-nlffigen as reported by
253    Nick Carter <nbc@andrew.cmu.edu>.
254    
255      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
256      2. Default names for generated files are now derived from the name of
257         the C file *without its directory*.  In particular, this causes generated
258         files to be placed locally even if the C file is in some system directory.
259      3. Default names for generated signatures and structures are also derived
260         from the C file name without its directory.  This avoids silly things
261         like "structure GL/GL".
262         (Other silly names are still possible because ml-nlffigen does not do
263          a thorough check of whether generated names are legal ML identifiers.
264          When in doubt, use command line arguments to force particular names.)
265    
266    ----------------------------------------------------------------------
267    Name: Matthias Blume
268    Date: 2001/06/21 12:25:00 EDT
269    Tag: blume-20010621-eXene
270    Description:
271    
272    eXene now compiles and (sort of) works again.
273    
274    The library name (for version > 110.33) is $/eXene.cm.
275    
276    I also added an new example in src/eXene/examples/nbody.  See the
277    README file there for details.
278    
279    ----------------------------------------------------------------------
280    Name: Matthias Blume
281    Date: 2001/06/20 16:40:00 EDT
282    Tag: blume-20010620-cml
283    Description:
284    
285    CML now compiles and works again.
286    
287    Libraries (for version > 110.33):
288    
289      $cml/cml.cm            Main CML library.
290      $cml/basis.cm          CML's version of $/basis.cm.
291      $cml/cml-internal.cm   Internal helper library.
292      $cml/core-cml.cm       Internal helper library.
293      $cml-lib/trace-cml.cm  Tracing facility.
294      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
295    
296    The installer (config/install.sh) has been taught how to properly
297    install this stuff.
298    
299    ----------------------------------------------------------------------
300    Name: Matthias Blume
301    Date: 2001/06/19 17:55:00 EDT
302    Tag: blume-20010619-instantiate
303    Description:
304    
305    This un-breaks the fix for bug 1432.
306    (The bug was originally fixed in 110.9 but I broke it again some
307    time after that.)
308    
309    ----------------------------------------------------------------------
310    Name: Matthias Blume
311    Date: 2001/06/19 17:25:00 EDT
312    Tag: blume-20010619-signals
313    Description:
314    
315    This should (hopefully) fix the long-standing signal handling bug.
316    (The runtime system was constructing a continuation record with an
317    incorrect descriptor which would cause the GC to drop data on the floor...)
318    
319    ----------------------------------------------------------------------
320    Name: Matthias Blume
321    Date: 2001/06/15 15:05:00 EDT
322    Tag: blume-20010615-moresparc
323    Description:
324    
325    Here is a short late-hour update related to Sparc c-calls:
326    
327     -- made handling of double-word arguments a bit smarter
328    
329     -- instruction selection phase tries to collapse certain clumsily
330        constructed ML-Trees; typical example:
331    
332            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
333    
334        This currently has no further impact on SML/NJ since mlriscGen does
335        not seem to generate such patterns in the first place, and c-calls
336        (which did generate them in the beginning) has meanwhile been fixed
337        so as to avoid them as well.
338    
339    ----------------------------------------------------------------------
340    Name: Matthias Blume
341    Date: 2001/06/15 15:05:00 EDT
342    Tag: blume-20010615-sparc
343    Description:
344    
345    The purpose of this update is to provide an implementation of NLFFI
346    on Sparc machines.
347    
348    Here are the changes in detail:
349    
350       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
351       the Sparc implementation of the c-calls API.
352       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
353       for accessing the ML frame.  Thus, we have a real frame pointer and
354       can freely modify %sp without need for an omit-frame-ptr phase.
355       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
356       this case.
357       * ml-nlffigen has been taught to produce code for different architectures
358       and calling conventions.
359       * In a way similar to what was done in the x86 case, the Sparc
360       backend uses its own specific extension to mltree.  (For example,
361       it needs to be able to generate UNIMP instructions which are part
362       of the calling convention.)
363       * ml-nlffi-lib was reorganized to make it more modular (in particular,
364       to make it easier to plug in new machine- and os-dependent parts).
365    
366    There are some other fairly unrelated bug fixes and cleanups as well:
367    
368       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
369       that they properly share their libraries with existing SML/NJ libraries.
370       * I fixed a minor cosmetic bug in CM, supressing certain spurious
371       follow-up error messages.
372       * Updates to CM/CMB documentation.
373    
374    TODO items:
375    
376       * MLRISC should use a different register as its asmTemp on the Sparc.
377         (The current %o2 is a really bad choice because it is part of the
378         calling conventions, so things might interfere in unexpected ways.)
379    
380    ----------------------------------------------------------------------
381    Name: Matthias Blume
382    Date: 2001/06/07
383    Tag: blume-20010607-calls
384    Description:
385    
386    A number of internal changes related to C calls and calling conventions:
387    
388    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
389       number of bytes popped implicitly (by the callee).  In most cases
390       this field is 0 but on x86/win32 it is some non-zero value.  This
391       is information provided for the benefit of the "omit-frameptr" pass.
392    2. The CALL instruction on the x86 carries a similar "pops" field.
393       The instruction selection phase copies its value from the ML-Tree
394       CALL statement.
395    3. On all other architectures, the instruction selection phase checks
396       whether "pops=0" and complains if not.
397    4. The c-calls implementation for x86 now accepts two calling conventions:
398       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
399       up after the call and pops is set to 0.  For "stdcall", the caller
400       does nothing, leaving the cleanup to the callee; pops is set to
401       the number of bytes that were pushed onto the stack.
402    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
403       distinguish between "ccall" and "stdcall".
404    6. The UNIMP instruction has been added to the supported Sparc instruction
405       set. (This is needed for implementing the official C calling convention
406       on this architecture.)
407    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
408       work with the latest CM.
409    
410    ----------------------------------------------------------------------
411    Name: Matthias Blume
412    Date: 2001/06/05 15:10:00 EDT
413    Tag: blume-20010605-cm-index
414    Description:
415    
416    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
417    
418    1. CM can now generate "index files".  These are human-readable files
419       that list on a per-.cm-file basis each toplevel symbol defined or
420       imported.  The location of the index file for
421       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
422       To enable index-file generation, set CM.Control.generate_index to true
423       or export an environment-symbol: export CM_GENERATE_INDEX=true.
424    
425       The CM manual has been updated accordingly.
426    
427    2. I made some slight modifications to the c-calls API in MLRISC.
428    
429         a) There is now a callback to support saving/restoring of
430            dedicated but caller-save registers around the actual call
431            instruction.
432         b) One can optionally specify a comment-annotation for the
433            call instruction.
434    
435    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
436       (For example, the comment annotation shows the C prototype of
437        the function being called.)
438    
439    ----------------------------------------------------------------------
440    Name: Matthias Blume
441    Date: 2001/06/01 13:30:00 EDT
442    Tag: blume-20010601-nlffi-cleanup
443    Description:
444    
445    This is mostly a cleanup of MLFFI stuff:
446    
447       - some signature files have been put into a more exposed place
448       - the ugly 'f type parameter is gone (simplifies types tremendously!)
449       - ml-nlffigen changed accordingly
450       - tutorial updated
451    
452    Other changes:
453    
454       - author's affiliation in CM manual(s) updated
455       - some more recognized keywords added to Allen's sml.sty
456    
457    ----------------------------------------------------------------------
458    Name: Matthias Blume
459    Date: 2001/05/25 15:30:00 EDT
460    Tag: blume-20010525-iptr
461    Description:
462    
463      - put the official 110.33-README (as it appears on the ftp server) under
464        CVS
465      - fixed a small bug related to incomplete pointer types in
466        ml-nlffigen
467      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
468        (it does not need the 'f type parameter)
469    
470    ----------------------------------------------------------------------
471    Name: Matthias Blume
472    Date: 2001/05/23 14:30:00 EDT
473    Tag: Release_110_33
474    Description:
475    
476    New version number (110.33).  New bootfiles.
477    
478    ----------------------------------------------------------------------
479    Name: Matthias Blume
480    Date: 2001/05/22 18:06:00 EDT
481    Tag: blume-20010522-targets
482    Description:
483    
484    Made install.sh use file config/targets.customized if it exists, falling
485    back to config/targets if it doesn't.  This way one can have a customized
486    version of the targets file without touching the "real thing", thus
487    eliminating the constant fear of accidentally checking something bogus
488    back into the CVS repository...  (File config/targets.customized must
489    not be added to the repository!)
490    
491    ----------------------------------------------------------------------
492    Name: Matthias Blume
493    Date: 2001/05/22 16:30:00 EDT
494    Tag: blume-20010522-minitut
495    Description:
496    
497    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
498       struct returns.
499    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
500       incomplete, preliminary documentation for NLFFI.
501    
502    ----------------------------------------------------------------------
503    Name: Matthias Blume
504    Date: 2001/05/14 11:30:00 EDT
505    Tag: blume-20010514-script
506    Description:
507    
508    Some bugs in install script fixed.
509    
510    In addition to that I also made a slight change to the NLFFI API:
511    Functors generated by ml-nlffigen now take the dynamic library as a
512    straight functor argument, not as a suspended one.  (The original
513    functor code used to force the suspension right away anyway, so there
514    was nothing gained by this complication of the interface.)
515    
516    ----------------------------------------------------------------------
517    Name: Matthias Blume
518    Date: 2001/05/11 14:35:00 EDT
519    Tag: blume-20010511-ml-nlffi
520    Description:
521    
522    I finally took the plunge and added my new FFI code to the main
523    repository.  For x86-linux it is now ready for prime-time.
524    
525    There are two new subdirectories of "src":
526    
527      - ml-nlffi-lib:
528           The utility library for programs using the FFI interface.
529           Here is the implementation of $/c.cm and its associated low-level
530           partners $/c-int.cm and $/memory.cm.
531      - ml-nlffigen:
532           A stand-alone program for generating ML glue code from C source
533           code.
534    
535    Building ml-nlffigen requires $/ckit-lib.cm.
536    
537    The config/install.sh script has been updates to do the Right Thing
538    (hopefully).
539    
540    Notice that the source tree for the C-Kit will not be put under "src"
541    but directly under the installation root directory.  (This is the
542    structure that currently exists on the CVS server when you check out
543    module "sml".)  Fortunately, config/install.sh knows about this oddity.
544    
545    Bugs: No documentation yet.
546    
547    ----------------------------------------------------------------------
548    Name: Matthias Blume
549    Date: 2001/05/09 16:35:00 EDT
550    Tag: blume-20010509-cpscontract
551    Description:
552    
553    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
554    wrapper/unwrapper elimination did not decrement usage counts and some
555    dead variables got overlooked by the dead-up logic.)
556    
557    ----------------------------------------------------------------------
558    Name: Lal George
559    Date: 2001/05/08  17:26:09 EDT
560    Tag: george-20010508-omit-frameptr
561    Description:
562    
563    Changes to implement the omit-frame-pointer optimization to support
564    raw C calls. For now, there is only support on the Intel x86, but
565    other architectures will follow as more experience is gained with this.
566    
567    
568    ----------------------------------------------------------------------
569    Name: Matthias Blume
570    Date: 2001/05/07 14:40:00 EDT
571    Tag: blume-20010507-proxies
572    Description:
573    
574    I made into "proxy libraries" all libraries that qualify for such a
575    change.  (A qualifying library is a library that has another library or
576    groups as its sole member and repeats that member's export list
577    verbatim.  A proxy library avoids this repetition by omitting its export
578    list, effectively inheriting the list that its (only) member exports.
579    See the CM manual for more explanation.)
580    The main effect is that explicit export lists for these libraries
581    do not have to be kepts in sync, making maintenance a bit easier.
582    
583    I also added copyright notices to many .cm-files.
584    
585    Last but not least, I made a new set of bootfiles.
586    
587    ----------------------------------------------------------------------
588    Name: Matthias Blume
589    Date: 2001/05/04 17:00:00 EDT
590    Tag: blume-20010504-cm-lsplit
591    Description:
592    
593    0. John merged pending changes to $/smlnj-lib.cm
594    
595    1. Allen's previous change accidentally backed out of one of Lal's
596       earlier changes.  I undid this mistake (re-introducing Lal's change).
597    
598    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
599       within the compiler where applicable.  There is some code simplification
600       because of that.
601    
602    3. The "split" phase (in FLINT) is now part of the default list of phases.
603       Compiler.Control.LambdaSplitting.* can be used to globally control the
604       lambda-splitting (cross-module-inlining) engine.  In addition to that,
605       it can now also be controlled on a per-source basis: CM has been taught
606       a new tool parameter applicable to ML source files.
607    
608       - To turn lambda-splitting off completely:
609            local open Compiler.Control.LambdaSplitting in
610                val _ = set Off
611            end
612       - To make "no lambda-splitting" the global default (but allow per-source
613         overriding); this is the initial setting:
614            local open Compiler.Control.LambdaSplitting in
615                val _ = set (Default NONE)
616            end
617       - To make "lambda-splitting with aggressiveness a" the global default
618         (and allow per-source overriding):
619            local open Compiler.Control.LambdaSplitting in
620                val _ = set (Default (SOME a))
621            end
622    
623       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
624         write (in the respective .cm-file):
625            a.sml (lambdasplitting:off)
626       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
627            a.sml (lambdasplitting:on)
628       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
629         <a> is a decimal non-negative integer):
630            a.sml (lambdasplitting:<a>)
631       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
632            a.sml (lambdasplitting:infinity)
633       - To use the global default for a.sml:
634            a.sml (lambdasplitting:default)
635         or simply
636            a.sml
637    
638    ----------------------------------------------------------------------
639    Name: Allen Leung
640    Date: 2001/05/04 01:57:00 EDT
641    Tag: leunga-20010504-sync
642    Description:
643    
644      MLRISC features.
645    
646      1. Fix to CMPXCHG instructions.
647      2. Changed RA interface to allow annotations in callbacks.
648      3. Added a new method to the stream interface to allow annotations updates.
649    
650    ----------------------------------------------------------------------
651    Name: Matthias Blume
652    Date: 2001/05/01 11:45:00 EDT
653    Tag: blume-20010501-pcedittmp
654    Description:
655    
656    Changed install.sh to use the current working directory instead of
657    /usr/tmp for a temporary file (pcedittmp).  The previous choice
658    of /usr/tmp caused trouble with MacOS X because of file premission
659    problems.
660    
661    ----------------------------------------------------------------------
662    Name: Matthias Blume
663    Date: 2001/04/20 11:10:00 EDT
664    Tag: blume-20010420-inMLflag
665    Description:
666    
667     - added vp_limitPtrMask to vproc-state.h
668       (for use by the raw-C-calls mechanism to implement proper interrupt
669        handling)
670     - made the ML compiler aware of various data-structure offsets so it
671       can generate code for accessing the vp_inML flag and vp_limitPtrMask
672     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
673       raw C-calls
674    
675    ----------------------------------------------------------------------
676    Name: Lal George
677    Date: 2001/04/20 09:15:28 EDT
678    Tag: george-20010420-macosX
679    Description:
680    
681     - Changes to port to Mac OS X; Darwin.
682    
683     - In the process I found that sqrt was broken on the PPC, because the
684       fsqrt instruction is not implemented.
685    
686    ----------------------------------------------------------------------
687    Name: Matthias Blume
688    Date: 2001/04/18 12:45:00 EDT
689    Tag: blume-20010418-ccalls
690    Description:
691    
692     - fixed two off-by-4 errors in the x86-specific c-calls implementation
693       (this bug prevented structure arguments containing pointers from being
694        passed correctly)
695     - changed the raw-C-call code in mlriscGen.sml in such a way that
696       structure arguments are represented as a pointer to the beginning
697       of the structure (instead of having a series of synthesized arguments,
698       one for each structure member)
699    
700     - made makeml script's verbosity level configurable via environment
701       variable (MAKEML_VERBOSITY)
702    
703     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
704       in rawmem-x86.sml; we are now using the real thing
705    
706    ----------------------------------------------------------------------
707    Name: Matthias Blume
708    Date: 2001/03/22 16:25:00 EST
709    Tag: blume-20010322-bootfiles
710    Description:
711    
712    Created a new set of bootfiles (for your automatic installation convenience).
713    
714    ----------------------------------------------------------------------
715    Name: Matthias Blume
716    Date: 2001/03/22 15:10:00 EST
717    Tag: blume-20010322-rawmem-parcm
718    Description:
719    
720    1. All "raw memory access" primitives for the new FFI are implemented now
721       (at least on the x86).
722    2. Some further cleanup of CM's parallel make mechanism.
723    
724    ----------------------------------------------------------------------
725    Name: Matthias Blume
726    Date: 2001/03/19 17:53:00 EST
727    Tag: blume-20010319-parallel
728    Description:
729    
730    Parallel make (using compile servers) now works again.
731    
732    To this end, CM.stabilize and CMB.make have been modified to work in
733    two passes when compile servers are attached:
734       1. Compile everything, do not perform stabilization; this pass
735          uses compile servers
736       2. Stabilize everything; this pass does not use compile servers
737    If there are no compile servers, the two passes are combined into one
738    (as before).  Splitting the passes increases the inherent parallelism
739    in the dependency graph because the entire graph including all
740    libraries is available at the same time.  This, in turn, improves
741    server utilization.  The downside is that the master process will
742    have to do some extra work after compilation is done (because for
743    technical reasons it must re-read all the binfiles during stabilization).
744    
745    ----------------------------------------------------------------------
746    Name: Matthias Blume
747    Date: 2001/03/16 12:22:00 EST
748    Tag: blume-20010316-bootfiles
749    Description:
750    
751    Created a new set of bootfiles (for your automatic installation convenience).
752    
753    ----------------------------------------------------------------------
754    Name: Matthias Blume
755    Date: 2001/03/16 11:00:00 EST
756    Tag: blume-20010316-MLTREE-fixup
757    Description:
758    
759    This is a minor fixup for an (untagged) earlier commit by Allen.
760    (A file was missing).
761    
762    ----------------------------------------------------------------------
763    Name: Allen Leung
764    Date: Mon Mar  5 18:54:57 EST 2001
765    Tag: leunga-20010305-cut-support
766    
767    1. New support for alternative control-flow in MLTREE.
768       Currently we support
769    
770          FLOW_TO(CALL ...., [k1,...,kn])
771    
772       This is needed for 'cuts to' in C-- and try/handle-like constructs
773       in Moby
774    
775       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
776    
777    2. Register Allocator
778    
779       Changes in interface [from Fermin, John]
780    
781    3. Alpha 8-bit SLL support [Fermin]
782    
783    4. All architectures
784    
785       A new module (ClusterExpandCopies) for expanding parallel copies.
786    
787    ----------------------------------------------------------------------
788    Name: Allen Leung
789    Date: 2001/02/27 23:07:00 EST
790    Tag: leunga-20010227-minor-stuff
791    
792    1. Alpha bug fix for CMOVNE
793    2. Handle mltree COND(..,FCMP ...,...)
794    3. Bug fix in simplifier
795    
796    ----------------------------------------------------------------------
797    Name: Matthias Blume
798    Date: 2001/01/30 17:50:00 EST
799    Tag: blume-20010130-sync
800    Description:
801    
802    This is just a minor update to sync my devel branch with the main brach.
803    The only visible change is the addition of some README files.
804    
805    ----------------------------------------------------------------------
806    Name: Matthias Blume
807    Date: 2001/01/12 23:30:00 JST
808    Tag: blume-20010112-bootfiles
809    Description:
810    
811    Made a new set of bootfiles that goes with the current state of the
812    repository.
813    
814    ----------------------------------------------------------------------
815    Name: Matthias Blume
816    Date: 2001/01/12 21:20:00 JST
817    Tag: blume-20010112-sync
818    Description:
819    
820    I am just flushing out some minor changes that had accumulated in
821    my private branch in order to sync with the main tree.  (This is
822    mainly because I had CVS trouble when trying to merge _into_ my
823    private branch.)
824    
825    Most people should be completely unaffected by this.
826    
827    ----------------------------------------------------------------------
828    Name: Allen Leung
829    Date: Thu Jan 11 21:03:00 EST 2001
830    Tag: leunga-20010111-labexp=mltree
831    Description:
832    
833    1.  Removed the type LabelExp and replace it by MLTree.
834    2.  Rewritten mltree-simplify with the pattern matcher tool.
835    3.  There were some bugs in alpha code generator which would break
836        64-bit code generation.
837    4.  Redo the tools to generate code with the
838    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
839        are now generated by perl scripts.
840    
841    ----------------------------------------------------------------------
842    Name: Matthias Blume
843    Date: 2001/01/10 21:55:00 JST
844    Tag: blume-20010110-rcc
845    Description:
846    
847    The RCC stuff now seems to work (but only on the x86).
848    This required hacking of the c-calls interface (and -implementation) in
849    MLRISC.
850    
851    Normal compiler users should be unaffected.
852    
853    ----------------------------------------------------------------------
854    Name: Matthias Blume
855    Date: 2001/01/09 01:20:00 JST
856    Tag: blume-20010109-rcc
857    Description:
858    
859    This is a fairly big patch, flushing out a large number of pending
860    changes that I made to my development copy over the last couple of days.
861    
862    Of practical relevance at this moment is a workaround for a pickling
863    bug that Allen ran into the other day.  The cause of the bug itself is
864    still unknown and it might be hard to fix it properly, but the
865    workaround has some merits of its own (namely somewhat reducing pickling
866    overhead for certain libraries).  Therefore, I think this solution should
867    be satisfactory at this time.
868    
869    The rest of the changes (i.e., the vast majority) has to do with my
870    ongoing efforts of providing direct support for C function calls from
871    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
872    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
873    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
874    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
875    handle RCC.
876    
877    The last part (the changes to mlriscGen.sml) are still known to be
878    wrong on the x86 and not implemented on all other architectures.  But
879    the infrastructure is in place. I had to change a few functor
880    signatures in the backend to be able to route the CCalls interface
881    from MLRISC there, and I had to specialize the mltree type (on the
882    x86) to include the necessary extensions. (The extensions themselves
883    were already there and redy to go in MLRISC/x86).
884    
885    Everything should be very happy as soon as someone helps me with
886    mlriscGen.sml...
887    
888    In any case, nothing of this should matter to anyone as long as the
889    new primop is not being used (which is going to be the case unless you
890    find it where I hid it :). The rest of the compiler is completely
891    unaffected.
892    
893    ----------------------------------------------------------------------
894    Name: Matthias Blume
895    Date: 2001/01/05 00:30:00 JST
896    Tag: blume-20010105-primops
897    Description:
898    
899    Added some experimental support for work that I am doing right now.
900    These changes mostly concern added primops, but there is also a new
901    experimental C library in the runtime system (but currently not enabled
902    anywhere except on Linux/X86).
903    
904    In the course of adding primops (and playing with them), I discovered that
905    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
906    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
907    To recover, I made type information in INL_PRIM mandatory and changed
908    prim.sml as well as built-in.sml accordingly.  The InLine structure now
909    has complete, correct type information (i.e., no bottom types).
910    
911    Since all these changes mean that we need new binfiles, I also bumped the
912    version number to 110.32.1.
913    
914    ----------------------------------------------------------------------
915    Name: Matthias Blume
916    Date: 2000/12/30 22:10:00 JST
917    Tag: blume-20001230-various
918    Description:
919    
920    Added proxy libraries for MLRISC and let MLRISC libraries refer
921    to each other using path anchors.  (See CM manual for explanation.)
922    
923    Updated CM documentation.
924    
925    Fixed some bugs in CM.
926    
927    Implemented "proxy" libraries (= syntactic sugar for CM).
928    
929    Added "-quiet" option to makeml and changed runtime system accordingly.
930    
931    Added cleanup handler for exportML to reset timers and compiler stats.
932    
933    ----------------------------------------------------------------------
934    Name: Lal George
935    Date: 2000/12/22 22:22:58 EST 2000
936    Tag: Release_110_32
937    Description:
938    
939            Infinite precision used throughout MLRISC.
940            see MLRISC/mltree/machine-int.sig
941    
942    ----------------------------------------------------------------------
943    Name: Matthias Blume
944    Date: 2000/12/22 23:16:00 JST
945    Tag: blume-20001222-warn
946    Description:
947    
948    Corrected wording and formatting of some CM warning message which I
949    broke in my previous patch.
950    
951  ----------------------------------------------------------------------  ----------------------------------------------------------------------
952  Name: Matthias Blume  Name: Matthias Blume
953  Date: 2000/12/22 21:20:00 JST  Date: 2000/12/22 21:20:00 JST
954  Tag: blume-20001222-anchorenv  Tag: blume-20001222-anchorenv
955  Description:  Description:
956    
957  Fixed a CM problem related to anchor environments and their interaction  Fixed CM's handling of anchor environments in connection with CMB.make.
 with CMB.make.  
958    
959  ----------------------------------------------------------------------  ----------------------------------------------------------------------
960  Name: Matthias Blume  Name: Matthias Blume

Legend:
Removed from v.759  
changed lines
  Added in v.890

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