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 800, Fri Mar 16 17:22:47 2001 UTC revision 961, Tue Oct 16 15:32:39 2001 UTC
# Line 14  Line 14 
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16  Name: Matthias Blume  Name: Matthias Blume
17    Date: 2001/10/16 11:32:00 EDT
18    Tag: blume-20011016-netbsd
19    Description:
20    
21    Underscore patch from Chris Richards (fixing problem with compiling
22    runtime system under recent NetBSD).
23    
24    ----------------------------------------------------------------------
25    Name: Allen Leung
26    Date: 2001/10/12 17:18:32 EDT 2001
27    Tag: leung-20011012-x86-printflowgraph
28    Description:
29    
30    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
31    
32    ----------------------------------------------------------------------
33    Name: Lal George
34    Date: 2001/10/11 23:51:34 EDT
35    Tag: george-20011011-too-many-instrs
36    Description:
37    
38    The representation of a program point never expected to see more
39    than 65536 instructions in a basic block!
40    
41    ----------------------------------------------------------------------
42    Name: Lal George
43    Date: 2001/10/09 09:41:37 EDT
44    Tag: george-20011008-mlrisc-labels
45    Description:
46    
47    Changed the machine description files to support printing of
48    local and global labels in assembly code, based on host assembler
49    conventions.
50    
51    ----------------------------------------------------------------------
52    Name: Matthias Blume
53    Date: 2001/09/25 15:25:00 EDT
54    Tag: blume-20010925-exninfo
55    Description:
56    
57    I provided a non-hook implementation of exnName (at the toplevel) and
58    made the "dummy" implementation of exnMessage (at the toplevel) more
59    useful: if nothing gets "hooked in", then at least you are going to
60    see the exception name and a message indicating why you don't see more.
61    
62    [For the time being, programs that need exnMessage and want to use
63    ml-build should either use General.exnMessage (strongly recommended) or
64    refer to structure General at some other point so that CM sees a
65    static dependency.]
66    
67    [Similar remarks go for "print" and "use":  If you want to use their
68    functionality in stand-alone programs generated by ml-build, then use
69    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
70    
71    ----------------------------------------------------------------------
72    Name: Matthias Blume
73    Date: 2001/09/20 17:28:00 EDT
74    Tag: blume-20010920-slowfp
75    Description:
76    
77    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
78    
79    ----------------------------------------------------------------------
80    Name: Matthias Blume
81    Date: 2001/09/20 17:20:00 EDT
82    Tag: blume-20010920-canonicalpaths
83    Description:
84    
85    0. Updated the BOOT file (something that I forgot to do earlier).
86    
87    1. Small internal change to CM so that it avoids "/../" in filenames
88       as much as possible (but only where it is safe).
89    
90    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
91       that arguments that contain delimiters are passed through correctly.
92       This change also means that all "special" arguments of the form
93       @SMLxxx... must come first.
94    
95    3. Changed install script to put relative anchor names for tool commands
96       into pathconfig.
97    
98    ----------------------------------------------------------------------
99    Name: Matthias Blume
100    >>>>>>> 1.169
101    Date: 2001/09/18 15:35:00 EDT
102    Tag: blume-20010918-readme11036
103    Description:
104    
105    Added README files.
106    
107    ----------------------------------------------------------------------
108    Name: Matthias Blume
109    Date: 2001/09/18 11:45:00 EDT
110    Tag: Release_110_36 (retag)
111    Description:
112    
113    Fixed mistake in config/preloads. Retagged as 110.36.
114    
115    ----------------------------------------------------------------------
116    Name: Matthias Blume
117    Date: 2001/09/18 09:40:00 EDT
118    Tag: Release_110_36_orig (tag changed)
119    Description:
120    
121    New version (110.36).  New bootfiles.
122    
123    ----------------------------------------------------------------------
124    Name: Matthias Blume
125    Date: 2001/09/14 16:15:00 EDT
126    Tag: blume-20010914-x86fastfp
127    Description:
128    
129    John committed some changes that Allen made, in particular a (hopefully)
130    correctly working version of the x86-fp module.
131    
132    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
133    flag to "true".  Everything seems to compile to a fixpoint ok, and
134    "mandelbrot" speeds up by about 15%.
135    
136    ----------------------------------------------------------------------
137    Name: Matthias Blume
138    Date: 2001/09/13 11:20:00 EDT
139    Tag: blume-20010913-minimal
140    Description:
141    
142    1. Stefan Monnier's patch to fix a miscompilation problem that
143       was brought to light by John Reppy's work on Moby.
144    
145    2. Implemented a minimal "structure Compiler" that contains just
146       "version" and "architecture".  The minimal version will be
147       available when the full version is not.  This is for backward-
148       compatibility with code that wants to test Compiler.version.
149    
150    ----------------------------------------------------------------------
151    Name: Matthias Blume
152    Date: 2001/08/28 14:03:00 EDT
153    Tag: blume-20010828-ml-lex
154    Description:
155    
156    Fix for bug 1581, received from Neophytos Michael.
157    
158    ----------------------------------------------------------------------
159    Name: Matthias Blume
160    Date: 2001/08/27 11:20:00 EDT
161    Tag: blume-20010827-readme11035
162    Description:
163    
164    Fleshed out the README file for 110.35.
165    
166    ----------------------------------------------------------------------
167    Name: Matthias Blume
168    Date: 2001/08/24 17:10:00 EDT
169    Tag: Release_110_35
170    Description:
171    
172    New version number (110.35).  New bootfiles.
173    
174    ----------------------------------------------------------------------
175    Name: Lal George
176    Date: 2001/08/24 13:47:18 EDT 2001
177    Tag: george-20010824-MLRISC-graphs
178    Description:
179    
180     removed clusters from MLRISC completely and replaced with graphs.
181    
182    ----------------------------------------------------------------------
183    Name: Matthias Blume
184    Date: 2001/08/23 17:50:00 EDT
185    Tag: blume-20010823-toplevel
186    Description:
187    
188    - some reorganization of the code that implements various kinds of
189      environments in the compiler (static, dynamic, symbolic, combined)
190    - re-implemented the EnvRef module so that evalStream works properly
191      (if the stream contains references to "use", "CM.make", etc.)
192    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
193    
194    ----------------------------------------------------------------------
195    Name: Matthias Blume
196    Date: 2001/08/20 15:50 EDT
197    Tag: blume20010820-slipup
198    Description:
199    
200    I forgot to commit a few files.  Here they are...
201    
202    ----------------------------------------------------------------------
203    Name: Matthias Blume
204    Date: 2001/08/20 15:35:00 EDT
205    Tag: blume-20010820-debugprof
206    Description:
207    
208    !!!! NEW BOOTFILES !!!!
209    
210    This is another round of reorganizing the compiler sources.  This
211    time the main goal was to factor out all the "instrumentation"
212    passes (for profiling and backtracing) into their own library.
213    The difficulty was to do it in such a way that it does not depend
214    on elaborate.cm but only on elabdata.cm.
215    
216    Therefore there have been further changes to both elaborate.cm and
217    elabdata.cm -- more "generic" things have been moved from the former
218    to the latter.  As a result, I was forced to split the assignment
219    of numbers indicating "primtyc"s into two portions: SML-generic and
220    SML/NJ-specific.  Since it would have been awkward to maintain,
221    I bit the bullet and actually _changed_ the mapping between these
222    numbers and primtycs.  The bottom line of this is that you need
223    a new set of bin- and bootfiles.
224    
225    I have built new bootfiles for all architectures, so doing a fresh
226    checkout and config/install.sh should be all you need.
227    
228    The newly created library's name is
229    
230        $smlnj/viscomp/debugprof.cm
231    
232    and its sources live under
233    
234        src/compiler/DebugProf
235    
236    ----------------------------------------------------------------------
237    Name: Matthias Blume
238    Date: 2001/08/15 17:15:00 EDT
239    Tag: blume-20010815-compreorg
240    Description:
241    
242    This is a first cut at reorganizing the CM libraries that make up the
243    core of the compiler.  The idea is to separate out pieces that could
244    be used independently by tools, e.g., the parser, the typechecker, etc.
245    
246    The current status is a step in this direction, but it is not quite
247    satisfactory yet.  Expect more changes in the future.
248    
249    Here is the current (new) organization...
250    
251        What used to be $smlnj/viscomp/core.cm is now divided into
252        six CM libraries:
253    
254             $smlnj/viscomp/basics.cm
255                           /parser.cm
256                           /elabdata.cm
257                           /elaborate.cm
258                           /execute.cm
259                           /core.cm
260    
261        The CM files for these libraries live under src/system/smlnj/viscomp.
262        All these libraries are proxy libraries that contain precisely
263        one CM library component.  Here are the locations of the components
264        (all within the src/compiler tree):
265    
266             Basics/basics.cm
267             Parse/parser.cm
268             ElabData/elabdata.cm
269             Elaborator/elaborate.cm
270             Execution/execute.cm
271             core.cm
272    
273         [This organization is the same that has been used already
274         for a while for the architecture-specific parts of the visible
275         compiler and for the old version of core.cm.]
276    
277         As you will notice, many source files have been moved from their
278         respective original locations to a new home in one of the above
279         subtrees.
280    
281         The division of labor between the new libraries is the following:
282    
283             basics.cm:
284                - Simple, basic definitions that pertain to many (or all) of
285                  the other libraries.
286             parser.cm:
287                - The SML parser, producing output of type Ast.dec.
288                - The type family for Ast is also defined and exported here.
289             elabdata.cm:
290                - The datatypes that describe input and output of the elaborator.
291                  This includes types, absyn, and static environments.
292             elaborator.cm:
293                - The SML/NJ type checker and elaborator.
294                  This maps an Ast.dec (with a given static environment) to
295                  an Absyn.dec (with a new static environment).
296                - This libraries implements certain modules that used to be
297                  structures as functors (to remove dependencies on FLINT).
298             execute.cm:
299                - Everything having to do with executing binary code objects.
300                - Dynamic environments.
301             core.cm:
302                - SML/NJ-specific instantiations of the elaborator and MLRISC.
303                - Top-level modules.
304                - FLINT (this should eventually become its own library)
305    
306    Notes:
307    
308    I am not 100% happy with the way I separated the elaborator (and its
309    data structures) from FLINT.  Two instances of the same problem:
310    
311        1. Data structures contain certain fields that carry FLINT-specific
312           information.  I hacked around this using exn and the property list
313           module from smlnj-lib.  But the fact that there are middle-end
314           specific fields around at all is a bit annoying.
315    
316        2. The elaborator calculates certain FLINT-related information.  I tried
317           to make this as abstract as I could using functorization, but, again,
318           the fact that the elaborator has to perform calculations on behalf
319           of the middle-end at all is not nice.
320    
321        3. Having to used exn and property lists is unfortunate because it
322           weakens type checking.  The other alternative (parameterizing
323           nearly *everything*) is not appealing, though.
324    
325    I removed the "rebinding =" warning hack because due to the new organization
326    it was awkward to maintain it.  As a result, the compiler now issues some of
327    these warnings when compiling init.cmi during bootstrap compilation. On
328    the plus side, you also get a warning when you do, for example:
329       val op = = Int32.+
330    which was not the case up to now.
331    
332    I placed "assign" and "deref" into the _Core structure so that the
333    code that deals with the "lazy" keyword can find them there.  This
334    removes the need for having access to the primitive environment
335    during elaboration.
336    
337    ----------------------------------------------------------------------
338    Name: Matthias Blume
339    Date: 2001/08/13
340    Tag: blume-20010813-closures
341    Description:
342    
343    This fix was sent to us by Zhong Shao.  It is supposed to improve the
344    performance of certain loops by avoiding needless closure allocation.
345    
346    ----------------------------------------------------------------------
347    Name: Lal George
348    Date: 2001/07/31 10:03:23 EDT 2001
349    Tag: george-20010731-x86-fmalloc
350    Description: Fixed bug in x86 calls
351    
352        There was a bug where call instructions would mysteriously
353        vanish. The call instruction had to be one that returned
354        a floating point value.
355    
356    ----------------------------------------------------------------------
357    Name: Lal George
358    Date: 2001/07/19 16:36:29 EDT 2001
359    Tag: george-20010719-simple-cells
360    Description:
361    
362    I have dramatically simplified the interface for CELLS in MLRISC.
363    
364    In summary, the cells interface is broken up into three parts:
365    
366      1. CellsBasis : CELLS_BASIS
367    
368            CellsBasis is a top level structure and common for all
369            architectures.  it contains the definitions of basic datatypes
370            and utility  functions over these types.
371    
372      2. functor Cells() : CELLS
373    
374            Cells generates an interface for CELLS that incorporates the
375            specific resources on the target architecture, such as the
376            presence of special register classes, their number and size,
377            and various useful substructures.
378    
379      3. <ARCH>CELLS
380    
381            e.g. SparcCells: SPARCCELLS
382    
383            <ARCH>CELLS usually contains additional bindings for special
384            registers  on the architecture, such as:
385    
386                    val r0 : cell           (* register zero *)
387                    val y : cell            (* Y register *)
388                    val psr : cell          (* processor status register *)
389                    ...
390    
391            The structure returned by applying the Cells functor is opened
392            in this interface.
393    
394    The main implication of all this is that the datatypes for cells is
395    split between CellsBasis and CELLS -- a fairly simple change for user
396    code.
397    
398    In the old scheme the CELLS interface had a definitional binding of
399    the form:
400    
401            signature CELLS = sig
402    
403               structure CellsBasis = CellsBasis
404    
405               ...
406    
407            end
408    
409    With all the sharing constraints that goes on in MLRISC, this old
410    design  quickly leads to errors such as:
411    
412            "structure definition spec inside of sharing ... "
413    
414    
415    and appears to require an unacceptable amount of sharing and where
416    constraint hackery.
417    
418    I think this error message (the interaction of definitional specs and
419    sharing) requires more explanation on our web page.
420    
421    ----------------------------------------------------------------------
422    Name: Matthias Blume
423    Date: 2001/07/19 15:00:00 EDT
424    Tag: blume-20010719-libreorg
425    Description:
426    
427    This update puts together a fairly extensive but straightforward change
428    to the way the libraries that implement the interactive system are
429    organized:
430    
431       The biggest change is the elimination of structure Compiler.  As a
432       replacement for this structure, there is now a CM library
433       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
434       that exports all the substructures of the original structure Compiler
435       directly.  So instead of saying Compiler.Foo.bar one now simply
436       says Foo.bar.  (The CM libraries actually export a collection of
437       structures that is richer than the collection of substructures of
438       structure Compiler.)
439    
440       To make the transition smooth, there is a separate library called
441       $smlnj/compiler/compiler.cm which puts together and exports the
442       original structure Compiler (or at least something very close to it).
443    
444       There are five members of the original structure Compiler
445       that are not exported directly but which instead became members
446       of a new structure Backend (described by signature BACKEND).  These are:
447       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
448       Interact (: INTERACT), structure Machine (: MACHINE), and val
449       architecture (: string).
450    
451       Structure Compiler.Version has become structure CompilerVersion.
452    
453       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
454       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
455       or x86, respectively.
456       Each of these exports the same frontend structures that
457       $smlnj/compiler.cm exports.  But they do not have a structure Backend
458       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
459       Hppa, PPC, Sparc, or X86, respectively.
460    
461       Library $smlnj/compiler/all.cm exports the union of the exports of
462       $smlnj/compiler/<arch>.cm
463    
464       There are no structures <Arch>Compiler anymore, use
465       $smlnj/compiler/<arch>.cm instead.
466    
467       Library host-compiler-0.cm is gone.  Instead, the internal library
468       that instantiates CM is now called cm0.cm.  Selection of the host
469       compiler (backend) is no longer done here but. (Responsibility for it
470       now lies with $smlnj/compiler/current.cm.  This seems to be more
471       logical.)
472    
473       Many individual files have been moved or renamed.  Some files have
474       been split into multiple files, and some "dead" files have been deleted.
475    
476    Aside from these changes to library organization, there are also changes
477    to the way the code itself is organized:
478    
479       Structure Binfile has been re-implemented in such a way that it no
480       longer needs any knowledge of the compiler.  It exclusively deals
481       with the details of binfile layout.  It no longer invokes the
482       compiler (for the purpose of creating new prospective binfile
483       content), and it no longer has any knowledge of how to interpret
484       pickles.
485    
486       Structure Compile (: COMPILE) has been stripped down to the bare
487       essentials of compilation.  It no longer deals with linking/execution.
488       The interface has been cleaned up considerably.
489    
490       Utility routines for dealing with linking and execution have been
491       moved into their own substructures.
492    
493       (The ultimate goal of these changes is to provide a light-weight
494       binfile loader/linker (at least for, e.g., stable libraries) that
495       does not require CM or the compiler to be present.)
496    
497    CM documentation has been updated to reflect the changes to library
498    organization.
499    
500    ----------------------------------------------------------------------
501    Name: Matthias Blume
502    Date: 2001/07/10 17:30:00 EDT
503    Tag: Release_110_34
504    Description:
505    
506    Minor tweak to 110.34 (re-tagged):
507    
508      - README.html file added to CVS repository
509      - runtime compiles properly under FreeBSD 3.X and 4.X
510    
511    ----------------------------------------------------------------------
512    Name: Matthias Blume
513    Date: 2001/07/10 17:30:00 EDT
514    Tag: Release_110_34
515    Description:
516    
517    New version number (110.34). New bootfiles.
518    
519    ----------------------------------------------------------------------
520    Name: Matthias Blume
521    Date: 2001/07/09 16:00:00 EDT
522    Tag: blume-20010709-more-varargs
523    Description:
524    
525    I changed the handling of varargs in ml-nlffigen again:
526    The ellipsis ... will now simply be ignored (with an accompanying warning).
527    
528    The immediate effect is that you can actually call a varargs function
529    from ML -- but you can't actually supply any arguments beyond the ones
530    specified explicitly.  (For example, you can call printf with its format
531    string, but you cannot pass additional arguments.)
532    
533    This behavior is only marginally more useful than the one before, but
534    it has the advantage that a function or, more importantly, a function
535    type never gets dropped on the floor, thus avoiding follow-up problems with
536    other types that refer to the offending one.
537    
538    ----------------------------------------------------------------------
539    Name: Matthias Blume
540    Date: 2001/07/09 11:25:00 EDT
541    Tag: blume-20010709-varargs
542    Description:
543    
544    1. ckit-lib.cm now exports structure Error
545    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
546       with a warning accompanied by a source location.  Moreover, it
547       merely skips the offending function or type and proceeds with the
548       rest of its work.u  As a result, one can safely feed C code containing
549       "..." to ml-nlffigen.
550    3. There are some internal improvements to CM, providing slightly
551       more general string substitutions in the tools subsystem.
552    
553    ----------------------------------------------------------------------
554    Name: Matthias Blume
555    Date: 2001/06/27 15:10:00 EDT
556    Tag: blume-20010627-concur
557    Description:
558    
559    Fixed a small bug in CM's handling of parallel compilation.
560    (You could observe the bug by Control-C-interrupting an ordinary
561    CMB.make or CM.stabilize and then attaching some compile servers.
562    The result was that all of a sudden the previously interrupted
563    compilation would continue on its own.  This was because of
564    an over-optimization: CM did not bother to clean out certain queues
565    when no servers were attached "anyway", resulting in the contents
566    of these queues to grab control when new servers did get attached.)
567    
568    There is also another minor update to the CM manual.
569    
570    ----------------------------------------------------------------------
571    Name: Matthias Blume
572    Date: 2001/06/26 16:15:00 EDT
573    Tag: blume-20010626-cmdoc
574    Description:
575    
576    Minor typo fixed in CM manual (syntax diagram for libraries).
577    
578    ----------------------------------------------------------------------
579    Name: Matthias Blume
580    Date: 2001/06/25 22:55:00 EDT
581    Tag: blume-20010625-x86pc
582    Description:
583    
584    Fixed a nasty bug in the X86 assembly code that caused signal
585    handlers to fail (crash) randomly.
586    
587    ----------------------------------------------------------------------
588    Name: Matthias Blume
589    Date: 2001/06/25 12:05:00 EDT
590    Tag: blume-20010625-nlffigen
591    Description:
592    
593    This update fixes a number of minor bugs in ml-nlffigen as reported by
594    Nick Carter <nbc@andrew.cmu.edu>.
595    
596      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
597      2. Default names for generated files are now derived from the name of
598         the C file *without its directory*.  In particular, this causes generated
599         files to be placed locally even if the C file is in some system directory.
600      3. Default names for generated signatures and structures are also derived
601         from the C file name without its directory.  This avoids silly things
602         like "structure GL/GL".
603         (Other silly names are still possible because ml-nlffigen does not do
604          a thorough check of whether generated names are legal ML identifiers.
605          When in doubt, use command line arguments to force particular names.)
606    
607    ----------------------------------------------------------------------
608    Name: Matthias Blume
609    Date: 2001/06/21 12:25:00 EDT
610    Tag: blume-20010621-eXene
611    Description:
612    
613    eXene now compiles and (sort of) works again.
614    
615    The library name (for version > 110.33) is $/eXene.cm.
616    
617    I also added an new example in src/eXene/examples/nbody.  See the
618    README file there for details.
619    
620    ----------------------------------------------------------------------
621    Name: Matthias Blume
622    Date: 2001/06/20 16:40:00 EDT
623    Tag: blume-20010620-cml
624    Description:
625    
626    CML now compiles and works again.
627    
628    Libraries (for version > 110.33):
629    
630      $cml/cml.cm            Main CML library.
631      $cml/basis.cm          CML's version of $/basis.cm.
632      $cml/cml-internal.cm   Internal helper library.
633      $cml/core-cml.cm       Internal helper library.
634      $cml-lib/trace-cml.cm  Tracing facility.
635      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
636    
637    The installer (config/install.sh) has been taught how to properly
638    install this stuff.
639    
640    ----------------------------------------------------------------------
641    Name: Matthias Blume
642    Date: 2001/06/19 17:55:00 EDT
643    Tag: blume-20010619-instantiate
644    Description:
645    
646    This un-breaks the fix for bug 1432.
647    (The bug was originally fixed in 110.9 but I broke it again some
648    time after that.)
649    
650    ----------------------------------------------------------------------
651    Name: Matthias Blume
652    Date: 2001/06/19 17:25:00 EDT
653    Tag: blume-20010619-signals
654    Description:
655    
656    This should (hopefully) fix the long-standing signal handling bug.
657    (The runtime system was constructing a continuation record with an
658    incorrect descriptor which would cause the GC to drop data on the floor...)
659    
660    ----------------------------------------------------------------------
661    Name: Matthias Blume
662    Date: 2001/06/15 15:05:00 EDT
663    Tag: blume-20010615-moresparc
664    Description:
665    
666    Here is a short late-hour update related to Sparc c-calls:
667    
668     -- made handling of double-word arguments a bit smarter
669    
670     -- instruction selection phase tries to collapse certain clumsily
671        constructed ML-Trees; typical example:
672    
673            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
674    
675        This currently has no further impact on SML/NJ since mlriscGen does
676        not seem to generate such patterns in the first place, and c-calls
677        (which did generate them in the beginning) has meanwhile been fixed
678        so as to avoid them as well.
679    
680    ----------------------------------------------------------------------
681    Name: Matthias Blume
682    Date: 2001/06/15 15:05:00 EDT
683    Tag: blume-20010615-sparc
684    Description:
685    
686    The purpose of this update is to provide an implementation of NLFFI
687    on Sparc machines.
688    
689    Here are the changes in detail:
690    
691       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
692       the Sparc implementation of the c-calls API.
693       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
694       for accessing the ML frame.  Thus, we have a real frame pointer and
695       can freely modify %sp without need for an omit-frame-ptr phase.
696       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
697       this case.
698       * ml-nlffigen has been taught to produce code for different architectures
699       and calling conventions.
700       * In a way similar to what was done in the x86 case, the Sparc
701       backend uses its own specific extension to mltree.  (For example,
702       it needs to be able to generate UNIMP instructions which are part
703       of the calling convention.)
704       * ml-nlffi-lib was reorganized to make it more modular (in particular,
705       to make it easier to plug in new machine- and os-dependent parts).
706    
707    There are some other fairly unrelated bug fixes and cleanups as well:
708    
709       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
710       that they properly share their libraries with existing SML/NJ libraries.
711       * I fixed a minor cosmetic bug in CM, supressing certain spurious
712       follow-up error messages.
713       * Updates to CM/CMB documentation.
714    
715    TODO items:
716    
717       * MLRISC should use a different register as its asmTemp on the Sparc.
718         (The current %o2 is a really bad choice because it is part of the
719         calling conventions, so things might interfere in unexpected ways.)
720    
721    ----------------------------------------------------------------------
722    Name: Matthias Blume
723    Date: 2001/06/07
724    Tag: blume-20010607-calls
725    Description:
726    
727    A number of internal changes related to C calls and calling conventions:
728    
729    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
730       number of bytes popped implicitly (by the callee).  In most cases
731       this field is 0 but on x86/win32 it is some non-zero value.  This
732       is information provided for the benefit of the "omit-frameptr" pass.
733    2. The CALL instruction on the x86 carries a similar "pops" field.
734       The instruction selection phase copies its value from the ML-Tree
735       CALL statement.
736    3. On all other architectures, the instruction selection phase checks
737       whether "pops=0" and complains if not.
738    4. The c-calls implementation for x86 now accepts two calling conventions:
739       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
740       up after the call and pops is set to 0.  For "stdcall", the caller
741       does nothing, leaving the cleanup to the callee; pops is set to
742       the number of bytes that were pushed onto the stack.
743    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
744       distinguish between "ccall" and "stdcall".
745    6. The UNIMP instruction has been added to the supported Sparc instruction
746       set. (This is needed for implementing the official C calling convention
747       on this architecture.)
748    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
749       work with the latest CM.
750    
751    ----------------------------------------------------------------------
752    Name: Matthias Blume
753    Date: 2001/06/05 15:10:00 EDT
754    Tag: blume-20010605-cm-index
755    Description:
756    
757    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
758    
759    1. CM can now generate "index files".  These are human-readable files
760       that list on a per-.cm-file basis each toplevel symbol defined or
761       imported.  The location of the index file for
762       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
763       To enable index-file generation, set CM.Control.generate_index to true
764       or export an environment-symbol: export CM_GENERATE_INDEX=true.
765    
766       The CM manual has been updated accordingly.
767    
768    2. I made some slight modifications to the c-calls API in MLRISC.
769    
770         a) There is now a callback to support saving/restoring of
771            dedicated but caller-save registers around the actual call
772            instruction.
773         b) One can optionally specify a comment-annotation for the
774            call instruction.
775    
776    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
777       (For example, the comment annotation shows the C prototype of
778        the function being called.)
779    
780    ----------------------------------------------------------------------
781    Name: Matthias Blume
782    Date: 2001/06/01 13:30:00 EDT
783    Tag: blume-20010601-nlffi-cleanup
784    Description:
785    
786    This is mostly a cleanup of MLFFI stuff:
787    
788       - some signature files have been put into a more exposed place
789       - the ugly 'f type parameter is gone (simplifies types tremendously!)
790       - ml-nlffigen changed accordingly
791       - tutorial updated
792    
793    Other changes:
794    
795       - author's affiliation in CM manual(s) updated
796       - some more recognized keywords added to Allen's sml.sty
797    
798    ----------------------------------------------------------------------
799    Name: Matthias Blume
800    Date: 2001/05/25 15:30:00 EDT
801    Tag: blume-20010525-iptr
802    Description:
803    
804      - put the official 110.33-README (as it appears on the ftp server) under
805        CVS
806      - fixed a small bug related to incomplete pointer types in
807        ml-nlffigen
808      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
809        (it does not need the 'f type parameter)
810    
811    ----------------------------------------------------------------------
812    Name: Matthias Blume
813    Date: 2001/05/23 14:30:00 EDT
814    Tag: Release_110_33
815    Description:
816    
817    New version number (110.33).  New bootfiles.
818    
819    ----------------------------------------------------------------------
820    Name: Matthias Blume
821    Date: 2001/05/22 18:06:00 EDT
822    Tag: blume-20010522-targets
823    Description:
824    
825    Made install.sh use file config/targets.customized if it exists, falling
826    back to config/targets if it doesn't.  This way one can have a customized
827    version of the targets file without touching the "real thing", thus
828    eliminating the constant fear of accidentally checking something bogus
829    back into the CVS repository...  (File config/targets.customized must
830    not be added to the repository!)
831    
832    ----------------------------------------------------------------------
833    Name: Matthias Blume
834    Date: 2001/05/22 16:30:00 EDT
835    Tag: blume-20010522-minitut
836    Description:
837    
838    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
839       struct returns.
840    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
841       incomplete, preliminary documentation for NLFFI.
842    
843    ----------------------------------------------------------------------
844    Name: Matthias Blume
845    Date: 2001/05/14 11:30:00 EDT
846    Tag: blume-20010514-script
847    Description:
848    
849    Some bugs in install script fixed.
850    
851    In addition to that I also made a slight change to the NLFFI API:
852    Functors generated by ml-nlffigen now take the dynamic library as a
853    straight functor argument, not as a suspended one.  (The original
854    functor code used to force the suspension right away anyway, so there
855    was nothing gained by this complication of the interface.)
856    
857    ----------------------------------------------------------------------
858    Name: Matthias Blume
859    Date: 2001/05/11 14:35:00 EDT
860    Tag: blume-20010511-ml-nlffi
861    Description:
862    
863    I finally took the plunge and added my new FFI code to the main
864    repository.  For x86-linux it is now ready for prime-time.
865    
866    There are two new subdirectories of "src":
867    
868      - ml-nlffi-lib:
869           The utility library for programs using the FFI interface.
870           Here is the implementation of $/c.cm and its associated low-level
871           partners $/c-int.cm and $/memory.cm.
872      - ml-nlffigen:
873           A stand-alone program for generating ML glue code from C source
874           code.
875    
876    Building ml-nlffigen requires $/ckit-lib.cm.
877    
878    The config/install.sh script has been updates to do the Right Thing
879    (hopefully).
880    
881    Notice that the source tree for the C-Kit will not be put under "src"
882    but directly under the installation root directory.  (This is the
883    structure that currently exists on the CVS server when you check out
884    module "sml".)  Fortunately, config/install.sh knows about this oddity.
885    
886    Bugs: No documentation yet.
887    
888    ----------------------------------------------------------------------
889    Name: Matthias Blume
890    Date: 2001/05/09 16:35:00 EDT
891    Tag: blume-20010509-cpscontract
892    Description:
893    
894    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
895    wrapper/unwrapper elimination did not decrement usage counts and some
896    dead variables got overlooked by the dead-up logic.)
897    
898    ----------------------------------------------------------------------
899    Name: Lal George
900    Date: 2001/05/08  17:26:09 EDT
901    Tag: george-20010508-omit-frameptr
902    Description:
903    
904    Changes to implement the omit-frame-pointer optimization to support
905    raw C calls. For now, there is only support on the Intel x86, but
906    other architectures will follow as more experience is gained with this.
907    
908    
909    ----------------------------------------------------------------------
910    Name: Matthias Blume
911    Date: 2001/05/07 14:40:00 EDT
912    Tag: blume-20010507-proxies
913    Description:
914    
915    I made into "proxy libraries" all libraries that qualify for such a
916    change.  (A qualifying library is a library that has another library or
917    groups as its sole member and repeats that member's export list
918    verbatim.  A proxy library avoids this repetition by omitting its export
919    list, effectively inheriting the list that its (only) member exports.
920    See the CM manual for more explanation.)
921    The main effect is that explicit export lists for these libraries
922    do not have to be kepts in sync, making maintenance a bit easier.
923    
924    I also added copyright notices to many .cm-files.
925    
926    Last but not least, I made a new set of bootfiles.
927    
928    ----------------------------------------------------------------------
929    Name: Matthias Blume
930    Date: 2001/05/04 17:00:00 EDT
931    Tag: blume-20010504-cm-lsplit
932    Description:
933    
934    0. John merged pending changes to $/smlnj-lib.cm
935    
936    1. Allen's previous change accidentally backed out of one of Lal's
937       earlier changes.  I undid this mistake (re-introducing Lal's change).
938    
939    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
940       within the compiler where applicable.  There is some code simplification
941       because of that.
942    
943    3. The "split" phase (in FLINT) is now part of the default list of phases.
944       Compiler.Control.LambdaSplitting.* can be used to globally control the
945       lambda-splitting (cross-module-inlining) engine.  In addition to that,
946       it can now also be controlled on a per-source basis: CM has been taught
947       a new tool parameter applicable to ML source files.
948    
949       - To turn lambda-splitting off completely:
950            local open Compiler.Control.LambdaSplitting in
951                val _ = set Off
952            end
953       - To make "no lambda-splitting" the global default (but allow per-source
954         overriding); this is the initial setting:
955            local open Compiler.Control.LambdaSplitting in
956                val _ = set (Default NONE)
957            end
958       - To make "lambda-splitting with aggressiveness a" the global default
959         (and allow per-source overriding):
960            local open Compiler.Control.LambdaSplitting in
961                val _ = set (Default (SOME a))
962            end
963    
964       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
965         write (in the respective .cm-file):
966            a.sml (lambdasplitting:off)
967       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
968            a.sml (lambdasplitting:on)
969       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
970         <a> is a decimal non-negative integer):
971            a.sml (lambdasplitting:<a>)
972       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
973            a.sml (lambdasplitting:infinity)
974       - To use the global default for a.sml:
975            a.sml (lambdasplitting:default)
976         or simply
977            a.sml
978    
979    ----------------------------------------------------------------------
980    Name: Allen Leung
981    Date: 2001/05/04 01:57:00 EDT
982    Tag: leunga-20010504-sync
983    Description:
984    
985      MLRISC features.
986    
987      1. Fix to CMPXCHG instructions.
988      2. Changed RA interface to allow annotations in callbacks.
989      3. Added a new method to the stream interface to allow annotations updates.
990    
991    ----------------------------------------------------------------------
992    Name: Matthias Blume
993    Date: 2001/05/01 11:45:00 EDT
994    Tag: blume-20010501-pcedittmp
995    Description:
996    
997    Changed install.sh to use the current working directory instead of
998    /usr/tmp for a temporary file (pcedittmp).  The previous choice
999    of /usr/tmp caused trouble with MacOS X because of file premission
1000    problems.
1001    
1002    ----------------------------------------------------------------------
1003    Name: Matthias Blume
1004    Date: 2001/04/20 11:10:00 EDT
1005    Tag: blume-20010420-inMLflag
1006    Description:
1007    
1008     - added vp_limitPtrMask to vproc-state.h
1009       (for use by the raw-C-calls mechanism to implement proper interrupt
1010        handling)
1011     - made the ML compiler aware of various data-structure offsets so it
1012       can generate code for accessing the vp_inML flag and vp_limitPtrMask
1013     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1014       raw C-calls
1015    
1016    ----------------------------------------------------------------------
1017    Name: Lal George
1018    Date: 2001/04/20 09:15:28 EDT
1019    Tag: george-20010420-macosX
1020    Description:
1021    
1022     - Changes to port to Mac OS X; Darwin.
1023    
1024     - In the process I found that sqrt was broken on the PPC, because the
1025       fsqrt instruction is not implemented.
1026    
1027    ----------------------------------------------------------------------
1028    Name: Matthias Blume
1029    Date: 2001/04/18 12:45:00 EDT
1030    Tag: blume-20010418-ccalls
1031    Description:
1032    
1033     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1034       (this bug prevented structure arguments containing pointers from being
1035        passed correctly)
1036     - changed the raw-C-call code in mlriscGen.sml in such a way that
1037       structure arguments are represented as a pointer to the beginning
1038       of the structure (instead of having a series of synthesized arguments,
1039       one for each structure member)
1040    
1041     - made makeml script's verbosity level configurable via environment
1042       variable (MAKEML_VERBOSITY)
1043    
1044     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1045       in rawmem-x86.sml; we are now using the real thing
1046    
1047    ----------------------------------------------------------------------
1048    Name: Matthias Blume
1049    Date: 2001/03/22 16:25:00 EST
1050    Tag: blume-20010322-bootfiles
1051    Description:
1052    
1053    Created a new set of bootfiles (for your automatic installation convenience).
1054    
1055    ----------------------------------------------------------------------
1056    Name: Matthias Blume
1057    Date: 2001/03/22 15:10:00 EST
1058    Tag: blume-20010322-rawmem-parcm
1059    Description:
1060    
1061    1. All "raw memory access" primitives for the new FFI are implemented now
1062       (at least on the x86).
1063    2. Some further cleanup of CM's parallel make mechanism.
1064    
1065    ----------------------------------------------------------------------
1066    Name: Matthias Blume
1067    Date: 2001/03/19 17:53:00 EST
1068    Tag: blume-20010319-parallel
1069    Description:
1070    
1071    Parallel make (using compile servers) now works again.
1072    
1073    To this end, CM.stabilize and CMB.make have been modified to work in
1074    two passes when compile servers are attached:
1075       1. Compile everything, do not perform stabilization; this pass
1076          uses compile servers
1077       2. Stabilize everything; this pass does not use compile servers
1078    If there are no compile servers, the two passes are combined into one
1079    (as before).  Splitting the passes increases the inherent parallelism
1080    in the dependency graph because the entire graph including all
1081    libraries is available at the same time.  This, in turn, improves
1082    server utilization.  The downside is that the master process will
1083    have to do some extra work after compilation is done (because for
1084    technical reasons it must re-read all the binfiles during stabilization).
1085    
1086    ----------------------------------------------------------------------
1087    Name: Matthias Blume
1088  Date: 2001/03/16 12:22:00 EST  Date: 2001/03/16 12:22:00 EST
1089  Tag: blume-20010316-bootfiles  Tag: blume-20010316-bootfiles
1090  Description:  Description:
# Line 2003  Line 3074 
3074       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3075       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3076       Please, don't be too harsh on me for messing with this code a bit more       Please, don't be too harsh on me for messing with this code a bit more
3077       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3078       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3079       Dave's hands. :)       Dave's hands. :)
3080    

Legend:
Removed from v.800  
changed lines
  Added in v.961

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