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 829, Mon May 14 15:30:56 2001 UTC revision 1030, Fri Jan 18 21:45:37 2002 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  Name: Matthias Blume
17    Date: 2002/01/18
18    Tag: blume-20020118-ready-for-new-release
19    Description:
20    
21    Made config/srcarchiveurl point to a new place.  (Will provide boot
22    files shortly.)
23    
24    Maybe we christen this to be 110.38?
25    
26    ----------------------------------------------------------------------
27    Name: Matthias Blume
28    Date: 2002/01/18
29    Tag: blume-20020118-more-ffifiddle
30    Description:
31    
32    Today's FFI fiddling:
33    
34      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
35        C.Set.* that use concrete (MLRep.*) arguments and results instead
36        of abstract ones.
37    
38      - Provided word-style bit operations etc. for "int" representation
39        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
40        and Long.
41    
42    ----------------------------------------------------------------------
43    Name: Matthias Blume
44    Date: 2002/01/18
45    Tag: blume-20020118-use-x86-fp
46    Description:
47    
48    Now that x86-fast-fp seems to be working, I turned it back on again
49    by default.  (Seems to work fine now, even with the FFI.)
50    
51    Other than that, I added some documentation about the FFI to
52    src/ml-nlffigen/README and updated the FFI test examples in
53    src/ml-nlffi-lib/Tests/*.
54    
55    ----------------------------------------------------------------------
56    Name: Allen Leung
57    Date: 2002/01/17
58    Tag: leunga-20020117-x86-fast-fp-call
59    Description:
60    
61       1. Fixed a problem with handling return fp values when x86's fast fp
62          mode is turned on.
63    
64       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
65          of %f32.
66    
67       3. Added a constructor INT32lit to the ast of MLRISC tools.
68    
69    ----------------------------------------------------------------------
70    Name: Matthias Blume
71    Date: 2002/01/16
72    Tag: blume-20020116-ffifiddle
73    Description:
74    
75    More fiddling with the FFI interface:
76    
77     - Make constness 'c instead of rw wherever possible.  This eliminates
78       the need for certain explicit coercions.  (However, due to ML's
79       value polymorphism, there will still be many cases where explicit
80       coercions are necessary.  Phantom types are not the whole answer
81       to modeling a subtyping relationship in ML.)
82    
83     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
84    
85     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
86       clumsy workaround for varargs problem.)
87    
88    ----------------------------------------------------------------------
89    Name: Lal George
90    Date: 2002/01/15
91    Tag: <none>
92    Description:
93    
94    1. Since COPY instructions are no longer native to the architecture,
95       a generic functor can be used to implement the expandCopies function.
96    
97    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
98       TEXT segment.
99    
100    ----------------------------------------------------------------------
101    Name: Matthias Blume
102    Date: 2002/01/15
103    Tag: blume-20020115-ffiupdates
104    Description:
105    
106    1. Fix for bug resulting in single-precision float values being returned
107       incorrectly from FFI calls.
108    
109    2. Small modifications to C FFI API:
110    
111        - memory-allocation routines return straight objects (no options)
112          and raise an exception in out-of-memory situations
113        - unsafe extensions to cast between function pointers and pointers
114          from/to ints
115        - added structure C_Debug as an alternative to structure C where
116          pointer-dereferencing (|*| and |*!) always check for null-pointers
117        - added open_lib' to DynLinkage;  open_lib' works like open_lib
118          but also takes a (possibly empty) list of existing library handles
119          that the current library depends on
120    
121    ----------------------------------------------------------------------
122    Name: Matthias Blume
123    Date: 2002/01/10
124    Tag: blume-20020110-newffigen
125    Description:
126    
127    1. Updates to portable graph code.
128    
129    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
130       more scalable now so that even huge interfaces such as the one
131       for GTK compile in finite time and space. :-)
132       See src/ml-nlffigen/README for details on what's new.
133    
134    ----------------------------------------------------------------------
135    Name: Lal George
136    Date: 2001/01/09 14:31:35 EST 2002
137    Tag: george-20011206-rm-native-copy
138    Description:
139    
140            Removed the native COPY and FCOPY instructions
141            from all the architectures and replaced it with the
142            explicit COPY instruction from the previous commit.
143    
144            It is now possible to simplify many of the optimizations
145            modules that manipulate copies. This has not been
146            done in this change.
147    
148    ----------------------------------------------------------------------
149    Name: Lal George
150    Date: 2001/12/06 16:50:13 EST 2001
151    Tag: george-20011206-mlrisc-instruction
152    Description:
153    
154    Changed the representation of instructions from being fully abstract
155    to being partially concrete. That is to say:
156    
157      from
158            type instruction
159    
160      to
161            type instr                              (* machine instruction *)
162    
163            datatype instruction =
164                LIVE of {regs: C.cellset, spilled: C.cellset}
165              | KILL of {regs: C.cellset, spilled: C.cellset}
166              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
167              | ANNOTATION of {i: instruction, a: Annotations.annotation}
168              | INSTR of instr
169    
170    This makes the handling of certain special instructions that appear on
171    all architectures easier and uniform.
172    
173    LIVE and KILL say that a list of registers are live or killed at the
174    program point where they appear. No spill code is generated when an
175    element of the 'regs' field is spilled, but the register is moved to
176    the 'spilled' (which is present, more for debugging than anything else).
177    
178    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
179    We used to generate:
180    
181            DEFFREG f1
182            f1 := f2 + f3
183            trapb
184    
185    but now generate:
186    
187            f1 := f2 + f3
188            trapb
189            LIVE {regs=[f1,f2,f3], spilled=[]}
190    
191    Furthermore, the DEFFREG (hack) required that all floating point instruction
192    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
193    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
194    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
195    this architecture).
196    
197    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
198    all the architectures. This will result in further simplification of the
199    register allocator that must be aware of them for coalescing purposes, and
200    will also simplify certain aspects of the machine description that provides
201    callbacks related to parallel copies.
202    
203    ANNOTATION should be obvious, and now INSTR represents the honest to God
204    machine instruction set!
205    
206    The <arch>/instructions/<arch>Instr.sml files define certain utility
207    functions for making porting easier -- essentially converting upper case
208    to lower case. All machine instructions (of type instr) are in upper case,
209    and the lower case form generates an MLRISC instruction. For example on
210    the alpha we have:
211    
212      datatype instr =
213         LDA of {r:cell, b:cell, d:operand}
214       | ...
215    
216      val lda : {r:cell, b:cell, d:operand} -> instruction
217        ...
218    
219    where lda is just (INSTR o LDA), etc.
220    
221    ----------------------------------------------------------------------
222    Name: Matthias Blume
223    Date: 2001/11/22 21:40:00 EST
224    Tag: Release_110_37
225    Description:
226    
227    Release 110.37.  This time for real.
228    
229    ----------------------------------------------------------------------
230    Name: Matthias Blume
231    Date: 2001/11/21 16:35:00 EST
232    Tag: blume-20011121-foot-in-mouth
233    Description:
234    
235    Removed the "Release_110_37" tag because of a serious bug.
236    This will be re-tagged once the bug is fixed.
237    
238    ----------------------------------------------------------------------
239    Name: Matthias Blume
240    Date: 2001/11/21 16:14:00 EST
241    Tag: blume-20011121-forgottenfile
242    Description:
243    
244    Forgot to add a file.  (Just a .tex-file -- part of
245    the CM manual source.)
246    
247    ----------------------------------------------------------------------
248    Name: Matthias Blume
249    Date: 2001/11/21 16:10:00 EST
250    Tag: blume-20011121-invalid_110_37
251    Description:
252    
253    Note: I removed the original tag "Release_110_37" from this commit
254          because we found a serious bug in all non-x86 backends.
255          - Matthias
256    
257    1. Modifications to the SML/NJ code generator and to the runtime system
258       so that code object name strings are directly inserted into code
259       objects at code generation time.  The only business the runtime system
260       has with this is now to read the name strings on occasions.
261       (The encoding of the name string has also changed somewhat.)
262    
263    2. CM now implements a simple "set calculus" for specifying export lists.
264       In particular, it is now possible to refer to the export lists of
265       other libraries/groups/sources and form unions as well as differences.
266       See the latest CM manual for details.
267    
268    3. An separate notion of "proxy" libraries has again be eliminated from
269       CM's model.  (Proxy libraries are now simply a special case of using
270       the export list calculus.)
271    
272    4. Some of the existing libraries now take advantage of the new set
273       calculus.
274       (Notice that not all libraries have been converted because some
275       of the existing .cm-files are supposed to be backward compatible
276       with 110.0.x.)
277    
278    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
279       "General.exnMessage"!  The former relies on a certain hook to be
280       initialized, and that often does not happen in the stand-alone case.)
281    
282    ----------------------------------------------------------------------
283    Name: Lal George
284    Date: 2001/11/21  13:56:18 EST
285    Tag: george-2001121-pseudo-ops
286    Description:
287    
288      Implemented a complete redesign of MLRISC pseudo-ops. Now there
289      ought to never be any question of incompatabilities with
290      pseudo-op syntax expected by host assemblers.
291    
292      For now, only modules supporting GAS syntax are implemented
293      but more should follow, such as MASM, and vendor assembler
294      syntax, e.g. IBM as, Sun as, etc.
295    
296    ----------------------------------------------------------------------
297    Name: Matthias Blume
298    Date: 2001/11/14 11:52:00 EST
299    Tag: blume-20011114-srcname
300    Description:
301    
302    1. Routed the name of the current source file to mlriscgen where it
303       should be directly emitted into the code object.  (This last part
304       is yet to be done.)
305    
306    2. Some cleanup of the pgraph code to make it match the proposal that
307       I put out the other day.  (The proposal notwithstanding, things are
308       still in flux here.)
309    
310    ----------------------------------------------------------------------
311    Name: Lal George
312    Date: 2001/11/14 09:44:04 EST
313    Tag:
314    Description:
315    
316      Fix for a backpatching bug reported by Allen.
317    
318      Because the boundary between short and long span-dependent
319      instructions is +/- 128, there are an astounding number of
320      span-dependent instructions whose size is over estimated.
321    
322      Allen came up with the idea of letting the size of span
323      dependent instructions be non-monotonic, for a maxIter
324      number of times, after which the size must be monotonically
325      increasing.
326    
327      This table shows the number of span-dependent instructions
328      whose size was over-estimated as a function of maxIter, for the
329      file Parse/parse/ml.grm.sml:
330    
331         maxIter            # of instructions:
332            10                      687
333            20                      438
334            30                      198
335            40                        0
336    
337      In compiling the compiler, there is no significant difference in
338      compilation speed between maxIter=10 and maxIter=40. Actually,
339      my measurements showed that maxIter=40 was a tad faster than
340      maxIter=10! Also 96% of the  files in the compiler reach a fix
341      point within 13 iterations, so fixing maxIter at 40, while high,
342      is okay.
343    
344    ----------------------------------------------------------------------
345    Name: Matthias Blume
346    Date: 2001/10/31 15:25:00 EST
347    Tag: blume-20011031-pgraph
348    Description:
349    
350    CKIT:
351    * Changed the "Function" constructor of type Ast.ctype to carry optional
352      argument identifiers.
353    * Changed the return type of TypeUtil.getFunction accordingly.
354    * Type equality ignores the argument names.
355    * TypeUtil.composite tries to preserve argument names but gives up quickly
356      if there is a mismatch.
357    
358    installation script:
359    * attempts to use "curl" if available (unless "wget" is available as well)
360    
361    CM:
362    * has an experimental implementation of "portable graphs" which I will
363      soon propose as an implementation-independent library format
364    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
365    
366    NLFFI-LIB:
367    * some cleanup (all cosmetic)
368    
369    NLFFIGEN:
370    * temporarily disabled the mechanism that suppresses ML output for
371      C definitions whose identifiers start with an underscore character
372    * generate val bindings for enum constants
373    * user can request that only one style (light or heavy) is being used;
374      default is to use both (command-line arguments: -heavy and -light)
375    * fixed bug in handling of function types involving incomplete pointers
376    * generate ML entry points that take record arguments (i.e., using
377      named arguments) for C functions that have a prototype with named
378      arguments
379      (see changes to CKIT)
380    
381    ----------------------------------------------------------------------
382    Name: Allen Leung
383    Date: 2001/10/27 20:34:00 EDT
384    Tag: leunga-20011027-x86-fast-fp-call
385    Description:
386    
387       Fixed the bug described in blume-20010920-slowfp.
388    
389       The fix involves
390          1. generating FCOPYs in FSTP in ia32-svid
391          2. marking a CALL with the appropriate annotation
392    
393    ----------------------------------------------------------------------
394    Name: Matthias Blume
395    Date: 2001/10/16 11:32:00 EDT
396    Tag: blume-20011016-netbsd
397    Description:
398    
399    Underscore patch from Chris Richards (fixing problem with compiling
400    runtime system under recent NetBSD).
401    
402    ----------------------------------------------------------------------
403    Name: Allen Leung
404    Date: 2001/10/12 17:18:32 EDT 2001
405    Tag: leung-20011012-x86-printflowgraph
406    Description:
407    
408    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
409    
410    ----------------------------------------------------------------------
411    Name: Lal George
412    Date: 2001/10/11 23:51:34 EDT
413    Tag: george-20011011-too-many-instrs
414    Description:
415    
416    The representation of a program point never expected to see more
417    than 65536 instructions in a basic block!
418    
419    ----------------------------------------------------------------------
420    Name: Lal George
421    Date: 2001/10/09 09:41:37 EDT
422    Tag: george-20011008-mlrisc-labels
423    Description:
424    
425    Changed the machine description files to support printing of
426    local and global labels in assembly code, based on host assembler
427    conventions.
428    
429    ----------------------------------------------------------------------
430    Name: Matthias Blume
431    Date: 2001/09/25 15:25:00 EDT
432    Tag: blume-20010925-exninfo
433    Description:
434    
435    I provided a non-hook implementation of exnName (at the toplevel) and
436    made the "dummy" implementation of exnMessage (at the toplevel) more
437    useful: if nothing gets "hooked in", then at least you are going to
438    see the exception name and a message indicating why you don't see more.
439    
440    [For the time being, programs that need exnMessage and want to use
441    ml-build should either use General.exnMessage (strongly recommended) or
442    refer to structure General at some other point so that CM sees a
443    static dependency.]
444    
445    [Similar remarks go for "print" and "use":  If you want to use their
446    functionality in stand-alone programs generated by ml-build, then use
447    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
448    
449    ----------------------------------------------------------------------
450    Name: Matthias Blume
451    Date: 2001/09/20 17:28:00 EDT
452    Tag: blume-20010920-slowfp
453    Description:
454    
455    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
456    
457    ----------------------------------------------------------------------
458    Name: Matthias Blume
459    Date: 2001/09/20 17:20:00 EDT
460    Tag: blume-20010920-canonicalpaths
461    Description:
462    
463    0. Updated the BOOT file (something that I forgot to do earlier).
464    
465    1. Small internal change to CM so that it avoids "/../" in filenames
466       as much as possible (but only where it is safe).
467    
468    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
469       that arguments that contain delimiters are passed through correctly.
470       This change also means that all "special" arguments of the form
471       @SMLxxx... must come first.
472    
473    3. Changed install script to put relative anchor names for tool commands
474       into pathconfig.
475    
476    ----------------------------------------------------------------------
477    Name: Matthias Blume
478    >>>>>>> 1.169
479    Date: 2001/09/18 15:35:00 EDT
480    Tag: blume-20010918-readme11036
481    Description:
482    
483    Added README files.
484    
485    ----------------------------------------------------------------------
486    Name: Matthias Blume
487    Date: 2001/09/18 11:45:00 EDT
488    Tag: Release_110_36 (retag)
489    Description:
490    
491    Fixed mistake in config/preloads. Retagged as 110.36.
492    
493    ----------------------------------------------------------------------
494    Name: Matthias Blume
495    Date: 2001/09/18 09:40:00 EDT
496    Tag: Release_110_36_orig (tag changed)
497    Description:
498    
499    New version (110.36).  New bootfiles.
500    
501    ----------------------------------------------------------------------
502    Name: Matthias Blume
503    Date: 2001/09/14 16:15:00 EDT
504    Tag: blume-20010914-x86fastfp
505    Description:
506    
507    John committed some changes that Allen made, in particular a (hopefully)
508    correctly working version of the x86-fp module.
509    
510    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
511    flag to "true".  Everything seems to compile to a fixpoint ok, and
512    "mandelbrot" speeds up by about 15%.
513    
514    ----------------------------------------------------------------------
515    Name: Matthias Blume
516    Date: 2001/09/13 11:20:00 EDT
517    Tag: blume-20010913-minimal
518    Description:
519    
520    1. Stefan Monnier's patch to fix a miscompilation problem that
521       was brought to light by John Reppy's work on Moby.
522    
523    2. Implemented a minimal "structure Compiler" that contains just
524       "version" and "architecture".  The minimal version will be
525       available when the full version is not.  This is for backward-
526       compatibility with code that wants to test Compiler.version.
527    
528    ----------------------------------------------------------------------
529    Name: Matthias Blume
530    Date: 2001/08/28 14:03:00 EDT
531    Tag: blume-20010828-ml-lex
532    Description:
533    
534    Fix for bug 1581, received from Neophytos Michael.
535    
536    ----------------------------------------------------------------------
537    Name: Matthias Blume
538    Date: 2001/08/27 11:20:00 EDT
539    Tag: blume-20010827-readme11035
540    Description:
541    
542    Fleshed out the README file for 110.35.
543    
544    ----------------------------------------------------------------------
545    Name: Matthias Blume
546    Date: 2001/08/24 17:10:00 EDT
547    Tag: Release_110_35
548    Description:
549    
550    New version number (110.35).  New bootfiles.
551    
552    ----------------------------------------------------------------------
553    Name: Lal George
554    Date: 2001/08/24 13:47:18 EDT 2001
555    Tag: george-20010824-MLRISC-graphs
556    Description:
557    
558     removed clusters from MLRISC completely and replaced with graphs.
559    
560    ----------------------------------------------------------------------
561    Name: Matthias Blume
562    Date: 2001/08/23 17:50:00 EDT
563    Tag: blume-20010823-toplevel
564    Description:
565    
566    - some reorganization of the code that implements various kinds of
567      environments in the compiler (static, dynamic, symbolic, combined)
568    - re-implemented the EnvRef module so that evalStream works properly
569      (if the stream contains references to "use", "CM.make", etc.)
570    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
571    
572    ----------------------------------------------------------------------
573    Name: Matthias Blume
574    Date: 2001/08/20 15:50 EDT
575    Tag: blume20010820-slipup
576    Description:
577    
578    I forgot to commit a few files.  Here they are...
579    
580    ----------------------------------------------------------------------
581    Name: Matthias Blume
582    Date: 2001/08/20 15:35:00 EDT
583    Tag: blume-20010820-debugprof
584    Description:
585    
586    !!!! NEW BOOTFILES !!!!
587    
588    This is another round of reorganizing the compiler sources.  This
589    time the main goal was to factor out all the "instrumentation"
590    passes (for profiling and backtracing) into their own library.
591    The difficulty was to do it in such a way that it does not depend
592    on elaborate.cm but only on elabdata.cm.
593    
594    Therefore there have been further changes to both elaborate.cm and
595    elabdata.cm -- more "generic" things have been moved from the former
596    to the latter.  As a result, I was forced to split the assignment
597    of numbers indicating "primtyc"s into two portions: SML-generic and
598    SML/NJ-specific.  Since it would have been awkward to maintain,
599    I bit the bullet and actually _changed_ the mapping between these
600    numbers and primtycs.  The bottom line of this is that you need
601    a new set of bin- and bootfiles.
602    
603    I have built new bootfiles for all architectures, so doing a fresh
604    checkout and config/install.sh should be all you need.
605    
606    The newly created library's name is
607    
608        $smlnj/viscomp/debugprof.cm
609    
610    and its sources live under
611    
612        src/compiler/DebugProf
613    
614    ----------------------------------------------------------------------
615    Name: Matthias Blume
616    Date: 2001/08/15 17:15:00 EDT
617    Tag: blume-20010815-compreorg
618    Description:
619    
620    This is a first cut at reorganizing the CM libraries that make up the
621    core of the compiler.  The idea is to separate out pieces that could
622    be used independently by tools, e.g., the parser, the typechecker, etc.
623    
624    The current status is a step in this direction, but it is not quite
625    satisfactory yet.  Expect more changes in the future.
626    
627    Here is the current (new) organization...
628    
629        What used to be $smlnj/viscomp/core.cm is now divided into
630        six CM libraries:
631    
632             $smlnj/viscomp/basics.cm
633                           /parser.cm
634                           /elabdata.cm
635                           /elaborate.cm
636                           /execute.cm
637                           /core.cm
638    
639        The CM files for these libraries live under src/system/smlnj/viscomp.
640        All these libraries are proxy libraries that contain precisely
641        one CM library component.  Here are the locations of the components
642        (all within the src/compiler tree):
643    
644             Basics/basics.cm
645             Parse/parser.cm
646             ElabData/elabdata.cm
647             Elaborator/elaborate.cm
648             Execution/execute.cm
649             core.cm
650    
651         [This organization is the same that has been used already
652         for a while for the architecture-specific parts of the visible
653         compiler and for the old version of core.cm.]
654    
655         As you will notice, many source files have been moved from their
656         respective original locations to a new home in one of the above
657         subtrees.
658    
659         The division of labor between the new libraries is the following:
660    
661             basics.cm:
662                - Simple, basic definitions that pertain to many (or all) of
663                  the other libraries.
664             parser.cm:
665                - The SML parser, producing output of type Ast.dec.
666                - The type family for Ast is also defined and exported here.
667             elabdata.cm:
668                - The datatypes that describe input and output of the elaborator.
669                  This includes types, absyn, and static environments.
670             elaborator.cm:
671                - The SML/NJ type checker and elaborator.
672                  This maps an Ast.dec (with a given static environment) to
673                  an Absyn.dec (with a new static environment).
674                - This libraries implements certain modules that used to be
675                  structures as functors (to remove dependencies on FLINT).
676             execute.cm:
677                - Everything having to do with executing binary code objects.
678                - Dynamic environments.
679             core.cm:
680                - SML/NJ-specific instantiations of the elaborator and MLRISC.
681                - Top-level modules.
682                - FLINT (this should eventually become its own library)
683    
684    Notes:
685    
686    I am not 100% happy with the way I separated the elaborator (and its
687    data structures) from FLINT.  Two instances of the same problem:
688    
689        1. Data structures contain certain fields that carry FLINT-specific
690           information.  I hacked around this using exn and the property list
691           module from smlnj-lib.  But the fact that there are middle-end
692           specific fields around at all is a bit annoying.
693    
694        2. The elaborator calculates certain FLINT-related information.  I tried
695           to make this as abstract as I could using functorization, but, again,
696           the fact that the elaborator has to perform calculations on behalf
697           of the middle-end at all is not nice.
698    
699        3. Having to used exn and property lists is unfortunate because it
700           weakens type checking.  The other alternative (parameterizing
701           nearly *everything*) is not appealing, though.
702    
703    I removed the "rebinding =" warning hack because due to the new organization
704    it was awkward to maintain it.  As a result, the compiler now issues some of
705    these warnings when compiling init.cmi during bootstrap compilation. On
706    the plus side, you also get a warning when you do, for example:
707       val op = = Int32.+
708    which was not the case up to now.
709    
710    I placed "assign" and "deref" into the _Core structure so that the
711    code that deals with the "lazy" keyword can find them there.  This
712    removes the need for having access to the primitive environment
713    during elaboration.
714    
715    ----------------------------------------------------------------------
716    Name: Matthias Blume
717    Date: 2001/08/13
718    Tag: blume-20010813-closures
719    Description:
720    
721    This fix was sent to us by Zhong Shao.  It is supposed to improve the
722    performance of certain loops by avoiding needless closure allocation.
723    
724    ----------------------------------------------------------------------
725    Name: Lal George
726    Date: 2001/07/31 10:03:23 EDT 2001
727    Tag: george-20010731-x86-fmalloc
728    Description: Fixed bug in x86 calls
729    
730        There was a bug where call instructions would mysteriously
731        vanish. The call instruction had to be one that returned
732        a floating point value.
733    
734    ----------------------------------------------------------------------
735    Name: Lal George
736    Date: 2001/07/19 16:36:29 EDT 2001
737    Tag: george-20010719-simple-cells
738    Description:
739    
740    I have dramatically simplified the interface for CELLS in MLRISC.
741    
742    In summary, the cells interface is broken up into three parts:
743    
744      1. CellsBasis : CELLS_BASIS
745    
746            CellsBasis is a top level structure and common for all
747            architectures.  it contains the definitions of basic datatypes
748            and utility  functions over these types.
749    
750      2. functor Cells() : CELLS
751    
752            Cells generates an interface for CELLS that incorporates the
753            specific resources on the target architecture, such as the
754            presence of special register classes, their number and size,
755            and various useful substructures.
756    
757      3. <ARCH>CELLS
758    
759            e.g. SparcCells: SPARCCELLS
760    
761            <ARCH>CELLS usually contains additional bindings for special
762            registers  on the architecture, such as:
763    
764                    val r0 : cell           (* register zero *)
765                    val y : cell            (* Y register *)
766                    val psr : cell          (* processor status register *)
767                    ...
768    
769            The structure returned by applying the Cells functor is opened
770            in this interface.
771    
772    The main implication of all this is that the datatypes for cells is
773    split between CellsBasis and CELLS -- a fairly simple change for user
774    code.
775    
776    In the old scheme the CELLS interface had a definitional binding of
777    the form:
778    
779            signature CELLS = sig
780    
781               structure CellsBasis = CellsBasis
782    
783               ...
784    
785            end
786    
787    With all the sharing constraints that goes on in MLRISC, this old
788    design  quickly leads to errors such as:
789    
790            "structure definition spec inside of sharing ... "
791    
792    
793    and appears to require an unacceptable amount of sharing and where
794    constraint hackery.
795    
796    I think this error message (the interaction of definitional specs and
797    sharing) requires more explanation on our web page.
798    
799    ----------------------------------------------------------------------
800    Name: Matthias Blume
801    Date: 2001/07/19 15:00:00 EDT
802    Tag: blume-20010719-libreorg
803    Description:
804    
805    This update puts together a fairly extensive but straightforward change
806    to the way the libraries that implement the interactive system are
807    organized:
808    
809       The biggest change is the elimination of structure Compiler.  As a
810       replacement for this structure, there is now a CM library
811       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
812       that exports all the substructures of the original structure Compiler
813       directly.  So instead of saying Compiler.Foo.bar one now simply
814       says Foo.bar.  (The CM libraries actually export a collection of
815       structures that is richer than the collection of substructures of
816       structure Compiler.)
817    
818       To make the transition smooth, there is a separate library called
819       $smlnj/compiler/compiler.cm which puts together and exports the
820       original structure Compiler (or at least something very close to it).
821    
822       There are five members of the original structure Compiler
823       that are not exported directly but which instead became members
824       of a new structure Backend (described by signature BACKEND).  These are:
825       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
826       Interact (: INTERACT), structure Machine (: MACHINE), and val
827       architecture (: string).
828    
829       Structure Compiler.Version has become structure CompilerVersion.
830    
831       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
832       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
833       or x86, respectively.
834       Each of these exports the same frontend structures that
835       $smlnj/compiler.cm exports.  But they do not have a structure Backend
836       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
837       Hppa, PPC, Sparc, or X86, respectively.
838    
839       Library $smlnj/compiler/all.cm exports the union of the exports of
840       $smlnj/compiler/<arch>.cm
841    
842       There are no structures <Arch>Compiler anymore, use
843       $smlnj/compiler/<arch>.cm instead.
844    
845       Library host-compiler-0.cm is gone.  Instead, the internal library
846       that instantiates CM is now called cm0.cm.  Selection of the host
847       compiler (backend) is no longer done here but. (Responsibility for it
848       now lies with $smlnj/compiler/current.cm.  This seems to be more
849       logical.)
850    
851       Many individual files have been moved or renamed.  Some files have
852       been split into multiple files, and some "dead" files have been deleted.
853    
854    Aside from these changes to library organization, there are also changes
855    to the way the code itself is organized:
856    
857       Structure Binfile has been re-implemented in such a way that it no
858       longer needs any knowledge of the compiler.  It exclusively deals
859       with the details of binfile layout.  It no longer invokes the
860       compiler (for the purpose of creating new prospective binfile
861       content), and it no longer has any knowledge of how to interpret
862       pickles.
863    
864       Structure Compile (: COMPILE) has been stripped down to the bare
865       essentials of compilation.  It no longer deals with linking/execution.
866       The interface has been cleaned up considerably.
867    
868       Utility routines for dealing with linking and execution have been
869       moved into their own substructures.
870    
871       (The ultimate goal of these changes is to provide a light-weight
872       binfile loader/linker (at least for, e.g., stable libraries) that
873       does not require CM or the compiler to be present.)
874    
875    CM documentation has been updated to reflect the changes to library
876    organization.
877    
878    ----------------------------------------------------------------------
879    Name: Matthias Blume
880    Date: 2001/07/10 17:30:00 EDT
881    Tag: Release_110_34
882    Description:
883    
884    Minor tweak to 110.34 (re-tagged):
885    
886      - README.html file added to CVS repository
887      - runtime compiles properly under FreeBSD 3.X and 4.X
888    
889    ----------------------------------------------------------------------
890    Name: Matthias Blume
891    Date: 2001/07/10 17:30:00 EDT
892    Tag: Release_110_34
893    Description:
894    
895    New version number (110.34). New bootfiles.
896    
897    ----------------------------------------------------------------------
898    Name: Matthias Blume
899    Date: 2001/07/09 16:00:00 EDT
900    Tag: blume-20010709-more-varargs
901    Description:
902    
903    I changed the handling of varargs in ml-nlffigen again:
904    The ellipsis ... will now simply be ignored (with an accompanying warning).
905    
906    The immediate effect is that you can actually call a varargs function
907    from ML -- but you can't actually supply any arguments beyond the ones
908    specified explicitly.  (For example, you can call printf with its format
909    string, but you cannot pass additional arguments.)
910    
911    This behavior is only marginally more useful than the one before, but
912    it has the advantage that a function or, more importantly, a function
913    type never gets dropped on the floor, thus avoiding follow-up problems with
914    other types that refer to the offending one.
915    
916    ----------------------------------------------------------------------
917    Name: Matthias Blume
918    Date: 2001/07/09 11:25:00 EDT
919    Tag: blume-20010709-varargs
920    Description:
921    
922    1. ckit-lib.cm now exports structure Error
923    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
924       with a warning accompanied by a source location.  Moreover, it
925       merely skips the offending function or type and proceeds with the
926       rest of its work.u  As a result, one can safely feed C code containing
927       "..." to ml-nlffigen.
928    3. There are some internal improvements to CM, providing slightly
929       more general string substitutions in the tools subsystem.
930    
931    ----------------------------------------------------------------------
932    Name: Matthias Blume
933    Date: 2001/06/27 15:10:00 EDT
934    Tag: blume-20010627-concur
935    Description:
936    
937    Fixed a small bug in CM's handling of parallel compilation.
938    (You could observe the bug by Control-C-interrupting an ordinary
939    CMB.make or CM.stabilize and then attaching some compile servers.
940    The result was that all of a sudden the previously interrupted
941    compilation would continue on its own.  This was because of
942    an over-optimization: CM did not bother to clean out certain queues
943    when no servers were attached "anyway", resulting in the contents
944    of these queues to grab control when new servers did get attached.)
945    
946    There is also another minor update to the CM manual.
947    
948    ----------------------------------------------------------------------
949    Name: Matthias Blume
950    Date: 2001/06/26 16:15:00 EDT
951    Tag: blume-20010626-cmdoc
952    Description:
953    
954    Minor typo fixed in CM manual (syntax diagram for libraries).
955    
956    ----------------------------------------------------------------------
957    Name: Matthias Blume
958    Date: 2001/06/25 22:55:00 EDT
959    Tag: blume-20010625-x86pc
960    Description:
961    
962    Fixed a nasty bug in the X86 assembly code that caused signal
963    handlers to fail (crash) randomly.
964    
965    ----------------------------------------------------------------------
966    Name: Matthias Blume
967    Date: 2001/06/25 12:05:00 EDT
968    Tag: blume-20010625-nlffigen
969    Description:
970    
971    This update fixes a number of minor bugs in ml-nlffigen as reported by
972    Nick Carter <nbc@andrew.cmu.edu>.
973    
974      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
975      2. Default names for generated files are now derived from the name of
976         the C file *without its directory*.  In particular, this causes generated
977         files to be placed locally even if the C file is in some system directory.
978      3. Default names for generated signatures and structures are also derived
979         from the C file name without its directory.  This avoids silly things
980         like "structure GL/GL".
981         (Other silly names are still possible because ml-nlffigen does not do
982          a thorough check of whether generated names are legal ML identifiers.
983          When in doubt, use command line arguments to force particular names.)
984    
985    ----------------------------------------------------------------------
986    Name: Matthias Blume
987    Date: 2001/06/21 12:25:00 EDT
988    Tag: blume-20010621-eXene
989    Description:
990    
991    eXene now compiles and (sort of) works again.
992    
993    The library name (for version > 110.33) is $/eXene.cm.
994    
995    I also added an new example in src/eXene/examples/nbody.  See the
996    README file there for details.
997    
998    ----------------------------------------------------------------------
999    Name: Matthias Blume
1000    Date: 2001/06/20 16:40:00 EDT
1001    Tag: blume-20010620-cml
1002    Description:
1003    
1004    CML now compiles and works again.
1005    
1006    Libraries (for version > 110.33):
1007    
1008      $cml/cml.cm            Main CML library.
1009      $cml/basis.cm          CML's version of $/basis.cm.
1010      $cml/cml-internal.cm   Internal helper library.
1011      $cml/core-cml.cm       Internal helper library.
1012      $cml-lib/trace-cml.cm  Tracing facility.
1013      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1014    
1015    The installer (config/install.sh) has been taught how to properly
1016    install this stuff.
1017    
1018    ----------------------------------------------------------------------
1019    Name: Matthias Blume
1020    Date: 2001/06/19 17:55:00 EDT
1021    Tag: blume-20010619-instantiate
1022    Description:
1023    
1024    This un-breaks the fix for bug 1432.
1025    (The bug was originally fixed in 110.9 but I broke it again some
1026    time after that.)
1027    
1028    ----------------------------------------------------------------------
1029    Name: Matthias Blume
1030    Date: 2001/06/19 17:25:00 EDT
1031    Tag: blume-20010619-signals
1032    Description:
1033    
1034    This should (hopefully) fix the long-standing signal handling bug.
1035    (The runtime system was constructing a continuation record with an
1036    incorrect descriptor which would cause the GC to drop data on the floor...)
1037    
1038    ----------------------------------------------------------------------
1039    Name: Matthias Blume
1040    Date: 2001/06/15 15:05:00 EDT
1041    Tag: blume-20010615-moresparc
1042    Description:
1043    
1044    Here is a short late-hour update related to Sparc c-calls:
1045    
1046     -- made handling of double-word arguments a bit smarter
1047    
1048     -- instruction selection phase tries to collapse certain clumsily
1049        constructed ML-Trees; typical example:
1050    
1051            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1052    
1053        This currently has no further impact on SML/NJ since mlriscGen does
1054        not seem to generate such patterns in the first place, and c-calls
1055        (which did generate them in the beginning) has meanwhile been fixed
1056        so as to avoid them as well.
1057    
1058    ----------------------------------------------------------------------
1059    Name: Matthias Blume
1060    Date: 2001/06/15 15:05:00 EDT
1061    Tag: blume-20010615-sparc
1062    Description:
1063    
1064    The purpose of this update is to provide an implementation of NLFFI
1065    on Sparc machines.
1066    
1067    Here are the changes in detail:
1068    
1069       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1070       the Sparc implementation of the c-calls API.
1071       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1072       for accessing the ML frame.  Thus, we have a real frame pointer and
1073       can freely modify %sp without need for an omit-frame-ptr phase.
1074       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1075       this case.
1076       * ml-nlffigen has been taught to produce code for different architectures
1077       and calling conventions.
1078       * In a way similar to what was done in the x86 case, the Sparc
1079       backend uses its own specific extension to mltree.  (For example,
1080       it needs to be able to generate UNIMP instructions which are part
1081       of the calling convention.)
1082       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1083       to make it easier to plug in new machine- and os-dependent parts).
1084    
1085    There are some other fairly unrelated bug fixes and cleanups as well:
1086    
1087       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1088       that they properly share their libraries with existing SML/NJ libraries.
1089       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1090       follow-up error messages.
1091       * Updates to CM/CMB documentation.
1092    
1093    TODO items:
1094    
1095       * MLRISC should use a different register as its asmTemp on the Sparc.
1096         (The current %o2 is a really bad choice because it is part of the
1097         calling conventions, so things might interfere in unexpected ways.)
1098    
1099    ----------------------------------------------------------------------
1100    Name: Matthias Blume
1101    Date: 2001/06/07
1102    Tag: blume-20010607-calls
1103    Description:
1104    
1105    A number of internal changes related to C calls and calling conventions:
1106    
1107    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1108       number of bytes popped implicitly (by the callee).  In most cases
1109       this field is 0 but on x86/win32 it is some non-zero value.  This
1110       is information provided for the benefit of the "omit-frameptr" pass.
1111    2. The CALL instruction on the x86 carries a similar "pops" field.
1112       The instruction selection phase copies its value from the ML-Tree
1113       CALL statement.
1114    3. On all other architectures, the instruction selection phase checks
1115       whether "pops=0" and complains if not.
1116    4. The c-calls implementation for x86 now accepts two calling conventions:
1117       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1118       up after the call and pops is set to 0.  For "stdcall", the caller
1119       does nothing, leaving the cleanup to the callee; pops is set to
1120       the number of bytes that were pushed onto the stack.
1121    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1122       distinguish between "ccall" and "stdcall".
1123    6. The UNIMP instruction has been added to the supported Sparc instruction
1124       set. (This is needed for implementing the official C calling convention
1125       on this architecture.)
1126    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1127       work with the latest CM.
1128    
1129    ----------------------------------------------------------------------
1130    Name: Matthias Blume
1131    Date: 2001/06/05 15:10:00 EDT
1132    Tag: blume-20010605-cm-index
1133    Description:
1134    
1135    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1136    
1137    1. CM can now generate "index files".  These are human-readable files
1138       that list on a per-.cm-file basis each toplevel symbol defined or
1139       imported.  The location of the index file for
1140       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1141       To enable index-file generation, set CM.Control.generate_index to true
1142       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1143    
1144       The CM manual has been updated accordingly.
1145    
1146    2. I made some slight modifications to the c-calls API in MLRISC.
1147    
1148         a) There is now a callback to support saving/restoring of
1149            dedicated but caller-save registers around the actual call
1150            instruction.
1151         b) One can optionally specify a comment-annotation for the
1152            call instruction.
1153    
1154    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1155       (For example, the comment annotation shows the C prototype of
1156        the function being called.)
1157    
1158    ----------------------------------------------------------------------
1159    Name: Matthias Blume
1160    Date: 2001/06/01 13:30:00 EDT
1161    Tag: blume-20010601-nlffi-cleanup
1162    Description:
1163    
1164    This is mostly a cleanup of MLFFI stuff:
1165    
1166       - some signature files have been put into a more exposed place
1167       - the ugly 'f type parameter is gone (simplifies types tremendously!)
1168       - ml-nlffigen changed accordingly
1169       - tutorial updated
1170    
1171    Other changes:
1172    
1173       - author's affiliation in CM manual(s) updated
1174       - some more recognized keywords added to Allen's sml.sty
1175    
1176    ----------------------------------------------------------------------
1177    Name: Matthias Blume
1178    Date: 2001/05/25 15:30:00 EDT
1179    Tag: blume-20010525-iptr
1180    Description:
1181    
1182      - put the official 110.33-README (as it appears on the ftp server) under
1183        CVS
1184      - fixed a small bug related to incomplete pointer types in
1185        ml-nlffigen
1186      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1187        (it does not need the 'f type parameter)
1188    
1189    ----------------------------------------------------------------------
1190    Name: Matthias Blume
1191    Date: 2001/05/23 14:30:00 EDT
1192    Tag: Release_110_33
1193    Description:
1194    
1195    New version number (110.33).  New bootfiles.
1196    
1197    ----------------------------------------------------------------------
1198    Name: Matthias Blume
1199    Date: 2001/05/22 18:06:00 EDT
1200    Tag: blume-20010522-targets
1201    Description:
1202    
1203    Made install.sh use file config/targets.customized if it exists, falling
1204    back to config/targets if it doesn't.  This way one can have a customized
1205    version of the targets file without touching the "real thing", thus
1206    eliminating the constant fear of accidentally checking something bogus
1207    back into the CVS repository...  (File config/targets.customized must
1208    not be added to the repository!)
1209    
1210    ----------------------------------------------------------------------
1211    Name: Matthias Blume
1212    Date: 2001/05/22 16:30:00 EDT
1213    Tag: blume-20010522-minitut
1214    Description:
1215    
1216    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1217       struct returns.
1218    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
1219       incomplete, preliminary documentation for NLFFI.
1220    
1221    ----------------------------------------------------------------------
1222    Name: Matthias Blume
1223  Date: 2001/05/14 11:30:00 EDT  Date: 2001/05/14 11:30:00 EDT
1224  Tag: blume-20010514-script  Tag: blume-20010514-script
1225  Description:  Description:
# Line 2246  Line 3452 
3452       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3453       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3454       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
3455       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3456       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3457       Dave's hands. :)       Dave's hands. :)
3458    

Legend:
Removed from v.829  
changed lines
  Added in v.1030

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