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 839, Thu Jun 7 20:28:44 2001 UTC revision 1044, Mon Jan 28 21:36:08 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: 2002/01/28
18    Tag: blume-20020128-allocParam
19    Description:
20    
21    John Reppy:  Changes c-calls API to accept client-callback for
22    allocating extra stack space.
23    me: Corresponding changes to mlriscGen (using a dummy argument that
24        does not change the current behavior).
25    
26    ----------------------------------------------------------------------
27    Name: Matthias Blume
28    Date: 2002/01/28 12:00:00
29    Tag: Release_110_38
30    Description:
31    
32    This time for real!!!
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume
36    Date: 2002/01/28 10:56:00 EST
37    Tag: blume-20020128-retraction
38    Description:
39    
40    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
41       with blume-Release_110_38-retracted.)
42    
43    1. Fixed a problem with incorrect rounding modes in real64.sml.
44       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
45    
46    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
47       and enums fixed.  The naming of corresponding ML identifiers should
48       now be consistent again.
49    
50    ----------------------------------------------------------------------
51    Name: Allen Leung
52    Date: 2002/01/27
53    Tag: leunga-20020127-nowhere
54    Description:
55    
56       Added a target called nowhere in the configuration scripts.
57       Enabling this will build the MLRISC 'nowhere' tool (for translating
58       programs with where-clauses into legal SML code) during installation.
59    
60    ----------------------------------------------------------------------
61    Name: Matthias Blume
62    Date: 2002/01/25 21:27:00 EST
63    Tag: blume-Release_110_38-retracted
64    Description:
65    
66    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
67    
68    README will be added later.
69    
70    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
71    (Release_110_38) removed.  Reason: Last-minute bug fixes.
72    
73    ----------------------------------------------------------------------
74    Name: Matthias Blume
75    Date: 2002/01/25
76    Tag: blume-20020125-ffi
77    Description:
78    
79    A large number of tweaks and improvements to ml-nlffi-lib and
80    ml-nlffigen:
81    
82       - ML represenation types have been streamlined
83       - getter and setter functions work with concrete values, not abstract
84         ones where possible
85       - ml-nlffigen command line more flexible (see README file there)
86       - some bugs have been fixed (hopefully)
87    
88    ----------------------------------------------------------------------
89    Name: Lal George
90    Date: 2002/01/24
91    Tag: george-20020124-risc-ra-interface
92    Description:
93    
94       There is a dramatic simplification in the interface to the
95       register allocator for RISC architectures as a result of making
96       parallel copy instructions explicit.
97    
98    ----------------------------------------------------------------------
99    Name: Matthias Blume
100    Date: 2002/01/22
101    Tag: blume-20020122-x86-ccalls
102    Description:
103    
104    Bug fix for c-calls on x86 (having to do with how char- and
105    short-arguments are being handled).
106    
107    ----------------------------------------------------------------------
108    Name: Matthias Blume
109    Date: 2002/01/21
110    Tag: blume-20020121-ff
111    Description:
112    
113    Another day of fiddling with the FFI...
114    
115    1. Bug fix/workaround:  CKIT does not complain about negative array
116       dimensions, so ml-nlffigen has to guard itself against this possibility.
117       (Otherwise a negative dimension would send it into an infinite loop.)
118    
119    2. Some of the abstract types (light objects, light pointers, most "base"
120       types) in structure C are now eqtypes.
121    
122    3. Added constructors and test functions for NULL function pointers.
123    
124    ----------------------------------------------------------------------
125    Name: Matthias Blume
126    Date: 2002/01/18
127    Tag: blume-20020118-ready-for-new-release
128    Description:
129    
130    Made config/srcarchiveurl point to a new place.  (Will provide boot
131    files shortly.)
132    
133    Maybe we christen this to be 110.38?
134    
135    ----------------------------------------------------------------------
136    Name: Matthias Blume
137    Date: 2002/01/18
138    Tag: blume-20020118-more-ffifiddle
139    Description:
140    
141    Today's FFI fiddling:
142    
143      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
144        C.Set.* that use concrete (MLRep.*) arguments and results instead
145        of abstract ones.
146    
147      - Provided word-style bit operations etc. for "int" representation
148        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
149        and Long.
150    
151    ----------------------------------------------------------------------
152    Name: Matthias Blume
153    Date: 2002/01/18
154    Tag: blume-20020118-use-x86-fp
155    Description:
156    
157    Now that x86-fast-fp seems to be working, I turned it back on again
158    by default.  (Seems to work fine now, even with the FFI.)
159    
160    Other than that, I added some documentation about the FFI to
161    src/ml-nlffigen/README and updated the FFI test examples in
162    src/ml-nlffi-lib/Tests/*.
163    
164    ----------------------------------------------------------------------
165    Name: Allen Leung
166    Date: 2002/01/17
167    Tag: leunga-20020117-x86-fast-fp-call
168    Description:
169    
170       1. Fixed a problem with handling return fp values when x86's fast fp
171          mode is turned on.
172    
173       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
174          of %f32.
175    
176       3. Added a constructor INT32lit to the ast of MLRISC tools.
177    
178    ----------------------------------------------------------------------
179    Name: Matthias Blume
180    Date: 2002/01/16
181    Tag: blume-20020116-ffifiddle
182    Description:
183    
184    More fiddling with the FFI interface:
185    
186     - Make constness 'c instead of rw wherever possible.  This eliminates
187       the need for certain explicit coercions.  (However, due to ML's
188       value polymorphism, there will still be many cases where explicit
189       coercions are necessary.  Phantom types are not the whole answer
190       to modeling a subtyping relationship in ML.)
191    
192     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
193    
194     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
195       clumsy workaround for varargs problem.)
196    
197    ----------------------------------------------------------------------
198    Name: Lal George
199    Date: 2002/01/15
200    Tag: <none>
201    Description:
202    
203    1. Since COPY instructions are no longer native to the architecture,
204       a generic functor can be used to implement the expandCopies function.
205    
206    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
207       TEXT segment.
208    
209    ----------------------------------------------------------------------
210    Name: Matthias Blume
211    Date: 2002/01/15
212    Tag: blume-20020115-ffiupdates
213    Description:
214    
215    1. Fix for bug resulting in single-precision float values being returned
216       incorrectly from FFI calls.
217    
218    2. Small modifications to C FFI API:
219    
220        - memory-allocation routines return straight objects (no options)
221          and raise an exception in out-of-memory situations
222        - unsafe extensions to cast between function pointers and pointers
223          from/to ints
224        - added structure C_Debug as an alternative to structure C where
225          pointer-dereferencing (|*| and |*!) always check for null-pointers
226        - added open_lib' to DynLinkage;  open_lib' works like open_lib
227          but also takes a (possibly empty) list of existing library handles
228          that the current library depends on
229    
230    ----------------------------------------------------------------------
231    Name: Matthias Blume
232    Date: 2002/01/10
233    Tag: blume-20020110-newffigen
234    Description:
235    
236    1. Updates to portable graph code.
237    
238    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
239       more scalable now so that even huge interfaces such as the one
240       for GTK compile in finite time and space. :-)
241       See src/ml-nlffigen/README for details on what's new.
242    
243    ----------------------------------------------------------------------
244    Name: Lal George
245    Date: 2001/01/09 14:31:35 EST 2002
246    Tag: george-20011206-rm-native-copy
247    Description:
248    
249            Removed the native COPY and FCOPY instructions
250            from all the architectures and replaced it with the
251            explicit COPY instruction from the previous commit.
252    
253            It is now possible to simplify many of the optimizations
254            modules that manipulate copies. This has not been
255            done in this change.
256    
257    ----------------------------------------------------------------------
258    Name: Lal George
259    Date: 2001/12/06 16:50:13 EST 2001
260    Tag: george-20011206-mlrisc-instruction
261    Description:
262    
263    Changed the representation of instructions from being fully abstract
264    to being partially concrete. That is to say:
265    
266      from
267            type instruction
268    
269      to
270            type instr                              (* machine instruction *)
271    
272            datatype instruction =
273                LIVE of {regs: C.cellset, spilled: C.cellset}
274              | KILL of {regs: C.cellset, spilled: C.cellset}
275              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
276              | ANNOTATION of {i: instruction, a: Annotations.annotation}
277              | INSTR of instr
278    
279    This makes the handling of certain special instructions that appear on
280    all architectures easier and uniform.
281    
282    LIVE and KILL say that a list of registers are live or killed at the
283    program point where they appear. No spill code is generated when an
284    element of the 'regs' field is spilled, but the register is moved to
285    the 'spilled' (which is present, more for debugging than anything else).
286    
287    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
288    We used to generate:
289    
290            DEFFREG f1
291            f1 := f2 + f3
292            trapb
293    
294    but now generate:
295    
296            f1 := f2 + f3
297            trapb
298            LIVE {regs=[f1,f2,f3], spilled=[]}
299    
300    Furthermore, the DEFFREG (hack) required that all floating point instruction
301    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
302    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
303    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
304    this architecture).
305    
306    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
307    all the architectures. This will result in further simplification of the
308    register allocator that must be aware of them for coalescing purposes, and
309    will also simplify certain aspects of the machine description that provides
310    callbacks related to parallel copies.
311    
312    ANNOTATION should be obvious, and now INSTR represents the honest to God
313    machine instruction set!
314    
315    The <arch>/instructions/<arch>Instr.sml files define certain utility
316    functions for making porting easier -- essentially converting upper case
317    to lower case. All machine instructions (of type instr) are in upper case,
318    and the lower case form generates an MLRISC instruction. For example on
319    the alpha we have:
320    
321      datatype instr =
322         LDA of {r:cell, b:cell, d:operand}
323       | ...
324    
325      val lda : {r:cell, b:cell, d:operand} -> instruction
326        ...
327    
328    where lda is just (INSTR o LDA), etc.
329    
330    ----------------------------------------------------------------------
331    Name: Matthias Blume
332    Date: 2001/11/22 21:40:00 EST
333    Tag: Release_110_37
334    Description:
335    
336    Release 110.37.  This time for real.
337    
338    ----------------------------------------------------------------------
339    Name: Matthias Blume
340    Date: 2001/11/21 16:35:00 EST
341    Tag: blume-20011121-foot-in-mouth
342    Description:
343    
344    Removed the "Release_110_37" tag because of a serious bug.
345    This will be re-tagged once the bug is fixed.
346    
347    ----------------------------------------------------------------------
348    Name: Matthias Blume
349    Date: 2001/11/21 16:14:00 EST
350    Tag: blume-20011121-forgottenfile
351    Description:
352    
353    Forgot to add a file.  (Just a .tex-file -- part of
354    the CM manual source.)
355    
356    ----------------------------------------------------------------------
357    Name: Matthias Blume
358    Date: 2001/11/21 16:10:00 EST
359    Tag: blume-20011121-invalid_110_37
360    Description:
361    
362    Note: I removed the original tag "Release_110_37" from this commit
363          because we found a serious bug in all non-x86 backends.
364          - Matthias
365    
366    1. Modifications to the SML/NJ code generator and to the runtime system
367       so that code object name strings are directly inserted into code
368       objects at code generation time.  The only business the runtime system
369       has with this is now to read the name strings on occasions.
370       (The encoding of the name string has also changed somewhat.)
371    
372    2. CM now implements a simple "set calculus" for specifying export lists.
373       In particular, it is now possible to refer to the export lists of
374       other libraries/groups/sources and form unions as well as differences.
375       See the latest CM manual for details.
376    
377    3. An separate notion of "proxy" libraries has again be eliminated from
378       CM's model.  (Proxy libraries are now simply a special case of using
379       the export list calculus.)
380    
381    4. Some of the existing libraries now take advantage of the new set
382       calculus.
383       (Notice that not all libraries have been converted because some
384       of the existing .cm-files are supposed to be backward compatible
385       with 110.0.x.)
386    
387    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
388       "General.exnMessage"!  The former relies on a certain hook to be
389       initialized, and that often does not happen in the stand-alone case.)
390    
391    ----------------------------------------------------------------------
392    Name: Lal George
393    Date: 2001/11/21  13:56:18 EST
394    Tag: george-2001121-pseudo-ops
395    Description:
396    
397      Implemented a complete redesign of MLRISC pseudo-ops. Now there
398      ought to never be any question of incompatabilities with
399      pseudo-op syntax expected by host assemblers.
400    
401      For now, only modules supporting GAS syntax are implemented
402      but more should follow, such as MASM, and vendor assembler
403      syntax, e.g. IBM as, Sun as, etc.
404    
405    ----------------------------------------------------------------------
406    Name: Matthias Blume
407    Date: 2001/11/14 11:52:00 EST
408    Tag: blume-20011114-srcname
409    Description:
410    
411    1. Routed the name of the current source file to mlriscgen where it
412       should be directly emitted into the code object.  (This last part
413       is yet to be done.)
414    
415    2. Some cleanup of the pgraph code to make it match the proposal that
416       I put out the other day.  (The proposal notwithstanding, things are
417       still in flux here.)
418    
419    ----------------------------------------------------------------------
420    Name: Lal George
421    Date: 2001/11/14 09:44:04 EST
422    Tag:
423    Description:
424    
425      Fix for a backpatching bug reported by Allen.
426    
427      Because the boundary between short and long span-dependent
428      instructions is +/- 128, there are an astounding number of
429      span-dependent instructions whose size is over estimated.
430    
431      Allen came up with the idea of letting the size of span
432      dependent instructions be non-monotonic, for a maxIter
433      number of times, after which the size must be monotonically
434      increasing.
435    
436      This table shows the number of span-dependent instructions
437      whose size was over-estimated as a function of maxIter, for the
438      file Parse/parse/ml.grm.sml:
439    
440         maxIter            # of instructions:
441            10                      687
442            20                      438
443            30                      198
444            40                        0
445    
446      In compiling the compiler, there is no significant difference in
447      compilation speed between maxIter=10 and maxIter=40. Actually,
448      my measurements showed that maxIter=40 was a tad faster than
449      maxIter=10! Also 96% of the  files in the compiler reach a fix
450      point within 13 iterations, so fixing maxIter at 40, while high,
451      is okay.
452    
453    ----------------------------------------------------------------------
454    Name: Matthias Blume
455    Date: 2001/10/31 15:25:00 EST
456    Tag: blume-20011031-pgraph
457    Description:
458    
459    CKIT:
460    * Changed the "Function" constructor of type Ast.ctype to carry optional
461      argument identifiers.
462    * Changed the return type of TypeUtil.getFunction accordingly.
463    * Type equality ignores the argument names.
464    * TypeUtil.composite tries to preserve argument names but gives up quickly
465      if there is a mismatch.
466    
467    installation script:
468    * attempts to use "curl" if available (unless "wget" is available as well)
469    
470    CM:
471    * has an experimental implementation of "portable graphs" which I will
472      soon propose as an implementation-independent library format
473    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
474    
475    NLFFI-LIB:
476    * some cleanup (all cosmetic)
477    
478    NLFFIGEN:
479    * temporarily disabled the mechanism that suppresses ML output for
480      C definitions whose identifiers start with an underscore character
481    * generate val bindings for enum constants
482    * user can request that only one style (light or heavy) is being used;
483      default is to use both (command-line arguments: -heavy and -light)
484    * fixed bug in handling of function types involving incomplete pointers
485    * generate ML entry points that take record arguments (i.e., using
486      named arguments) for C functions that have a prototype with named
487      arguments
488      (see changes to CKIT)
489    
490    ----------------------------------------------------------------------
491    Name: Allen Leung
492    Date: 2001/10/27 20:34:00 EDT
493    Tag: leunga-20011027-x86-fast-fp-call
494    Description:
495    
496       Fixed the bug described in blume-20010920-slowfp.
497    
498       The fix involves
499          1. generating FCOPYs in FSTP in ia32-svid
500          2. marking a CALL with the appropriate annotation
501    
502    ----------------------------------------------------------------------
503    Name: Matthias Blume
504    Date: 2001/10/16 11:32:00 EDT
505    Tag: blume-20011016-netbsd
506    Description:
507    
508    Underscore patch from Chris Richards (fixing problem with compiling
509    runtime system under recent NetBSD).
510    
511    ----------------------------------------------------------------------
512    Name: Allen Leung
513    Date: 2001/10/12 17:18:32 EDT 2001
514    Tag: leung-20011012-x86-printflowgraph
515    Description:
516    
517    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
518    
519    ----------------------------------------------------------------------
520    Name: Lal George
521    Date: 2001/10/11 23:51:34 EDT
522    Tag: george-20011011-too-many-instrs
523    Description:
524    
525    The representation of a program point never expected to see more
526    than 65536 instructions in a basic block!
527    
528    ----------------------------------------------------------------------
529    Name: Lal George
530    Date: 2001/10/09 09:41:37 EDT
531    Tag: george-20011008-mlrisc-labels
532    Description:
533    
534    Changed the machine description files to support printing of
535    local and global labels in assembly code, based on host assembler
536    conventions.
537    
538    ----------------------------------------------------------------------
539    Name: Matthias Blume
540    Date: 2001/09/25 15:25:00 EDT
541    Tag: blume-20010925-exninfo
542    Description:
543    
544    I provided a non-hook implementation of exnName (at the toplevel) and
545    made the "dummy" implementation of exnMessage (at the toplevel) more
546    useful: if nothing gets "hooked in", then at least you are going to
547    see the exception name and a message indicating why you don't see more.
548    
549    [For the time being, programs that need exnMessage and want to use
550    ml-build should either use General.exnMessage (strongly recommended) or
551    refer to structure General at some other point so that CM sees a
552    static dependency.]
553    
554    [Similar remarks go for "print" and "use":  If you want to use their
555    functionality in stand-alone programs generated by ml-build, then use
556    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
557    
558    ----------------------------------------------------------------------
559    Name: Matthias Blume
560    Date: 2001/09/20 17:28:00 EDT
561    Tag: blume-20010920-slowfp
562    Description:
563    
564    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
565    
566    ----------------------------------------------------------------------
567    Name: Matthias Blume
568    Date: 2001/09/20 17:20:00 EDT
569    Tag: blume-20010920-canonicalpaths
570    Description:
571    
572    0. Updated the BOOT file (something that I forgot to do earlier).
573    
574    1. Small internal change to CM so that it avoids "/../" in filenames
575       as much as possible (but only where it is safe).
576    
577    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
578       that arguments that contain delimiters are passed through correctly.
579       This change also means that all "special" arguments of the form
580       @SMLxxx... must come first.
581    
582    3. Changed install script to put relative anchor names for tool commands
583       into pathconfig.
584    
585    ----------------------------------------------------------------------
586    Name: Matthias Blume
587    >>>>>>> 1.169
588    Date: 2001/09/18 15:35:00 EDT
589    Tag: blume-20010918-readme11036
590    Description:
591    
592    Added README files.
593    
594    ----------------------------------------------------------------------
595    Name: Matthias Blume
596    Date: 2001/09/18 11:45:00 EDT
597    Tag: Release_110_36 (retag)
598    Description:
599    
600    Fixed mistake in config/preloads. Retagged as 110.36.
601    
602    ----------------------------------------------------------------------
603    Name: Matthias Blume
604    Date: 2001/09/18 09:40:00 EDT
605    Tag: Release_110_36_orig (tag changed)
606    Description:
607    
608    New version (110.36).  New bootfiles.
609    
610    ----------------------------------------------------------------------
611    Name: Matthias Blume
612    Date: 2001/09/14 16:15:00 EDT
613    Tag: blume-20010914-x86fastfp
614    Description:
615    
616    John committed some changes that Allen made, in particular a (hopefully)
617    correctly working version of the x86-fp module.
618    
619    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
620    flag to "true".  Everything seems to compile to a fixpoint ok, and
621    "mandelbrot" speeds up by about 15%.
622    
623    ----------------------------------------------------------------------
624    Name: Matthias Blume
625    Date: 2001/09/13 11:20:00 EDT
626    Tag: blume-20010913-minimal
627    Description:
628    
629    1. Stefan Monnier's patch to fix a miscompilation problem that
630       was brought to light by John Reppy's work on Moby.
631    
632    2. Implemented a minimal "structure Compiler" that contains just
633       "version" and "architecture".  The minimal version will be
634       available when the full version is not.  This is for backward-
635       compatibility with code that wants to test Compiler.version.
636    
637    ----------------------------------------------------------------------
638    Name: Matthias Blume
639    Date: 2001/08/28 14:03:00 EDT
640    Tag: blume-20010828-ml-lex
641    Description:
642    
643    Fix for bug 1581, received from Neophytos Michael.
644    
645    ----------------------------------------------------------------------
646    Name: Matthias Blume
647    Date: 2001/08/27 11:20:00 EDT
648    Tag: blume-20010827-readme11035
649    Description:
650    
651    Fleshed out the README file for 110.35.
652    
653    ----------------------------------------------------------------------
654    Name: Matthias Blume
655    Date: 2001/08/24 17:10:00 EDT
656    Tag: Release_110_35
657    Description:
658    
659    New version number (110.35).  New bootfiles.
660    
661    ----------------------------------------------------------------------
662    Name: Lal George
663    Date: 2001/08/24 13:47:18 EDT 2001
664    Tag: george-20010824-MLRISC-graphs
665    Description:
666    
667     removed clusters from MLRISC completely and replaced with graphs.
668    
669    ----------------------------------------------------------------------
670    Name: Matthias Blume
671    Date: 2001/08/23 17:50:00 EDT
672    Tag: blume-20010823-toplevel
673    Description:
674    
675    - some reorganization of the code that implements various kinds of
676      environments in the compiler (static, dynamic, symbolic, combined)
677    - re-implemented the EnvRef module so that evalStream works properly
678      (if the stream contains references to "use", "CM.make", etc.)
679    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
680    
681    ----------------------------------------------------------------------
682    Name: Matthias Blume
683    Date: 2001/08/20 15:50 EDT
684    Tag: blume20010820-slipup
685    Description:
686    
687    I forgot to commit a few files.  Here they are...
688    
689    ----------------------------------------------------------------------
690    Name: Matthias Blume
691    Date: 2001/08/20 15:35:00 EDT
692    Tag: blume-20010820-debugprof
693    Description:
694    
695    !!!! NEW BOOTFILES !!!!
696    
697    This is another round of reorganizing the compiler sources.  This
698    time the main goal was to factor out all the "instrumentation"
699    passes (for profiling and backtracing) into their own library.
700    The difficulty was to do it in such a way that it does not depend
701    on elaborate.cm but only on elabdata.cm.
702    
703    Therefore there have been further changes to both elaborate.cm and
704    elabdata.cm -- more "generic" things have been moved from the former
705    to the latter.  As a result, I was forced to split the assignment
706    of numbers indicating "primtyc"s into two portions: SML-generic and
707    SML/NJ-specific.  Since it would have been awkward to maintain,
708    I bit the bullet and actually _changed_ the mapping between these
709    numbers and primtycs.  The bottom line of this is that you need
710    a new set of bin- and bootfiles.
711    
712    I have built new bootfiles for all architectures, so doing a fresh
713    checkout and config/install.sh should be all you need.
714    
715    The newly created library's name is
716    
717        $smlnj/viscomp/debugprof.cm
718    
719    and its sources live under
720    
721        src/compiler/DebugProf
722    
723    ----------------------------------------------------------------------
724    Name: Matthias Blume
725    Date: 2001/08/15 17:15:00 EDT
726    Tag: blume-20010815-compreorg
727    Description:
728    
729    This is a first cut at reorganizing the CM libraries that make up the
730    core of the compiler.  The idea is to separate out pieces that could
731    be used independently by tools, e.g., the parser, the typechecker, etc.
732    
733    The current status is a step in this direction, but it is not quite
734    satisfactory yet.  Expect more changes in the future.
735    
736    Here is the current (new) organization...
737    
738        What used to be $smlnj/viscomp/core.cm is now divided into
739        six CM libraries:
740    
741             $smlnj/viscomp/basics.cm
742                           /parser.cm
743                           /elabdata.cm
744                           /elaborate.cm
745                           /execute.cm
746                           /core.cm
747    
748        The CM files for these libraries live under src/system/smlnj/viscomp.
749        All these libraries are proxy libraries that contain precisely
750        one CM library component.  Here are the locations of the components
751        (all within the src/compiler tree):
752    
753             Basics/basics.cm
754             Parse/parser.cm
755             ElabData/elabdata.cm
756             Elaborator/elaborate.cm
757             Execution/execute.cm
758             core.cm
759    
760         [This organization is the same that has been used already
761         for a while for the architecture-specific parts of the visible
762         compiler and for the old version of core.cm.]
763    
764         As you will notice, many source files have been moved from their
765         respective original locations to a new home in one of the above
766         subtrees.
767    
768         The division of labor between the new libraries is the following:
769    
770             basics.cm:
771                - Simple, basic definitions that pertain to many (or all) of
772                  the other libraries.
773             parser.cm:
774                - The SML parser, producing output of type Ast.dec.
775                - The type family for Ast is also defined and exported here.
776             elabdata.cm:
777                - The datatypes that describe input and output of the elaborator.
778                  This includes types, absyn, and static environments.
779             elaborator.cm:
780                - The SML/NJ type checker and elaborator.
781                  This maps an Ast.dec (with a given static environment) to
782                  an Absyn.dec (with a new static environment).
783                - This libraries implements certain modules that used to be
784                  structures as functors (to remove dependencies on FLINT).
785             execute.cm:
786                - Everything having to do with executing binary code objects.
787                - Dynamic environments.
788             core.cm:
789                - SML/NJ-specific instantiations of the elaborator and MLRISC.
790                - Top-level modules.
791                - FLINT (this should eventually become its own library)
792    
793    Notes:
794    
795    I am not 100% happy with the way I separated the elaborator (and its
796    data structures) from FLINT.  Two instances of the same problem:
797    
798        1. Data structures contain certain fields that carry FLINT-specific
799           information.  I hacked around this using exn and the property list
800           module from smlnj-lib.  But the fact that there are middle-end
801           specific fields around at all is a bit annoying.
802    
803        2. The elaborator calculates certain FLINT-related information.  I tried
804           to make this as abstract as I could using functorization, but, again,
805           the fact that the elaborator has to perform calculations on behalf
806           of the middle-end at all is not nice.
807    
808        3. Having to used exn and property lists is unfortunate because it
809           weakens type checking.  The other alternative (parameterizing
810           nearly *everything*) is not appealing, though.
811    
812    I removed the "rebinding =" warning hack because due to the new organization
813    it was awkward to maintain it.  As a result, the compiler now issues some of
814    these warnings when compiling init.cmi during bootstrap compilation. On
815    the plus side, you also get a warning when you do, for example:
816       val op = = Int32.+
817    which was not the case up to now.
818    
819    I placed "assign" and "deref" into the _Core structure so that the
820    code that deals with the "lazy" keyword can find them there.  This
821    removes the need for having access to the primitive environment
822    during elaboration.
823    
824    ----------------------------------------------------------------------
825    Name: Matthias Blume
826    Date: 2001/08/13
827    Tag: blume-20010813-closures
828    Description:
829    
830    This fix was sent to us by Zhong Shao.  It is supposed to improve the
831    performance of certain loops by avoiding needless closure allocation.
832    
833    ----------------------------------------------------------------------
834    Name: Lal George
835    Date: 2001/07/31 10:03:23 EDT 2001
836    Tag: george-20010731-x86-fmalloc
837    Description: Fixed bug in x86 calls
838    
839        There was a bug where call instructions would mysteriously
840        vanish. The call instruction had to be one that returned
841        a floating point value.
842    
843    ----------------------------------------------------------------------
844    Name: Lal George
845    Date: 2001/07/19 16:36:29 EDT 2001
846    Tag: george-20010719-simple-cells
847    Description:
848    
849    I have dramatically simplified the interface for CELLS in MLRISC.
850    
851    In summary, the cells interface is broken up into three parts:
852    
853      1. CellsBasis : CELLS_BASIS
854    
855            CellsBasis is a top level structure and common for all
856            architectures.  it contains the definitions of basic datatypes
857            and utility  functions over these types.
858    
859      2. functor Cells() : CELLS
860    
861            Cells generates an interface for CELLS that incorporates the
862            specific resources on the target architecture, such as the
863            presence of special register classes, their number and size,
864            and various useful substructures.
865    
866      3. <ARCH>CELLS
867    
868            e.g. SparcCells: SPARCCELLS
869    
870            <ARCH>CELLS usually contains additional bindings for special
871            registers  on the architecture, such as:
872    
873                    val r0 : cell           (* register zero *)
874                    val y : cell            (* Y register *)
875                    val psr : cell          (* processor status register *)
876                    ...
877    
878            The structure returned by applying the Cells functor is opened
879            in this interface.
880    
881    The main implication of all this is that the datatypes for cells is
882    split between CellsBasis and CELLS -- a fairly simple change for user
883    code.
884    
885    In the old scheme the CELLS interface had a definitional binding of
886    the form:
887    
888            signature CELLS = sig
889    
890               structure CellsBasis = CellsBasis
891    
892               ...
893    
894            end
895    
896    With all the sharing constraints that goes on in MLRISC, this old
897    design  quickly leads to errors such as:
898    
899            "structure definition spec inside of sharing ... "
900    
901    
902    and appears to require an unacceptable amount of sharing and where
903    constraint hackery.
904    
905    I think this error message (the interaction of definitional specs and
906    sharing) requires more explanation on our web page.
907    
908    ----------------------------------------------------------------------
909    Name: Matthias Blume
910    Date: 2001/07/19 15:00:00 EDT
911    Tag: blume-20010719-libreorg
912    Description:
913    
914    This update puts together a fairly extensive but straightforward change
915    to the way the libraries that implement the interactive system are
916    organized:
917    
918       The biggest change is the elimination of structure Compiler.  As a
919       replacement for this structure, there is now a CM library
920       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
921       that exports all the substructures of the original structure Compiler
922       directly.  So instead of saying Compiler.Foo.bar one now simply
923       says Foo.bar.  (The CM libraries actually export a collection of
924       structures that is richer than the collection of substructures of
925       structure Compiler.)
926    
927       To make the transition smooth, there is a separate library called
928       $smlnj/compiler/compiler.cm which puts together and exports the
929       original structure Compiler (or at least something very close to it).
930    
931       There are five members of the original structure Compiler
932       that are not exported directly but which instead became members
933       of a new structure Backend (described by signature BACKEND).  These are:
934       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
935       Interact (: INTERACT), structure Machine (: MACHINE), and val
936       architecture (: string).
937    
938       Structure Compiler.Version has become structure CompilerVersion.
939    
940       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
941       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
942       or x86, respectively.
943       Each of these exports the same frontend structures that
944       $smlnj/compiler.cm exports.  But they do not have a structure Backend
945       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
946       Hppa, PPC, Sparc, or X86, respectively.
947    
948       Library $smlnj/compiler/all.cm exports the union of the exports of
949       $smlnj/compiler/<arch>.cm
950    
951       There are no structures <Arch>Compiler anymore, use
952       $smlnj/compiler/<arch>.cm instead.
953    
954       Library host-compiler-0.cm is gone.  Instead, the internal library
955       that instantiates CM is now called cm0.cm.  Selection of the host
956       compiler (backend) is no longer done here but. (Responsibility for it
957       now lies with $smlnj/compiler/current.cm.  This seems to be more
958       logical.)
959    
960       Many individual files have been moved or renamed.  Some files have
961       been split into multiple files, and some "dead" files have been deleted.
962    
963    Aside from these changes to library organization, there are also changes
964    to the way the code itself is organized:
965    
966       Structure Binfile has been re-implemented in such a way that it no
967       longer needs any knowledge of the compiler.  It exclusively deals
968       with the details of binfile layout.  It no longer invokes the
969       compiler (for the purpose of creating new prospective binfile
970       content), and it no longer has any knowledge of how to interpret
971       pickles.
972    
973       Structure Compile (: COMPILE) has been stripped down to the bare
974       essentials of compilation.  It no longer deals with linking/execution.
975       The interface has been cleaned up considerably.
976    
977       Utility routines for dealing with linking and execution have been
978       moved into their own substructures.
979    
980       (The ultimate goal of these changes is to provide a light-weight
981       binfile loader/linker (at least for, e.g., stable libraries) that
982       does not require CM or the compiler to be present.)
983    
984    CM documentation has been updated to reflect the changes to library
985    organization.
986    
987    ----------------------------------------------------------------------
988    Name: Matthias Blume
989    Date: 2001/07/10 17:30:00 EDT
990    Tag: Release_110_34
991    Description:
992    
993    Minor tweak to 110.34 (re-tagged):
994    
995      - README.html file added to CVS repository
996      - runtime compiles properly under FreeBSD 3.X and 4.X
997    
998    ----------------------------------------------------------------------
999    Name: Matthias Blume
1000    Date: 2001/07/10 17:30:00 EDT
1001    Tag: Release_110_34
1002    Description:
1003    
1004    New version number (110.34). New bootfiles.
1005    
1006    ----------------------------------------------------------------------
1007    Name: Matthias Blume
1008    Date: 2001/07/09 16:00:00 EDT
1009    Tag: blume-20010709-more-varargs
1010    Description:
1011    
1012    I changed the handling of varargs in ml-nlffigen again:
1013    The ellipsis ... will now simply be ignored (with an accompanying warning).
1014    
1015    The immediate effect is that you can actually call a varargs function
1016    from ML -- but you can't actually supply any arguments beyond the ones
1017    specified explicitly.  (For example, you can call printf with its format
1018    string, but you cannot pass additional arguments.)
1019    
1020    This behavior is only marginally more useful than the one before, but
1021    it has the advantage that a function or, more importantly, a function
1022    type never gets dropped on the floor, thus avoiding follow-up problems with
1023    other types that refer to the offending one.
1024    
1025    ----------------------------------------------------------------------
1026    Name: Matthias Blume
1027    Date: 2001/07/09 11:25:00 EDT
1028    Tag: blume-20010709-varargs
1029    Description:
1030    
1031    1. ckit-lib.cm now exports structure Error
1032    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1033       with a warning accompanied by a source location.  Moreover, it
1034       merely skips the offending function or type and proceeds with the
1035       rest of its work.u  As a result, one can safely feed C code containing
1036       "..." to ml-nlffigen.
1037    3. There are some internal improvements to CM, providing slightly
1038       more general string substitutions in the tools subsystem.
1039    
1040    ----------------------------------------------------------------------
1041    Name: Matthias Blume
1042    Date: 2001/06/27 15:10:00 EDT
1043    Tag: blume-20010627-concur
1044    Description:
1045    
1046    Fixed a small bug in CM's handling of parallel compilation.
1047    (You could observe the bug by Control-C-interrupting an ordinary
1048    CMB.make or CM.stabilize and then attaching some compile servers.
1049    The result was that all of a sudden the previously interrupted
1050    compilation would continue on its own.  This was because of
1051    an over-optimization: CM did not bother to clean out certain queues
1052    when no servers were attached "anyway", resulting in the contents
1053    of these queues to grab control when new servers did get attached.)
1054    
1055    There is also another minor update to the CM manual.
1056    
1057    ----------------------------------------------------------------------
1058    Name: Matthias Blume
1059    Date: 2001/06/26 16:15:00 EDT
1060    Tag: blume-20010626-cmdoc
1061    Description:
1062    
1063    Minor typo fixed in CM manual (syntax diagram for libraries).
1064    
1065    ----------------------------------------------------------------------
1066    Name: Matthias Blume
1067    Date: 2001/06/25 22:55:00 EDT
1068    Tag: blume-20010625-x86pc
1069    Description:
1070    
1071    Fixed a nasty bug in the X86 assembly code that caused signal
1072    handlers to fail (crash) randomly.
1073    
1074    ----------------------------------------------------------------------
1075    Name: Matthias Blume
1076    Date: 2001/06/25 12:05:00 EDT
1077    Tag: blume-20010625-nlffigen
1078    Description:
1079    
1080    This update fixes a number of minor bugs in ml-nlffigen as reported by
1081    Nick Carter <nbc@andrew.cmu.edu>.
1082    
1083      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1084      2. Default names for generated files are now derived from the name of
1085         the C file *without its directory*.  In particular, this causes generated
1086         files to be placed locally even if the C file is in some system directory.
1087      3. Default names for generated signatures and structures are also derived
1088         from the C file name without its directory.  This avoids silly things
1089         like "structure GL/GL".
1090         (Other silly names are still possible because ml-nlffigen does not do
1091          a thorough check of whether generated names are legal ML identifiers.
1092          When in doubt, use command line arguments to force particular names.)
1093    
1094    ----------------------------------------------------------------------
1095    Name: Matthias Blume
1096    Date: 2001/06/21 12:25:00 EDT
1097    Tag: blume-20010621-eXene
1098    Description:
1099    
1100    eXene now compiles and (sort of) works again.
1101    
1102    The library name (for version > 110.33) is $/eXene.cm.
1103    
1104    I also added an new example in src/eXene/examples/nbody.  See the
1105    README file there for details.
1106    
1107    ----------------------------------------------------------------------
1108    Name: Matthias Blume
1109    Date: 2001/06/20 16:40:00 EDT
1110    Tag: blume-20010620-cml
1111    Description:
1112    
1113    CML now compiles and works again.
1114    
1115    Libraries (for version > 110.33):
1116    
1117      $cml/cml.cm            Main CML library.
1118      $cml/basis.cm          CML's version of $/basis.cm.
1119      $cml/cml-internal.cm   Internal helper library.
1120      $cml/core-cml.cm       Internal helper library.
1121      $cml-lib/trace-cml.cm  Tracing facility.
1122      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1123    
1124    The installer (config/install.sh) has been taught how to properly
1125    install this stuff.
1126    
1127    ----------------------------------------------------------------------
1128    Name: Matthias Blume
1129    Date: 2001/06/19 17:55:00 EDT
1130    Tag: blume-20010619-instantiate
1131    Description:
1132    
1133    This un-breaks the fix for bug 1432.
1134    (The bug was originally fixed in 110.9 but I broke it again some
1135    time after that.)
1136    
1137    ----------------------------------------------------------------------
1138    Name: Matthias Blume
1139    Date: 2001/06/19 17:25:00 EDT
1140    Tag: blume-20010619-signals
1141    Description:
1142    
1143    This should (hopefully) fix the long-standing signal handling bug.
1144    (The runtime system was constructing a continuation record with an
1145    incorrect descriptor which would cause the GC to drop data on the floor...)
1146    
1147    ----------------------------------------------------------------------
1148    Name: Matthias Blume
1149    Date: 2001/06/15 15:05:00 EDT
1150    Tag: blume-20010615-moresparc
1151    Description:
1152    
1153    Here is a short late-hour update related to Sparc c-calls:
1154    
1155     -- made handling of double-word arguments a bit smarter
1156    
1157     -- instruction selection phase tries to collapse certain clumsily
1158        constructed ML-Trees; typical example:
1159    
1160            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1161    
1162        This currently has no further impact on SML/NJ since mlriscGen does
1163        not seem to generate such patterns in the first place, and c-calls
1164        (which did generate them in the beginning) has meanwhile been fixed
1165        so as to avoid them as well.
1166    
1167    ----------------------------------------------------------------------
1168    Name: Matthias Blume
1169    Date: 2001/06/15 15:05:00 EDT
1170    Tag: blume-20010615-sparc
1171    Description:
1172    
1173    The purpose of this update is to provide an implementation of NLFFI
1174    on Sparc machines.
1175    
1176    Here are the changes in detail:
1177    
1178       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1179       the Sparc implementation of the c-calls API.
1180       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1181       for accessing the ML frame.  Thus, we have a real frame pointer and
1182       can freely modify %sp without need for an omit-frame-ptr phase.
1183       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1184       this case.
1185       * ml-nlffigen has been taught to produce code for different architectures
1186       and calling conventions.
1187       * In a way similar to what was done in the x86 case, the Sparc
1188       backend uses its own specific extension to mltree.  (For example,
1189       it needs to be able to generate UNIMP instructions which are part
1190       of the calling convention.)
1191       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1192       to make it easier to plug in new machine- and os-dependent parts).
1193    
1194    There are some other fairly unrelated bug fixes and cleanups as well:
1195    
1196       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1197       that they properly share their libraries with existing SML/NJ libraries.
1198       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1199       follow-up error messages.
1200       * Updates to CM/CMB documentation.
1201    
1202    TODO items:
1203    
1204       * MLRISC should use a different register as its asmTemp on the Sparc.
1205         (The current %o2 is a really bad choice because it is part of the
1206         calling conventions, so things might interfere in unexpected ways.)
1207    
1208    ----------------------------------------------------------------------
1209    Name: Matthias Blume
1210  Date: 2001/06/07  Date: 2001/06/07
1211  Tag: blume-20010607-calls  Tag: blume-20010607-calls
1212  Description:  Description:
# Line 2368  Line 3561 
3561       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3562       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3563       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
3564       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3565       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3566       Dave's hands. :)       Dave's hands. :)
3567    

Legend:
Removed from v.839  
changed lines
  Added in v.1044

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