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 721, Mon Nov 6 14:03:39 2000 UTC revision 1068, Fri Feb 15 19:18:00 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
17    Date: 2002/02/15 14:17:00 EST
18    Tag: blume-20020215-showbindings
19    Description:
20    
21    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
22    the latter can be useful for exploring what bindings are available at
23    the interactive prompt.  (The first function returns only the list
24    of symbols that are really bound, the second prints those but also the
25    ones that CM's autoloading mechanism knows about.)
26    
27    ----------------------------------------------------------------------
28    Name: Matthias Blume
29    Date: 2002/02/15 12:08:00 EST
30    Tag: blume-20020215-iptrs
31    Description:
32    
33    Two improvements to ml-nlffigen:
34    
35      1. Write files only if they do not exist or if their current contents
36         do not coincide with what's being written.  (That is, avoid messing
37         with the time stamps unless absolutely necessary.)
38    
39      2. Implement a "repository" mechanism for generated files related
40         to "incomplete pointer types".   See the README file for details.
41    
42    ----------------------------------------------------------------------
43    Name: Matthias Blume
44    Date: 2002/02/14 11:50:00 EST
45    Tag: blume-20020214-quote
46    Description:
47    
48    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
49    because of the new and improved tag generation scheme.  (Thanks to Allen
50    Leung for pointing it out.)
51    
52    ----------------------------------------------------------------------
53    Name: Lal George
54    Date: 2002/02/14 09:55:27 EST 2002
55    Tag: george-20020214-isabelle-bug
56    Description:
57    
58    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
59    of Isabelle on the x86.
60    
61    From Allen:
62    -----------
63     I've found the problem:
64    
65         in ra-core.sml, I use the counter "blocked" to keep track of the
66         true number of elements in the freeze queue.  When the counter goes
67         to zero, I skip examining the queue.  But I've messed up the
68         bookkeeping in combine():
69    
70             else ();
71             case !ucol of
72               PSEUDO => (if !cntv > 0 then
73                     (if !cntu > 0 then blocked := !blocked - 1 else ();
74                                        ^^^^^^^^^^^^^^^^^^^^^^^
75                      moveu := mergeMoveList(!movev, !moveu)
76                     )
77                  else ();
78    
79         combine() is called to coalesce two nodes u and v.
80         I think I was thinking that if the move counts of u and v are both
81         greater than zero then after they are coalesced then one node is
82         removed from the freeze queue.  Apparently I was thinking that
83         both u and v are of low degree, but that's clearly not necessarily true.
84    
85    
86    02/12/2002:
87        Here's the patch.  HOL now compiles.
88    
89        I don't know how this impact on performance (compile
90        time or runtime).  This bug caused the RA (especially on the x86)
91        to go thru the potential spill phase when there are still nodes on the
92        freeze queue.
93    
94    
95    
96    
97    ----------------------------------------------------------------------
98    Name: Matthias Blume
99    Date: 2002/02/13 22:40:00 EST
100    Tag: blume-20020213-fptr-rtti
101    Description:
102    
103    Fixed a bug in ml-nlffigen that was introduced with one of the previous
104    updates.
105    
106    ----------------------------------------------------------------------
107    Name: Matthias Blume
108    Date: 2002/02/13 16:41:00 EST
109    Tag: blume-20020213-cmlpq
110    Description:
111    
112    Added new priority queue export symbols (which have just been added to
113    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
114    not compile and the installer would choke.)
115    
116    ----------------------------------------------------------------------
117    Name: Matthias Blume
118    Date: 2002/02/13 16:15:00 EST
119    Tag: blume-20020213-various
120    Description:
121    
122    1. More tweaks to ml-nlffigen:
123    
124       - better internal datastructures (resulting in slight speedup)
125       - "-match" option requires exact match
126       - "localized" gensym counters (untagged structs/unions nested within
127         other structs/unions or within typedefs get a fresh counter; their
128         tag will be prefixed by a concatenation of their parents' tags)
129       - bug fixes (related to calculation of transitive closure of types
130         to be included in the output)
131    
132    2. Minor Basis updates:
133    
134       - added implementations for List.collate and Option.app
135    
136    ----------------------------------------------------------------------
137    Name: Matthias Blume
138    Date: 2002/02/11 15:55:00 EST
139    Tag: blume-20020211-gensym
140    Description:
141    
142    Added a "-gensym" option to command line of ml-nlffigen.  This can be
143    used to specify a "stem" -- a string that is inserted in all "gensym'd"
144    names (ML structure names that correspond to unnamed C structs, unions,
145    and enums), so that separate runs of ml-nlffigen do not clash.
146    
147    ----------------------------------------------------------------------
148    Name: Matthias Blume
149    Date: 2002/02/11 12:05:00 EST
150    Tag: blume-20020211-gensml
151    Description:
152    
153    A quick fix for a problem with GenSML (in the pgraph-util library):
154    Make generation of toplevel "local" optional.  (Strictly speaking,
155    signature definitions within "local" are not legal SML.)
156    
157    Other than that: updates to INSTALL and cm/TODO.
158    
159    ----------------------------------------------------------------------
160    Name: Matthias Blume
161    Date: 2002/02/08 15:00:00 EST
162    Tag: blume-20020208-uniquepid
163    Description:
164    
165    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
166    
167    1. The installer (config/install.sh) has gotten smarter:
168    
169         - Configuration options are a bit easier to specify now
170           (in config/targets).
171         - Bug in recognizing .tar.bz2 files fixed.
172         - Installer automatically resolves dependencies between
173           configuration options (e.g., if you ask for eXene, you will
174           also get cml -- regardless whether you asked for it or not).
175         - Installer can run in "quieter mode" by setting the environment
176           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
177           "completely silent", though.
178         - Build HashCons library as part of smlnj-lib.
179    
180    2. A new scheme for assigning persistent identifiers to compilation
181       units (and, by extension, to types etc.) has been put into place.
182       This fixes a long-standing bug where types and even dynamic values
183       can get internally confused, thereby compromising type safety
184       (abstraction) and dynamic correctness.  See
185    
186         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
187    
188       for an example of how things could go wrong until now.
189    
190       The downside of the new scheme is that pids are not quite as
191       persistent as they used to be: CM will generate a fresh pid
192       for every compilation unit that it thinks it sees for the first
193       time.  That means that if you compile starting from a clean, fresh
194       source tree at two different times, you end up with different
195       binaries.
196    
197       Cutoff recompilation, however, has not been compromised because
198       CM keeps pid information in special caches between runs.
199    
200    ----------------------------------------------------------------------
201    Name: Lal George
202    Date: 2002/02/07 15:34:13 EST 2002
203    Tag: <none>
204    Description:
205    
206    Compilers that generate assembly code may produce  global labels
207    whose value is resolved at link time. The various peephole optimization
208    modules did not take this in account.
209    
210    TODO. The Labels.addrOf function should really return an option
211    type so that clients are forced to deal with this issue, rather
212    than an exception being raised.
213    
214    ----------------------------------------------------------------------
215    Name: Lal George
216    Date: 2002/02/06 13:55:02 EST
217    Tag: george-20020206-ra-breakup
218    Description:
219    
220    1. A bug fix from Allen.
221    
222        A typo causes extra fstp %st(0)'s to be generated at compensation
223        edges, which might cause stack underflow traps at runtime.  This
224        occurs in fft where there are extraneous fstps right before the 'into'
225        trap instruction (in this case they are harmless since none of the
226        integers overflow.)
227    
228    2. Pulled out various utility modules that were embedded in the modules
229       of the register allocator. I need these modules for other purposes, but
230       they are not complete enough to put into a library (just yet).
231    ----------------------------------------------------------------------
232    Name: Matthias Blume
233    Date: 2002/01/31 16:05:00 EST
234    Tag: blume-20020131-sparc-ccalls
235    Description:
236    
237    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
238       of extra stack space by mistake.  Fixed.
239    
240    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
241    
242    ----------------------------------------------------------------------
243    Name: Allen Leung
244    Date: 2002/01/30
245    Tag: leunga-20020130-nowhere-bug-fix
246    Description:
247    
248       MLRISC bug fixes:
249       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
250       2. MachineInt.fromString was negating its value.
251    
252    ----------------------------------------------------------------------
253    Name: Matthias Blume
254    Date: 2002/01/29
255    Tag: blume-20020129-INSTALL
256    Description:
257    
258    - Added somewhat detailed installation instructions (file INSTALL).
259    - Fixed curl-detection bug in config/install.sh.
260    - It is now possible to select the URL getter using the URLGETTER
261      environment variable:
262    
263          not set / "unknown"      --> automatic detection (script tries wget,
264                                       curl, and lynx)
265          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
266                                       how to properly invoke them)
267          other                    --> use $URLGETTER directly, it must take
268                                       precisely two command-line arguments
269                                       (source URL and destination file name)
270    
271    ----------------------------------------------------------------------
272    Name: Matthias Blume
273    Date: 2002/01/28
274    Tag: blume-20020128-sparc-ccalls
275    Description:
276    
277    - Fixed problem with calculation of "used" registers in sparc-c-calls.
278    - Make use of the allocParam argument in sparc-c-calls.
279    
280    ----------------------------------------------------------------------
281    Name: Matthias Blume
282    Date: 2002/01/28
283    Tag: blume-20020128-allocParam
284    Description:
285    
286    John Reppy:  Changes c-calls API to accept client-callback for
287    allocating extra stack space.
288    me: Corresponding changes to mlriscGen (using a dummy argument that
289        does not change the current behavior).
290    
291    ----------------------------------------------------------------------
292    Name: Matthias Blume
293    Date: 2002/01/28 12:00:00
294    Tag: Release_110_38
295    Description:
296    
297    This time for real!!!
298    
299    ----------------------------------------------------------------------
300    Name: Matthias Blume
301    Date: 2002/01/28 10:56:00 EST
302    Tag: blume-20020128-retraction
303    Description:
304    
305    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
306       with blume-Release_110_38-retracted.)
307    
308    1. Fixed a problem with incorrect rounding modes in real64.sml.
309       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
310    
311    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
312       and enums fixed.  The naming of corresponding ML identifiers should
313       now be consistent again.
314    
315    ----------------------------------------------------------------------
316    Name: Allen Leung
317    Date: 2002/01/27
318    Tag: leunga-20020127-nowhere
319    Description:
320    
321       Added a target called nowhere in the configuration scripts.
322       Enabling this will build the MLRISC 'nowhere' tool (for translating
323       programs with where-clauses into legal SML code) during installation.
324    
325    ----------------------------------------------------------------------
326    Name: Matthias Blume
327    Date: 2002/01/25 21:27:00 EST
328    Tag: blume-Release_110_38-retracted
329    Description:
330    
331    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
332    
333    README will be added later.
334    
335    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
336    (Release_110_38) removed.  Reason: Last-minute bug fixes.
337    
338    ----------------------------------------------------------------------
339    Name: Matthias Blume
340    Date: 2002/01/25
341    Tag: blume-20020125-ffi
342    Description:
343    
344    A large number of tweaks and improvements to ml-nlffi-lib and
345    ml-nlffigen:
346    
347       - ML represenation types have been streamlined
348       - getter and setter functions work with concrete values, not abstract
349         ones where possible
350       - ml-nlffigen command line more flexible (see README file there)
351       - some bugs have been fixed (hopefully)
352    
353    ----------------------------------------------------------------------
354    Name: Lal George
355    Date: 2002/01/24
356    Tag: george-20020124-risc-ra-interface
357    Description:
358    
359       There is a dramatic simplification in the interface to the
360       register allocator for RISC architectures as a result of making
361       parallel copy instructions explicit.
362    
363    ----------------------------------------------------------------------
364    Name: Matthias Blume
365    Date: 2002/01/22
366    Tag: blume-20020122-x86-ccalls
367    Description:
368    
369    Bug fix for c-calls on x86 (having to do with how char- and
370    short-arguments are being handled).
371    
372    ----------------------------------------------------------------------
373    Name: Matthias Blume
374    Date: 2002/01/21
375    Tag: blume-20020121-ff
376    Description:
377    
378    Another day of fiddling with the FFI...
379    
380    1. Bug fix/workaround:  CKIT does not complain about negative array
381       dimensions, so ml-nlffigen has to guard itself against this possibility.
382       (Otherwise a negative dimension would send it into an infinite loop.)
383    
384    2. Some of the abstract types (light objects, light pointers, most "base"
385       types) in structure C are now eqtypes.
386    
387    3. Added constructors and test functions for NULL function pointers.
388    
389    ----------------------------------------------------------------------
390    Name: Matthias Blume
391    Date: 2002/01/18
392    Tag: blume-20020118-ready-for-new-release
393    Description:
394    
395    Made config/srcarchiveurl point to a new place.  (Will provide boot
396    files shortly.)
397    
398    Maybe we christen this to be 110.38?
399    
400    ----------------------------------------------------------------------
401    Name: Matthias Blume
402    Date: 2002/01/18
403    Tag: blume-20020118-more-ffifiddle
404    Description:
405    
406    Today's FFI fiddling:
407    
408      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
409        C.Set.* that use concrete (MLRep.*) arguments and results instead
410        of abstract ones.
411    
412      - Provided word-style bit operations etc. for "int" representation
413        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
414        and Long.
415    
416    ----------------------------------------------------------------------
417    Name: Matthias Blume
418    Date: 2002/01/18
419    Tag: blume-20020118-use-x86-fp
420    Description:
421    
422    Now that x86-fast-fp seems to be working, I turned it back on again
423    by default.  (Seems to work fine now, even with the FFI.)
424    
425    Other than that, I added some documentation about the FFI to
426    src/ml-nlffigen/README and updated the FFI test examples in
427    src/ml-nlffi-lib/Tests/*.
428    
429    ----------------------------------------------------------------------
430    Name: Allen Leung
431    Date: 2002/01/17
432    Tag: leunga-20020117-x86-fast-fp-call
433    Description:
434    
435       1. Fixed a problem with handling return fp values when x86's fast fp
436          mode is turned on.
437    
438       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
439          of %f32.
440    
441       3. Added a constructor INT32lit to the ast of MLRISC tools.
442    
443    ----------------------------------------------------------------------
444    Name: Matthias Blume
445    Date: 2002/01/16
446    Tag: blume-20020116-ffifiddle
447    Description:
448    
449    More fiddling with the FFI interface:
450    
451     - Make constness 'c instead of rw wherever possible.  This eliminates
452       the need for certain explicit coercions.  (However, due to ML's
453       value polymorphism, there will still be many cases where explicit
454       coercions are necessary.  Phantom types are not the whole answer
455       to modeling a subtyping relationship in ML.)
456    
457     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
458    
459     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
460       clumsy workaround for varargs problem.)
461    
462    ----------------------------------------------------------------------
463    Name: Lal George
464    Date: 2002/01/15
465    Tag: <none>
466    Description:
467    
468    1. Since COPY instructions are no longer native to the architecture,
469       a generic functor can be used to implement the expandCopies function.
470    
471    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
472       TEXT segment.
473    
474    ----------------------------------------------------------------------
475    Name: Matthias Blume
476    Date: 2002/01/15
477    Tag: blume-20020115-ffiupdates
478    Description:
479    
480    1. Fix for bug resulting in single-precision float values being returned
481       incorrectly from FFI calls.
482    
483    2. Small modifications to C FFI API:
484    
485        - memory-allocation routines return straight objects (no options)
486          and raise an exception in out-of-memory situations
487        - unsafe extensions to cast between function pointers and pointers
488          from/to ints
489        - added structure C_Debug as an alternative to structure C where
490          pointer-dereferencing (|*| and |*!) always check for null-pointers
491        - added open_lib' to DynLinkage;  open_lib' works like open_lib
492          but also takes a (possibly empty) list of existing library handles
493          that the current library depends on
494    
495    ----------------------------------------------------------------------
496    Name: Matthias Blume
497    Date: 2002/01/10
498    Tag: blume-20020110-newffigen
499    Description:
500    
501    1. Updates to portable graph code.
502    
503    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
504       more scalable now so that even huge interfaces such as the one
505       for GTK compile in finite time and space. :-)
506       See src/ml-nlffigen/README for details on what's new.
507    
508    ----------------------------------------------------------------------
509    Name: Lal George
510    Date: 2001/01/09 14:31:35 EST 2002
511    Tag: george-20011206-rm-native-copy
512    Description:
513    
514            Removed the native COPY and FCOPY instructions
515            from all the architectures and replaced it with the
516            explicit COPY instruction from the previous commit.
517    
518            It is now possible to simplify many of the optimizations
519            modules that manipulate copies. This has not been
520            done in this change.
521    
522    ----------------------------------------------------------------------
523    Name: Lal George
524    Date: 2001/12/06 16:50:13 EST 2001
525    Tag: george-20011206-mlrisc-instruction
526    Description:
527    
528    Changed the representation of instructions from being fully abstract
529    to being partially concrete. That is to say:
530    
531      from
532            type instruction
533    
534      to
535            type instr                              (* machine instruction *)
536    
537            datatype instruction =
538                LIVE of {regs: C.cellset, spilled: C.cellset}
539              | KILL of {regs: C.cellset, spilled: C.cellset}
540              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
541              | ANNOTATION of {i: instruction, a: Annotations.annotation}
542              | INSTR of instr
543    
544    This makes the handling of certain special instructions that appear on
545    all architectures easier and uniform.
546    
547    LIVE and KILL say that a list of registers are live or killed at the
548    program point where they appear. No spill code is generated when an
549    element of the 'regs' field is spilled, but the register is moved to
550    the 'spilled' (which is present, more for debugging than anything else).
551    
552    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
553    We used to generate:
554    
555            DEFFREG f1
556            f1 := f2 + f3
557            trapb
558    
559    but now generate:
560    
561            f1 := f2 + f3
562            trapb
563            LIVE {regs=[f1,f2,f3], spilled=[]}
564    
565    Furthermore, the DEFFREG (hack) required that all floating point instruction
566    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
567    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
568    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
569    this architecture).
570    
571    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
572    all the architectures. This will result in further simplification of the
573    register allocator that must be aware of them for coalescing purposes, and
574    will also simplify certain aspects of the machine description that provides
575    callbacks related to parallel copies.
576    
577    ANNOTATION should be obvious, and now INSTR represents the honest to God
578    machine instruction set!
579    
580    The <arch>/instructions/<arch>Instr.sml files define certain utility
581    functions for making porting easier -- essentially converting upper case
582    to lower case. All machine instructions (of type instr) are in upper case,
583    and the lower case form generates an MLRISC instruction. For example on
584    the alpha we have:
585    
586      datatype instr =
587         LDA of {r:cell, b:cell, d:operand}
588       | ...
589    
590      val lda : {r:cell, b:cell, d:operand} -> instruction
591        ...
592    
593    where lda is just (INSTR o LDA), etc.
594    
595    ----------------------------------------------------------------------
596    Name: Matthias Blume
597    Date: 2001/11/22 21:40:00 EST
598    Tag: Release_110_37
599    Description:
600    
601    Release 110.37.  This time for real.
602    
603    ----------------------------------------------------------------------
604    Name: Matthias Blume
605    Date: 2001/11/21 16:35:00 EST
606    Tag: blume-20011121-foot-in-mouth
607    Description:
608    
609    Removed the "Release_110_37" tag because of a serious bug.
610    This will be re-tagged once the bug is fixed.
611    
612    ----------------------------------------------------------------------
613    Name: Matthias Blume
614    Date: 2001/11/21 16:14:00 EST
615    Tag: blume-20011121-forgottenfile
616    Description:
617    
618    Forgot to add a file.  (Just a .tex-file -- part of
619    the CM manual source.)
620    
621    ----------------------------------------------------------------------
622    Name: Matthias Blume
623    Date: 2001/11/21 16:10:00 EST
624    Tag: blume-20011121-invalid_110_37
625    Description:
626    
627    Note: I removed the original tag "Release_110_37" from this commit
628          because we found a serious bug in all non-x86 backends.
629          - Matthias
630    
631    1. Modifications to the SML/NJ code generator and to the runtime system
632       so that code object name strings are directly inserted into code
633       objects at code generation time.  The only business the runtime system
634       has with this is now to read the name strings on occasions.
635       (The encoding of the name string has also changed somewhat.)
636    
637    2. CM now implements a simple "set calculus" for specifying export lists.
638       In particular, it is now possible to refer to the export lists of
639       other libraries/groups/sources and form unions as well as differences.
640       See the latest CM manual for details.
641    
642    3. An separate notion of "proxy" libraries has again be eliminated from
643       CM's model.  (Proxy libraries are now simply a special case of using
644       the export list calculus.)
645    
646    4. Some of the existing libraries now take advantage of the new set
647       calculus.
648       (Notice that not all libraries have been converted because some
649       of the existing .cm-files are supposed to be backward compatible
650       with 110.0.x.)
651    
652    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
653       "General.exnMessage"!  The former relies on a certain hook to be
654       initialized, and that often does not happen in the stand-alone case.)
655    
656    ----------------------------------------------------------------------
657    Name: Lal George
658    Date: 2001/11/21  13:56:18 EST
659    Tag: george-2001121-pseudo-ops
660    Description:
661    
662      Implemented a complete redesign of MLRISC pseudo-ops. Now there
663      ought to never be any question of incompatabilities with
664      pseudo-op syntax expected by host assemblers.
665    
666      For now, only modules supporting GAS syntax are implemented
667      but more should follow, such as MASM, and vendor assembler
668      syntax, e.g. IBM as, Sun as, etc.
669    
670    ----------------------------------------------------------------------
671    Name: Matthias Blume
672    Date: 2001/11/14 11:52:00 EST
673    Tag: blume-20011114-srcname
674    Description:
675    
676    1. Routed the name of the current source file to mlriscgen where it
677       should be directly emitted into the code object.  (This last part
678       is yet to be done.)
679    
680    2. Some cleanup of the pgraph code to make it match the proposal that
681       I put out the other day.  (The proposal notwithstanding, things are
682       still in flux here.)
683    
684    ----------------------------------------------------------------------
685    Name: Lal George
686    Date: 2001/11/14 09:44:04 EST
687    Tag:
688    Description:
689    
690      Fix for a backpatching bug reported by Allen.
691    
692      Because the boundary between short and long span-dependent
693      instructions is +/- 128, there are an astounding number of
694      span-dependent instructions whose size is over estimated.
695    
696      Allen came up with the idea of letting the size of span
697      dependent instructions be non-monotonic, for a maxIter
698      number of times, after which the size must be monotonically
699      increasing.
700    
701      This table shows the number of span-dependent instructions
702      whose size was over-estimated as a function of maxIter, for the
703      file Parse/parse/ml.grm.sml:
704    
705         maxIter            # of instructions:
706            10                      687
707            20                      438
708            30                      198
709            40                        0
710    
711      In compiling the compiler, there is no significant difference in
712      compilation speed between maxIter=10 and maxIter=40. Actually,
713      my measurements showed that maxIter=40 was a tad faster than
714      maxIter=10! Also 96% of the  files in the compiler reach a fix
715      point within 13 iterations, so fixing maxIter at 40, while high,
716      is okay.
717    
718    ----------------------------------------------------------------------
719    Name: Matthias Blume
720    Date: 2001/10/31 15:25:00 EST
721    Tag: blume-20011031-pgraph
722    Description:
723    
724    CKIT:
725    * Changed the "Function" constructor of type Ast.ctype to carry optional
726      argument identifiers.
727    * Changed the return type of TypeUtil.getFunction accordingly.
728    * Type equality ignores the argument names.
729    * TypeUtil.composite tries to preserve argument names but gives up quickly
730      if there is a mismatch.
731    
732    installation script:
733    * attempts to use "curl" if available (unless "wget" is available as well)
734    
735    CM:
736    * has an experimental implementation of "portable graphs" which I will
737      soon propose as an implementation-independent library format
738    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
739    
740    NLFFI-LIB:
741    * some cleanup (all cosmetic)
742    
743    NLFFIGEN:
744    * temporarily disabled the mechanism that suppresses ML output for
745      C definitions whose identifiers start with an underscore character
746    * generate val bindings for enum constants
747    * user can request that only one style (light or heavy) is being used;
748      default is to use both (command-line arguments: -heavy and -light)
749    * fixed bug in handling of function types involving incomplete pointers
750    * generate ML entry points that take record arguments (i.e., using
751      named arguments) for C functions that have a prototype with named
752      arguments
753      (see changes to CKIT)
754    
755    ----------------------------------------------------------------------
756    Name: Allen Leung
757    Date: 2001/10/27 20:34:00 EDT
758    Tag: leunga-20011027-x86-fast-fp-call
759    Description:
760    
761       Fixed the bug described in blume-20010920-slowfp.
762    
763       The fix involves
764          1. generating FCOPYs in FSTP in ia32-svid
765          2. marking a CALL with the appropriate annotation
766    
767    ----------------------------------------------------------------------
768    Name: Matthias Blume
769    Date: 2001/10/16 11:32:00 EDT
770    Tag: blume-20011016-netbsd
771    Description:
772    
773    Underscore patch from Chris Richards (fixing problem with compiling
774    runtime system under recent NetBSD).
775    
776    ----------------------------------------------------------------------
777    Name: Allen Leung
778    Date: 2001/10/12 17:18:32 EDT 2001
779    Tag: leung-20011012-x86-printflowgraph
780    Description:
781    
782    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
783    
784    ----------------------------------------------------------------------
785    Name: Lal George
786    Date: 2001/10/11 23:51:34 EDT
787    Tag: george-20011011-too-many-instrs
788    Description:
789    
790    The representation of a program point never expected to see more
791    than 65536 instructions in a basic block!
792    
793    ----------------------------------------------------------------------
794    Name: Lal George
795    Date: 2001/10/09 09:41:37 EDT
796    Tag: george-20011008-mlrisc-labels
797    Description:
798    
799    Changed the machine description files to support printing of
800    local and global labels in assembly code, based on host assembler
801    conventions.
802    
803    ----------------------------------------------------------------------
804    Name: Matthias Blume
805    Date: 2001/09/25 15:25:00 EDT
806    Tag: blume-20010925-exninfo
807    Description:
808    
809    I provided a non-hook implementation of exnName (at the toplevel) and
810    made the "dummy" implementation of exnMessage (at the toplevel) more
811    useful: if nothing gets "hooked in", then at least you are going to
812    see the exception name and a message indicating why you don't see more.
813    
814    [For the time being, programs that need exnMessage and want to use
815    ml-build should either use General.exnMessage (strongly recommended) or
816    refer to structure General at some other point so that CM sees a
817    static dependency.]
818    
819    [Similar remarks go for "print" and "use":  If you want to use their
820    functionality in stand-alone programs generated by ml-build, then use
821    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
822    
823    ----------------------------------------------------------------------
824    Name: Matthias Blume
825    Date: 2001/09/20 17:28:00 EDT
826    Tag: blume-20010920-slowfp
827    Description:
828    
829    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
830    
831    ----------------------------------------------------------------------
832    Name: Matthias Blume
833    Date: 2001/09/20 17:20:00 EDT
834    Tag: blume-20010920-canonicalpaths
835    Description:
836    
837    0. Updated the BOOT file (something that I forgot to do earlier).
838    
839    1. Small internal change to CM so that it avoids "/../" in filenames
840       as much as possible (but only where it is safe).
841    
842    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
843       that arguments that contain delimiters are passed through correctly.
844       This change also means that all "special" arguments of the form
845       @SMLxxx... must come first.
846    
847    3. Changed install script to put relative anchor names for tool commands
848       into pathconfig.
849    
850    ----------------------------------------------------------------------
851    Name: Matthias Blume
852    Date: 2001/09/18 15:35:00 EDT
853    Tag: blume-20010918-readme11036
854    Description:
855    
856    Added README files.
857    
858    ----------------------------------------------------------------------
859    Name: Matthias Blume
860    Date: 2001/09/18 11:45:00 EDT
861    Tag: Release_110_36 (retag)
862    Description:
863    
864    Fixed mistake in config/preloads. Retagged as 110.36.
865    
866    ----------------------------------------------------------------------
867    Name: Matthias Blume
868    Date: 2001/09/18 09:40:00 EDT
869    Tag: Release_110_36_orig (tag changed)
870    Description:
871    
872    New version (110.36).  New bootfiles.
873    
874    ----------------------------------------------------------------------
875    Name: Matthias Blume
876    Date: 2001/09/14 16:15:00 EDT
877    Tag: blume-20010914-x86fastfp
878    Description:
879    
880    John committed some changes that Allen made, in particular a (hopefully)
881    correctly working version of the x86-fp module.
882    
883    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
884    flag to "true".  Everything seems to compile to a fixpoint ok, and
885    "mandelbrot" speeds up by about 15%.
886    
887    ----------------------------------------------------------------------
888    Name: Matthias Blume
889    Date: 2001/09/13 11:20:00 EDT
890    Tag: blume-20010913-minimal
891    Description:
892    
893    1. Stefan Monnier's patch to fix a miscompilation problem that
894       was brought to light by John Reppy's work on Moby.
895    
896    2. Implemented a minimal "structure Compiler" that contains just
897       "version" and "architecture".  The minimal version will be
898       available when the full version is not.  This is for backward-
899       compatibility with code that wants to test Compiler.version.
900    
901    ----------------------------------------------------------------------
902    Name: Matthias Blume
903    Date: 2001/08/28 14:03:00 EDT
904    Tag: blume-20010828-ml-lex
905    Description:
906    
907    Fix for bug 1581, received from Neophytos Michael.
908    
909    ----------------------------------------------------------------------
910    Name: Matthias Blume
911    Date: 2001/08/27 11:20:00 EDT
912    Tag: blume-20010827-readme11035
913    Description:
914    
915    Fleshed out the README file for 110.35.
916    
917    ----------------------------------------------------------------------
918    Name: Matthias Blume
919    Date: 2001/08/24 17:10:00 EDT
920    Tag: Release_110_35
921    Description:
922    
923    New version number (110.35).  New bootfiles.
924    
925    ----------------------------------------------------------------------
926    Name: Lal George
927    Date: 2001/08/24 13:47:18 EDT 2001
928    Tag: george-20010824-MLRISC-graphs
929    Description:
930    
931     removed clusters from MLRISC completely and replaced with graphs.
932    
933    ----------------------------------------------------------------------
934    Name: Matthias Blume
935    Date: 2001/08/23 17:50:00 EDT
936    Tag: blume-20010823-toplevel
937    Description:
938    
939    - some reorganization of the code that implements various kinds of
940      environments in the compiler (static, dynamic, symbolic, combined)
941    - re-implemented the EnvRef module so that evalStream works properly
942      (if the stream contains references to "use", "CM.make", etc.)
943    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
944    
945    ----------------------------------------------------------------------
946    Name: Matthias Blume
947    Date: 2001/08/20 15:50 EDT
948    Tag: blume20010820-slipup
949    Description:
950    
951    I forgot to commit a few files.  Here they are...
952    
953    ----------------------------------------------------------------------
954    Name: Matthias Blume
955    Date: 2001/08/20 15:35:00 EDT
956    Tag: blume-20010820-debugprof
957    Description:
958    
959    !!!! NEW BOOTFILES !!!!
960    
961    This is another round of reorganizing the compiler sources.  This
962    time the main goal was to factor out all the "instrumentation"
963    passes (for profiling and backtracing) into their own library.
964    The difficulty was to do it in such a way that it does not depend
965    on elaborate.cm but only on elabdata.cm.
966    
967    Therefore there have been further changes to both elaborate.cm and
968    elabdata.cm -- more "generic" things have been moved from the former
969    to the latter.  As a result, I was forced to split the assignment
970    of numbers indicating "primtyc"s into two portions: SML-generic and
971    SML/NJ-specific.  Since it would have been awkward to maintain,
972    I bit the bullet and actually _changed_ the mapping between these
973    numbers and primtycs.  The bottom line of this is that you need
974    a new set of bin- and bootfiles.
975    
976    I have built new bootfiles for all architectures, so doing a fresh
977    checkout and config/install.sh should be all you need.
978    
979    The newly created library's name is
980    
981        $smlnj/viscomp/debugprof.cm
982    
983    and its sources live under
984    
985        src/compiler/DebugProf
986    
987    ----------------------------------------------------------------------
988    Name: Matthias Blume
989    Date: 2001/08/15 17:15:00 EDT
990    Tag: blume-20010815-compreorg
991    Description:
992    
993    This is a first cut at reorganizing the CM libraries that make up the
994    core of the compiler.  The idea is to separate out pieces that could
995    be used independently by tools, e.g., the parser, the typechecker, etc.
996    
997    The current status is a step in this direction, but it is not quite
998    satisfactory yet.  Expect more changes in the future.
999    
1000    Here is the current (new) organization...
1001    
1002        What used to be $smlnj/viscomp/core.cm is now divided into
1003        six CM libraries:
1004    
1005             $smlnj/viscomp/basics.cm
1006                           /parser.cm
1007                           /elabdata.cm
1008                           /elaborate.cm
1009                           /execute.cm
1010                           /core.cm
1011    
1012        The CM files for these libraries live under src/system/smlnj/viscomp.
1013        All these libraries are proxy libraries that contain precisely
1014        one CM library component.  Here are the locations of the components
1015        (all within the src/compiler tree):
1016    
1017             Basics/basics.cm
1018             Parse/parser.cm
1019             ElabData/elabdata.cm
1020             Elaborator/elaborate.cm
1021             Execution/execute.cm
1022             core.cm
1023    
1024         [This organization is the same that has been used already
1025         for a while for the architecture-specific parts of the visible
1026         compiler and for the old version of core.cm.]
1027    
1028         As you will notice, many source files have been moved from their
1029         respective original locations to a new home in one of the above
1030         subtrees.
1031    
1032         The division of labor between the new libraries is the following:
1033    
1034             basics.cm:
1035                - Simple, basic definitions that pertain to many (or all) of
1036                  the other libraries.
1037             parser.cm:
1038                - The SML parser, producing output of type Ast.dec.
1039                - The type family for Ast is also defined and exported here.
1040             elabdata.cm:
1041                - The datatypes that describe input and output of the elaborator.
1042                  This includes types, absyn, and static environments.
1043             elaborator.cm:
1044                - The SML/NJ type checker and elaborator.
1045                  This maps an Ast.dec (with a given static environment) to
1046                  an Absyn.dec (with a new static environment).
1047                - This libraries implements certain modules that used to be
1048                  structures as functors (to remove dependencies on FLINT).
1049             execute.cm:
1050                - Everything having to do with executing binary code objects.
1051                - Dynamic environments.
1052             core.cm:
1053                - SML/NJ-specific instantiations of the elaborator and MLRISC.
1054                - Top-level modules.
1055                - FLINT (this should eventually become its own library)
1056    
1057    Notes:
1058    
1059    I am not 100% happy with the way I separated the elaborator (and its
1060    data structures) from FLINT.  Two instances of the same problem:
1061    
1062        1. Data structures contain certain fields that carry FLINT-specific
1063           information.  I hacked around this using exn and the property list
1064           module from smlnj-lib.  But the fact that there are middle-end
1065           specific fields around at all is a bit annoying.
1066    
1067        2. The elaborator calculates certain FLINT-related information.  I tried
1068           to make this as abstract as I could using functorization, but, again,
1069           the fact that the elaborator has to perform calculations on behalf
1070           of the middle-end at all is not nice.
1071    
1072        3. Having to used exn and property lists is unfortunate because it
1073           weakens type checking.  The other alternative (parameterizing
1074           nearly *everything*) is not appealing, though.
1075    
1076    I removed the "rebinding =" warning hack because due to the new organization
1077    it was awkward to maintain it.  As a result, the compiler now issues some of
1078    these warnings when compiling init.cmi during bootstrap compilation. On
1079    the plus side, you also get a warning when you do, for example:
1080       val op = = Int32.+
1081    which was not the case up to now.
1082    
1083    I placed "assign" and "deref" into the _Core structure so that the
1084    code that deals with the "lazy" keyword can find them there.  This
1085    removes the need for having access to the primitive environment
1086    during elaboration.
1087    
1088    ----------------------------------------------------------------------
1089    Name: Matthias Blume
1090    Date: 2001/08/13
1091    Tag: blume-20010813-closures
1092    Description:
1093    
1094    This fix was sent to us by Zhong Shao.  It is supposed to improve the
1095    performance of certain loops by avoiding needless closure allocation.
1096    
1097    ----------------------------------------------------------------------
1098    Name: Lal George
1099    Date: 2001/07/31 10:03:23 EDT 2001
1100    Tag: george-20010731-x86-fmalloc
1101    Description: Fixed bug in x86 calls
1102    
1103        There was a bug where call instructions would mysteriously
1104        vanish. The call instruction had to be one that returned
1105        a floating point value.
1106    
1107    ----------------------------------------------------------------------
1108    Name: Lal George
1109    Date: 2001/07/19 16:36:29 EDT 2001
1110    Tag: george-20010719-simple-cells
1111    Description:
1112    
1113    I have dramatically simplified the interface for CELLS in MLRISC.
1114    
1115    In summary, the cells interface is broken up into three parts:
1116    
1117      1. CellsBasis : CELLS_BASIS
1118    
1119            CellsBasis is a top level structure and common for all
1120            architectures.  it contains the definitions of basic datatypes
1121            and utility  functions over these types.
1122    
1123      2. functor Cells() : CELLS
1124    
1125            Cells generates an interface for CELLS that incorporates the
1126            specific resources on the target architecture, such as the
1127            presence of special register classes, their number and size,
1128            and various useful substructures.
1129    
1130      3. <ARCH>CELLS
1131    
1132            e.g. SparcCells: SPARCCELLS
1133    
1134            <ARCH>CELLS usually contains additional bindings for special
1135            registers  on the architecture, such as:
1136    
1137                    val r0 : cell           (* register zero *)
1138                    val y : cell            (* Y register *)
1139                    val psr : cell          (* processor status register *)
1140                    ...
1141    
1142            The structure returned by applying the Cells functor is opened
1143            in this interface.
1144    
1145    The main implication of all this is that the datatypes for cells is
1146    split between CellsBasis and CELLS -- a fairly simple change for user
1147    code.
1148    
1149    In the old scheme the CELLS interface had a definitional binding of
1150    the form:
1151    
1152            signature CELLS = sig
1153    
1154               structure CellsBasis = CellsBasis
1155    
1156               ...
1157    
1158            end
1159    
1160    With all the sharing constraints that goes on in MLRISC, this old
1161    design  quickly leads to errors such as:
1162    
1163            "structure definition spec inside of sharing ... "
1164    
1165    
1166    and appears to require an unacceptable amount of sharing and where
1167    constraint hackery.
1168    
1169    I think this error message (the interaction of definitional specs and
1170    sharing) requires more explanation on our web page.
1171    
1172    ----------------------------------------------------------------------
1173    Name: Matthias Blume
1174    Date: 2001/07/19 15:00:00 EDT
1175    Tag: blume-20010719-libreorg
1176    Description:
1177    
1178    This update puts together a fairly extensive but straightforward change
1179    to the way the libraries that implement the interactive system are
1180    organized:
1181    
1182       The biggest change is the elimination of structure Compiler.  As a
1183       replacement for this structure, there is now a CM library
1184       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1185       that exports all the substructures of the original structure Compiler
1186       directly.  So instead of saying Compiler.Foo.bar one now simply
1187       says Foo.bar.  (The CM libraries actually export a collection of
1188       structures that is richer than the collection of substructures of
1189       structure Compiler.)
1190    
1191       To make the transition smooth, there is a separate library called
1192       $smlnj/compiler/compiler.cm which puts together and exports the
1193       original structure Compiler (or at least something very close to it).
1194    
1195       There are five members of the original structure Compiler
1196       that are not exported directly but which instead became members
1197       of a new structure Backend (described by signature BACKEND).  These are:
1198       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1199       Interact (: INTERACT), structure Machine (: MACHINE), and val
1200       architecture (: string).
1201    
1202       Structure Compiler.Version has become structure CompilerVersion.
1203    
1204       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1205       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1206       or x86, respectively.
1207       Each of these exports the same frontend structures that
1208       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1209       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1210       Hppa, PPC, Sparc, or X86, respectively.
1211    
1212       Library $smlnj/compiler/all.cm exports the union of the exports of
1213       $smlnj/compiler/<arch>.cm
1214    
1215       There are no structures <Arch>Compiler anymore, use
1216       $smlnj/compiler/<arch>.cm instead.
1217    
1218       Library host-compiler-0.cm is gone.  Instead, the internal library
1219       that instantiates CM is now called cm0.cm.  Selection of the host
1220       compiler (backend) is no longer done here but. (Responsibility for it
1221       now lies with $smlnj/compiler/current.cm.  This seems to be more
1222       logical.)
1223    
1224       Many individual files have been moved or renamed.  Some files have
1225       been split into multiple files, and some "dead" files have been deleted.
1226    
1227    Aside from these changes to library organization, there are also changes
1228    to the way the code itself is organized:
1229    
1230       Structure Binfile has been re-implemented in such a way that it no
1231       longer needs any knowledge of the compiler.  It exclusively deals
1232       with the details of binfile layout.  It no longer invokes the
1233       compiler (for the purpose of creating new prospective binfile
1234       content), and it no longer has any knowledge of how to interpret
1235       pickles.
1236    
1237       Structure Compile (: COMPILE) has been stripped down to the bare
1238       essentials of compilation.  It no longer deals with linking/execution.
1239       The interface has been cleaned up considerably.
1240    
1241       Utility routines for dealing with linking and execution have been
1242       moved into their own substructures.
1243    
1244       (The ultimate goal of these changes is to provide a light-weight
1245       binfile loader/linker (at least for, e.g., stable libraries) that
1246       does not require CM or the compiler to be present.)
1247    
1248    CM documentation has been updated to reflect the changes to library
1249    organization.
1250    
1251    ----------------------------------------------------------------------
1252    Name: Matthias Blume
1253    Date: 2001/07/10 17:30:00 EDT
1254    Tag: Release_110_34
1255    Description:
1256    
1257    Minor tweak to 110.34 (re-tagged):
1258    
1259      - README.html file added to CVS repository
1260      - runtime compiles properly under FreeBSD 3.X and 4.X
1261    
1262    ----------------------------------------------------------------------
1263    Name: Matthias Blume
1264    Date: 2001/07/10 17:30:00 EDT
1265    Tag: Release_110_34
1266    Description:
1267    
1268    New version number (110.34). New bootfiles.
1269    
1270    ----------------------------------------------------------------------
1271    Name: Matthias Blume
1272    Date: 2001/07/09 16:00:00 EDT
1273    Tag: blume-20010709-more-varargs
1274    Description:
1275    
1276    I changed the handling of varargs in ml-nlffigen again:
1277    The ellipsis ... will now simply be ignored (with an accompanying warning).
1278    
1279    The immediate effect is that you can actually call a varargs function
1280    from ML -- but you can't actually supply any arguments beyond the ones
1281    specified explicitly.  (For example, you can call printf with its format
1282    string, but you cannot pass additional arguments.)
1283    
1284    This behavior is only marginally more useful than the one before, but
1285    it has the advantage that a function or, more importantly, a function
1286    type never gets dropped on the floor, thus avoiding follow-up problems with
1287    other types that refer to the offending one.
1288    
1289    ----------------------------------------------------------------------
1290    Name: Matthias Blume
1291    Date: 2001/07/09 11:25:00 EDT
1292    Tag: blume-20010709-varargs
1293    Description:
1294    
1295    1. ckit-lib.cm now exports structure Error
1296    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1297       with a warning accompanied by a source location.  Moreover, it
1298       merely skips the offending function or type and proceeds with the
1299       rest of its work.u  As a result, one can safely feed C code containing
1300       "..." to ml-nlffigen.
1301    3. There are some internal improvements to CM, providing slightly
1302       more general string substitutions in the tools subsystem.
1303    
1304    ----------------------------------------------------------------------
1305    Name: Matthias Blume
1306    Date: 2001/06/27 15:10:00 EDT
1307    Tag: blume-20010627-concur
1308    Description:
1309    
1310    Fixed a small bug in CM's handling of parallel compilation.
1311    (You could observe the bug by Control-C-interrupting an ordinary
1312    CMB.make or CM.stabilize and then attaching some compile servers.
1313    The result was that all of a sudden the previously interrupted
1314    compilation would continue on its own.  This was because of
1315    an over-optimization: CM did not bother to clean out certain queues
1316    when no servers were attached "anyway", resulting in the contents
1317    of these queues to grab control when new servers did get attached.)
1318    
1319    There is also another minor update to the CM manual.
1320    
1321    ----------------------------------------------------------------------
1322    Name: Matthias Blume
1323    Date: 2001/06/26 16:15:00 EDT
1324    Tag: blume-20010626-cmdoc
1325    Description:
1326    
1327    Minor typo fixed in CM manual (syntax diagram for libraries).
1328    
1329    ----------------------------------------------------------------------
1330    Name: Matthias Blume
1331    Date: 2001/06/25 22:55:00 EDT
1332    Tag: blume-20010625-x86pc
1333    Description:
1334    
1335    Fixed a nasty bug in the X86 assembly code that caused signal
1336    handlers to fail (crash) randomly.
1337    
1338    ----------------------------------------------------------------------
1339    Name: Matthias Blume
1340    Date: 2001/06/25 12:05:00 EDT
1341    Tag: blume-20010625-nlffigen
1342    Description:
1343    
1344    This update fixes a number of minor bugs in ml-nlffigen as reported by
1345    Nick Carter <nbc@andrew.cmu.edu>.
1346    
1347      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1348      2. Default names for generated files are now derived from the name of
1349         the C file *without its directory*.  In particular, this causes generated
1350         files to be placed locally even if the C file is in some system directory.
1351      3. Default names for generated signatures and structures are also derived
1352         from the C file name without its directory.  This avoids silly things
1353         like "structure GL/GL".
1354         (Other silly names are still possible because ml-nlffigen does not do
1355          a thorough check of whether generated names are legal ML identifiers.
1356          When in doubt, use command line arguments to force particular names.)
1357    
1358    ----------------------------------------------------------------------
1359    Name: Matthias Blume
1360    Date: 2001/06/21 12:25:00 EDT
1361    Tag: blume-20010621-eXene
1362    Description:
1363    
1364    eXene now compiles and (sort of) works again.
1365    
1366    The library name (for version > 110.33) is $/eXene.cm.
1367    
1368    I also added an new example in src/eXene/examples/nbody.  See the
1369    README file there for details.
1370    
1371    ----------------------------------------------------------------------
1372    Name: Matthias Blume
1373    Date: 2001/06/20 16:40:00 EDT
1374    Tag: blume-20010620-cml
1375    Description:
1376    
1377    CML now compiles and works again.
1378    
1379    Libraries (for version > 110.33):
1380    
1381      $cml/cml.cm            Main CML library.
1382      $cml/basis.cm          CML's version of $/basis.cm.
1383      $cml/cml-internal.cm   Internal helper library.
1384      $cml/core-cml.cm       Internal helper library.
1385      $cml-lib/trace-cml.cm  Tracing facility.
1386      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1387    
1388    The installer (config/install.sh) has been taught how to properly
1389    install this stuff.
1390    
1391    ----------------------------------------------------------------------
1392    Name: Matthias Blume
1393    Date: 2001/06/19 17:55:00 EDT
1394    Tag: blume-20010619-instantiate
1395    Description:
1396    
1397    This un-breaks the fix for bug 1432.
1398    (The bug was originally fixed in 110.9 but I broke it again some
1399    time after that.)
1400    
1401    ----------------------------------------------------------------------
1402    Name: Matthias Blume
1403    Date: 2001/06/19 17:25:00 EDT
1404    Tag: blume-20010619-signals
1405    Description:
1406    
1407    This should (hopefully) fix the long-standing signal handling bug.
1408    (The runtime system was constructing a continuation record with an
1409    incorrect descriptor which would cause the GC to drop data on the floor...)
1410    
1411    ----------------------------------------------------------------------
1412    Name: Matthias Blume
1413    Date: 2001/06/15 15:05:00 EDT
1414    Tag: blume-20010615-moresparc
1415    Description:
1416    
1417    Here is a short late-hour update related to Sparc c-calls:
1418    
1419     -- made handling of double-word arguments a bit smarter
1420    
1421     -- instruction selection phase tries to collapse certain clumsily
1422        constructed ML-Trees; typical example:
1423    
1424            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1425    
1426        This currently has no further impact on SML/NJ since mlriscGen does
1427        not seem to generate such patterns in the first place, and c-calls
1428        (which did generate them in the beginning) has meanwhile been fixed
1429        so as to avoid them as well.
1430    
1431    ----------------------------------------------------------------------
1432    Name: Matthias Blume
1433    Date: 2001/06/15 15:05:00 EDT
1434    Tag: blume-20010615-sparc
1435    Description:
1436    
1437    The purpose of this update is to provide an implementation of NLFFI
1438    on Sparc machines.
1439    
1440    Here are the changes in detail:
1441    
1442       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1443       the Sparc implementation of the c-calls API.
1444       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1445       for accessing the ML frame.  Thus, we have a real frame pointer and
1446       can freely modify %sp without need for an omit-frame-ptr phase.
1447       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1448       this case.
1449       * ml-nlffigen has been taught to produce code for different architectures
1450       and calling conventions.
1451       * In a way similar to what was done in the x86 case, the Sparc
1452       backend uses its own specific extension to mltree.  (For example,
1453       it needs to be able to generate UNIMP instructions which are part
1454       of the calling convention.)
1455       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1456       to make it easier to plug in new machine- and os-dependent parts).
1457    
1458    There are some other fairly unrelated bug fixes and cleanups as well:
1459    
1460       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1461       that they properly share their libraries with existing SML/NJ libraries.
1462       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1463       follow-up error messages.
1464       * Updates to CM/CMB documentation.
1465    
1466    TODO items:
1467    
1468       * MLRISC should use a different register as its asmTemp on the Sparc.
1469         (The current %o2 is a really bad choice because it is part of the
1470         calling conventions, so things might interfere in unexpected ways.)
1471    
1472    ----------------------------------------------------------------------
1473    Name: Matthias Blume
1474    Date: 2001/06/07
1475    Tag: blume-20010607-calls
1476    Description:
1477    
1478    A number of internal changes related to C calls and calling conventions:
1479    
1480    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1481       number of bytes popped implicitly (by the callee).  In most cases
1482       this field is 0 but on x86/win32 it is some non-zero value.  This
1483       is information provided for the benefit of the "omit-frameptr" pass.
1484    2. The CALL instruction on the x86 carries a similar "pops" field.
1485       The instruction selection phase copies its value from the ML-Tree
1486       CALL statement.
1487    3. On all other architectures, the instruction selection phase checks
1488       whether "pops=0" and complains if not.
1489    4. The c-calls implementation for x86 now accepts two calling conventions:
1490       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1491       up after the call and pops is set to 0.  For "stdcall", the caller
1492       does nothing, leaving the cleanup to the callee; pops is set to
1493       the number of bytes that were pushed onto the stack.
1494    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1495       distinguish between "ccall" and "stdcall".
1496    6. The UNIMP instruction has been added to the supported Sparc instruction
1497       set. (This is needed for implementing the official C calling convention
1498       on this architecture.)
1499    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1500       work with the latest CM.
1501    
1502    ----------------------------------------------------------------------
1503    Name: Matthias Blume
1504    Date: 2001/06/05 15:10:00 EDT
1505    Tag: blume-20010605-cm-index
1506    Description:
1507    
1508    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1509    
1510    1. CM can now generate "index files".  These are human-readable files
1511       that list on a per-.cm-file basis each toplevel symbol defined or
1512       imported.  The location of the index file for
1513       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1514       To enable index-file generation, set CM.Control.generate_index to true
1515       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1516    
1517       The CM manual has been updated accordingly.
1518    
1519    2. I made some slight modifications to the c-calls API in MLRISC.
1520    
1521         a) There is now a callback to support saving/restoring of
1522            dedicated but caller-save registers around the actual call
1523            instruction.
1524         b) One can optionally specify a comment-annotation for the
1525            call instruction.
1526    
1527    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1528       (For example, the comment annotation shows the C prototype of
1529        the function being called.)
1530    
1531    ----------------------------------------------------------------------
1532    Name: Matthias Blume
1533    Date: 2001/06/01 13:30:00 EDT
1534    Tag: blume-20010601-nlffi-cleanup
1535    Description:
1536    
1537    This is mostly a cleanup of MLFFI stuff:
1538    
1539       - some signature files have been put into a more exposed place
1540       - the ugly 'f type parameter is gone (simplifies types tremendously!)
1541       - ml-nlffigen changed accordingly
1542       - tutorial updated
1543    
1544    Other changes:
1545    
1546       - author's affiliation in CM manual(s) updated
1547       - some more recognized keywords added to Allen's sml.sty
1548    
1549    ----------------------------------------------------------------------
1550    Name: Matthias Blume
1551    Date: 2001/05/25 15:30:00 EDT
1552    Tag: blume-20010525-iptr
1553    Description:
1554    
1555      - put the official 110.33-README (as it appears on the ftp server) under
1556        CVS
1557      - fixed a small bug related to incomplete pointer types in
1558        ml-nlffigen
1559      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1560        (it does not need the 'f type parameter)
1561    
1562    ----------------------------------------------------------------------
1563    Name: Matthias Blume
1564    Date: 2001/05/23 14:30:00 EDT
1565    Tag: Release_110_33
1566    Description:
1567    
1568    New version number (110.33).  New bootfiles.
1569    
1570    ----------------------------------------------------------------------
1571    Name: Matthias Blume
1572    Date: 2001/05/22 18:06:00 EDT
1573    Tag: blume-20010522-targets
1574    Description:
1575    
1576    Made install.sh use file config/targets.customized if it exists, falling
1577    back to config/targets if it doesn't.  This way one can have a customized
1578    version of the targets file without touching the "real thing", thus
1579    eliminating the constant fear of accidentally checking something bogus
1580    back into the CVS repository...  (File config/targets.customized must
1581    not be added to the repository!)
1582    
1583    ----------------------------------------------------------------------
1584    Name: Matthias Blume
1585    Date: 2001/05/22 16:30:00 EDT
1586    Tag: blume-20010522-minitut
1587    Description:
1588    
1589    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1590       struct returns.
1591    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
1592       incomplete, preliminary documentation for NLFFI.
1593    
1594    ----------------------------------------------------------------------
1595    Name: Matthias Blume
1596    Date: 2001/05/14 11:30:00 EDT
1597    Tag: blume-20010514-script
1598    Description:
1599    
1600    Some bugs in install script fixed.
1601    
1602    In addition to that I also made a slight change to the NLFFI API:
1603    Functors generated by ml-nlffigen now take the dynamic library as a
1604    straight functor argument, not as a suspended one.  (The original
1605    functor code used to force the suspension right away anyway, so there
1606    was nothing gained by this complication of the interface.)
1607    
1608    ----------------------------------------------------------------------
1609    Name: Matthias Blume
1610    Date: 2001/05/11 14:35:00 EDT
1611    Tag: blume-20010511-ml-nlffi
1612    Description:
1613    
1614    I finally took the plunge and added my new FFI code to the main
1615    repository.  For x86-linux it is now ready for prime-time.
1616    
1617    There are two new subdirectories of "src":
1618    
1619      - ml-nlffi-lib:
1620           The utility library for programs using the FFI interface.
1621           Here is the implementation of $/c.cm and its associated low-level
1622           partners $/c-int.cm and $/memory.cm.
1623      - ml-nlffigen:
1624           A stand-alone program for generating ML glue code from C source
1625           code.
1626    
1627    Building ml-nlffigen requires $/ckit-lib.cm.
1628    
1629    The config/install.sh script has been updates to do the Right Thing
1630    (hopefully).
1631    
1632    Notice that the source tree for the C-Kit will not be put under "src"
1633    but directly under the installation root directory.  (This is the
1634    structure that currently exists on the CVS server when you check out
1635    module "sml".)  Fortunately, config/install.sh knows about this oddity.
1636    
1637    Bugs: No documentation yet.
1638    
1639    ----------------------------------------------------------------------
1640    Name: Matthias Blume
1641    Date: 2001/05/09 16:35:00 EDT
1642    Tag: blume-20010509-cpscontract
1643    Description:
1644    
1645    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
1646    wrapper/unwrapper elimination did not decrement usage counts and some
1647    dead variables got overlooked by the dead-up logic.)
1648    
1649    ----------------------------------------------------------------------
1650    Name: Lal George
1651    Date: 2001/05/08  17:26:09 EDT
1652    Tag: george-20010508-omit-frameptr
1653    Description:
1654    
1655    Changes to implement the omit-frame-pointer optimization to support
1656    raw C calls. For now, there is only support on the Intel x86, but
1657    other architectures will follow as more experience is gained with this.
1658    
1659    
1660    ----------------------------------------------------------------------
1661    Name: Matthias Blume
1662    Date: 2001/05/07 14:40:00 EDT
1663    Tag: blume-20010507-proxies
1664    Description:
1665    
1666    I made into "proxy libraries" all libraries that qualify for such a
1667    change.  (A qualifying library is a library that has another library or
1668    groups as its sole member and repeats that member's export list
1669    verbatim.  A proxy library avoids this repetition by omitting its export
1670    list, effectively inheriting the list that its (only) member exports.
1671    See the CM manual for more explanation.)
1672    The main effect is that explicit export lists for these libraries
1673    do not have to be kepts in sync, making maintenance a bit easier.
1674    
1675    I also added copyright notices to many .cm-files.
1676    
1677    Last but not least, I made a new set of bootfiles.
1678    
1679    ----------------------------------------------------------------------
1680    Name: Matthias Blume
1681    Date: 2001/05/04 17:00:00 EDT
1682    Tag: blume-20010504-cm-lsplit
1683    Description:
1684    
1685    0. John merged pending changes to $/smlnj-lib.cm
1686    
1687    1. Allen's previous change accidentally backed out of one of Lal's
1688       earlier changes.  I undid this mistake (re-introducing Lal's change).
1689    
1690    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1691       within the compiler where applicable.  There is some code simplification
1692       because of that.
1693    
1694    3. The "split" phase (in FLINT) is now part of the default list of phases.
1695       Compiler.Control.LambdaSplitting.* can be used to globally control the
1696       lambda-splitting (cross-module-inlining) engine.  In addition to that,
1697       it can now also be controlled on a per-source basis: CM has been taught
1698       a new tool parameter applicable to ML source files.
1699    
1700       - To turn lambda-splitting off completely:
1701            local open Compiler.Control.LambdaSplitting in
1702                val _ = set Off
1703            end
1704       - To make "no lambda-splitting" the global default (but allow per-source
1705         overriding); this is the initial setting:
1706            local open Compiler.Control.LambdaSplitting in
1707                val _ = set (Default NONE)
1708            end
1709       - To make "lambda-splitting with aggressiveness a" the global default
1710         (and allow per-source overriding):
1711            local open Compiler.Control.LambdaSplitting in
1712                val _ = set (Default (SOME a))
1713            end
1714    
1715       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
1716         write (in the respective .cm-file):
1717            a.sml (lambdasplitting:off)
1718       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
1719            a.sml (lambdasplitting:on)
1720       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1721         <a> is a decimal non-negative integer):
1722            a.sml (lambdasplitting:<a>)
1723       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
1724            a.sml (lambdasplitting:infinity)
1725       - To use the global default for a.sml:
1726            a.sml (lambdasplitting:default)
1727         or simply
1728            a.sml
1729    
1730    ----------------------------------------------------------------------
1731    Name: Allen Leung
1732    Date: 2001/05/04 01:57:00 EDT
1733    Tag: leunga-20010504-sync
1734    Description:
1735    
1736      MLRISC features.
1737    
1738      1. Fix to CMPXCHG instructions.
1739      2. Changed RA interface to allow annotations in callbacks.
1740      3. Added a new method to the stream interface to allow annotations updates.
1741    
1742    ----------------------------------------------------------------------
1743    Name: Matthias Blume
1744    Date: 2001/05/01 11:45:00 EDT
1745    Tag: blume-20010501-pcedittmp
1746    Description:
1747    
1748    Changed install.sh to use the current working directory instead of
1749    /usr/tmp for a temporary file (pcedittmp).  The previous choice
1750    of /usr/tmp caused trouble with MacOS X because of file premission
1751    problems.
1752    
1753    ----------------------------------------------------------------------
1754    Name: Matthias Blume
1755    Date: 2001/04/20 11:10:00 EDT
1756    Tag: blume-20010420-inMLflag
1757    Description:
1758    
1759     - added vp_limitPtrMask to vproc-state.h
1760       (for use by the raw-C-calls mechanism to implement proper interrupt
1761        handling)
1762     - made the ML compiler aware of various data-structure offsets so it
1763       can generate code for accessing the vp_inML flag and vp_limitPtrMask
1764     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1765       raw C-calls
1766    
1767    ----------------------------------------------------------------------
1768    Name: Lal George
1769    Date: 2001/04/20 09:15:28 EDT
1770    Tag: george-20010420-macosX
1771    Description:
1772    
1773     - Changes to port to Mac OS X; Darwin.
1774    
1775     - In the process I found that sqrt was broken on the PPC, because the
1776       fsqrt instruction is not implemented.
1777    
1778    ----------------------------------------------------------------------
1779    Name: Matthias Blume
1780    Date: 2001/04/18 12:45:00 EDT
1781    Tag: blume-20010418-ccalls
1782    Description:
1783    
1784     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1785       (this bug prevented structure arguments containing pointers from being
1786        passed correctly)
1787     - changed the raw-C-call code in mlriscGen.sml in such a way that
1788       structure arguments are represented as a pointer to the beginning
1789       of the structure (instead of having a series of synthesized arguments,
1790       one for each structure member)
1791    
1792     - made makeml script's verbosity level configurable via environment
1793       variable (MAKEML_VERBOSITY)
1794    
1795     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1796       in rawmem-x86.sml; we are now using the real thing
1797    
1798    ----------------------------------------------------------------------
1799    Name: Matthias Blume
1800    Date: 2001/03/22 16:25:00 EST
1801    Tag: blume-20010322-bootfiles
1802    Description:
1803    
1804    Created a new set of bootfiles (for your automatic installation convenience).
1805    
1806    ----------------------------------------------------------------------
1807    Name: Matthias Blume
1808    Date: 2001/03/22 15:10:00 EST
1809    Tag: blume-20010322-rawmem-parcm
1810    Description:
1811    
1812    1. All "raw memory access" primitives for the new FFI are implemented now
1813       (at least on the x86).
1814    2. Some further cleanup of CM's parallel make mechanism.
1815    
1816    ----------------------------------------------------------------------
1817    Name: Matthias Blume
1818    Date: 2001/03/19 17:53:00 EST
1819    Tag: blume-20010319-parallel
1820    Description:
1821    
1822    Parallel make (using compile servers) now works again.
1823    
1824    To this end, CM.stabilize and CMB.make have been modified to work in
1825    two passes when compile servers are attached:
1826       1. Compile everything, do not perform stabilization; this pass
1827          uses compile servers
1828       2. Stabilize everything; this pass does not use compile servers
1829    If there are no compile servers, the two passes are combined into one
1830    (as before).  Splitting the passes increases the inherent parallelism
1831    in the dependency graph because the entire graph including all
1832    libraries is available at the same time.  This, in turn, improves
1833    server utilization.  The downside is that the master process will
1834    have to do some extra work after compilation is done (because for
1835    technical reasons it must re-read all the binfiles during stabilization).
1836    
1837    ----------------------------------------------------------------------
1838    Name: Matthias Blume
1839    Date: 2001/03/16 12:22:00 EST
1840    Tag: blume-20010316-bootfiles
1841    Description:
1842    
1843    Created a new set of bootfiles (for your automatic installation convenience).
1844    
1845    ----------------------------------------------------------------------
1846    Name: Matthias Blume
1847    Date: 2001/03/16 11:00:00 EST
1848    Tag: blume-20010316-MLTREE-fixup
1849    Description:
1850    
1851    This is a minor fixup for an (untagged) earlier commit by Allen.
1852    (A file was missing).
1853    
1854    ----------------------------------------------------------------------
1855    Name: Allen Leung
1856    Date: Mon Mar  5 18:54:57 EST 2001
1857    Tag: leunga-20010305-cut-support
1858    
1859    1. New support for alternative control-flow in MLTREE.
1860       Currently we support
1861    
1862          FLOW_TO(CALL ...., [k1,...,kn])
1863    
1864       This is needed for 'cuts to' in C-- and try/handle-like constructs
1865       in Moby
1866    
1867       New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
1868    
1869    2. Register Allocator
1870    
1871       Changes in interface [from Fermin, John]
1872    
1873    3. Alpha 8-bit SLL support [Fermin]
1874    
1875    4. All architectures
1876    
1877       A new module (ClusterExpandCopies) for expanding parallel copies.
1878    
1879    ----------------------------------------------------------------------
1880    Name: Allen Leung
1881    Date: 2001/02/27 23:07:00 EST
1882    Tag: leunga-20010227-minor-stuff
1883    
1884    1. Alpha bug fix for CMOVNE
1885    2. Handle mltree COND(..,FCMP ...,...)
1886    3. Bug fix in simplifier
1887    
1888    ----------------------------------------------------------------------
1889    Name: Matthias Blume
1890    Date: 2001/01/30 17:50:00 EST
1891    Tag: blume-20010130-sync
1892    Description:
1893    
1894    This is just a minor update to sync my devel branch with the main brach.
1895    The only visible change is the addition of some README files.
1896    
1897    ----------------------------------------------------------------------
1898    Name: Matthias Blume
1899    Date: 2001/01/12 23:30:00 JST
1900    Tag: blume-20010112-bootfiles
1901    Description:
1902    
1903    Made a new set of bootfiles that goes with the current state of the
1904    repository.
1905    
1906    ----------------------------------------------------------------------
1907    Name: Matthias Blume
1908    Date: 2001/01/12 21:20:00 JST
1909    Tag: blume-20010112-sync
1910    Description:
1911    
1912    I am just flushing out some minor changes that had accumulated in
1913    my private branch in order to sync with the main tree.  (This is
1914    mainly because I had CVS trouble when trying to merge _into_ my
1915    private branch.)
1916    
1917    Most people should be completely unaffected by this.
1918    
1919    ----------------------------------------------------------------------
1920    Name: Allen Leung
1921    Date: Thu Jan 11 21:03:00 EST 2001
1922    Tag: leunga-20010111-labexp=mltree
1923    Description:
1924    
1925    1.  Removed the type LabelExp and replace it by MLTree.
1926    2.  Rewritten mltree-simplify with the pattern matcher tool.
1927    3.  There were some bugs in alpha code generator which would break
1928        64-bit code generation.
1929    4.  Redo the tools to generate code with the
1930    5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
1931        are now generated by perl scripts.
1932    
1933    ----------------------------------------------------------------------
1934    Name: Matthias Blume
1935    Date: 2001/01/10 21:55:00 JST
1936    Tag: blume-20010110-rcc
1937    Description:
1938    
1939    The RCC stuff now seems to work (but only on the x86).
1940    This required hacking of the c-calls interface (and -implementation) in
1941    MLRISC.
1942    
1943    Normal compiler users should be unaffected.
1944    
1945    ----------------------------------------------------------------------
1946    Name: Matthias Blume
1947    Date: 2001/01/09 01:20:00 JST
1948    Tag: blume-20010109-rcc
1949    Description:
1950    
1951    This is a fairly big patch, flushing out a large number of pending
1952    changes that I made to my development copy over the last couple of days.
1953    
1954    Of practical relevance at this moment is a workaround for a pickling
1955    bug that Allen ran into the other day.  The cause of the bug itself is
1956    still unknown and it might be hard to fix it properly, but the
1957    workaround has some merits of its own (namely somewhat reducing pickling
1958    overhead for certain libraries).  Therefore, I think this solution should
1959    be satisfactory at this time.
1960    
1961    The rest of the changes (i.e., the vast majority) has to do with my
1962    ongoing efforts of providing direct support for C function calls from
1963    ML.  At the moment there is a new primop "RAW_CCALL", typing magic
1964    in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
1965    case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
1966    translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
1967    handle RCC.
1968    
1969    The last part (the changes to mlriscGen.sml) are still known to be
1970    wrong on the x86 and not implemented on all other architectures.  But
1971    the infrastructure is in place. I had to change a few functor
1972    signatures in the backend to be able to route the CCalls interface
1973    from MLRISC there, and I had to specialize the mltree type (on the
1974    x86) to include the necessary extensions. (The extensions themselves
1975    were already there and redy to go in MLRISC/x86).
1976    
1977    Everything should be very happy as soon as someone helps me with
1978    mlriscGen.sml...
1979    
1980    In any case, nothing of this should matter to anyone as long as the
1981    new primop is not being used (which is going to be the case unless you
1982    find it where I hid it :). The rest of the compiler is completely
1983    unaffected.
1984    
1985    ----------------------------------------------------------------------
1986    Name: Matthias Blume
1987    Date: 2001/01/05 00:30:00 JST
1988    Tag: blume-20010105-primops
1989    Description:
1990    
1991    Added some experimental support for work that I am doing right now.
1992    These changes mostly concern added primops, but there is also a new
1993    experimental C library in the runtime system (but currently not enabled
1994    anywhere except on Linux/X86).
1995    
1996    In the course of adding primops (and playing with them), I discovered that
1997    Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
1998    broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
1999    To recover, I made type information in INL_PRIM mandatory and changed
2000    prim.sml as well as built-in.sml accordingly.  The InLine structure now
2001    has complete, correct type information (i.e., no bottom types).
2002    
2003    Since all these changes mean that we need new binfiles, I also bumped the
2004    version number to 110.32.1.
2005    
2006    ----------------------------------------------------------------------
2007    Name: Matthias Blume
2008    Date: 2000/12/30 22:10:00 JST
2009    Tag: blume-20001230-various
2010    Description:
2011    
2012    Added proxy libraries for MLRISC and let MLRISC libraries refer
2013    to each other using path anchors.  (See CM manual for explanation.)
2014    
2015    Updated CM documentation.
2016    
2017    Fixed some bugs in CM.
2018    
2019    Implemented "proxy" libraries (= syntactic sugar for CM).
2020    
2021    Added "-quiet" option to makeml and changed runtime system accordingly.
2022    
2023    Added cleanup handler for exportML to reset timers and compiler stats.
2024    
2025    ----------------------------------------------------------------------
2026    Name: Lal George
2027    Date: 2000/12/22 22:22:58 EST 2000
2028    Tag: Release_110_32
2029    Description:
2030    
2031            Infinite precision used throughout MLRISC.
2032            see MLRISC/mltree/machine-int.sig
2033    
2034    ----------------------------------------------------------------------
2035    Name: Matthias Blume
2036    Date: 2000/12/22 23:16:00 JST
2037    Tag: blume-20001222-warn
2038    Description:
2039    
2040    Corrected wording and formatting of some CM warning message which I
2041    broke in my previous patch.
2042    
2043    ----------------------------------------------------------------------
2044    Name: Matthias Blume
2045    Date: 2000/12/22 21:20:00 JST
2046    Tag: blume-20001222-anchorenv
2047    Description:
2048    
2049    Fixed CM's handling of anchor environments in connection with CMB.make.
2050    
2051    ----------------------------------------------------------------------
2052    Name: Matthias Blume
2053    Date: 2000/12/22 13:15:00 JST
2054    Tag: blume-20001222-cleanup
2055    Description:
2056    
2057    Removed src/cm/ffi which does not (and did not) belong here.
2058    
2059    ----------------------------------------------------------------------
2060    Name: Matthias Blume
2061    Date: 2000/12/21 23:55:00 JST
2062    Tag: blume-20001221-exn
2063    Description:
2064    
2065    Probably most important: CM no longer silently swallows all exceptions
2066    in the compiler.
2067    Plus: some other minor CM changes.  For example, CM now reports some
2068    sizes for generated binfiles (code, data, envpickle, lambdapickle).
2069    
2070    ----------------------------------------------------------------------
2071    Name: Matthias Blume
2072    Date: 2000/12/15 00:01:05 JST
2073    Tag: blume-20001215-dirtool
2074    Description:
2075    
2076    - "dir" tool added.
2077    - improvements and cleanup to Tools structure
2078    - documentation updates
2079    
2080    ----------------------------------------------------------------------
2081    Name: Allen Leung
2082    Date: Thu Dec 14 03:45:24 EST 2000
2083    Description:
2084    Tag:  leunga-20001214-int-inf
2085    Description:
2086    
2087       In IntInf, added these standard functions, which are missing from our
2088    implementation:
2089    
2090        andb : int * int -> int
2091        xorb : int * int -> int
2092        orb  : int * int -> int
2093        notb : int -> int
2094         <<   : int * word -> int
2095        ~>>  : int * word -> int
2096    
2097       Not tested, I hope they are correct.
2098    
2099    ----------------------------------------------------------------------
2100    Name: Allen Leung
2101    Date: Fri Dec  8 19:23:26 EST 2000
2102    Description:
2103    Tag:  leunga-20001208-nowhere
2104    Description:
2105    
2106      Slight improvements to the 'nowhere' tool to handle OR-patterns,
2107    to generate better error messages etc.  Plus a brief manual.
2108    
2109    ----------------------------------------------------------------------
2110    Name: Lal George
2111    Date: 2000/12/08 09:54:02 EST 2000
2112    Tag: Release_110_31
2113    Description:
2114    
2115    - Version 110.31
2116    ----------------------------------------------------------------------
2117    Name: Allen Leung
2118    Date: Thu Dec  7 22:01:04 EST 2000
2119    Tag:  leunga-20001207-cell-monster-hack
2120    Description:
2121    
2122    Major MLRISC internal changes.  Affect all clients.
2123    Summary:
2124    
2125    1.  Type CELLS.cell = int is now replaced by a datatype.
2126        As a result, the old regmap is now gone.  Almost all interfaces
2127        in MLRISC change as a consequence.
2128    
2129    2.  A new brand version of machine description tool (v3.0) that generates
2130        modules expecting the new interface.  The old version is removed.
2131    
2132    3.  The RA interface has been further abstracted into two new functors.
2133        RISC_RA and X86RA.  These functors have much simpler interfaces.
2134        [See also directory MLRISC/demo.]
2135    
2136    4.  Some other new source->source code generation tools are available:
2137    
2138        a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
2139        b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
2140           I use this tool to generate the peephole optimizers---with the new
2141           cell type changes, peephole rules are becoming difficult to write
2142           without conditional pattern matching.
2143    
2144    5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
2145        cover the entire MLRISC source tree so many things broke.
2146    
2147    6.  CM files have been moved to the subdirectory MLRISC/cm.
2148        They are moved because there are a lot of them and they clutter up the
2149        root dir.
2150    
2151    7.  More detailed documentation to come...
2152    
2153        NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
2154        a makeml -rebuild first.  This is because of other other
2155        changes that Matthias has made (see below).
2156    
2157    
2158    ----------------------------------------------------------------------
2159    Name: Matthias Blume
2160    Date: 2000/11/30 23:12:00 JST
2161    Tag: blume-20001130-filereorg
2162    Description:
2163    
2164    Some manual updates and some file reorganizations in CM.
2165    
2166    ----------------------------------------------------------------------
2167    Name: Matthias Blume
2168    Date: 2000/11/24 17:45:00 JST
2169    Tag: blume-20001124-link
2170    Description:
2171    
2172    Drastically improved link traversal code for the case that the dynamic
2173    value was already loaded at bootstrap time.  As a result, CM and CMB
2174    now both load blazingly fast -- even on a very slow machine.  Also,
2175    memory consumption has been further reduced by this.
2176    
2177    Warning: The format of the PIDMAP file has changed.  THerefore, to
2178    bootstrap you have to do this:
2179    
2180    1. Run CMB.make
2181    2. Make a symbolic link for the boot directory:
2182         ln -s sml.boot.ARCH-OS xxx
2183    3. "Rebuild" the boot directory:
2184         ./makeml -boot xxx -rebuild sml ; rm xxx
2185    4. Boot normally:
2186          ./makeml
2187    
2188    ----------------------------------------------------------------------
2189    Name: Matthias Blume
2190    Date: 2000/11/21 21:20:00 JST
2191    Tag: blume-20001121-tools
2192    Description:
2193    
2194    Continued hacking on autoloading problem -- with success this time.
2195    Also changed tool-plugin mechanism.  See new CM manual.
2196    
2197    ----------------------------------------------------------------------
2198    Name: Matthias Blume
2199    Date: 2000/11/19 14:30:00 JST
2200    Tag:  blume-20001119-autoload
2201    Description:
2202    
2203    Some hacking to make autoloading faster.  Success for CMB, no success
2204    so far for CM.  There is a reduced structure CM' that autoloads faster.
2205    (This is a temporary, non-documented hack to be eliminated again when
2206    the general problem is solved.)
2207    
2208    ----------------------------------------------------------------------
2209    Name: Matthias Blume
2210    Date: 2000/11/17 14:10:00 JST
2211    Tag: blume-20001117-pickle-lib
2212    Description:
2213    
2214    1. Eliminated comp-lib.cm
2215    2. Made pickle-lib.cm
2216    3. Eliminated all uses of intset.sml (from comp-lib.cm)
2217    4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
2218       equivalent constructs from smlnj-lib.cm (INtHashTable).
2219    5. Point 4. also goes for those uses of intmap.* in MLRISC.
2220       Duplicated intmap modules thrown out.
2221    6. Hunted down all duplicated SCC code and replaced it with
2222       equivalent stuff (GraphSCCFn from smlnj-lib.cm).
2223    7. Rewrote Feedback module.
2224    8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
2225       should be thrown out and equivalent modules from smlnj-lib.cm
2226       should be used (IntRedBlackSet, IntListSet, ...).
2227    
2228    Confirmed that compiler compiles to fixpoint.
2229    
2230    ----------------------------------------------------------------------
2231    Name: Allen Leung
2232    Date: 2000/11/10 18:00:00
2233    Tag: leunga-20001110-new-x86-fp
2234    
2235    A new x86 floating point code generator has been added.
2236    By default this is turned off.  To turn this on, do:
2237    
2238        CM.autoload "$smlnj/compiler.cm";
2239        Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
2240    
2241    Changes:
2242    
2243    1.  Changed FTAN to FPTAN so that the assembly output is correct.
2244    2.  Changed the extension callback for FTANGENT to generate:
2245    
2246              fptan
2247              fstp  %st(0)
2248        instead of
2249              fptan
2250              fstpl ftempmem
2251    
2252    3.  Numerous assembly fixes for x86.
2253    
2254    5.  Cleaned up the machine code output module x86/x86MC.sml and added
2255        support for a whole bunch of instructions and addressing modes:
2256    
2257          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
2258          faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
2259          fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
2260          fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
2261          fxch %st(n)
2262          fld %st(n)
2263          fst %st(n)
2264          fst mem
2265          fstp %st(n)
2266          fucom %st(n)
2267          fucomp %st(n)
2268    
2269        All these are now generated when the fast fp mode is turned on.
2270    
2271    6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
2272    
2273    ----------------------------------------------------------------------
2274    Name: Matthias Blume
2275    Date: 2000/11/09 11:20:00 JST
2276    Tag: blume-20001109-scc
2277    Description:
2278    
2279    Eliminated some code duplication:
2280    
2281    1. Added "where" clause to GraphSCCFn in SML/NJ Library.
2282       (Otherwise the functor is useless.)
2283    2. Used GraphSCCFn where SCCUtilFun was used previously.
2284    3. Got rid of SCCUtilFun (in comp-lib.cm).
2285    
2286    ----------------------------------------------------------------------
2287  Name: Lal George  Name: Lal George
2288  Date: 2000/11/06 09:02:21 EST 2000  Date: 2000/11/06 09:02:21 EST 2000
2289  Tag: Release_110_30  Tag: Release_110_30
# Line 1554  Line 3825 
3825       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3826       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3827       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
3828       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3829       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3830       Dave's hands. :)       Dave's hands. :)
3831    

Legend:
Removed from v.721  
changed lines
  Added in v.1068

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