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 812, Fri Apr 20 15:07:30 2001 UTC revision 1064, Thu Feb 14 03:40:24 2002 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16  Name: Matthias Blume  Name: Matthias Blume
17    Date: 2002/02/13 22:40:00 EST
18    Tag: blume-20020213-fptr-rtti
19    Description:
20    
21    Fixed a bug in ml-nlffigen that was introduced with one of the previous
22    updates.
23    
24    ----------------------------------------------------------------------
25    Name: Matthias Blume
26    Date: 2002/02/13 16:41:00 EST
27    Tag: blume-20020213-cmlpq
28    Description:
29    
30    Added new priority queue export symbols (which have just been added to
31    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
32    not compile and the installer would choke.)
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume
36    Date: 2002/02/13 16:15:00 EST
37    Tag: blume-20020213-various
38    Description:
39    
40    1. More tweaks to ml-nlffigen:
41    
42       - better internal datastructures (resulting in slight speedup)
43       - "-match" option requires exact match
44       - "localized" gensym counters (untagged structs/unions nested within
45         other structs/unions or within typedefs get a fresh counter; their
46         tag will be prefixed by a concatenation of their parents' tags)
47       - bug fixes (related to calculation of transitive closure of types
48         to be included in the output)
49    
50    2. Minor Basis updates:
51    
52       - added implementations for List.collate and Option.app
53    
54    ----------------------------------------------------------------------
55    Name: Matthias Blume
56    Date: 2002/02/11 15:55:00 EST
57    Tag: blume-20020211-gensym
58    Description:
59    
60    Added a "-gensym" option to command line of ml-nlffigen.  This can be
61    used to specify a "stem" -- a string that is inserted in all "gensym'd"
62    names (ML structure names that correspond to unnamed C structs, unions,
63    and enums), so that separate runs of ml-nlffigen do not clash.
64    
65    ----------------------------------------------------------------------
66    Name: Matthias Blume
67    Date: 2002/02/11 12:05:00 EST
68    Tag: blume-20020211-gensml
69    Description:
70    
71    A quick fix for a problem with GenSML (in the pgraph-util library):
72    Make generation of toplevel "local" optional.  (Strictly speaking,
73    signature definitions within "local" are not legal SML.)
74    
75    Other than that: updates to INSTALL and cm/TODO.
76    
77    ----------------------------------------------------------------------
78    Name: Matthias Blume
79    Date: 2002/02/08 15:00:00 EST
80    Tag: blume-20020208-uniquepid
81    Description:
82    
83    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
84    
85    1. The installer (config/install.sh) has gotten smarter:
86    
87         - Configuration options are a bit easier to specify now
88           (in config/targets).
89         - Bug in recognizing .tar.bz2 files fixed.
90         - Installer automatically resolves dependencies between
91           configuration options (e.g., if you ask for eXene, you will
92           also get cml -- regardless whether you asked for it or not).
93         - Installer can run in "quieter mode" by setting the environment
94           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
95           "completely silent", though.
96         - Build HashCons library as part of smlnj-lib.
97    
98    2. A new scheme for assigning persistent identifiers to compilation
99       units (and, by extension, to types etc.) has been put into place.
100       This fixes a long-standing bug where types and even dynamic values
101       can get internally confused, thereby compromising type safety
102       (abstraction) and dynamic correctness.  See
103    
104         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
105    
106       for an example of how things could go wrong until now.
107    
108       The downside of the new scheme is that pids are not quite as
109       persistent as they used to be: CM will generate a fresh pid
110       for every compilation unit that it thinks it sees for the first
111       time.  That means that if you compile starting from a clean, fresh
112       source tree at two different times, you end up with different
113       binaries.
114    
115       Cutoff recompilation, however, has not been compromised because
116       CM keeps pid information in special caches between runs.
117    
118    ----------------------------------------------------------------------
119    Name: Lal George
120    Date: 2002/02/07 15:34:13 EST 2002
121    Tag: <none>
122    Description:
123    
124    Compilers that generate assembly code may produce  global labels
125    whose value is resolved at link time. The various peephole optimization
126    modules did not take this in account.
127    
128    TODO. The Labels.addrOf function should really return an option
129    type so that clients are forced to deal with this issue, rather
130    than an exception being raised.
131    
132    ----------------------------------------------------------------------
133    Name: Lal George
134    Date: 2002/02/06 13:55:02 EST
135    Tag: george-20020206-ra-breakup
136    Description:
137    
138    1. A bug fix from Allen.
139    
140        A typo causes extra fstp %st(0)'s to be generated at compensation
141        edges, which might cause stack underflow traps at runtime.  This
142        occurs in fft where there are extraneous fstps right before the 'into'
143        trap instruction (in this case they are harmless since none of the
144        integers overflow.)
145    
146    2. Pulled out various utility modules that were embedded in the modules
147       of the register allocator. I need these modules for other purposes, but
148       they are not complete enough to put into a library (just yet).
149    ----------------------------------------------------------------------
150    Name: Matthias Blume
151    Date: 2002/01/31 16:05:00 EST
152    Tag: blume-20020131-sparc-ccalls
153    Description:
154    
155    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
156       of extra stack space by mistake.  Fixed.
157    
158    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
159    
160    ----------------------------------------------------------------------
161    Name: Allen Leung
162    Date: 2002/01/30
163    Tag: leunga-20020130-nowhere-bug-fix
164    Description:
165    
166       MLRISC bug fixes:
167       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
168       2. MachineInt.fromString was negating its value.
169    
170    ----------------------------------------------------------------------
171    Name: Matthias Blume
172    Date: 2002/01/29
173    Tag: blume-20020129-INSTALL
174    Description:
175    
176    - Added somewhat detailed installation instructions (file INSTALL).
177    - Fixed curl-detection bug in config/install.sh.
178    - It is now possible to select the URL getter using the URLGETTER
179      environment variable:
180    
181          not set / "unknown"      --> automatic detection (script tries wget,
182                                       curl, and lynx)
183          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
184                                       how to properly invoke them)
185          other                    --> use $URLGETTER directly, it must take
186                                       precisely two command-line arguments
187                                       (source URL and destination file name)
188    
189    ----------------------------------------------------------------------
190    Name: Matthias Blume
191    Date: 2002/01/28
192    Tag: blume-20020128-sparc-ccalls
193    Description:
194    
195    - Fixed problem with calculation of "used" registers in sparc-c-calls.
196    - Make use of the allocParam argument in sparc-c-calls.
197    
198    ----------------------------------------------------------------------
199    Name: Matthias Blume
200    Date: 2002/01/28
201    Tag: blume-20020128-allocParam
202    Description:
203    
204    John Reppy:  Changes c-calls API to accept client-callback for
205    allocating extra stack space.
206    me: Corresponding changes to mlriscGen (using a dummy argument that
207        does not change the current behavior).
208    
209    ----------------------------------------------------------------------
210    Name: Matthias Blume
211    Date: 2002/01/28 12:00:00
212    Tag: Release_110_38
213    Description:
214    
215    This time for real!!!
216    
217    ----------------------------------------------------------------------
218    Name: Matthias Blume
219    Date: 2002/01/28 10:56:00 EST
220    Tag: blume-20020128-retraction
221    Description:
222    
223    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
224       with blume-Release_110_38-retracted.)
225    
226    1. Fixed a problem with incorrect rounding modes in real64.sml.
227       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
228    
229    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
230       and enums fixed.  The naming of corresponding ML identifiers should
231       now be consistent again.
232    
233    ----------------------------------------------------------------------
234    Name: Allen Leung
235    Date: 2002/01/27
236    Tag: leunga-20020127-nowhere
237    Description:
238    
239       Added a target called nowhere in the configuration scripts.
240       Enabling this will build the MLRISC 'nowhere' tool (for translating
241       programs with where-clauses into legal SML code) during installation.
242    
243    ----------------------------------------------------------------------
244    Name: Matthias Blume
245    Date: 2002/01/25 21:27:00 EST
246    Tag: blume-Release_110_38-retracted
247    Description:
248    
249    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
250    
251    README will be added later.
252    
253    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
254    (Release_110_38) removed.  Reason: Last-minute bug fixes.
255    
256    ----------------------------------------------------------------------
257    Name: Matthias Blume
258    Date: 2002/01/25
259    Tag: blume-20020125-ffi
260    Description:
261    
262    A large number of tweaks and improvements to ml-nlffi-lib and
263    ml-nlffigen:
264    
265       - ML represenation types have been streamlined
266       - getter and setter functions work with concrete values, not abstract
267         ones where possible
268       - ml-nlffigen command line more flexible (see README file there)
269       - some bugs have been fixed (hopefully)
270    
271    ----------------------------------------------------------------------
272    Name: Lal George
273    Date: 2002/01/24
274    Tag: george-20020124-risc-ra-interface
275    Description:
276    
277       There is a dramatic simplification in the interface to the
278       register allocator for RISC architectures as a result of making
279       parallel copy instructions explicit.
280    
281    ----------------------------------------------------------------------
282    Name: Matthias Blume
283    Date: 2002/01/22
284    Tag: blume-20020122-x86-ccalls
285    Description:
286    
287    Bug fix for c-calls on x86 (having to do with how char- and
288    short-arguments are being handled).
289    
290    ----------------------------------------------------------------------
291    Name: Matthias Blume
292    Date: 2002/01/21
293    Tag: blume-20020121-ff
294    Description:
295    
296    Another day of fiddling with the FFI...
297    
298    1. Bug fix/workaround:  CKIT does not complain about negative array
299       dimensions, so ml-nlffigen has to guard itself against this possibility.
300       (Otherwise a negative dimension would send it into an infinite loop.)
301    
302    2. Some of the abstract types (light objects, light pointers, most "base"
303       types) in structure C are now eqtypes.
304    
305    3. Added constructors and test functions for NULL function pointers.
306    
307    ----------------------------------------------------------------------
308    Name: Matthias Blume
309    Date: 2002/01/18
310    Tag: blume-20020118-ready-for-new-release
311    Description:
312    
313    Made config/srcarchiveurl point to a new place.  (Will provide boot
314    files shortly.)
315    
316    Maybe we christen this to be 110.38?
317    
318    ----------------------------------------------------------------------
319    Name: Matthias Blume
320    Date: 2002/01/18
321    Tag: blume-20020118-more-ffifiddle
322    Description:
323    
324    Today's FFI fiddling:
325    
326      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
327        C.Set.* that use concrete (MLRep.*) arguments and results instead
328        of abstract ones.
329    
330      - Provided word-style bit operations etc. for "int" representation
331        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
332        and Long.
333    
334    ----------------------------------------------------------------------
335    Name: Matthias Blume
336    Date: 2002/01/18
337    Tag: blume-20020118-use-x86-fp
338    Description:
339    
340    Now that x86-fast-fp seems to be working, I turned it back on again
341    by default.  (Seems to work fine now, even with the FFI.)
342    
343    Other than that, I added some documentation about the FFI to
344    src/ml-nlffigen/README and updated the FFI test examples in
345    src/ml-nlffi-lib/Tests/*.
346    
347    ----------------------------------------------------------------------
348    Name: Allen Leung
349    Date: 2002/01/17
350    Tag: leunga-20020117-x86-fast-fp-call
351    Description:
352    
353       1. Fixed a problem with handling return fp values when x86's fast fp
354          mode is turned on.
355    
356       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
357          of %f32.
358    
359       3. Added a constructor INT32lit to the ast of MLRISC tools.
360    
361    ----------------------------------------------------------------------
362    Name: Matthias Blume
363    Date: 2002/01/16
364    Tag: blume-20020116-ffifiddle
365    Description:
366    
367    More fiddling with the FFI interface:
368    
369     - Make constness 'c instead of rw wherever possible.  This eliminates
370       the need for certain explicit coercions.  (However, due to ML's
371       value polymorphism, there will still be many cases where explicit
372       coercions are necessary.  Phantom types are not the whole answer
373       to modeling a subtyping relationship in ML.)
374    
375     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
376    
377     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
378       clumsy workaround for varargs problem.)
379    
380    ----------------------------------------------------------------------
381    Name: Lal George
382    Date: 2002/01/15
383    Tag: <none>
384    Description:
385    
386    1. Since COPY instructions are no longer native to the architecture,
387       a generic functor can be used to implement the expandCopies function.
388    
389    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
390       TEXT segment.
391    
392    ----------------------------------------------------------------------
393    Name: Matthias Blume
394    Date: 2002/01/15
395    Tag: blume-20020115-ffiupdates
396    Description:
397    
398    1. Fix for bug resulting in single-precision float values being returned
399       incorrectly from FFI calls.
400    
401    2. Small modifications to C FFI API:
402    
403        - memory-allocation routines return straight objects (no options)
404          and raise an exception in out-of-memory situations
405        - unsafe extensions to cast between function pointers and pointers
406          from/to ints
407        - added structure C_Debug as an alternative to structure C where
408          pointer-dereferencing (|*| and |*!) always check for null-pointers
409        - added open_lib' to DynLinkage;  open_lib' works like open_lib
410          but also takes a (possibly empty) list of existing library handles
411          that the current library depends on
412    
413    ----------------------------------------------------------------------
414    Name: Matthias Blume
415    Date: 2002/01/10
416    Tag: blume-20020110-newffigen
417    Description:
418    
419    1. Updates to portable graph code.
420    
421    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
422       more scalable now so that even huge interfaces such as the one
423       for GTK compile in finite time and space. :-)
424       See src/ml-nlffigen/README for details on what's new.
425    
426    ----------------------------------------------------------------------
427    Name: Lal George
428    Date: 2001/01/09 14:31:35 EST 2002
429    Tag: george-20011206-rm-native-copy
430    Description:
431    
432            Removed the native COPY and FCOPY instructions
433            from all the architectures and replaced it with the
434            explicit COPY instruction from the previous commit.
435    
436            It is now possible to simplify many of the optimizations
437            modules that manipulate copies. This has not been
438            done in this change.
439    
440    ----------------------------------------------------------------------
441    Name: Lal George
442    Date: 2001/12/06 16:50:13 EST 2001
443    Tag: george-20011206-mlrisc-instruction
444    Description:
445    
446    Changed the representation of instructions from being fully abstract
447    to being partially concrete. That is to say:
448    
449      from
450            type instruction
451    
452      to
453            type instr                              (* machine instruction *)
454    
455            datatype instruction =
456                LIVE of {regs: C.cellset, spilled: C.cellset}
457              | KILL of {regs: C.cellset, spilled: C.cellset}
458              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
459              | ANNOTATION of {i: instruction, a: Annotations.annotation}
460              | INSTR of instr
461    
462    This makes the handling of certain special instructions that appear on
463    all architectures easier and uniform.
464    
465    LIVE and KILL say that a list of registers are live or killed at the
466    program point where they appear. No spill code is generated when an
467    element of the 'regs' field is spilled, but the register is moved to
468    the 'spilled' (which is present, more for debugging than anything else).
469    
470    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
471    We used to generate:
472    
473            DEFFREG f1
474            f1 := f2 + f3
475            trapb
476    
477    but now generate:
478    
479            f1 := f2 + f3
480            trapb
481            LIVE {regs=[f1,f2,f3], spilled=[]}
482    
483    Furthermore, the DEFFREG (hack) required that all floating point instruction
484    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
485    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
486    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
487    this architecture).
488    
489    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
490    all the architectures. This will result in further simplification of the
491    register allocator that must be aware of them for coalescing purposes, and
492    will also simplify certain aspects of the machine description that provides
493    callbacks related to parallel copies.
494    
495    ANNOTATION should be obvious, and now INSTR represents the honest to God
496    machine instruction set!
497    
498    The <arch>/instructions/<arch>Instr.sml files define certain utility
499    functions for making porting easier -- essentially converting upper case
500    to lower case. All machine instructions (of type instr) are in upper case,
501    and the lower case form generates an MLRISC instruction. For example on
502    the alpha we have:
503    
504      datatype instr =
505         LDA of {r:cell, b:cell, d:operand}
506       | ...
507    
508      val lda : {r:cell, b:cell, d:operand} -> instruction
509        ...
510    
511    where lda is just (INSTR o LDA), etc.
512    
513    ----------------------------------------------------------------------
514    Name: Matthias Blume
515    Date: 2001/11/22 21:40:00 EST
516    Tag: Release_110_37
517    Description:
518    
519    Release 110.37.  This time for real.
520    
521    ----------------------------------------------------------------------
522    Name: Matthias Blume
523    Date: 2001/11/21 16:35:00 EST
524    Tag: blume-20011121-foot-in-mouth
525    Description:
526    
527    Removed the "Release_110_37" tag because of a serious bug.
528    This will be re-tagged once the bug is fixed.
529    
530    ----------------------------------------------------------------------
531    Name: Matthias Blume
532    Date: 2001/11/21 16:14:00 EST
533    Tag: blume-20011121-forgottenfile
534    Description:
535    
536    Forgot to add a file.  (Just a .tex-file -- part of
537    the CM manual source.)
538    
539    ----------------------------------------------------------------------
540    Name: Matthias Blume
541    Date: 2001/11/21 16:10:00 EST
542    Tag: blume-20011121-invalid_110_37
543    Description:
544    
545    Note: I removed the original tag "Release_110_37" from this commit
546          because we found a serious bug in all non-x86 backends.
547          - Matthias
548    
549    1. Modifications to the SML/NJ code generator and to the runtime system
550       so that code object name strings are directly inserted into code
551       objects at code generation time.  The only business the runtime system
552       has with this is now to read the name strings on occasions.
553       (The encoding of the name string has also changed somewhat.)
554    
555    2. CM now implements a simple "set calculus" for specifying export lists.
556       In particular, it is now possible to refer to the export lists of
557       other libraries/groups/sources and form unions as well as differences.
558       See the latest CM manual for details.
559    
560    3. An separate notion of "proxy" libraries has again be eliminated from
561       CM's model.  (Proxy libraries are now simply a special case of using
562       the export list calculus.)
563    
564    4. Some of the existing libraries now take advantage of the new set
565       calculus.
566       (Notice that not all libraries have been converted because some
567       of the existing .cm-files are supposed to be backward compatible
568       with 110.0.x.)
569    
570    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
571       "General.exnMessage"!  The former relies on a certain hook to be
572       initialized, and that often does not happen in the stand-alone case.)
573    
574    ----------------------------------------------------------------------
575    Name: Lal George
576    Date: 2001/11/21  13:56:18 EST
577    Tag: george-2001121-pseudo-ops
578    Description:
579    
580      Implemented a complete redesign of MLRISC pseudo-ops. Now there
581      ought to never be any question of incompatabilities with
582      pseudo-op syntax expected by host assemblers.
583    
584      For now, only modules supporting GAS syntax are implemented
585      but more should follow, such as MASM, and vendor assembler
586      syntax, e.g. IBM as, Sun as, etc.
587    
588    ----------------------------------------------------------------------
589    Name: Matthias Blume
590    Date: 2001/11/14 11:52:00 EST
591    Tag: blume-20011114-srcname
592    Description:
593    
594    1. Routed the name of the current source file to mlriscgen where it
595       should be directly emitted into the code object.  (This last part
596       is yet to be done.)
597    
598    2. Some cleanup of the pgraph code to make it match the proposal that
599       I put out the other day.  (The proposal notwithstanding, things are
600       still in flux here.)
601    
602    ----------------------------------------------------------------------
603    Name: Lal George
604    Date: 2001/11/14 09:44:04 EST
605    Tag:
606    Description:
607    
608      Fix for a backpatching bug reported by Allen.
609    
610      Because the boundary between short and long span-dependent
611      instructions is +/- 128, there are an astounding number of
612      span-dependent instructions whose size is over estimated.
613    
614      Allen came up with the idea of letting the size of span
615      dependent instructions be non-monotonic, for a maxIter
616      number of times, after which the size must be monotonically
617      increasing.
618    
619      This table shows the number of span-dependent instructions
620      whose size was over-estimated as a function of maxIter, for the
621      file Parse/parse/ml.grm.sml:
622    
623         maxIter            # of instructions:
624            10                      687
625            20                      438
626            30                      198
627            40                        0
628    
629      In compiling the compiler, there is no significant difference in
630      compilation speed between maxIter=10 and maxIter=40. Actually,
631      my measurements showed that maxIter=40 was a tad faster than
632      maxIter=10! Also 96% of the  files in the compiler reach a fix
633      point within 13 iterations, so fixing maxIter at 40, while high,
634      is okay.
635    
636    ----------------------------------------------------------------------
637    Name: Matthias Blume
638    Date: 2001/10/31 15:25:00 EST
639    Tag: blume-20011031-pgraph
640    Description:
641    
642    CKIT:
643    * Changed the "Function" constructor of type Ast.ctype to carry optional
644      argument identifiers.
645    * Changed the return type of TypeUtil.getFunction accordingly.
646    * Type equality ignores the argument names.
647    * TypeUtil.composite tries to preserve argument names but gives up quickly
648      if there is a mismatch.
649    
650    installation script:
651    * attempts to use "curl" if available (unless "wget" is available as well)
652    
653    CM:
654    * has an experimental implementation of "portable graphs" which I will
655      soon propose as an implementation-independent library format
656    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
657    
658    NLFFI-LIB:
659    * some cleanup (all cosmetic)
660    
661    NLFFIGEN:
662    * temporarily disabled the mechanism that suppresses ML output for
663      C definitions whose identifiers start with an underscore character
664    * generate val bindings for enum constants
665    * user can request that only one style (light or heavy) is being used;
666      default is to use both (command-line arguments: -heavy and -light)
667    * fixed bug in handling of function types involving incomplete pointers
668    * generate ML entry points that take record arguments (i.e., using
669      named arguments) for C functions that have a prototype with named
670      arguments
671      (see changes to CKIT)
672    
673    ----------------------------------------------------------------------
674    Name: Allen Leung
675    Date: 2001/10/27 20:34:00 EDT
676    Tag: leunga-20011027-x86-fast-fp-call
677    Description:
678    
679       Fixed the bug described in blume-20010920-slowfp.
680    
681       The fix involves
682          1. generating FCOPYs in FSTP in ia32-svid
683          2. marking a CALL with the appropriate annotation
684    
685    ----------------------------------------------------------------------
686    Name: Matthias Blume
687    Date: 2001/10/16 11:32:00 EDT
688    Tag: blume-20011016-netbsd
689    Description:
690    
691    Underscore patch from Chris Richards (fixing problem with compiling
692    runtime system under recent NetBSD).
693    
694    ----------------------------------------------------------------------
695    Name: Allen Leung
696    Date: 2001/10/12 17:18:32 EDT 2001
697    Tag: leung-20011012-x86-printflowgraph
698    Description:
699    
700    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
701    
702    ----------------------------------------------------------------------
703    Name: Lal George
704    Date: 2001/10/11 23:51:34 EDT
705    Tag: george-20011011-too-many-instrs
706    Description:
707    
708    The representation of a program point never expected to see more
709    than 65536 instructions in a basic block!
710    
711    ----------------------------------------------------------------------
712    Name: Lal George
713    Date: 2001/10/09 09:41:37 EDT
714    Tag: george-20011008-mlrisc-labels
715    Description:
716    
717    Changed the machine description files to support printing of
718    local and global labels in assembly code, based on host assembler
719    conventions.
720    
721    ----------------------------------------------------------------------
722    Name: Matthias Blume
723    Date: 2001/09/25 15:25:00 EDT
724    Tag: blume-20010925-exninfo
725    Description:
726    
727    I provided a non-hook implementation of exnName (at the toplevel) and
728    made the "dummy" implementation of exnMessage (at the toplevel) more
729    useful: if nothing gets "hooked in", then at least you are going to
730    see the exception name and a message indicating why you don't see more.
731    
732    [For the time being, programs that need exnMessage and want to use
733    ml-build should either use General.exnMessage (strongly recommended) or
734    refer to structure General at some other point so that CM sees a
735    static dependency.]
736    
737    [Similar remarks go for "print" and "use":  If you want to use their
738    functionality in stand-alone programs generated by ml-build, then use
739    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
740    
741    ----------------------------------------------------------------------
742    Name: Matthias Blume
743    Date: 2001/09/20 17:28:00 EDT
744    Tag: blume-20010920-slowfp
745    Description:
746    
747    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
748    
749    ----------------------------------------------------------------------
750    Name: Matthias Blume
751    Date: 2001/09/20 17:20:00 EDT
752    Tag: blume-20010920-canonicalpaths
753    Description:
754    
755    0. Updated the BOOT file (something that I forgot to do earlier).
756    
757    1. Small internal change to CM so that it avoids "/../" in filenames
758       as much as possible (but only where it is safe).
759    
760    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
761       that arguments that contain delimiters are passed through correctly.
762       This change also means that all "special" arguments of the form
763       @SMLxxx... must come first.
764    
765    3. Changed install script to put relative anchor names for tool commands
766       into pathconfig.
767    
768    ----------------------------------------------------------------------
769    Name: Matthias Blume
770    Date: 2001/09/18 15:35:00 EDT
771    Tag: blume-20010918-readme11036
772    Description:
773    
774    Added README files.
775    
776    ----------------------------------------------------------------------
777    Name: Matthias Blume
778    Date: 2001/09/18 11:45:00 EDT
779    Tag: Release_110_36 (retag)
780    Description:
781    
782    Fixed mistake in config/preloads. Retagged as 110.36.
783    
784    ----------------------------------------------------------------------
785    Name: Matthias Blume
786    Date: 2001/09/18 09:40:00 EDT
787    Tag: Release_110_36_orig (tag changed)
788    Description:
789    
790    New version (110.36).  New bootfiles.
791    
792    ----------------------------------------------------------------------
793    Name: Matthias Blume
794    Date: 2001/09/14 16:15:00 EDT
795    Tag: blume-20010914-x86fastfp
796    Description:
797    
798    John committed some changes that Allen made, in particular a (hopefully)
799    correctly working version of the x86-fp module.
800    
801    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
802    flag to "true".  Everything seems to compile to a fixpoint ok, and
803    "mandelbrot" speeds up by about 15%.
804    
805    ----------------------------------------------------------------------
806    Name: Matthias Blume
807    Date: 2001/09/13 11:20:00 EDT
808    Tag: blume-20010913-minimal
809    Description:
810    
811    1. Stefan Monnier's patch to fix a miscompilation problem that
812       was brought to light by John Reppy's work on Moby.
813    
814    2. Implemented a minimal "structure Compiler" that contains just
815       "version" and "architecture".  The minimal version will be
816       available when the full version is not.  This is for backward-
817       compatibility with code that wants to test Compiler.version.
818    
819    ----------------------------------------------------------------------
820    Name: Matthias Blume
821    Date: 2001/08/28 14:03:00 EDT
822    Tag: blume-20010828-ml-lex
823    Description:
824    
825    Fix for bug 1581, received from Neophytos Michael.
826    
827    ----------------------------------------------------------------------
828    Name: Matthias Blume
829    Date: 2001/08/27 11:20:00 EDT
830    Tag: blume-20010827-readme11035
831    Description:
832    
833    Fleshed out the README file for 110.35.
834    
835    ----------------------------------------------------------------------
836    Name: Matthias Blume
837    Date: 2001/08/24 17:10:00 EDT
838    Tag: Release_110_35
839    Description:
840    
841    New version number (110.35).  New bootfiles.
842    
843    ----------------------------------------------------------------------
844    Name: Lal George
845    Date: 2001/08/24 13:47:18 EDT 2001
846    Tag: george-20010824-MLRISC-graphs
847    Description:
848    
849     removed clusters from MLRISC completely and replaced with graphs.
850    
851    ----------------------------------------------------------------------
852    Name: Matthias Blume
853    Date: 2001/08/23 17:50:00 EDT
854    Tag: blume-20010823-toplevel
855    Description:
856    
857    - some reorganization of the code that implements various kinds of
858      environments in the compiler (static, dynamic, symbolic, combined)
859    - re-implemented the EnvRef module so that evalStream works properly
860      (if the stream contains references to "use", "CM.make", etc.)
861    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
862    
863    ----------------------------------------------------------------------
864    Name: Matthias Blume
865    Date: 2001/08/20 15:50 EDT
866    Tag: blume20010820-slipup
867    Description:
868    
869    I forgot to commit a few files.  Here they are...
870    
871    ----------------------------------------------------------------------
872    Name: Matthias Blume
873    Date: 2001/08/20 15:35:00 EDT
874    Tag: blume-20010820-debugprof
875    Description:
876    
877    !!!! NEW BOOTFILES !!!!
878    
879    This is another round of reorganizing the compiler sources.  This
880    time the main goal was to factor out all the "instrumentation"
881    passes (for profiling and backtracing) into their own library.
882    The difficulty was to do it in such a way that it does not depend
883    on elaborate.cm but only on elabdata.cm.
884    
885    Therefore there have been further changes to both elaborate.cm and
886    elabdata.cm -- more "generic" things have been moved from the former
887    to the latter.  As a result, I was forced to split the assignment
888    of numbers indicating "primtyc"s into two portions: SML-generic and
889    SML/NJ-specific.  Since it would have been awkward to maintain,
890    I bit the bullet and actually _changed_ the mapping between these
891    numbers and primtycs.  The bottom line of this is that you need
892    a new set of bin- and bootfiles.
893    
894    I have built new bootfiles for all architectures, so doing a fresh
895    checkout and config/install.sh should be all you need.
896    
897    The newly created library's name is
898    
899        $smlnj/viscomp/debugprof.cm
900    
901    and its sources live under
902    
903        src/compiler/DebugProf
904    
905    ----------------------------------------------------------------------
906    Name: Matthias Blume
907    Date: 2001/08/15 17:15:00 EDT
908    Tag: blume-20010815-compreorg
909    Description:
910    
911    This is a first cut at reorganizing the CM libraries that make up the
912    core of the compiler.  The idea is to separate out pieces that could
913    be used independently by tools, e.g., the parser, the typechecker, etc.
914    
915    The current status is a step in this direction, but it is not quite
916    satisfactory yet.  Expect more changes in the future.
917    
918    Here is the current (new) organization...
919    
920        What used to be $smlnj/viscomp/core.cm is now divided into
921        six CM libraries:
922    
923             $smlnj/viscomp/basics.cm
924                           /parser.cm
925                           /elabdata.cm
926                           /elaborate.cm
927                           /execute.cm
928                           /core.cm
929    
930        The CM files for these libraries live under src/system/smlnj/viscomp.
931        All these libraries are proxy libraries that contain precisely
932        one CM library component.  Here are the locations of the components
933        (all within the src/compiler tree):
934    
935             Basics/basics.cm
936             Parse/parser.cm
937             ElabData/elabdata.cm
938             Elaborator/elaborate.cm
939             Execution/execute.cm
940             core.cm
941    
942         [This organization is the same that has been used already
943         for a while for the architecture-specific parts of the visible
944         compiler and for the old version of core.cm.]
945    
946         As you will notice, many source files have been moved from their
947         respective original locations to a new home in one of the above
948         subtrees.
949    
950         The division of labor between the new libraries is the following:
951    
952             basics.cm:
953                - Simple, basic definitions that pertain to many (or all) of
954                  the other libraries.
955             parser.cm:
956                - The SML parser, producing output of type Ast.dec.
957                - The type family for Ast is also defined and exported here.
958             elabdata.cm:
959                - The datatypes that describe input and output of the elaborator.
960                  This includes types, absyn, and static environments.
961             elaborator.cm:
962                - The SML/NJ type checker and elaborator.
963                  This maps an Ast.dec (with a given static environment) to
964                  an Absyn.dec (with a new static environment).
965                - This libraries implements certain modules that used to be
966                  structures as functors (to remove dependencies on FLINT).
967             execute.cm:
968                - Everything having to do with executing binary code objects.
969                - Dynamic environments.
970             core.cm:
971                - SML/NJ-specific instantiations of the elaborator and MLRISC.
972                - Top-level modules.
973                - FLINT (this should eventually become its own library)
974    
975    Notes:
976    
977    I am not 100% happy with the way I separated the elaborator (and its
978    data structures) from FLINT.  Two instances of the same problem:
979    
980        1. Data structures contain certain fields that carry FLINT-specific
981           information.  I hacked around this using exn and the property list
982           module from smlnj-lib.  But the fact that there are middle-end
983           specific fields around at all is a bit annoying.
984    
985        2. The elaborator calculates certain FLINT-related information.  I tried
986           to make this as abstract as I could using functorization, but, again,
987           the fact that the elaborator has to perform calculations on behalf
988           of the middle-end at all is not nice.
989    
990        3. Having to used exn and property lists is unfortunate because it
991           weakens type checking.  The other alternative (parameterizing
992           nearly *everything*) is not appealing, though.
993    
994    I removed the "rebinding =" warning hack because due to the new organization
995    it was awkward to maintain it.  As a result, the compiler now issues some of
996    these warnings when compiling init.cmi during bootstrap compilation. On
997    the plus side, you also get a warning when you do, for example:
998       val op = = Int32.+
999    which was not the case up to now.
1000    
1001    I placed "assign" and "deref" into the _Core structure so that the
1002    code that deals with the "lazy" keyword can find them there.  This
1003    removes the need for having access to the primitive environment
1004    during elaboration.
1005    
1006    ----------------------------------------------------------------------
1007    Name: Matthias Blume
1008    Date: 2001/08/13
1009    Tag: blume-20010813-closures
1010    Description:
1011    
1012    This fix was sent to us by Zhong Shao.  It is supposed to improve the
1013    performance of certain loops by avoiding needless closure allocation.
1014    
1015    ----------------------------------------------------------------------
1016    Name: Lal George
1017    Date: 2001/07/31 10:03:23 EDT 2001
1018    Tag: george-20010731-x86-fmalloc
1019    Description: Fixed bug in x86 calls
1020    
1021        There was a bug where call instructions would mysteriously
1022        vanish. The call instruction had to be one that returned
1023        a floating point value.
1024    
1025    ----------------------------------------------------------------------
1026    Name: Lal George
1027    Date: 2001/07/19 16:36:29 EDT 2001
1028    Tag: george-20010719-simple-cells
1029    Description:
1030    
1031    I have dramatically simplified the interface for CELLS in MLRISC.
1032    
1033    In summary, the cells interface is broken up into three parts:
1034    
1035      1. CellsBasis : CELLS_BASIS
1036    
1037            CellsBasis is a top level structure and common for all
1038            architectures.  it contains the definitions of basic datatypes
1039            and utility  functions over these types.
1040    
1041      2. functor Cells() : CELLS
1042    
1043            Cells generates an interface for CELLS that incorporates the
1044            specific resources on the target architecture, such as the
1045            presence of special register classes, their number and size,
1046            and various useful substructures.
1047    
1048      3. <ARCH>CELLS
1049    
1050            e.g. SparcCells: SPARCCELLS
1051    
1052            <ARCH>CELLS usually contains additional bindings for special
1053            registers  on the architecture, such as:
1054    
1055                    val r0 : cell           (* register zero *)
1056                    val y : cell            (* Y register *)
1057                    val psr : cell          (* processor status register *)
1058                    ...
1059    
1060            The structure returned by applying the Cells functor is opened
1061            in this interface.
1062    
1063    The main implication of all this is that the datatypes for cells is
1064    split between CellsBasis and CELLS -- a fairly simple change for user
1065    code.
1066    
1067    In the old scheme the CELLS interface had a definitional binding of
1068    the form:
1069    
1070            signature CELLS = sig
1071    
1072               structure CellsBasis = CellsBasis
1073    
1074               ...
1075    
1076            end
1077    
1078    With all the sharing constraints that goes on in MLRISC, this old
1079    design  quickly leads to errors such as:
1080    
1081            "structure definition spec inside of sharing ... "
1082    
1083    
1084    and appears to require an unacceptable amount of sharing and where
1085    constraint hackery.
1086    
1087    I think this error message (the interaction of definitional specs and
1088    sharing) requires more explanation on our web page.
1089    
1090    ----------------------------------------------------------------------
1091    Name: Matthias Blume
1092    Date: 2001/07/19 15:00:00 EDT
1093    Tag: blume-20010719-libreorg
1094    Description:
1095    
1096    This update puts together a fairly extensive but straightforward change
1097    to the way the libraries that implement the interactive system are
1098    organized:
1099    
1100       The biggest change is the elimination of structure Compiler.  As a
1101       replacement for this structure, there is now a CM library
1102       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1103       that exports all the substructures of the original structure Compiler
1104       directly.  So instead of saying Compiler.Foo.bar one now simply
1105       says Foo.bar.  (The CM libraries actually export a collection of
1106       structures that is richer than the collection of substructures of
1107       structure Compiler.)
1108    
1109       To make the transition smooth, there is a separate library called
1110       $smlnj/compiler/compiler.cm which puts together and exports the
1111       original structure Compiler (or at least something very close to it).
1112    
1113       There are five members of the original structure Compiler
1114       that are not exported directly but which instead became members
1115       of a new structure Backend (described by signature BACKEND).  These are:
1116       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1117       Interact (: INTERACT), structure Machine (: MACHINE), and val
1118       architecture (: string).
1119    
1120       Structure Compiler.Version has become structure CompilerVersion.
1121    
1122       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1123       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1124       or x86, respectively.
1125       Each of these exports the same frontend structures that
1126       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1127       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1128       Hppa, PPC, Sparc, or X86, respectively.
1129    
1130       Library $smlnj/compiler/all.cm exports the union of the exports of
1131       $smlnj/compiler/<arch>.cm
1132    
1133       There are no structures <Arch>Compiler anymore, use
1134       $smlnj/compiler/<arch>.cm instead.
1135    
1136       Library host-compiler-0.cm is gone.  Instead, the internal library
1137       that instantiates CM is now called cm0.cm.  Selection of the host
1138       compiler (backend) is no longer done here but. (Responsibility for it
1139       now lies with $smlnj/compiler/current.cm.  This seems to be more
1140       logical.)
1141    
1142       Many individual files have been moved or renamed.  Some files have
1143       been split into multiple files, and some "dead" files have been deleted.
1144    
1145    Aside from these changes to library organization, there are also changes
1146    to the way the code itself is organized:
1147    
1148       Structure Binfile has been re-implemented in such a way that it no
1149       longer needs any knowledge of the compiler.  It exclusively deals
1150       with the details of binfile layout.  It no longer invokes the
1151       compiler (for the purpose of creating new prospective binfile
1152       content), and it no longer has any knowledge of how to interpret
1153       pickles.
1154    
1155       Structure Compile (: COMPILE) has been stripped down to the bare
1156       essentials of compilation.  It no longer deals with linking/execution.
1157       The interface has been cleaned up considerably.
1158    
1159       Utility routines for dealing with linking and execution have been
1160       moved into their own substructures.
1161    
1162       (The ultimate goal of these changes is to provide a light-weight
1163       binfile loader/linker (at least for, e.g., stable libraries) that
1164       does not require CM or the compiler to be present.)
1165    
1166    CM documentation has been updated to reflect the changes to library
1167    organization.
1168    
1169    ----------------------------------------------------------------------
1170    Name: Matthias Blume
1171    Date: 2001/07/10 17:30:00 EDT
1172    Tag: Release_110_34
1173    Description:
1174    
1175    Minor tweak to 110.34 (re-tagged):
1176    
1177      - README.html file added to CVS repository
1178      - runtime compiles properly under FreeBSD 3.X and 4.X
1179    
1180    ----------------------------------------------------------------------
1181    Name: Matthias Blume
1182    Date: 2001/07/10 17:30:00 EDT
1183    Tag: Release_110_34
1184    Description:
1185    
1186    New version number (110.34). New bootfiles.
1187    
1188    ----------------------------------------------------------------------
1189    Name: Matthias Blume
1190    Date: 2001/07/09 16:00:00 EDT
1191    Tag: blume-20010709-more-varargs
1192    Description:
1193    
1194    I changed the handling of varargs in ml-nlffigen again:
1195    The ellipsis ... will now simply be ignored (with an accompanying warning).
1196    
1197    The immediate effect is that you can actually call a varargs function
1198    from ML -- but you can't actually supply any arguments beyond the ones
1199    specified explicitly.  (For example, you can call printf with its format
1200    string, but you cannot pass additional arguments.)
1201    
1202    This behavior is only marginally more useful than the one before, but
1203    it has the advantage that a function or, more importantly, a function
1204    type never gets dropped on the floor, thus avoiding follow-up problems with
1205    other types that refer to the offending one.
1206    
1207    ----------------------------------------------------------------------
1208    Name: Matthias Blume
1209    Date: 2001/07/09 11:25:00 EDT
1210    Tag: blume-20010709-varargs
1211    Description:
1212    
1213    1. ckit-lib.cm now exports structure Error
1214    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1215       with a warning accompanied by a source location.  Moreover, it
1216       merely skips the offending function or type and proceeds with the
1217       rest of its work.u  As a result, one can safely feed C code containing
1218       "..." to ml-nlffigen.
1219    3. There are some internal improvements to CM, providing slightly
1220       more general string substitutions in the tools subsystem.
1221    
1222    ----------------------------------------------------------------------
1223    Name: Matthias Blume
1224    Date: 2001/06/27 15:10:00 EDT
1225    Tag: blume-20010627-concur
1226    Description:
1227    
1228    Fixed a small bug in CM's handling of parallel compilation.
1229    (You could observe the bug by Control-C-interrupting an ordinary
1230    CMB.make or CM.stabilize and then attaching some compile servers.
1231    The result was that all of a sudden the previously interrupted
1232    compilation would continue on its own.  This was because of
1233    an over-optimization: CM did not bother to clean out certain queues
1234    when no servers were attached "anyway", resulting in the contents
1235    of these queues to grab control when new servers did get attached.)
1236    
1237    There is also another minor update to the CM manual.
1238    
1239    ----------------------------------------------------------------------
1240    Name: Matthias Blume
1241    Date: 2001/06/26 16:15:00 EDT
1242    Tag: blume-20010626-cmdoc
1243    Description:
1244    
1245    Minor typo fixed in CM manual (syntax diagram for libraries).
1246    
1247    ----------------------------------------------------------------------
1248    Name: Matthias Blume
1249    Date: 2001/06/25 22:55:00 EDT
1250    Tag: blume-20010625-x86pc
1251    Description:
1252    
1253    Fixed a nasty bug in the X86 assembly code that caused signal
1254    handlers to fail (crash) randomly.
1255    
1256    ----------------------------------------------------------------------
1257    Name: Matthias Blume
1258    Date: 2001/06/25 12:05:00 EDT
1259    Tag: blume-20010625-nlffigen
1260    Description:
1261    
1262    This update fixes a number of minor bugs in ml-nlffigen as reported by
1263    Nick Carter <nbc@andrew.cmu.edu>.
1264    
1265      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1266      2. Default names for generated files are now derived from the name of
1267         the C file *without its directory*.  In particular, this causes generated
1268         files to be placed locally even if the C file is in some system directory.
1269      3. Default names for generated signatures and structures are also derived
1270         from the C file name without its directory.  This avoids silly things
1271         like "structure GL/GL".
1272         (Other silly names are still possible because ml-nlffigen does not do
1273          a thorough check of whether generated names are legal ML identifiers.
1274          When in doubt, use command line arguments to force particular names.)
1275    
1276    ----------------------------------------------------------------------
1277    Name: Matthias Blume
1278    Date: 2001/06/21 12:25:00 EDT
1279    Tag: blume-20010621-eXene
1280    Description:
1281    
1282    eXene now compiles and (sort of) works again.
1283    
1284    The library name (for version > 110.33) is $/eXene.cm.
1285    
1286    I also added an new example in src/eXene/examples/nbody.  See the
1287    README file there for details.
1288    
1289    ----------------------------------------------------------------------
1290    Name: Matthias Blume
1291    Date: 2001/06/20 16:40:00 EDT
1292    Tag: blume-20010620-cml
1293    Description:
1294    
1295    CML now compiles and works again.
1296    
1297    Libraries (for version > 110.33):
1298    
1299      $cml/cml.cm            Main CML library.
1300      $cml/basis.cm          CML's version of $/basis.cm.
1301      $cml/cml-internal.cm   Internal helper library.
1302      $cml/core-cml.cm       Internal helper library.
1303      $cml-lib/trace-cml.cm  Tracing facility.
1304      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1305    
1306    The installer (config/install.sh) has been taught how to properly
1307    install this stuff.
1308    
1309    ----------------------------------------------------------------------
1310    Name: Matthias Blume
1311    Date: 2001/06/19 17:55:00 EDT
1312    Tag: blume-20010619-instantiate
1313    Description:
1314    
1315    This un-breaks the fix for bug 1432.
1316    (The bug was originally fixed in 110.9 but I broke it again some
1317    time after that.)
1318    
1319    ----------------------------------------------------------------------
1320    Name: Matthias Blume
1321    Date: 2001/06/19 17:25:00 EDT
1322    Tag: blume-20010619-signals
1323    Description:
1324    
1325    This should (hopefully) fix the long-standing signal handling bug.
1326    (The runtime system was constructing a continuation record with an
1327    incorrect descriptor which would cause the GC to drop data on the floor...)
1328    
1329    ----------------------------------------------------------------------
1330    Name: Matthias Blume
1331    Date: 2001/06/15 15:05:00 EDT
1332    Tag: blume-20010615-moresparc
1333    Description:
1334    
1335    Here is a short late-hour update related to Sparc c-calls:
1336    
1337     -- made handling of double-word arguments a bit smarter
1338    
1339     -- instruction selection phase tries to collapse certain clumsily
1340        constructed ML-Trees; typical example:
1341    
1342            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1343    
1344        This currently has no further impact on SML/NJ since mlriscGen does
1345        not seem to generate such patterns in the first place, and c-calls
1346        (which did generate them in the beginning) has meanwhile been fixed
1347        so as to avoid them as well.
1348    
1349    ----------------------------------------------------------------------
1350    Name: Matthias Blume
1351    Date: 2001/06/15 15:05:00 EDT
1352    Tag: blume-20010615-sparc
1353    Description:
1354    
1355    The purpose of this update is to provide an implementation of NLFFI
1356    on Sparc machines.
1357    
1358    Here are the changes in detail:
1359    
1360       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1361       the Sparc implementation of the c-calls API.
1362       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1363       for accessing the ML frame.  Thus, we have a real frame pointer and
1364       can freely modify %sp without need for an omit-frame-ptr phase.
1365       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1366       this case.
1367       * ml-nlffigen has been taught to produce code for different architectures
1368       and calling conventions.
1369       * In a way similar to what was done in the x86 case, the Sparc
1370       backend uses its own specific extension to mltree.  (For example,
1371       it needs to be able to generate UNIMP instructions which are part
1372       of the calling convention.)
1373       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1374       to make it easier to plug in new machine- and os-dependent parts).
1375    
1376    There are some other fairly unrelated bug fixes and cleanups as well:
1377    
1378       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1379       that they properly share their libraries with existing SML/NJ libraries.
1380       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1381       follow-up error messages.
1382       * Updates to CM/CMB documentation.
1383    
1384    TODO items:
1385    
1386       * MLRISC should use a different register as its asmTemp on the Sparc.
1387         (The current %o2 is a really bad choice because it is part of the
1388         calling conventions, so things might interfere in unexpected ways.)
1389    
1390    ----------------------------------------------------------------------
1391    Name: Matthias Blume
1392    Date: 2001/06/07
1393    Tag: blume-20010607-calls
1394    Description:
1395    
1396    A number of internal changes related to C calls and calling conventions:
1397    
1398    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1399       number of bytes popped implicitly (by the callee).  In most cases
1400       this field is 0 but on x86/win32 it is some non-zero value.  This
1401       is information provided for the benefit of the "omit-frameptr" pass.
1402    2. The CALL instruction on the x86 carries a similar "pops" field.
1403       The instruction selection phase copies its value from the ML-Tree
1404       CALL statement.
1405    3. On all other architectures, the instruction selection phase checks
1406       whether "pops=0" and complains if not.
1407    4. The c-calls implementation for x86 now accepts two calling conventions:
1408       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1409       up after the call and pops is set to 0.  For "stdcall", the caller
1410       does nothing, leaving the cleanup to the callee; pops is set to
1411       the number of bytes that were pushed onto the stack.
1412    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1413       distinguish between "ccall" and "stdcall".
1414    6. The UNIMP instruction has been added to the supported Sparc instruction
1415       set. (This is needed for implementing the official C calling convention
1416       on this architecture.)
1417    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1418       work with the latest CM.
1419    
1420    ----------------------------------------------------------------------
1421    Name: Matthias Blume
1422    Date: 2001/06/05 15:10:00 EDT
1423    Tag: blume-20010605-cm-index
1424    Description:
1425    
1426    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1427    
1428    1. CM can now generate "index files".  These are human-readable files
1429       that list on a per-.cm-file basis each toplevel symbol defined or
1430       imported.  The location of the index file for
1431       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1432       To enable index-file generation, set CM.Control.generate_index to true
1433       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1434    
1435       The CM manual has been updated accordingly.
1436    
1437    2. I made some slight modifications to the c-calls API in MLRISC.
1438    
1439         a) There is now a callback to support saving/restoring of
1440            dedicated but caller-save registers around the actual call
1441            instruction.
1442         b) One can optionally specify a comment-annotation for the
1443            call instruction.
1444    
1445    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1446       (For example, the comment annotation shows the C prototype of
1447        the function being called.)
1448    
1449    ----------------------------------------------------------------------
1450    Name: Matthias Blume
1451    Date: 2001/06/01 13:30:00 EDT
1452    Tag: blume-20010601-nlffi-cleanup
1453    Description:
1454    
1455    This is mostly a cleanup of MLFFI stuff:
1456    
1457       - some signature files have been put into a more exposed place
1458       - the ugly 'f type parameter is gone (simplifies types tremendously!)
1459       - ml-nlffigen changed accordingly
1460       - tutorial updated
1461    
1462    Other changes:
1463    
1464       - author's affiliation in CM manual(s) updated
1465       - some more recognized keywords added to Allen's sml.sty
1466    
1467    ----------------------------------------------------------------------
1468    Name: Matthias Blume
1469    Date: 2001/05/25 15:30:00 EDT
1470    Tag: blume-20010525-iptr
1471    Description:
1472    
1473      - put the official 110.33-README (as it appears on the ftp server) under
1474        CVS
1475      - fixed a small bug related to incomplete pointer types in
1476        ml-nlffigen
1477      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1478        (it does not need the 'f type parameter)
1479    
1480    ----------------------------------------------------------------------
1481    Name: Matthias Blume
1482    Date: 2001/05/23 14:30:00 EDT
1483    Tag: Release_110_33
1484    Description:
1485    
1486    New version number (110.33).  New bootfiles.
1487    
1488    ----------------------------------------------------------------------
1489    Name: Matthias Blume
1490    Date: 2001/05/22 18:06:00 EDT
1491    Tag: blume-20010522-targets
1492    Description:
1493    
1494    Made install.sh use file config/targets.customized if it exists, falling
1495    back to config/targets if it doesn't.  This way one can have a customized
1496    version of the targets file without touching the "real thing", thus
1497    eliminating the constant fear of accidentally checking something bogus
1498    back into the CVS repository...  (File config/targets.customized must
1499    not be added to the repository!)
1500    
1501    ----------------------------------------------------------------------
1502    Name: Matthias Blume
1503    Date: 2001/05/22 16:30:00 EDT
1504    Tag: blume-20010522-minitut
1505    Description:
1506    
1507    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1508       struct returns.
1509    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
1510       incomplete, preliminary documentation for NLFFI.
1511    
1512    ----------------------------------------------------------------------
1513    Name: Matthias Blume
1514    Date: 2001/05/14 11:30:00 EDT
1515    Tag: blume-20010514-script
1516    Description:
1517    
1518    Some bugs in install script fixed.
1519    
1520    In addition to that I also made a slight change to the NLFFI API:
1521    Functors generated by ml-nlffigen now take the dynamic library as a
1522    straight functor argument, not as a suspended one.  (The original
1523    functor code used to force the suspension right away anyway, so there
1524    was nothing gained by this complication of the interface.)
1525    
1526    ----------------------------------------------------------------------
1527    Name: Matthias Blume
1528    Date: 2001/05/11 14:35:00 EDT
1529    Tag: blume-20010511-ml-nlffi
1530    Description:
1531    
1532    I finally took the plunge and added my new FFI code to the main
1533    repository.  For x86-linux it is now ready for prime-time.
1534    
1535    There are two new subdirectories of "src":
1536    
1537      - ml-nlffi-lib:
1538           The utility library for programs using the FFI interface.
1539           Here is the implementation of $/c.cm and its associated low-level
1540           partners $/c-int.cm and $/memory.cm.
1541      - ml-nlffigen:
1542           A stand-alone program for generating ML glue code from C source
1543           code.
1544    
1545    Building ml-nlffigen requires $/ckit-lib.cm.
1546    
1547    The config/install.sh script has been updates to do the Right Thing
1548    (hopefully).
1549    
1550    Notice that the source tree for the C-Kit will not be put under "src"
1551    but directly under the installation root directory.  (This is the
1552    structure that currently exists on the CVS server when you check out
1553    module "sml".)  Fortunately, config/install.sh knows about this oddity.
1554    
1555    Bugs: No documentation yet.
1556    
1557    ----------------------------------------------------------------------
1558    Name: Matthias Blume
1559    Date: 2001/05/09 16:35:00 EDT
1560    Tag: blume-20010509-cpscontract
1561    Description:
1562    
1563    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
1564    wrapper/unwrapper elimination did not decrement usage counts and some
1565    dead variables got overlooked by the dead-up logic.)
1566    
1567    ----------------------------------------------------------------------
1568    Name: Lal George
1569    Date: 2001/05/08  17:26:09 EDT
1570    Tag: george-20010508-omit-frameptr
1571    Description:
1572    
1573    Changes to implement the omit-frame-pointer optimization to support
1574    raw C calls. For now, there is only support on the Intel x86, but
1575    other architectures will follow as more experience is gained with this.
1576    
1577    
1578    ----------------------------------------------------------------------
1579    Name: Matthias Blume
1580    Date: 2001/05/07 14:40:00 EDT
1581    Tag: blume-20010507-proxies
1582    Description:
1583    
1584    I made into "proxy libraries" all libraries that qualify for such a
1585    change.  (A qualifying library is a library that has another library or
1586    groups as its sole member and repeats that member's export list
1587    verbatim.  A proxy library avoids this repetition by omitting its export
1588    list, effectively inheriting the list that its (only) member exports.
1589    See the CM manual for more explanation.)
1590    The main effect is that explicit export lists for these libraries
1591    do not have to be kepts in sync, making maintenance a bit easier.
1592    
1593    I also added copyright notices to many .cm-files.
1594    
1595    Last but not least, I made a new set of bootfiles.
1596    
1597    ----------------------------------------------------------------------
1598    Name: Matthias Blume
1599    Date: 2001/05/04 17:00:00 EDT
1600    Tag: blume-20010504-cm-lsplit
1601    Description:
1602    
1603    0. John merged pending changes to $/smlnj-lib.cm
1604    
1605    1. Allen's previous change accidentally backed out of one of Lal's
1606       earlier changes.  I undid this mistake (re-introducing Lal's change).
1607    
1608    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1609       within the compiler where applicable.  There is some code simplification
1610       because of that.
1611    
1612    3. The "split" phase (in FLINT) is now part of the default list of phases.
1613       Compiler.Control.LambdaSplitting.* can be used to globally control the
1614       lambda-splitting (cross-module-inlining) engine.  In addition to that,
1615       it can now also be controlled on a per-source basis: CM has been taught
1616       a new tool parameter applicable to ML source files.
1617    
1618       - To turn lambda-splitting off completely:
1619            local open Compiler.Control.LambdaSplitting in
1620                val _ = set Off
1621            end
1622       - To make "no lambda-splitting" the global default (but allow per-source
1623         overriding); this is the initial setting:
1624            local open Compiler.Control.LambdaSplitting in
1625                val _ = set (Default NONE)
1626            end
1627       - To make "lambda-splitting with aggressiveness a" the global default
1628         (and allow per-source overriding):
1629            local open Compiler.Control.LambdaSplitting in
1630                val _ = set (Default (SOME a))
1631            end
1632    
1633       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
1634         write (in the respective .cm-file):
1635            a.sml (lambdasplitting:off)
1636       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
1637            a.sml (lambdasplitting:on)
1638       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1639         <a> is a decimal non-negative integer):
1640            a.sml (lambdasplitting:<a>)
1641       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
1642            a.sml (lambdasplitting:infinity)
1643       - To use the global default for a.sml:
1644            a.sml (lambdasplitting:default)
1645         or simply
1646            a.sml
1647    
1648    ----------------------------------------------------------------------
1649    Name: Allen Leung
1650    Date: 2001/05/04 01:57:00 EDT
1651    Tag: leunga-20010504-sync
1652    Description:
1653    
1654      MLRISC features.
1655    
1656      1. Fix to CMPXCHG instructions.
1657      2. Changed RA interface to allow annotations in callbacks.
1658      3. Added a new method to the stream interface to allow annotations updates.
1659    
1660    ----------------------------------------------------------------------
1661    Name: Matthias Blume
1662    Date: 2001/05/01 11:45:00 EDT
1663    Tag: blume-20010501-pcedittmp
1664    Description:
1665    
1666    Changed install.sh to use the current working directory instead of
1667    /usr/tmp for a temporary file (pcedittmp).  The previous choice
1668    of /usr/tmp caused trouble with MacOS X because of file premission
1669    problems.
1670    
1671    ----------------------------------------------------------------------
1672    Name: Matthias Blume
1673  Date: 2001/04/20 11:10:00 EDT  Date: 2001/04/20 11:10:00 EDT
1674  Tag: blume-20010420-inMLflag  Tag: blume-20010420-inMLflag
1675  Description:  Description:
# Line 54  Line 1710 
1710   - made makeml script's verbosity level configurable via environment   - made makeml script's verbosity level configurable via environment
1711     variable (MAKEML_VERBOSITY)     variable (MAKEML_VERBOSITY)
1712    
1713   - eliminteded placeholder implementations for f32l, w16s, i16s, and f32s   - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1714     in rawmem-x86.sml; we are now using the real thing     in rawmem-x86.sml; we are now using the real thing
1715    
1716  ----------------------------------------------------------------------  ----------------------------------------------------------------------
# Line 2087  Line 3743 
3743       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3744       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3745       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
3746       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3747       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3748       Dave's hands. :)       Dave's hands. :)
3749    

Legend:
Removed from v.812  
changed lines
  Added in v.1064

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