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 832, Tue May 22 20:30:36 2001 UTC revision 1069, Fri Feb 15 21:00:05 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: 2002/02/15 16:00:00 EST
18    Tag: Release_110_39
19    Description:
20    
21    Working release 110.39.  New bootfiles.
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume
25    Date: 2002/02/15 14:17:00 EST
26    Tag: blume-20020215-showbindings
27    Description:
28    
29    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
30    the latter can be useful for exploring what bindings are available at
31    the interactive prompt.  (The first function returns only the list
32    of symbols that are really bound, the second prints those but also the
33    ones that CM's autoloading mechanism knows about.)
34    
35    ----------------------------------------------------------------------
36    Name: Matthias Blume
37    Date: 2002/02/15 12:08:00 EST
38    Tag: blume-20020215-iptrs
39    Description:
40    
41    Two improvements to ml-nlffigen:
42    
43      1. Write files only if they do not exist or if their current contents
44         do not coincide with what's being written.  (That is, avoid messing
45         with the time stamps unless absolutely necessary.)
46    
47      2. Implement a "repository" mechanism for generated files related
48         to "incomplete pointer types".   See the README file for details.
49    
50    ----------------------------------------------------------------------
51    Name: Matthias Blume
52    Date: 2002/02/14 11:50:00 EST
53    Tag: blume-20020214-quote
54    Description:
55    
56    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
57    because of the new and improved tag generation scheme.  (Thanks to Allen
58    Leung for pointing it out.)
59    
60    ----------------------------------------------------------------------
61    Name: Lal George
62    Date: 2002/02/14 09:55:27 EST 2002
63    Tag: george-20020214-isabelle-bug
64    Description:
65    
66    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
67    of Isabelle on the x86.
68    
69    From Allen:
70    -----------
71     I've found the problem:
72    
73         in ra-core.sml, I use the counter "blocked" to keep track of the
74         true number of elements in the freeze queue.  When the counter goes
75         to zero, I skip examining the queue.  But I've messed up the
76         bookkeeping in combine():
77    
78             else ();
79             case !ucol of
80               PSEUDO => (if !cntv > 0 then
81                     (if !cntu > 0 then blocked := !blocked - 1 else ();
82                                        ^^^^^^^^^^^^^^^^^^^^^^^
83                      moveu := mergeMoveList(!movev, !moveu)
84                     )
85                  else ();
86    
87         combine() is called to coalesce two nodes u and v.
88         I think I was thinking that if the move counts of u and v are both
89         greater than zero then after they are coalesced then one node is
90         removed from the freeze queue.  Apparently I was thinking that
91         both u and v are of low degree, but that's clearly not necessarily true.
92    
93    
94    02/12/2002:
95        Here's the patch.  HOL now compiles.
96    
97        I don't know how this impact on performance (compile
98        time or runtime).  This bug caused the RA (especially on the x86)
99        to go thru the potential spill phase when there are still nodes on the
100        freeze queue.
101    
102    
103    
104    
105    ----------------------------------------------------------------------
106    Name: Matthias Blume
107    Date: 2002/02/13 22:40:00 EST
108    Tag: blume-20020213-fptr-rtti
109    Description:
110    
111    Fixed a bug in ml-nlffigen that was introduced with one of the previous
112    updates.
113    
114    ----------------------------------------------------------------------
115    Name: Matthias Blume
116    Date: 2002/02/13 16:41:00 EST
117    Tag: blume-20020213-cmlpq
118    Description:
119    
120    Added new priority queue export symbols (which have just been added to
121    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
122    not compile and the installer would choke.)
123    
124    ----------------------------------------------------------------------
125    Name: Matthias Blume
126    Date: 2002/02/13 16:15:00 EST
127    Tag: blume-20020213-various
128    Description:
129    
130    1. More tweaks to ml-nlffigen:
131    
132       - better internal datastructures (resulting in slight speedup)
133       - "-match" option requires exact match
134       - "localized" gensym counters (untagged structs/unions nested within
135         other structs/unions or within typedefs get a fresh counter; their
136         tag will be prefixed by a concatenation of their parents' tags)
137       - bug fixes (related to calculation of transitive closure of types
138         to be included in the output)
139    
140    2. Minor Basis updates:
141    
142       - added implementations for List.collate and Option.app
143    
144    ----------------------------------------------------------------------
145    Name: Matthias Blume
146    Date: 2002/02/11 15:55:00 EST
147    Tag: blume-20020211-gensym
148    Description:
149    
150    Added a "-gensym" option to command line of ml-nlffigen.  This can be
151    used to specify a "stem" -- a string that is inserted in all "gensym'd"
152    names (ML structure names that correspond to unnamed C structs, unions,
153    and enums), so that separate runs of ml-nlffigen do not clash.
154    
155    ----------------------------------------------------------------------
156    Name: Matthias Blume
157    Date: 2002/02/11 12:05:00 EST
158    Tag: blume-20020211-gensml
159    Description:
160    
161    A quick fix for a problem with GenSML (in the pgraph-util library):
162    Make generation of toplevel "local" optional.  (Strictly speaking,
163    signature definitions within "local" are not legal SML.)
164    
165    Other than that: updates to INSTALL and cm/TODO.
166    
167    ----------------------------------------------------------------------
168    Name: Matthias Blume
169    Date: 2002/02/08 15:00:00 EST
170    Tag: blume-20020208-uniquepid
171    Description:
172    
173    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
174    
175    1. The installer (config/install.sh) has gotten smarter:
176    
177         - Configuration options are a bit easier to specify now
178           (in config/targets).
179         - Bug in recognizing .tar.bz2 files fixed.
180         - Installer automatically resolves dependencies between
181           configuration options (e.g., if you ask for eXene, you will
182           also get cml -- regardless whether you asked for it or not).
183         - Installer can run in "quieter mode" by setting the environment
184           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
185           "completely silent", though.
186         - Build HashCons library as part of smlnj-lib.
187    
188    2. A new scheme for assigning persistent identifiers to compilation
189       units (and, by extension, to types etc.) has been put into place.
190       This fixes a long-standing bug where types and even dynamic values
191       can get internally confused, thereby compromising type safety
192       (abstraction) and dynamic correctness.  See
193    
194         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
195    
196       for an example of how things could go wrong until now.
197    
198       The downside of the new scheme is that pids are not quite as
199       persistent as they used to be: CM will generate a fresh pid
200       for every compilation unit that it thinks it sees for the first
201       time.  That means that if you compile starting from a clean, fresh
202       source tree at two different times, you end up with different
203       binaries.
204    
205       Cutoff recompilation, however, has not been compromised because
206       CM keeps pid information in special caches between runs.
207    
208    ----------------------------------------------------------------------
209    Name: Lal George
210    Date: 2002/02/07 15:34:13 EST 2002
211    Tag: <none>
212    Description:
213    
214    Compilers that generate assembly code may produce  global labels
215    whose value is resolved at link time. The various peephole optimization
216    modules did not take this in account.
217    
218    TODO. The Labels.addrOf function should really return an option
219    type so that clients are forced to deal with this issue, rather
220    than an exception being raised.
221    
222    ----------------------------------------------------------------------
223    Name: Lal George
224    Date: 2002/02/06 13:55:02 EST
225    Tag: george-20020206-ra-breakup
226    Description:
227    
228    1. A bug fix from Allen.
229    
230        A typo causes extra fstp %st(0)'s to be generated at compensation
231        edges, which might cause stack underflow traps at runtime.  This
232        occurs in fft where there are extraneous fstps right before the 'into'
233        trap instruction (in this case they are harmless since none of the
234        integers overflow.)
235    
236    2. Pulled out various utility modules that were embedded in the modules
237       of the register allocator. I need these modules for other purposes, but
238       they are not complete enough to put into a library (just yet).
239    ----------------------------------------------------------------------
240    Name: Matthias Blume
241    Date: 2002/01/31 16:05:00 EST
242    Tag: blume-20020131-sparc-ccalls
243    Description:
244    
245    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
246       of extra stack space by mistake.  Fixed.
247    
248    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
249    
250    ----------------------------------------------------------------------
251    Name: Allen Leung
252    Date: 2002/01/30
253    Tag: leunga-20020130-nowhere-bug-fix
254    Description:
255    
256       MLRISC bug fixes:
257       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
258       2. MachineInt.fromString was negating its value.
259    
260    ----------------------------------------------------------------------
261    Name: Matthias Blume
262    Date: 2002/01/29
263    Tag: blume-20020129-INSTALL
264    Description:
265    
266    - Added somewhat detailed installation instructions (file INSTALL).
267    - Fixed curl-detection bug in config/install.sh.
268    - It is now possible to select the URL getter using the URLGETTER
269      environment variable:
270    
271          not set / "unknown"      --> automatic detection (script tries wget,
272                                       curl, and lynx)
273          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
274                                       how to properly invoke them)
275          other                    --> use $URLGETTER directly, it must take
276                                       precisely two command-line arguments
277                                       (source URL and destination file name)
278    
279    ----------------------------------------------------------------------
280    Name: Matthias Blume
281    Date: 2002/01/28
282    Tag: blume-20020128-sparc-ccalls
283    Description:
284    
285    - Fixed problem with calculation of "used" registers in sparc-c-calls.
286    - Make use of the allocParam argument in sparc-c-calls.
287    
288    ----------------------------------------------------------------------
289    Name: Matthias Blume
290    Date: 2002/01/28
291    Tag: blume-20020128-allocParam
292    Description:
293    
294    John Reppy:  Changes c-calls API to accept client-callback for
295    allocating extra stack space.
296    me: Corresponding changes to mlriscGen (using a dummy argument that
297        does not change the current behavior).
298    
299    ----------------------------------------------------------------------
300    Name: Matthias Blume
301    Date: 2002/01/28 12:00:00
302    Tag: Release_110_38
303    Description:
304    
305    This time for real!!!
306    
307    ----------------------------------------------------------------------
308    Name: Matthias Blume
309    Date: 2002/01/28 10:56:00 EST
310    Tag: blume-20020128-retraction
311    Description:
312    
313    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
314       with blume-Release_110_38-retracted.)
315    
316    1. Fixed a problem with incorrect rounding modes in real64.sml.
317       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
318    
319    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
320       and enums fixed.  The naming of corresponding ML identifiers should
321       now be consistent again.
322    
323    ----------------------------------------------------------------------
324    Name: Allen Leung
325    Date: 2002/01/27
326    Tag: leunga-20020127-nowhere
327    Description:
328    
329       Added a target called nowhere in the configuration scripts.
330       Enabling this will build the MLRISC 'nowhere' tool (for translating
331       programs with where-clauses into legal SML code) during installation.
332    
333    ----------------------------------------------------------------------
334    Name: Matthias Blume
335    Date: 2002/01/25 21:27:00 EST
336    Tag: blume-Release_110_38-retracted
337    Description:
338    
339    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
340    
341    README will be added later.
342    
343    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
344    (Release_110_38) removed.  Reason: Last-minute bug fixes.
345    
346    ----------------------------------------------------------------------
347    Name: Matthias Blume
348    Date: 2002/01/25
349    Tag: blume-20020125-ffi
350    Description:
351    
352    A large number of tweaks and improvements to ml-nlffi-lib and
353    ml-nlffigen:
354    
355       - ML represenation types have been streamlined
356       - getter and setter functions work with concrete values, not abstract
357         ones where possible
358       - ml-nlffigen command line more flexible (see README file there)
359       - some bugs have been fixed (hopefully)
360    
361    ----------------------------------------------------------------------
362    Name: Lal George
363    Date: 2002/01/24
364    Tag: george-20020124-risc-ra-interface
365    Description:
366    
367       There is a dramatic simplification in the interface to the
368       register allocator for RISC architectures as a result of making
369       parallel copy instructions explicit.
370    
371    ----------------------------------------------------------------------
372    Name: Matthias Blume
373    Date: 2002/01/22
374    Tag: blume-20020122-x86-ccalls
375    Description:
376    
377    Bug fix for c-calls on x86 (having to do with how char- and
378    short-arguments are being handled).
379    
380    ----------------------------------------------------------------------
381    Name: Matthias Blume
382    Date: 2002/01/21
383    Tag: blume-20020121-ff
384    Description:
385    
386    Another day of fiddling with the FFI...
387    
388    1. Bug fix/workaround:  CKIT does not complain about negative array
389       dimensions, so ml-nlffigen has to guard itself against this possibility.
390       (Otherwise a negative dimension would send it into an infinite loop.)
391    
392    2. Some of the abstract types (light objects, light pointers, most "base"
393       types) in structure C are now eqtypes.
394    
395    3. Added constructors and test functions for NULL function pointers.
396    
397    ----------------------------------------------------------------------
398    Name: Matthias Blume
399    Date: 2002/01/18
400    Tag: blume-20020118-ready-for-new-release
401    Description:
402    
403    Made config/srcarchiveurl point to a new place.  (Will provide boot
404    files shortly.)
405    
406    Maybe we christen this to be 110.38?
407    
408    ----------------------------------------------------------------------
409    Name: Matthias Blume
410    Date: 2002/01/18
411    Tag: blume-20020118-more-ffifiddle
412    Description:
413    
414    Today's FFI fiddling:
415    
416      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
417        C.Set.* that use concrete (MLRep.*) arguments and results instead
418        of abstract ones.
419    
420      - Provided word-style bit operations etc. for "int" representation
421        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
422        and Long.
423    
424    ----------------------------------------------------------------------
425    Name: Matthias Blume
426    Date: 2002/01/18
427    Tag: blume-20020118-use-x86-fp
428    Description:
429    
430    Now that x86-fast-fp seems to be working, I turned it back on again
431    by default.  (Seems to work fine now, even with the FFI.)
432    
433    Other than that, I added some documentation about the FFI to
434    src/ml-nlffigen/README and updated the FFI test examples in
435    src/ml-nlffi-lib/Tests/*.
436    
437    ----------------------------------------------------------------------
438    Name: Allen Leung
439    Date: 2002/01/17
440    Tag: leunga-20020117-x86-fast-fp-call
441    Description:
442    
443       1. Fixed a problem with handling return fp values when x86's fast fp
444          mode is turned on.
445    
446       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
447          of %f32.
448    
449       3. Added a constructor INT32lit to the ast of MLRISC tools.
450    
451    ----------------------------------------------------------------------
452    Name: Matthias Blume
453    Date: 2002/01/16
454    Tag: blume-20020116-ffifiddle
455    Description:
456    
457    More fiddling with the FFI interface:
458    
459     - Make constness 'c instead of rw wherever possible.  This eliminates
460       the need for certain explicit coercions.  (However, due to ML's
461       value polymorphism, there will still be many cases where explicit
462       coercions are necessary.  Phantom types are not the whole answer
463       to modeling a subtyping relationship in ML.)
464    
465     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
466    
467     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
468       clumsy workaround for varargs problem.)
469    
470    ----------------------------------------------------------------------
471    Name: Lal George
472    Date: 2002/01/15
473    Tag: <none>
474    Description:
475    
476    1. Since COPY instructions are no longer native to the architecture,
477       a generic functor can be used to implement the expandCopies function.
478    
479    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
480       TEXT segment.
481    
482    ----------------------------------------------------------------------
483    Name: Matthias Blume
484    Date: 2002/01/15
485    Tag: blume-20020115-ffiupdates
486    Description:
487    
488    1. Fix for bug resulting in single-precision float values being returned
489       incorrectly from FFI calls.
490    
491    2. Small modifications to C FFI API:
492    
493        - memory-allocation routines return straight objects (no options)
494          and raise an exception in out-of-memory situations
495        - unsafe extensions to cast between function pointers and pointers
496          from/to ints
497        - added structure C_Debug as an alternative to structure C where
498          pointer-dereferencing (|*| and |*!) always check for null-pointers
499        - added open_lib' to DynLinkage;  open_lib' works like open_lib
500          but also takes a (possibly empty) list of existing library handles
501          that the current library depends on
502    
503    ----------------------------------------------------------------------
504    Name: Matthias Blume
505    Date: 2002/01/10
506    Tag: blume-20020110-newffigen
507    Description:
508    
509    1. Updates to portable graph code.
510    
511    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
512       more scalable now so that even huge interfaces such as the one
513       for GTK compile in finite time and space. :-)
514       See src/ml-nlffigen/README for details on what's new.
515    
516    ----------------------------------------------------------------------
517    Name: Lal George
518    Date: 2001/01/09 14:31:35 EST 2002
519    Tag: george-20011206-rm-native-copy
520    Description:
521    
522            Removed the native COPY and FCOPY instructions
523            from all the architectures and replaced it with the
524            explicit COPY instruction from the previous commit.
525    
526            It is now possible to simplify many of the optimizations
527            modules that manipulate copies. This has not been
528            done in this change.
529    
530    ----------------------------------------------------------------------
531    Name: Lal George
532    Date: 2001/12/06 16:50:13 EST 2001
533    Tag: george-20011206-mlrisc-instruction
534    Description:
535    
536    Changed the representation of instructions from being fully abstract
537    to being partially concrete. That is to say:
538    
539      from
540            type instruction
541    
542      to
543            type instr                              (* machine instruction *)
544    
545            datatype instruction =
546                LIVE of {regs: C.cellset, spilled: C.cellset}
547              | KILL of {regs: C.cellset, spilled: C.cellset}
548              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
549              | ANNOTATION of {i: instruction, a: Annotations.annotation}
550              | INSTR of instr
551    
552    This makes the handling of certain special instructions that appear on
553    all architectures easier and uniform.
554    
555    LIVE and KILL say that a list of registers are live or killed at the
556    program point where they appear. No spill code is generated when an
557    element of the 'regs' field is spilled, but the register is moved to
558    the 'spilled' (which is present, more for debugging than anything else).
559    
560    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
561    We used to generate:
562    
563            DEFFREG f1
564            f1 := f2 + f3
565            trapb
566    
567    but now generate:
568    
569            f1 := f2 + f3
570            trapb
571            LIVE {regs=[f1,f2,f3], spilled=[]}
572    
573    Furthermore, the DEFFREG (hack) required that all floating point instruction
574    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
575    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
576    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
577    this architecture).
578    
579    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
580    all the architectures. This will result in further simplification of the
581    register allocator that must be aware of them for coalescing purposes, and
582    will also simplify certain aspects of the machine description that provides
583    callbacks related to parallel copies.
584    
585    ANNOTATION should be obvious, and now INSTR represents the honest to God
586    machine instruction set!
587    
588    The <arch>/instructions/<arch>Instr.sml files define certain utility
589    functions for making porting easier -- essentially converting upper case
590    to lower case. All machine instructions (of type instr) are in upper case,
591    and the lower case form generates an MLRISC instruction. For example on
592    the alpha we have:
593    
594      datatype instr =
595         LDA of {r:cell, b:cell, d:operand}
596       | ...
597    
598      val lda : {r:cell, b:cell, d:operand} -> instruction
599        ...
600    
601    where lda is just (INSTR o LDA), etc.
602    
603    ----------------------------------------------------------------------
604    Name: Matthias Blume
605    Date: 2001/11/22 21:40:00 EST
606    Tag: Release_110_37
607    Description:
608    
609    Release 110.37.  This time for real.
610    
611    ----------------------------------------------------------------------
612    Name: Matthias Blume
613    Date: 2001/11/21 16:35:00 EST
614    Tag: blume-20011121-foot-in-mouth
615    Description:
616    
617    Removed the "Release_110_37" tag because of a serious bug.
618    This will be re-tagged once the bug is fixed.
619    
620    ----------------------------------------------------------------------
621    Name: Matthias Blume
622    Date: 2001/11/21 16:14:00 EST
623    Tag: blume-20011121-forgottenfile
624    Description:
625    
626    Forgot to add a file.  (Just a .tex-file -- part of
627    the CM manual source.)
628    
629    ----------------------------------------------------------------------
630    Name: Matthias Blume
631    Date: 2001/11/21 16:10:00 EST
632    Tag: blume-20011121-invalid_110_37
633    Description:
634    
635    Note: I removed the original tag "Release_110_37" from this commit
636          because we found a serious bug in all non-x86 backends.
637          - Matthias
638    
639    1. Modifications to the SML/NJ code generator and to the runtime system
640       so that code object name strings are directly inserted into code
641       objects at code generation time.  The only business the runtime system
642       has with this is now to read the name strings on occasions.
643       (The encoding of the name string has also changed somewhat.)
644    
645    2. CM now implements a simple "set calculus" for specifying export lists.
646       In particular, it is now possible to refer to the export lists of
647       other libraries/groups/sources and form unions as well as differences.
648       See the latest CM manual for details.
649    
650    3. An separate notion of "proxy" libraries has again be eliminated from
651       CM's model.  (Proxy libraries are now simply a special case of using
652       the export list calculus.)
653    
654    4. Some of the existing libraries now take advantage of the new set
655       calculus.
656       (Notice that not all libraries have been converted because some
657       of the existing .cm-files are supposed to be backward compatible
658       with 110.0.x.)
659    
660    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
661       "General.exnMessage"!  The former relies on a certain hook to be
662       initialized, and that often does not happen in the stand-alone case.)
663    
664    ----------------------------------------------------------------------
665    Name: Lal George
666    Date: 2001/11/21  13:56:18 EST
667    Tag: george-2001121-pseudo-ops
668    Description:
669    
670      Implemented a complete redesign of MLRISC pseudo-ops. Now there
671      ought to never be any question of incompatabilities with
672      pseudo-op syntax expected by host assemblers.
673    
674      For now, only modules supporting GAS syntax are implemented
675      but more should follow, such as MASM, and vendor assembler
676      syntax, e.g. IBM as, Sun as, etc.
677    
678    ----------------------------------------------------------------------
679    Name: Matthias Blume
680    Date: 2001/11/14 11:52:00 EST
681    Tag: blume-20011114-srcname
682    Description:
683    
684    1. Routed the name of the current source file to mlriscgen where it
685       should be directly emitted into the code object.  (This last part
686       is yet to be done.)
687    
688    2. Some cleanup of the pgraph code to make it match the proposal that
689       I put out the other day.  (The proposal notwithstanding, things are
690       still in flux here.)
691    
692    ----------------------------------------------------------------------
693    Name: Lal George
694    Date: 2001/11/14 09:44:04 EST
695    Tag:
696    Description:
697    
698      Fix for a backpatching bug reported by Allen.
699    
700      Because the boundary between short and long span-dependent
701      instructions is +/- 128, there are an astounding number of
702      span-dependent instructions whose size is over estimated.
703    
704      Allen came up with the idea of letting the size of span
705      dependent instructions be non-monotonic, for a maxIter
706      number of times, after which the size must be monotonically
707      increasing.
708    
709      This table shows the number of span-dependent instructions
710      whose size was over-estimated as a function of maxIter, for the
711      file Parse/parse/ml.grm.sml:
712    
713         maxIter            # of instructions:
714            10                      687
715            20                      438
716            30                      198
717            40                        0
718    
719      In compiling the compiler, there is no significant difference in
720      compilation speed between maxIter=10 and maxIter=40. Actually,
721      my measurements showed that maxIter=40 was a tad faster than
722      maxIter=10! Also 96% of the  files in the compiler reach a fix
723      point within 13 iterations, so fixing maxIter at 40, while high,
724      is okay.
725    
726    ----------------------------------------------------------------------
727    Name: Matthias Blume
728    Date: 2001/10/31 15:25:00 EST
729    Tag: blume-20011031-pgraph
730    Description:
731    
732    CKIT:
733    * Changed the "Function" constructor of type Ast.ctype to carry optional
734      argument identifiers.
735    * Changed the return type of TypeUtil.getFunction accordingly.
736    * Type equality ignores the argument names.
737    * TypeUtil.composite tries to preserve argument names but gives up quickly
738      if there is a mismatch.
739    
740    installation script:
741    * attempts to use "curl" if available (unless "wget" is available as well)
742    
743    CM:
744    * has an experimental implementation of "portable graphs" which I will
745      soon propose as an implementation-independent library format
746    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
747    
748    NLFFI-LIB:
749    * some cleanup (all cosmetic)
750    
751    NLFFIGEN:
752    * temporarily disabled the mechanism that suppresses ML output for
753      C definitions whose identifiers start with an underscore character
754    * generate val bindings for enum constants
755    * user can request that only one style (light or heavy) is being used;
756      default is to use both (command-line arguments: -heavy and -light)
757    * fixed bug in handling of function types involving incomplete pointers
758    * generate ML entry points that take record arguments (i.e., using
759      named arguments) for C functions that have a prototype with named
760      arguments
761      (see changes to CKIT)
762    
763    ----------------------------------------------------------------------
764    Name: Allen Leung
765    Date: 2001/10/27 20:34:00 EDT
766    Tag: leunga-20011027-x86-fast-fp-call
767    Description:
768    
769       Fixed the bug described in blume-20010920-slowfp.
770    
771       The fix involves
772          1. generating FCOPYs in FSTP in ia32-svid
773          2. marking a CALL with the appropriate annotation
774    
775    ----------------------------------------------------------------------
776    Name: Matthias Blume
777    Date: 2001/10/16 11:32:00 EDT
778    Tag: blume-20011016-netbsd
779    Description:
780    
781    Underscore patch from Chris Richards (fixing problem with compiling
782    runtime system under recent NetBSD).
783    
784    ----------------------------------------------------------------------
785    Name: Allen Leung
786    Date: 2001/10/12 17:18:32 EDT 2001
787    Tag: leung-20011012-x86-printflowgraph
788    Description:
789    
790    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
791    
792    ----------------------------------------------------------------------
793    Name: Lal George
794    Date: 2001/10/11 23:51:34 EDT
795    Tag: george-20011011-too-many-instrs
796    Description:
797    
798    The representation of a program point never expected to see more
799    than 65536 instructions in a basic block!
800    
801    ----------------------------------------------------------------------
802    Name: Lal George
803    Date: 2001/10/09 09:41:37 EDT
804    Tag: george-20011008-mlrisc-labels
805    Description:
806    
807    Changed the machine description files to support printing of
808    local and global labels in assembly code, based on host assembler
809    conventions.
810    
811    ----------------------------------------------------------------------
812    Name: Matthias Blume
813    Date: 2001/09/25 15:25:00 EDT
814    Tag: blume-20010925-exninfo
815    Description:
816    
817    I provided a non-hook implementation of exnName (at the toplevel) and
818    made the "dummy" implementation of exnMessage (at the toplevel) more
819    useful: if nothing gets "hooked in", then at least you are going to
820    see the exception name and a message indicating why you don't see more.
821    
822    [For the time being, programs that need exnMessage and want to use
823    ml-build should either use General.exnMessage (strongly recommended) or
824    refer to structure General at some other point so that CM sees a
825    static dependency.]
826    
827    [Similar remarks go for "print" and "use":  If you want to use their
828    functionality in stand-alone programs generated by ml-build, then use
829    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
830    
831    ----------------------------------------------------------------------
832    Name: Matthias Blume
833    Date: 2001/09/20 17:28:00 EDT
834    Tag: blume-20010920-slowfp
835    Description:
836    
837    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
838    
839    ----------------------------------------------------------------------
840    Name: Matthias Blume
841    Date: 2001/09/20 17:20:00 EDT
842    Tag: blume-20010920-canonicalpaths
843    Description:
844    
845    0. Updated the BOOT file (something that I forgot to do earlier).
846    
847    1. Small internal change to CM so that it avoids "/../" in filenames
848       as much as possible (but only where it is safe).
849    
850    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
851       that arguments that contain delimiters are passed through correctly.
852       This change also means that all "special" arguments of the form
853       @SMLxxx... must come first.
854    
855    3. Changed install script to put relative anchor names for tool commands
856       into pathconfig.
857    
858    ----------------------------------------------------------------------
859    Name: Matthias Blume
860    Date: 2001/09/18 15:35:00 EDT
861    Tag: blume-20010918-readme11036
862    Description:
863    
864    Added README files.
865    
866    ----------------------------------------------------------------------
867    Name: Matthias Blume
868    Date: 2001/09/18 11:45:00 EDT
869    Tag: Release_110_36 (retag)
870    Description:
871    
872    Fixed mistake in config/preloads. Retagged as 110.36.
873    
874    ----------------------------------------------------------------------
875    Name: Matthias Blume
876    Date: 2001/09/18 09:40:00 EDT
877    Tag: Release_110_36_orig (tag changed)
878    Description:
879    
880    New version (110.36).  New bootfiles.
881    
882    ----------------------------------------------------------------------
883    Name: Matthias Blume
884    Date: 2001/09/14 16:15:00 EDT
885    Tag: blume-20010914-x86fastfp
886    Description:
887    
888    John committed some changes that Allen made, in particular a (hopefully)
889    correctly working version of the x86-fp module.
890    
891    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
892    flag to "true".  Everything seems to compile to a fixpoint ok, and
893    "mandelbrot" speeds up by about 15%.
894    
895    ----------------------------------------------------------------------
896    Name: Matthias Blume
897    Date: 2001/09/13 11:20:00 EDT
898    Tag: blume-20010913-minimal
899    Description:
900    
901    1. Stefan Monnier's patch to fix a miscompilation problem that
902       was brought to light by John Reppy's work on Moby.
903    
904    2. Implemented a minimal "structure Compiler" that contains just
905       "version" and "architecture".  The minimal version will be
906       available when the full version is not.  This is for backward-
907       compatibility with code that wants to test Compiler.version.
908    
909    ----------------------------------------------------------------------
910    Name: Matthias Blume
911    Date: 2001/08/28 14:03:00 EDT
912    Tag: blume-20010828-ml-lex
913    Description:
914    
915    Fix for bug 1581, received from Neophytos Michael.
916    
917    ----------------------------------------------------------------------
918    Name: Matthias Blume
919    Date: 2001/08/27 11:20:00 EDT
920    Tag: blume-20010827-readme11035
921    Description:
922    
923    Fleshed out the README file for 110.35.
924    
925    ----------------------------------------------------------------------
926    Name: Matthias Blume
927    Date: 2001/08/24 17:10:00 EDT
928    Tag: Release_110_35
929    Description:
930    
931    New version number (110.35).  New bootfiles.
932    
933    ----------------------------------------------------------------------
934    Name: Lal George
935    Date: 2001/08/24 13:47:18 EDT 2001
936    Tag: george-20010824-MLRISC-graphs
937    Description:
938    
939     removed clusters from MLRISC completely and replaced with graphs.
940    
941    ----------------------------------------------------------------------
942    Name: Matthias Blume
943    Date: 2001/08/23 17:50:00 EDT
944    Tag: blume-20010823-toplevel
945    Description:
946    
947    - some reorganization of the code that implements various kinds of
948      environments in the compiler (static, dynamic, symbolic, combined)
949    - re-implemented the EnvRef module so that evalStream works properly
950      (if the stream contains references to "use", "CM.make", etc.)
951    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
952    
953    ----------------------------------------------------------------------
954    Name: Matthias Blume
955    Date: 2001/08/20 15:50 EDT
956    Tag: blume20010820-slipup
957    Description:
958    
959    I forgot to commit a few files.  Here they are...
960    
961    ----------------------------------------------------------------------
962    Name: Matthias Blume
963    Date: 2001/08/20 15:35:00 EDT
964    Tag: blume-20010820-debugprof
965    Description:
966    
967    !!!! NEW BOOTFILES !!!!
968    
969    This is another round of reorganizing the compiler sources.  This
970    time the main goal was to factor out all the "instrumentation"
971    passes (for profiling and backtracing) into their own library.
972    The difficulty was to do it in such a way that it does not depend
973    on elaborate.cm but only on elabdata.cm.
974    
975    Therefore there have been further changes to both elaborate.cm and
976    elabdata.cm -- more "generic" things have been moved from the former
977    to the latter.  As a result, I was forced to split the assignment
978    of numbers indicating "primtyc"s into two portions: SML-generic and
979    SML/NJ-specific.  Since it would have been awkward to maintain,
980    I bit the bullet and actually _changed_ the mapping between these
981    numbers and primtycs.  The bottom line of this is that you need
982    a new set of bin- and bootfiles.
983    
984    I have built new bootfiles for all architectures, so doing a fresh
985    checkout and config/install.sh should be all you need.
986    
987    The newly created library's name is
988    
989        $smlnj/viscomp/debugprof.cm
990    
991    and its sources live under
992    
993        src/compiler/DebugProf
994    
995    ----------------------------------------------------------------------
996    Name: Matthias Blume
997    Date: 2001/08/15 17:15:00 EDT
998    Tag: blume-20010815-compreorg
999    Description:
1000    
1001    This is a first cut at reorganizing the CM libraries that make up the
1002    core of the compiler.  The idea is to separate out pieces that could
1003    be used independently by tools, e.g., the parser, the typechecker, etc.
1004    
1005    The current status is a step in this direction, but it is not quite
1006    satisfactory yet.  Expect more changes in the future.
1007    
1008    Here is the current (new) organization...
1009    
1010        What used to be $smlnj/viscomp/core.cm is now divided into
1011        six CM libraries:
1012    
1013             $smlnj/viscomp/basics.cm
1014                           /parser.cm
1015                           /elabdata.cm
1016                           /elaborate.cm
1017                           /execute.cm
1018                           /core.cm
1019    
1020        The CM files for these libraries live under src/system/smlnj/viscomp.
1021        All these libraries are proxy libraries that contain precisely
1022        one CM library component.  Here are the locations of the components
1023        (all within the src/compiler tree):
1024    
1025             Basics/basics.cm
1026             Parse/parser.cm
1027             ElabData/elabdata.cm
1028             Elaborator/elaborate.cm
1029             Execution/execute.cm
1030             core.cm
1031    
1032         [This organization is the same that has been used already
1033         for a while for the architecture-specific parts of the visible
1034         compiler and for the old version of core.cm.]
1035    
1036         As you will notice, many source files have been moved from their
1037         respective original locations to a new home in one of the above
1038         subtrees.
1039    
1040         The division of labor between the new libraries is the following:
1041    
1042             basics.cm:
1043                - Simple, basic definitions that pertain to many (or all) of
1044                  the other libraries.
1045             parser.cm:
1046                - The SML parser, producing output of type Ast.dec.
1047                - The type family for Ast is also defined and exported here.
1048             elabdata.cm:
1049                - The datatypes that describe input and output of the elaborator.
1050                  This includes types, absyn, and static environments.
1051             elaborator.cm:
1052                - The SML/NJ type checker and elaborator.
1053                  This maps an Ast.dec (with a given static environment) to
1054                  an Absyn.dec (with a new static environment).
1055                - This libraries implements certain modules that used to be
1056                  structures as functors (to remove dependencies on FLINT).
1057             execute.cm:
1058                - Everything having to do with executing binary code objects.
1059                - Dynamic environments.
1060             core.cm:
1061                - SML/NJ-specific instantiations of the elaborator and MLRISC.
1062                - Top-level modules.
1063                - FLINT (this should eventually become its own library)
1064    
1065    Notes:
1066    
1067    I am not 100% happy with the way I separated the elaborator (and its
1068    data structures) from FLINT.  Two instances of the same problem:
1069    
1070        1. Data structures contain certain fields that carry FLINT-specific
1071           information.  I hacked around this using exn and the property list
1072           module from smlnj-lib.  But the fact that there are middle-end
1073           specific fields around at all is a bit annoying.
1074    
1075        2. The elaborator calculates certain FLINT-related information.  I tried
1076           to make this as abstract as I could using functorization, but, again,
1077           the fact that the elaborator has to perform calculations on behalf
1078           of the middle-end at all is not nice.
1079    
1080        3. Having to used exn and property lists is unfortunate because it
1081           weakens type checking.  The other alternative (parameterizing
1082           nearly *everything*) is not appealing, though.
1083    
1084    I removed the "rebinding =" warning hack because due to the new organization
1085    it was awkward to maintain it.  As a result, the compiler now issues some of
1086    these warnings when compiling init.cmi during bootstrap compilation. On
1087    the plus side, you also get a warning when you do, for example:
1088       val op = = Int32.+
1089    which was not the case up to now.
1090    
1091    I placed "assign" and "deref" into the _Core structure so that the
1092    code that deals with the "lazy" keyword can find them there.  This
1093    removes the need for having access to the primitive environment
1094    during elaboration.
1095    
1096    ----------------------------------------------------------------------
1097    Name: Matthias Blume
1098    Date: 2001/08/13
1099    Tag: blume-20010813-closures
1100    Description:
1101    
1102    This fix was sent to us by Zhong Shao.  It is supposed to improve the
1103    performance of certain loops by avoiding needless closure allocation.
1104    
1105    ----------------------------------------------------------------------
1106    Name: Lal George
1107    Date: 2001/07/31 10:03:23 EDT 2001
1108    Tag: george-20010731-x86-fmalloc
1109    Description: Fixed bug in x86 calls
1110    
1111        There was a bug where call instructions would mysteriously
1112        vanish. The call instruction had to be one that returned
1113        a floating point value.
1114    
1115    ----------------------------------------------------------------------
1116    Name: Lal George
1117    Date: 2001/07/19 16:36:29 EDT 2001
1118    Tag: george-20010719-simple-cells
1119    Description:
1120    
1121    I have dramatically simplified the interface for CELLS in MLRISC.
1122    
1123    In summary, the cells interface is broken up into three parts:
1124    
1125      1. CellsBasis : CELLS_BASIS
1126    
1127            CellsBasis is a top level structure and common for all
1128            architectures.  it contains the definitions of basic datatypes
1129            and utility  functions over these types.
1130    
1131      2. functor Cells() : CELLS
1132    
1133            Cells generates an interface for CELLS that incorporates the
1134            specific resources on the target architecture, such as the
1135            presence of special register classes, their number and size,
1136            and various useful substructures.
1137    
1138      3. <ARCH>CELLS
1139    
1140            e.g. SparcCells: SPARCCELLS
1141    
1142            <ARCH>CELLS usually contains additional bindings for special
1143            registers  on the architecture, such as:
1144    
1145                    val r0 : cell           (* register zero *)
1146                    val y : cell            (* Y register *)
1147                    val psr : cell          (* processor status register *)
1148                    ...
1149    
1150            The structure returned by applying the Cells functor is opened
1151            in this interface.
1152    
1153    The main implication of all this is that the datatypes for cells is
1154    split between CellsBasis and CELLS -- a fairly simple change for user
1155    code.
1156    
1157    In the old scheme the CELLS interface had a definitional binding of
1158    the form:
1159    
1160            signature CELLS = sig
1161    
1162               structure CellsBasis = CellsBasis
1163    
1164               ...
1165    
1166            end
1167    
1168    With all the sharing constraints that goes on in MLRISC, this old
1169    design  quickly leads to errors such as:
1170    
1171            "structure definition spec inside of sharing ... "
1172    
1173    
1174    and appears to require an unacceptable amount of sharing and where
1175    constraint hackery.
1176    
1177    I think this error message (the interaction of definitional specs and
1178    sharing) requires more explanation on our web page.
1179    
1180    ----------------------------------------------------------------------
1181    Name: Matthias Blume
1182    Date: 2001/07/19 15:00:00 EDT
1183    Tag: blume-20010719-libreorg
1184    Description:
1185    
1186    This update puts together a fairly extensive but straightforward change
1187    to the way the libraries that implement the interactive system are
1188    organized:
1189    
1190       The biggest change is the elimination of structure Compiler.  As a
1191       replacement for this structure, there is now a CM library
1192       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1193       that exports all the substructures of the original structure Compiler
1194       directly.  So instead of saying Compiler.Foo.bar one now simply
1195       says Foo.bar.  (The CM libraries actually export a collection of
1196       structures that is richer than the collection of substructures of
1197       structure Compiler.)
1198    
1199       To make the transition smooth, there is a separate library called
1200       $smlnj/compiler/compiler.cm which puts together and exports the
1201       original structure Compiler (or at least something very close to it).
1202    
1203       There are five members of the original structure Compiler
1204       that are not exported directly but which instead became members
1205       of a new structure Backend (described by signature BACKEND).  These are:
1206       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1207       Interact (: INTERACT), structure Machine (: MACHINE), and val
1208       architecture (: string).
1209    
1210       Structure Compiler.Version has become structure CompilerVersion.
1211    
1212       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1213       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1214       or x86, respectively.
1215       Each of these exports the same frontend structures that
1216       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1217       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1218       Hppa, PPC, Sparc, or X86, respectively.
1219    
1220       Library $smlnj/compiler/all.cm exports the union of the exports of
1221       $smlnj/compiler/<arch>.cm
1222    
1223       There are no structures <Arch>Compiler anymore, use
1224       $smlnj/compiler/<arch>.cm instead.
1225    
1226       Library host-compiler-0.cm is gone.  Instead, the internal library
1227       that instantiates CM is now called cm0.cm.  Selection of the host
1228       compiler (backend) is no longer done here but. (Responsibility for it
1229       now lies with $smlnj/compiler/current.cm.  This seems to be more
1230       logical.)
1231    
1232       Many individual files have been moved or renamed.  Some files have
1233       been split into multiple files, and some "dead" files have been deleted.
1234    
1235    Aside from these changes to library organization, there are also changes
1236    to the way the code itself is organized:
1237    
1238       Structure Binfile has been re-implemented in such a way that it no
1239       longer needs any knowledge of the compiler.  It exclusively deals
1240       with the details of binfile layout.  It no longer invokes the
1241       compiler (for the purpose of creating new prospective binfile
1242       content), and it no longer has any knowledge of how to interpret
1243       pickles.
1244    
1245       Structure Compile (: COMPILE) has been stripped down to the bare
1246       essentials of compilation.  It no longer deals with linking/execution.
1247       The interface has been cleaned up considerably.
1248    
1249       Utility routines for dealing with linking and execution have been
1250       moved into their own substructures.
1251    
1252       (The ultimate goal of these changes is to provide a light-weight
1253       binfile loader/linker (at least for, e.g., stable libraries) that
1254       does not require CM or the compiler to be present.)
1255    
1256    CM documentation has been updated to reflect the changes to library
1257    organization.
1258    
1259    ----------------------------------------------------------------------
1260    Name: Matthias Blume
1261    Date: 2001/07/10 17:30:00 EDT
1262    Tag: Release_110_34
1263    Description:
1264    
1265    Minor tweak to 110.34 (re-tagged):
1266    
1267      - README.html file added to CVS repository
1268      - runtime compiles properly under FreeBSD 3.X and 4.X
1269    
1270    ----------------------------------------------------------------------
1271    Name: Matthias Blume
1272    Date: 2001/07/10 17:30:00 EDT
1273    Tag: Release_110_34
1274    Description:
1275    
1276    New version number (110.34). New bootfiles.
1277    
1278    ----------------------------------------------------------------------
1279    Name: Matthias Blume
1280    Date: 2001/07/09 16:00:00 EDT
1281    Tag: blume-20010709-more-varargs
1282    Description:
1283    
1284    I changed the handling of varargs in ml-nlffigen again:
1285    The ellipsis ... will now simply be ignored (with an accompanying warning).
1286    
1287    The immediate effect is that you can actually call a varargs function
1288    from ML -- but you can't actually supply any arguments beyond the ones
1289    specified explicitly.  (For example, you can call printf with its format
1290    string, but you cannot pass additional arguments.)
1291    
1292    This behavior is only marginally more useful than the one before, but
1293    it has the advantage that a function or, more importantly, a function
1294    type never gets dropped on the floor, thus avoiding follow-up problems with
1295    other types that refer to the offending one.
1296    
1297    ----------------------------------------------------------------------
1298    Name: Matthias Blume
1299    Date: 2001/07/09 11:25:00 EDT
1300    Tag: blume-20010709-varargs
1301    Description:
1302    
1303    1. ckit-lib.cm now exports structure Error
1304    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1305       with a warning accompanied by a source location.  Moreover, it
1306       merely skips the offending function or type and proceeds with the
1307       rest of its work.u  As a result, one can safely feed C code containing
1308       "..." to ml-nlffigen.
1309    3. There are some internal improvements to CM, providing slightly
1310       more general string substitutions in the tools subsystem.
1311    
1312    ----------------------------------------------------------------------
1313    Name: Matthias Blume
1314    Date: 2001/06/27 15:10:00 EDT
1315    Tag: blume-20010627-concur
1316    Description:
1317    
1318    Fixed a small bug in CM's handling of parallel compilation.
1319    (You could observe the bug by Control-C-interrupting an ordinary
1320    CMB.make or CM.stabilize and then attaching some compile servers.
1321    The result was that all of a sudden the previously interrupted
1322    compilation would continue on its own.  This was because of
1323    an over-optimization: CM did not bother to clean out certain queues
1324    when no servers were attached "anyway", resulting in the contents
1325    of these queues to grab control when new servers did get attached.)
1326    
1327    There is also another minor update to the CM manual.
1328    
1329    ----------------------------------------------------------------------
1330    Name: Matthias Blume
1331    Date: 2001/06/26 16:15:00 EDT
1332    Tag: blume-20010626-cmdoc
1333    Description:
1334    
1335    Minor typo fixed in CM manual (syntax diagram for libraries).
1336    
1337    ----------------------------------------------------------------------
1338    Name: Matthias Blume
1339    Date: 2001/06/25 22:55:00 EDT
1340    Tag: blume-20010625-x86pc
1341    Description:
1342    
1343    Fixed a nasty bug in the X86 assembly code that caused signal
1344    handlers to fail (crash) randomly.
1345    
1346    ----------------------------------------------------------------------
1347    Name: Matthias Blume
1348    Date: 2001/06/25 12:05:00 EDT
1349    Tag: blume-20010625-nlffigen
1350    Description:
1351    
1352    This update fixes a number of minor bugs in ml-nlffigen as reported by
1353    Nick Carter <nbc@andrew.cmu.edu>.
1354    
1355      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1356      2. Default names for generated files are now derived from the name of
1357         the C file *without its directory*.  In particular, this causes generated
1358         files to be placed locally even if the C file is in some system directory.
1359      3. Default names for generated signatures and structures are also derived
1360         from the C file name without its directory.  This avoids silly things
1361         like "structure GL/GL".
1362         (Other silly names are still possible because ml-nlffigen does not do
1363          a thorough check of whether generated names are legal ML identifiers.
1364          When in doubt, use command line arguments to force particular names.)
1365    
1366    ----------------------------------------------------------------------
1367    Name: Matthias Blume
1368    Date: 2001/06/21 12:25:00 EDT
1369    Tag: blume-20010621-eXene
1370    Description:
1371    
1372    eXene now compiles and (sort of) works again.
1373    
1374    The library name (for version > 110.33) is $/eXene.cm.
1375    
1376    I also added an new example in src/eXene/examples/nbody.  See the
1377    README file there for details.
1378    
1379    ----------------------------------------------------------------------
1380    Name: Matthias Blume
1381    Date: 2001/06/20 16:40:00 EDT
1382    Tag: blume-20010620-cml
1383    Description:
1384    
1385    CML now compiles and works again.
1386    
1387    Libraries (for version > 110.33):
1388    
1389      $cml/cml.cm            Main CML library.
1390      $cml/basis.cm          CML's version of $/basis.cm.
1391      $cml/cml-internal.cm   Internal helper library.
1392      $cml/core-cml.cm       Internal helper library.
1393      $cml-lib/trace-cml.cm  Tracing facility.
1394      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1395    
1396    The installer (config/install.sh) has been taught how to properly
1397    install this stuff.
1398    
1399    ----------------------------------------------------------------------
1400    Name: Matthias Blume
1401    Date: 2001/06/19 17:55:00 EDT
1402    Tag: blume-20010619-instantiate
1403    Description:
1404    
1405    This un-breaks the fix for bug 1432.
1406    (The bug was originally fixed in 110.9 but I broke it again some
1407    time after that.)
1408    
1409    ----------------------------------------------------------------------
1410    Name: Matthias Blume
1411    Date: 2001/06/19 17:25:00 EDT
1412    Tag: blume-20010619-signals
1413    Description:
1414    
1415    This should (hopefully) fix the long-standing signal handling bug.
1416    (The runtime system was constructing a continuation record with an
1417    incorrect descriptor which would cause the GC to drop data on the floor...)
1418    
1419    ----------------------------------------------------------------------
1420    Name: Matthias Blume
1421    Date: 2001/06/15 15:05:00 EDT
1422    Tag: blume-20010615-moresparc
1423    Description:
1424    
1425    Here is a short late-hour update related to Sparc c-calls:
1426    
1427     -- made handling of double-word arguments a bit smarter
1428    
1429     -- instruction selection phase tries to collapse certain clumsily
1430        constructed ML-Trees; typical example:
1431    
1432            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1433    
1434        This currently has no further impact on SML/NJ since mlriscGen does
1435        not seem to generate such patterns in the first place, and c-calls
1436        (which did generate them in the beginning) has meanwhile been fixed
1437        so as to avoid them as well.
1438    
1439    ----------------------------------------------------------------------
1440    Name: Matthias Blume
1441    Date: 2001/06/15 15:05:00 EDT
1442    Tag: blume-20010615-sparc
1443    Description:
1444    
1445    The purpose of this update is to provide an implementation of NLFFI
1446    on Sparc machines.
1447    
1448    Here are the changes in detail:
1449    
1450       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1451       the Sparc implementation of the c-calls API.
1452       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1453       for accessing the ML frame.  Thus, we have a real frame pointer and
1454       can freely modify %sp without need for an omit-frame-ptr phase.
1455       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1456       this case.
1457       * ml-nlffigen has been taught to produce code for different architectures
1458       and calling conventions.
1459       * In a way similar to what was done in the x86 case, the Sparc
1460       backend uses its own specific extension to mltree.  (For example,
1461       it needs to be able to generate UNIMP instructions which are part
1462       of the calling convention.)
1463       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1464       to make it easier to plug in new machine- and os-dependent parts).
1465    
1466    There are some other fairly unrelated bug fixes and cleanups as well:
1467    
1468       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1469       that they properly share their libraries with existing SML/NJ libraries.
1470       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1471       follow-up error messages.
1472       * Updates to CM/CMB documentation.
1473    
1474    TODO items:
1475    
1476       * MLRISC should use a different register as its asmTemp on the Sparc.
1477         (The current %o2 is a really bad choice because it is part of the
1478         calling conventions, so things might interfere in unexpected ways.)
1479    
1480    ----------------------------------------------------------------------
1481    Name: Matthias Blume
1482    Date: 2001/06/07
1483    Tag: blume-20010607-calls
1484    Description:
1485    
1486    A number of internal changes related to C calls and calling conventions:
1487    
1488    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1489       number of bytes popped implicitly (by the callee).  In most cases
1490       this field is 0 but on x86/win32 it is some non-zero value.  This
1491       is information provided for the benefit of the "omit-frameptr" pass.
1492    2. The CALL instruction on the x86 carries a similar "pops" field.
1493       The instruction selection phase copies its value from the ML-Tree
1494       CALL statement.
1495    3. On all other architectures, the instruction selection phase checks
1496       whether "pops=0" and complains if not.
1497    4. The c-calls implementation for x86 now accepts two calling conventions:
1498       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1499       up after the call and pops is set to 0.  For "stdcall", the caller
1500       does nothing, leaving the cleanup to the callee; pops is set to
1501       the number of bytes that were pushed onto the stack.
1502    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1503       distinguish between "ccall" and "stdcall".
1504    6. The UNIMP instruction has been added to the supported Sparc instruction
1505       set. (This is needed for implementing the official C calling convention
1506       on this architecture.)
1507    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1508       work with the latest CM.
1509    
1510    ----------------------------------------------------------------------
1511    Name: Matthias Blume
1512    Date: 2001/06/05 15:10:00 EDT
1513    Tag: blume-20010605-cm-index
1514    Description:
1515    
1516    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1517    
1518    1. CM can now generate "index files".  These are human-readable files
1519       that list on a per-.cm-file basis each toplevel symbol defined or
1520       imported.  The location of the index file for
1521       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1522       To enable index-file generation, set CM.Control.generate_index to true
1523       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1524    
1525       The CM manual has been updated accordingly.
1526    
1527    2. I made some slight modifications to the c-calls API in MLRISC.
1528    
1529         a) There is now a callback to support saving/restoring of
1530            dedicated but caller-save registers around the actual call
1531            instruction.
1532         b) One can optionally specify a comment-annotation for the
1533            call instruction.
1534    
1535    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1536       (For example, the comment annotation shows the C prototype of
1537        the function being called.)
1538    
1539    ----------------------------------------------------------------------
1540    Name: Matthias Blume
1541    Date: 2001/06/01 13:30:00 EDT
1542    Tag: blume-20010601-nlffi-cleanup
1543    Description:
1544    
1545    This is mostly a cleanup of MLFFI stuff:
1546    
1547       - some signature files have been put into a more exposed place
1548       - the ugly 'f type parameter is gone (simplifies types tremendously!)
1549       - ml-nlffigen changed accordingly
1550       - tutorial updated
1551    
1552    Other changes:
1553    
1554       - author's affiliation in CM manual(s) updated
1555       - some more recognized keywords added to Allen's sml.sty
1556    
1557    ----------------------------------------------------------------------
1558    Name: Matthias Blume
1559    Date: 2001/05/25 15:30:00 EDT
1560    Tag: blume-20010525-iptr
1561    Description:
1562    
1563      - put the official 110.33-README (as it appears on the ftp server) under
1564        CVS
1565      - fixed a small bug related to incomplete pointer types in
1566        ml-nlffigen
1567      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1568        (it does not need the 'f type parameter)
1569    
1570    ----------------------------------------------------------------------
1571    Name: Matthias Blume
1572    Date: 2001/05/23 14:30:00 EDT
1573    Tag: Release_110_33
1574    Description:
1575    
1576    New version number (110.33).  New bootfiles.
1577    
1578    ----------------------------------------------------------------------
1579    Name: Matthias Blume
1580    Date: 2001/05/22 18:06:00 EDT
1581    Tag: blume-20010522-targets
1582    Description:
1583    
1584    Made install.sh use file config/targets.customized if it exists, falling
1585    back to config/targets if it doesn't.  This way one can have a customized
1586    version of the targets file without touching the "real thing", thus
1587    eliminating the constant fear of accidentally checking something bogus
1588    back into the CVS repository...  (File config/targets.customized must
1589    not be added to the repository!)
1590    
1591    ----------------------------------------------------------------------
1592    Name: Matthias Blume
1593  Date: 2001/05/22 16:30:00 EDT  Date: 2001/05/22 16:30:00 EDT
1594  Tag: blume-20010522-minitut  Tag: blume-20010522-minitut
1595  Description:  Description:
# Line 2257  Line 3833 
3833       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3834       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3835       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
3836       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3837       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3838       Dave's hands. :)       Dave's hands. :)
3839    

Legend:
Removed from v.832  
changed lines
  Added in v.1069

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