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 837, Fri Jun 1 17:27:54 2001 UTC revision 1073, Fri Feb 15 22:07:38 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 17:05:00
18    Tag: Release_110_39
19    Description:
20    
21    Last-minute retagging is becoming a tradition... :-(
22    
23    This is the working release 110.39.
24    
25    ----------------------------------------------------------------------
26    Name: Matthias Blume
27    Date: 2002/02/15 16:00:00 EST
28    Tag: Release_110_39-orig
29    Description:
30    
31    Working release 110.39.  New bootfiles.
32    
33    (Update: There was a small bug in the installer so it wouldn't work
34    with all shells.  So I retagged. -Matthias)
35    
36    ----------------------------------------------------------------------
37    Name: Matthias Blume
38    Date: 2002/02/15 14:17:00 EST
39    Tag: blume-20020215-showbindings
40    Description:
41    
42    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
43    the latter can be useful for exploring what bindings are available at
44    the interactive prompt.  (The first function returns only the list
45    of symbols that are really bound, the second prints those but also the
46    ones that CM's autoloading mechanism knows about.)
47    
48    ----------------------------------------------------------------------
49    Name: Matthias Blume
50    Date: 2002/02/15 12:08:00 EST
51    Tag: blume-20020215-iptrs
52    Description:
53    
54    Two improvements to ml-nlffigen:
55    
56      1. Write files only if they do not exist or if their current contents
57         do not coincide with what's being written.  (That is, avoid messing
58         with the time stamps unless absolutely necessary.)
59    
60      2. Implement a "repository" mechanism for generated files related
61         to "incomplete pointer types".   See the README file for details.
62    
63    ----------------------------------------------------------------------
64    Name: Matthias Blume
65    Date: 2002/02/14 11:50:00 EST
66    Tag: blume-20020214-quote
67    Description:
68    
69    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
70    because of the new and improved tag generation scheme.  (Thanks to Allen
71    Leung for pointing it out.)
72    
73    ----------------------------------------------------------------------
74    Name: Lal George
75    Date: 2002/02/14 09:55:27 EST 2002
76    Tag: george-20020214-isabelle-bug
77    Description:
78    
79    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
80    of Isabelle on the x86.
81    
82    From Allen:
83    -----------
84     I've found the problem:
85    
86         in ra-core.sml, I use the counter "blocked" to keep track of the
87         true number of elements in the freeze queue.  When the counter goes
88         to zero, I skip examining the queue.  But I've messed up the
89         bookkeeping in combine():
90    
91             else ();
92             case !ucol of
93               PSEUDO => (if !cntv > 0 then
94                     (if !cntu > 0 then blocked := !blocked - 1 else ();
95                                        ^^^^^^^^^^^^^^^^^^^^^^^
96                      moveu := mergeMoveList(!movev, !moveu)
97                     )
98                  else ();
99    
100         combine() is called to coalesce two nodes u and v.
101         I think I was thinking that if the move counts of u and v are both
102         greater than zero then after they are coalesced then one node is
103         removed from the freeze queue.  Apparently I was thinking that
104         both u and v are of low degree, but that's clearly not necessarily true.
105    
106    
107    02/12/2002:
108        Here's the patch.  HOL now compiles.
109    
110        I don't know how this impact on performance (compile
111        time or runtime).  This bug caused the RA (especially on the x86)
112        to go thru the potential spill phase when there are still nodes on the
113        freeze queue.
114    
115    
116    
117    
118    ----------------------------------------------------------------------
119    Name: Matthias Blume
120    Date: 2002/02/13 22:40:00 EST
121    Tag: blume-20020213-fptr-rtti
122    Description:
123    
124    Fixed a bug in ml-nlffigen that was introduced with one of the previous
125    updates.
126    
127    ----------------------------------------------------------------------
128    Name: Matthias Blume
129    Date: 2002/02/13 16:41:00 EST
130    Tag: blume-20020213-cmlpq
131    Description:
132    
133    Added new priority queue export symbols (which have just been added to
134    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
135    not compile and the installer would choke.)
136    
137    ----------------------------------------------------------------------
138    Name: Matthias Blume
139    Date: 2002/02/13 16:15:00 EST
140    Tag: blume-20020213-various
141    Description:
142    
143    1. More tweaks to ml-nlffigen:
144    
145       - better internal datastructures (resulting in slight speedup)
146       - "-match" option requires exact match
147       - "localized" gensym counters (untagged structs/unions nested within
148         other structs/unions or within typedefs get a fresh counter; their
149         tag will be prefixed by a concatenation of their parents' tags)
150       - bug fixes (related to calculation of transitive closure of types
151         to be included in the output)
152    
153    2. Minor Basis updates:
154    
155       - added implementations for List.collate and Option.app
156    
157    ----------------------------------------------------------------------
158    Name: Matthias Blume
159    Date: 2002/02/11 15:55:00 EST
160    Tag: blume-20020211-gensym
161    Description:
162    
163    Added a "-gensym" option to command line of ml-nlffigen.  This can be
164    used to specify a "stem" -- a string that is inserted in all "gensym'd"
165    names (ML structure names that correspond to unnamed C structs, unions,
166    and enums), so that separate runs of ml-nlffigen do not clash.
167    
168    ----------------------------------------------------------------------
169    Name: Matthias Blume
170    Date: 2002/02/11 12:05:00 EST
171    Tag: blume-20020211-gensml
172    Description:
173    
174    A quick fix for a problem with GenSML (in the pgraph-util library):
175    Make generation of toplevel "local" optional.  (Strictly speaking,
176    signature definitions within "local" are not legal SML.)
177    
178    Other than that: updates to INSTALL and cm/TODO.
179    
180    ----------------------------------------------------------------------
181    Name: Matthias Blume
182    Date: 2002/02/08 15:00:00 EST
183    Tag: blume-20020208-uniquepid
184    Description:
185    
186    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
187    
188    1. The installer (config/install.sh) has gotten smarter:
189    
190         - Configuration options are a bit easier to specify now
191           (in config/targets).
192         - Bug in recognizing .tar.bz2 files fixed.
193         - Installer automatically resolves dependencies between
194           configuration options (e.g., if you ask for eXene, you will
195           also get cml -- regardless whether you asked for it or not).
196         - Installer can run in "quieter mode" by setting the environment
197           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
198           "completely silent", though.
199         - Build HashCons library as part of smlnj-lib.
200    
201    2. A new scheme for assigning persistent identifiers to compilation
202       units (and, by extension, to types etc.) has been put into place.
203       This fixes a long-standing bug where types and even dynamic values
204       can get internally confused, thereby compromising type safety
205       (abstraction) and dynamic correctness.  See
206    
207         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
208    
209       for an example of how things could go wrong until now.
210    
211       The downside of the new scheme is that pids are not quite as
212       persistent as they used to be: CM will generate a fresh pid
213       for every compilation unit that it thinks it sees for the first
214       time.  That means that if you compile starting from a clean, fresh
215       source tree at two different times, you end up with different
216       binaries.
217    
218       Cutoff recompilation, however, has not been compromised because
219       CM keeps pid information in special caches between runs.
220    
221    ----------------------------------------------------------------------
222    Name: Lal George
223    Date: 2002/02/07 15:34:13 EST 2002
224    Tag: <none>
225    Description:
226    
227    Compilers that generate assembly code may produce  global labels
228    whose value is resolved at link time. The various peephole optimization
229    modules did not take this in account.
230    
231    TODO. The Labels.addrOf function should really return an option
232    type so that clients are forced to deal with this issue, rather
233    than an exception being raised.
234    
235    ----------------------------------------------------------------------
236    Name: Lal George
237    Date: 2002/02/06 13:55:02 EST
238    Tag: george-20020206-ra-breakup
239    Description:
240    
241    1. A bug fix from Allen.
242    
243        A typo causes extra fstp %st(0)'s to be generated at compensation
244        edges, which might cause stack underflow traps at runtime.  This
245        occurs in fft where there are extraneous fstps right before the 'into'
246        trap instruction (in this case they are harmless since none of the
247        integers overflow.)
248    
249    2. Pulled out various utility modules that were embedded in the modules
250       of the register allocator. I need these modules for other purposes, but
251       they are not complete enough to put into a library (just yet).
252    ----------------------------------------------------------------------
253    Name: Matthias Blume
254    Date: 2002/01/31 16:05:00 EST
255    Tag: blume-20020131-sparc-ccalls
256    Description:
257    
258    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
259       of extra stack space by mistake.  Fixed.
260    
261    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
262    
263    ----------------------------------------------------------------------
264    Name: Allen Leung
265    Date: 2002/01/30
266    Tag: leunga-20020130-nowhere-bug-fix
267    Description:
268    
269       MLRISC bug fixes:
270       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
271       2. MachineInt.fromString was negating its value.
272    
273    ----------------------------------------------------------------------
274    Name: Matthias Blume
275    Date: 2002/01/29
276    Tag: blume-20020129-INSTALL
277    Description:
278    
279    - Added somewhat detailed installation instructions (file INSTALL).
280    - Fixed curl-detection bug in config/install.sh.
281    - It is now possible to select the URL getter using the URLGETTER
282      environment variable:
283    
284          not set / "unknown"      --> automatic detection (script tries wget,
285                                       curl, and lynx)
286          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
287                                       how to properly invoke them)
288          other                    --> use $URLGETTER directly, it must take
289                                       precisely two command-line arguments
290                                       (source URL and destination file name)
291    
292    ----------------------------------------------------------------------
293    Name: Matthias Blume
294    Date: 2002/01/28
295    Tag: blume-20020128-sparc-ccalls
296    Description:
297    
298    - Fixed problem with calculation of "used" registers in sparc-c-calls.
299    - Make use of the allocParam argument in sparc-c-calls.
300    
301    ----------------------------------------------------------------------
302    Name: Matthias Blume
303    Date: 2002/01/28
304    Tag: blume-20020128-allocParam
305    Description:
306    
307    John Reppy:  Changes c-calls API to accept client-callback for
308    allocating extra stack space.
309    me: Corresponding changes to mlriscGen (using a dummy argument that
310        does not change the current behavior).
311    
312    ----------------------------------------------------------------------
313    Name: Matthias Blume
314    Date: 2002/01/28 12:00:00
315    Tag: Release_110_38
316    Description:
317    
318    This time for real!!!
319    
320    ----------------------------------------------------------------------
321    Name: Matthias Blume
322    Date: 2002/01/28 10:56:00 EST
323    Tag: blume-20020128-retraction
324    Description:
325    
326    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
327       with blume-Release_110_38-retracted.)
328    
329    1. Fixed a problem with incorrect rounding modes in real64.sml.
330       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
331    
332    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
333       and enums fixed.  The naming of corresponding ML identifiers should
334       now be consistent again.
335    
336    ----------------------------------------------------------------------
337    Name: Allen Leung
338    Date: 2002/01/27
339    Tag: leunga-20020127-nowhere
340    Description:
341    
342       Added a target called nowhere in the configuration scripts.
343       Enabling this will build the MLRISC 'nowhere' tool (for translating
344       programs with where-clauses into legal SML code) during installation.
345    
346    ----------------------------------------------------------------------
347    Name: Matthias Blume
348    Date: 2002/01/25 21:27:00 EST
349    Tag: blume-Release_110_38-retracted
350    Description:
351    
352    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
353    
354    README will be added later.
355    
356    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
357    (Release_110_38) removed.  Reason: Last-minute bug fixes.
358    
359    ----------------------------------------------------------------------
360    Name: Matthias Blume
361    Date: 2002/01/25
362    Tag: blume-20020125-ffi
363    Description:
364    
365    A large number of tweaks and improvements to ml-nlffi-lib and
366    ml-nlffigen:
367    
368       - ML represenation types have been streamlined
369       - getter and setter functions work with concrete values, not abstract
370         ones where possible
371       - ml-nlffigen command line more flexible (see README file there)
372       - some bugs have been fixed (hopefully)
373    
374    ----------------------------------------------------------------------
375    Name: Lal George
376    Date: 2002/01/24
377    Tag: george-20020124-risc-ra-interface
378    Description:
379    
380       There is a dramatic simplification in the interface to the
381       register allocator for RISC architectures as a result of making
382       parallel copy instructions explicit.
383    
384    ----------------------------------------------------------------------
385    Name: Matthias Blume
386    Date: 2002/01/22
387    Tag: blume-20020122-x86-ccalls
388    Description:
389    
390    Bug fix for c-calls on x86 (having to do with how char- and
391    short-arguments are being handled).
392    
393    ----------------------------------------------------------------------
394    Name: Matthias Blume
395    Date: 2002/01/21
396    Tag: blume-20020121-ff
397    Description:
398    
399    Another day of fiddling with the FFI...
400    
401    1. Bug fix/workaround:  CKIT does not complain about negative array
402       dimensions, so ml-nlffigen has to guard itself against this possibility.
403       (Otherwise a negative dimension would send it into an infinite loop.)
404    
405    2. Some of the abstract types (light objects, light pointers, most "base"
406       types) in structure C are now eqtypes.
407    
408    3. Added constructors and test functions for NULL function pointers.
409    
410    ----------------------------------------------------------------------
411    Name: Matthias Blume
412    Date: 2002/01/18
413    Tag: blume-20020118-ready-for-new-release
414    Description:
415    
416    Made config/srcarchiveurl point to a new place.  (Will provide boot
417    files shortly.)
418    
419    Maybe we christen this to be 110.38?
420    
421    ----------------------------------------------------------------------
422    Name: Matthias Blume
423    Date: 2002/01/18
424    Tag: blume-20020118-more-ffifiddle
425    Description:
426    
427    Today's FFI fiddling:
428    
429      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
430        C.Set.* that use concrete (MLRep.*) arguments and results instead
431        of abstract ones.
432    
433      - Provided word-style bit operations etc. for "int" representation
434        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
435        and Long.
436    
437    ----------------------------------------------------------------------
438    Name: Matthias Blume
439    Date: 2002/01/18
440    Tag: blume-20020118-use-x86-fp
441    Description:
442    
443    Now that x86-fast-fp seems to be working, I turned it back on again
444    by default.  (Seems to work fine now, even with the FFI.)
445    
446    Other than that, I added some documentation about the FFI to
447    src/ml-nlffigen/README and updated the FFI test examples in
448    src/ml-nlffi-lib/Tests/*.
449    
450    ----------------------------------------------------------------------
451    Name: Allen Leung
452    Date: 2002/01/17
453    Tag: leunga-20020117-x86-fast-fp-call
454    Description:
455    
456       1. Fixed a problem with handling return fp values when x86's fast fp
457          mode is turned on.
458    
459       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
460          of %f32.
461    
462       3. Added a constructor INT32lit to the ast of MLRISC tools.
463    
464    ----------------------------------------------------------------------
465    Name: Matthias Blume
466    Date: 2002/01/16
467    Tag: blume-20020116-ffifiddle
468    Description:
469    
470    More fiddling with the FFI interface:
471    
472     - Make constness 'c instead of rw wherever possible.  This eliminates
473       the need for certain explicit coercions.  (However, due to ML's
474       value polymorphism, there will still be many cases where explicit
475       coercions are necessary.  Phantom types are not the whole answer
476       to modeling a subtyping relationship in ML.)
477    
478     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
479    
480     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
481       clumsy workaround for varargs problem.)
482    
483    ----------------------------------------------------------------------
484    Name: Lal George
485    Date: 2002/01/15
486    Tag: <none>
487    Description:
488    
489    1. Since COPY instructions are no longer native to the architecture,
490       a generic functor can be used to implement the expandCopies function.
491    
492    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
493       TEXT segment.
494    
495    ----------------------------------------------------------------------
496    Name: Matthias Blume
497    Date: 2002/01/15
498    Tag: blume-20020115-ffiupdates
499    Description:
500    
501    1. Fix for bug resulting in single-precision float values being returned
502       incorrectly from FFI calls.
503    
504    2. Small modifications to C FFI API:
505    
506        - memory-allocation routines return straight objects (no options)
507          and raise an exception in out-of-memory situations
508        - unsafe extensions to cast between function pointers and pointers
509          from/to ints
510        - added structure C_Debug as an alternative to structure C where
511          pointer-dereferencing (|*| and |*!) always check for null-pointers
512        - added open_lib' to DynLinkage;  open_lib' works like open_lib
513          but also takes a (possibly empty) list of existing library handles
514          that the current library depends on
515    
516    ----------------------------------------------------------------------
517    Name: Matthias Blume
518    Date: 2002/01/10
519    Tag: blume-20020110-newffigen
520    Description:
521    
522    1. Updates to portable graph code.
523    
524    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
525       more scalable now so that even huge interfaces such as the one
526       for GTK compile in finite time and space. :-)
527       See src/ml-nlffigen/README for details on what's new.
528    
529    ----------------------------------------------------------------------
530    Name: Lal George
531    Date: 2001/01/09 14:31:35 EST 2002
532    Tag: george-20011206-rm-native-copy
533    Description:
534    
535            Removed the native COPY and FCOPY instructions
536            from all the architectures and replaced it with the
537            explicit COPY instruction from the previous commit.
538    
539            It is now possible to simplify many of the optimizations
540            modules that manipulate copies. This has not been
541            done in this change.
542    
543    ----------------------------------------------------------------------
544    Name: Lal George
545    Date: 2001/12/06 16:50:13 EST 2001
546    Tag: george-20011206-mlrisc-instruction
547    Description:
548    
549    Changed the representation of instructions from being fully abstract
550    to being partially concrete. That is to say:
551    
552      from
553            type instruction
554    
555      to
556            type instr                              (* machine instruction *)
557    
558            datatype instruction =
559                LIVE of {regs: C.cellset, spilled: C.cellset}
560              | KILL of {regs: C.cellset, spilled: C.cellset}
561              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
562              | ANNOTATION of {i: instruction, a: Annotations.annotation}
563              | INSTR of instr
564    
565    This makes the handling of certain special instructions that appear on
566    all architectures easier and uniform.
567    
568    LIVE and KILL say that a list of registers are live or killed at the
569    program point where they appear. No spill code is generated when an
570    element of the 'regs' field is spilled, but the register is moved to
571    the 'spilled' (which is present, more for debugging than anything else).
572    
573    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
574    We used to generate:
575    
576            DEFFREG f1
577            f1 := f2 + f3
578            trapb
579    
580    but now generate:
581    
582            f1 := f2 + f3
583            trapb
584            LIVE {regs=[f1,f2,f3], spilled=[]}
585    
586    Furthermore, the DEFFREG (hack) required that all floating point instruction
587    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
588    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
589    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
590    this architecture).
591    
592    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
593    all the architectures. This will result in further simplification of the
594    register allocator that must be aware of them for coalescing purposes, and
595    will also simplify certain aspects of the machine description that provides
596    callbacks related to parallel copies.
597    
598    ANNOTATION should be obvious, and now INSTR represents the honest to God
599    machine instruction set!
600    
601    The <arch>/instructions/<arch>Instr.sml files define certain utility
602    functions for making porting easier -- essentially converting upper case
603    to lower case. All machine instructions (of type instr) are in upper case,
604    and the lower case form generates an MLRISC instruction. For example on
605    the alpha we have:
606    
607      datatype instr =
608         LDA of {r:cell, b:cell, d:operand}
609       | ...
610    
611      val lda : {r:cell, b:cell, d:operand} -> instruction
612        ...
613    
614    where lda is just (INSTR o LDA), etc.
615    
616    ----------------------------------------------------------------------
617    Name: Matthias Blume
618    Date: 2001/11/22 21:40:00 EST
619    Tag: Release_110_37
620    Description:
621    
622    Release 110.37.  This time for real.
623    
624    ----------------------------------------------------------------------
625    Name: Matthias Blume
626    Date: 2001/11/21 16:35:00 EST
627    Tag: blume-20011121-foot-in-mouth
628    Description:
629    
630    Removed the "Release_110_37" tag because of a serious bug.
631    This will be re-tagged once the bug is fixed.
632    
633    ----------------------------------------------------------------------
634    Name: Matthias Blume
635    Date: 2001/11/21 16:14:00 EST
636    Tag: blume-20011121-forgottenfile
637    Description:
638    
639    Forgot to add a file.  (Just a .tex-file -- part of
640    the CM manual source.)
641    
642    ----------------------------------------------------------------------
643    Name: Matthias Blume
644    Date: 2001/11/21 16:10:00 EST
645    Tag: blume-20011121-invalid_110_37
646    Description:
647    
648    Note: I removed the original tag "Release_110_37" from this commit
649          because we found a serious bug in all non-x86 backends.
650          - Matthias
651    
652    1. Modifications to the SML/NJ code generator and to the runtime system
653       so that code object name strings are directly inserted into code
654       objects at code generation time.  The only business the runtime system
655       has with this is now to read the name strings on occasions.
656       (The encoding of the name string has also changed somewhat.)
657    
658    2. CM now implements a simple "set calculus" for specifying export lists.
659       In particular, it is now possible to refer to the export lists of
660       other libraries/groups/sources and form unions as well as differences.
661       See the latest CM manual for details.
662    
663    3. An separate notion of "proxy" libraries has again be eliminated from
664       CM's model.  (Proxy libraries are now simply a special case of using
665       the export list calculus.)
666    
667    4. Some of the existing libraries now take advantage of the new set
668       calculus.
669       (Notice that not all libraries have been converted because some
670       of the existing .cm-files are supposed to be backward compatible
671       with 110.0.x.)
672    
673    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
674       "General.exnMessage"!  The former relies on a certain hook to be
675       initialized, and that often does not happen in the stand-alone case.)
676    
677    ----------------------------------------------------------------------
678    Name: Lal George
679    Date: 2001/11/21  13:56:18 EST
680    Tag: george-2001121-pseudo-ops
681    Description:
682    
683      Implemented a complete redesign of MLRISC pseudo-ops. Now there
684      ought to never be any question of incompatabilities with
685      pseudo-op syntax expected by host assemblers.
686    
687      For now, only modules supporting GAS syntax are implemented
688      but more should follow, such as MASM, and vendor assembler
689      syntax, e.g. IBM as, Sun as, etc.
690    
691    ----------------------------------------------------------------------
692    Name: Matthias Blume
693    Date: 2001/11/14 11:52:00 EST
694    Tag: blume-20011114-srcname
695    Description:
696    
697    1. Routed the name of the current source file to mlriscgen where it
698       should be directly emitted into the code object.  (This last part
699       is yet to be done.)
700    
701    2. Some cleanup of the pgraph code to make it match the proposal that
702       I put out the other day.  (The proposal notwithstanding, things are
703       still in flux here.)
704    
705    ----------------------------------------------------------------------
706    Name: Lal George
707    Date: 2001/11/14 09:44:04 EST
708    Tag:
709    Description:
710    
711      Fix for a backpatching bug reported by Allen.
712    
713      Because the boundary between short and long span-dependent
714      instructions is +/- 128, there are an astounding number of
715      span-dependent instructions whose size is over estimated.
716    
717      Allen came up with the idea of letting the size of span
718      dependent instructions be non-monotonic, for a maxIter
719      number of times, after which the size must be monotonically
720      increasing.
721    
722      This table shows the number of span-dependent instructions
723      whose size was over-estimated as a function of maxIter, for the
724      file Parse/parse/ml.grm.sml:
725    
726         maxIter            # of instructions:
727            10                      687
728            20                      438
729            30                      198
730            40                        0
731    
732      In compiling the compiler, there is no significant difference in
733      compilation speed between maxIter=10 and maxIter=40. Actually,
734      my measurements showed that maxIter=40 was a tad faster than
735      maxIter=10! Also 96% of the  files in the compiler reach a fix
736      point within 13 iterations, so fixing maxIter at 40, while high,
737      is okay.
738    
739    ----------------------------------------------------------------------
740    Name: Matthias Blume
741    Date: 2001/10/31 15:25:00 EST
742    Tag: blume-20011031-pgraph
743    Description:
744    
745    CKIT:
746    * Changed the "Function" constructor of type Ast.ctype to carry optional
747      argument identifiers.
748    * Changed the return type of TypeUtil.getFunction accordingly.
749    * Type equality ignores the argument names.
750    * TypeUtil.composite tries to preserve argument names but gives up quickly
751      if there is a mismatch.
752    
753    installation script:
754    * attempts to use "curl" if available (unless "wget" is available as well)
755    
756    CM:
757    * has an experimental implementation of "portable graphs" which I will
758      soon propose as an implementation-independent library format
759    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
760    
761    NLFFI-LIB:
762    * some cleanup (all cosmetic)
763    
764    NLFFIGEN:
765    * temporarily disabled the mechanism that suppresses ML output for
766      C definitions whose identifiers start with an underscore character
767    * generate val bindings for enum constants
768    * user can request that only one style (light or heavy) is being used;
769      default is to use both (command-line arguments: -heavy and -light)
770    * fixed bug in handling of function types involving incomplete pointers
771    * generate ML entry points that take record arguments (i.e., using
772      named arguments) for C functions that have a prototype with named
773      arguments
774      (see changes to CKIT)
775    
776    ----------------------------------------------------------------------
777    Name: Allen Leung
778    Date: 2001/10/27 20:34:00 EDT
779    Tag: leunga-20011027-x86-fast-fp-call
780    Description:
781    
782       Fixed the bug described in blume-20010920-slowfp.
783    
784       The fix involves
785          1. generating FCOPYs in FSTP in ia32-svid
786          2. marking a CALL with the appropriate annotation
787    
788    ----------------------------------------------------------------------
789    Name: Matthias Blume
790    Date: 2001/10/16 11:32:00 EDT
791    Tag: blume-20011016-netbsd
792    Description:
793    
794    Underscore patch from Chris Richards (fixing problem with compiling
795    runtime system under recent NetBSD).
796    
797    ----------------------------------------------------------------------
798    Name: Allen Leung
799    Date: 2001/10/12 17:18:32 EDT 2001
800    Tag: leung-20011012-x86-printflowgraph
801    Description:
802    
803    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
804    
805    ----------------------------------------------------------------------
806    Name: Lal George
807    Date: 2001/10/11 23:51:34 EDT
808    Tag: george-20011011-too-many-instrs
809    Description:
810    
811    The representation of a program point never expected to see more
812    than 65536 instructions in a basic block!
813    
814    ----------------------------------------------------------------------
815    Name: Lal George
816    Date: 2001/10/09 09:41:37 EDT
817    Tag: george-20011008-mlrisc-labels
818    Description:
819    
820    Changed the machine description files to support printing of
821    local and global labels in assembly code, based on host assembler
822    conventions.
823    
824    ----------------------------------------------------------------------
825    Name: Matthias Blume
826    Date: 2001/09/25 15:25:00 EDT
827    Tag: blume-20010925-exninfo
828    Description:
829    
830    I provided a non-hook implementation of exnName (at the toplevel) and
831    made the "dummy" implementation of exnMessage (at the toplevel) more
832    useful: if nothing gets "hooked in", then at least you are going to
833    see the exception name and a message indicating why you don't see more.
834    
835    [For the time being, programs that need exnMessage and want to use
836    ml-build should either use General.exnMessage (strongly recommended) or
837    refer to structure General at some other point so that CM sees a
838    static dependency.]
839    
840    [Similar remarks go for "print" and "use":  If you want to use their
841    functionality in stand-alone programs generated by ml-build, then use
842    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
843    
844    ----------------------------------------------------------------------
845    Name: Matthias Blume
846    Date: 2001/09/20 17:28:00 EDT
847    Tag: blume-20010920-slowfp
848    Description:
849    
850    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume
854    Date: 2001/09/20 17:20:00 EDT
855    Tag: blume-20010920-canonicalpaths
856    Description:
857    
858    0. Updated the BOOT file (something that I forgot to do earlier).
859    
860    1. Small internal change to CM so that it avoids "/../" in filenames
861       as much as possible (but only where it is safe).
862    
863    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
864       that arguments that contain delimiters are passed through correctly.
865       This change also means that all "special" arguments of the form
866       @SMLxxx... must come first.
867    
868    3. Changed install script to put relative anchor names for tool commands
869       into pathconfig.
870    
871    ----------------------------------------------------------------------
872    Name: Matthias Blume
873    Date: 2001/09/18 15:35:00 EDT
874    Tag: blume-20010918-readme11036
875    Description:
876    
877    Added README files.
878    
879    ----------------------------------------------------------------------
880    Name: Matthias Blume
881    Date: 2001/09/18 11:45:00 EDT
882    Tag: Release_110_36 (retag)
883    Description:
884    
885    Fixed mistake in config/preloads. Retagged as 110.36.
886    
887    ----------------------------------------------------------------------
888    Name: Matthias Blume
889    Date: 2001/09/18 09:40:00 EDT
890    Tag: Release_110_36_orig (tag changed)
891    Description:
892    
893    New version (110.36).  New bootfiles.
894    
895    ----------------------------------------------------------------------
896    Name: Matthias Blume
897    Date: 2001/09/14 16:15:00 EDT
898    Tag: blume-20010914-x86fastfp
899    Description:
900    
901    John committed some changes that Allen made, in particular a (hopefully)
902    correctly working version of the x86-fp module.
903    
904    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
905    flag to "true".  Everything seems to compile to a fixpoint ok, and
906    "mandelbrot" speeds up by about 15%.
907    
908    ----------------------------------------------------------------------
909    Name: Matthias Blume
910    Date: 2001/09/13 11:20:00 EDT
911    Tag: blume-20010913-minimal
912    Description:
913    
914    1. Stefan Monnier's patch to fix a miscompilation problem that
915       was brought to light by John Reppy's work on Moby.
916    
917    2. Implemented a minimal "structure Compiler" that contains just
918       "version" and "architecture".  The minimal version will be
919       available when the full version is not.  This is for backward-
920       compatibility with code that wants to test Compiler.version.
921    
922    ----------------------------------------------------------------------
923    Name: Matthias Blume
924    Date: 2001/08/28 14:03:00 EDT
925    Tag: blume-20010828-ml-lex
926    Description:
927    
928    Fix for bug 1581, received from Neophytos Michael.
929    
930    ----------------------------------------------------------------------
931    Name: Matthias Blume
932    Date: 2001/08/27 11:20:00 EDT
933    Tag: blume-20010827-readme11035
934    Description:
935    
936    Fleshed out the README file for 110.35.
937    
938    ----------------------------------------------------------------------
939    Name: Matthias Blume
940    Date: 2001/08/24 17:10:00 EDT
941    Tag: Release_110_35
942    Description:
943    
944    New version number (110.35).  New bootfiles.
945    
946    ----------------------------------------------------------------------
947    Name: Lal George
948    Date: 2001/08/24 13:47:18 EDT 2001
949    Tag: george-20010824-MLRISC-graphs
950    Description:
951    
952     removed clusters from MLRISC completely and replaced with graphs.
953    
954    ----------------------------------------------------------------------
955    Name: Matthias Blume
956    Date: 2001/08/23 17:50:00 EDT
957    Tag: blume-20010823-toplevel
958    Description:
959    
960    - some reorganization of the code that implements various kinds of
961      environments in the compiler (static, dynamic, symbolic, combined)
962    - re-implemented the EnvRef module so that evalStream works properly
963      (if the stream contains references to "use", "CM.make", etc.)
964    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
965    
966    ----------------------------------------------------------------------
967    Name: Matthias Blume
968    Date: 2001/08/20 15:50 EDT
969    Tag: blume20010820-slipup
970    Description:
971    
972    I forgot to commit a few files.  Here they are...
973    
974    ----------------------------------------------------------------------
975    Name: Matthias Blume
976    Date: 2001/08/20 15:35:00 EDT
977    Tag: blume-20010820-debugprof
978    Description:
979    
980    !!!! NEW BOOTFILES !!!!
981    
982    This is another round of reorganizing the compiler sources.  This
983    time the main goal was to factor out all the "instrumentation"
984    passes (for profiling and backtracing) into their own library.
985    The difficulty was to do it in such a way that it does not depend
986    on elaborate.cm but only on elabdata.cm.
987    
988    Therefore there have been further changes to both elaborate.cm and
989    elabdata.cm -- more "generic" things have been moved from the former
990    to the latter.  As a result, I was forced to split the assignment
991    of numbers indicating "primtyc"s into two portions: SML-generic and
992    SML/NJ-specific.  Since it would have been awkward to maintain,
993    I bit the bullet and actually _changed_ the mapping between these
994    numbers and primtycs.  The bottom line of this is that you need
995    a new set of bin- and bootfiles.
996    
997    I have built new bootfiles for all architectures, so doing a fresh
998    checkout and config/install.sh should be all you need.
999    
1000    The newly created library's name is
1001    
1002        $smlnj/viscomp/debugprof.cm
1003    
1004    and its sources live under
1005    
1006        src/compiler/DebugProf
1007    
1008    ----------------------------------------------------------------------
1009    Name: Matthias Blume
1010    Date: 2001/08/15 17:15:00 EDT
1011    Tag: blume-20010815-compreorg
1012    Description:
1013    
1014    This is a first cut at reorganizing the CM libraries that make up the
1015    core of the compiler.  The idea is to separate out pieces that could
1016    be used independently by tools, e.g., the parser, the typechecker, etc.
1017    
1018    The current status is a step in this direction, but it is not quite
1019    satisfactory yet.  Expect more changes in the future.
1020    
1021    Here is the current (new) organization...
1022    
1023        What used to be $smlnj/viscomp/core.cm is now divided into
1024        six CM libraries:
1025    
1026             $smlnj/viscomp/basics.cm
1027                           /parser.cm
1028                           /elabdata.cm
1029                           /elaborate.cm
1030                           /execute.cm
1031                           /core.cm
1032    
1033        The CM files for these libraries live under src/system/smlnj/viscomp.
1034        All these libraries are proxy libraries that contain precisely
1035        one CM library component.  Here are the locations of the components
1036        (all within the src/compiler tree):
1037    
1038             Basics/basics.cm
1039             Parse/parser.cm
1040             ElabData/elabdata.cm
1041             Elaborator/elaborate.cm
1042             Execution/execute.cm
1043             core.cm
1044    
1045         [This organization is the same that has been used already
1046         for a while for the architecture-specific parts of the visible
1047         compiler and for the old version of core.cm.]
1048    
1049         As you will notice, many source files have been moved from their
1050         respective original locations to a new home in one of the above
1051         subtrees.
1052    
1053         The division of labor between the new libraries is the following:
1054    
1055             basics.cm:
1056                - Simple, basic definitions that pertain to many (or all) of
1057                  the other libraries.
1058             parser.cm:
1059                - The SML parser, producing output of type Ast.dec.
1060                - The type family for Ast is also defined and exported here.
1061             elabdata.cm:
1062                - The datatypes that describe input and output of the elaborator.
1063                  This includes types, absyn, and static environments.
1064             elaborator.cm:
1065                - The SML/NJ type checker and elaborator.
1066                  This maps an Ast.dec (with a given static environment) to
1067                  an Absyn.dec (with a new static environment).
1068                - This libraries implements certain modules that used to be
1069                  structures as functors (to remove dependencies on FLINT).
1070             execute.cm:
1071                - Everything having to do with executing binary code objects.
1072                - Dynamic environments.
1073             core.cm:
1074                - SML/NJ-specific instantiations of the elaborator and MLRISC.
1075                - Top-level modules.
1076                - FLINT (this should eventually become its own library)
1077    
1078    Notes:
1079    
1080    I am not 100% happy with the way I separated the elaborator (and its
1081    data structures) from FLINT.  Two instances of the same problem:
1082    
1083        1. Data structures contain certain fields that carry FLINT-specific
1084           information.  I hacked around this using exn and the property list
1085           module from smlnj-lib.  But the fact that there are middle-end
1086           specific fields around at all is a bit annoying.
1087    
1088        2. The elaborator calculates certain FLINT-related information.  I tried
1089           to make this as abstract as I could using functorization, but, again,
1090           the fact that the elaborator has to perform calculations on behalf
1091           of the middle-end at all is not nice.
1092    
1093        3. Having to used exn and property lists is unfortunate because it
1094           weakens type checking.  The other alternative (parameterizing
1095           nearly *everything*) is not appealing, though.
1096    
1097    I removed the "rebinding =" warning hack because due to the new organization
1098    it was awkward to maintain it.  As a result, the compiler now issues some of
1099    these warnings when compiling init.cmi during bootstrap compilation. On
1100    the plus side, you also get a warning when you do, for example:
1101       val op = = Int32.+
1102    which was not the case up to now.
1103    
1104    I placed "assign" and "deref" into the _Core structure so that the
1105    code that deals with the "lazy" keyword can find them there.  This
1106    removes the need for having access to the primitive environment
1107    during elaboration.
1108    
1109    ----------------------------------------------------------------------
1110    Name: Matthias Blume
1111    Date: 2001/08/13
1112    Tag: blume-20010813-closures
1113    Description:
1114    
1115    This fix was sent to us by Zhong Shao.  It is supposed to improve the
1116    performance of certain loops by avoiding needless closure allocation.
1117    
1118    ----------------------------------------------------------------------
1119    Name: Lal George
1120    Date: 2001/07/31 10:03:23 EDT 2001
1121    Tag: george-20010731-x86-fmalloc
1122    Description: Fixed bug in x86 calls
1123    
1124        There was a bug where call instructions would mysteriously
1125        vanish. The call instruction had to be one that returned
1126        a floating point value.
1127    
1128    ----------------------------------------------------------------------
1129    Name: Lal George
1130    Date: 2001/07/19 16:36:29 EDT 2001
1131    Tag: george-20010719-simple-cells
1132    Description:
1133    
1134    I have dramatically simplified the interface for CELLS in MLRISC.
1135    
1136    In summary, the cells interface is broken up into three parts:
1137    
1138      1. CellsBasis : CELLS_BASIS
1139    
1140            CellsBasis is a top level structure and common for all
1141            architectures.  it contains the definitions of basic datatypes
1142            and utility  functions over these types.
1143    
1144      2. functor Cells() : CELLS
1145    
1146            Cells generates an interface for CELLS that incorporates the
1147            specific resources on the target architecture, such as the
1148            presence of special register classes, their number and size,
1149            and various useful substructures.
1150    
1151      3. <ARCH>CELLS
1152    
1153            e.g. SparcCells: SPARCCELLS
1154    
1155            <ARCH>CELLS usually contains additional bindings for special
1156            registers  on the architecture, such as:
1157    
1158                    val r0 : cell           (* register zero *)
1159                    val y : cell            (* Y register *)
1160                    val psr : cell          (* processor status register *)
1161                    ...
1162    
1163            The structure returned by applying the Cells functor is opened
1164            in this interface.
1165    
1166    The main implication of all this is that the datatypes for cells is
1167    split between CellsBasis and CELLS -- a fairly simple change for user
1168    code.
1169    
1170    In the old scheme the CELLS interface had a definitional binding of
1171    the form:
1172    
1173            signature CELLS = sig
1174    
1175               structure CellsBasis = CellsBasis
1176    
1177               ...
1178    
1179            end
1180    
1181    With all the sharing constraints that goes on in MLRISC, this old
1182    design  quickly leads to errors such as:
1183    
1184            "structure definition spec inside of sharing ... "
1185    
1186    
1187    and appears to require an unacceptable amount of sharing and where
1188    constraint hackery.
1189    
1190    I think this error message (the interaction of definitional specs and
1191    sharing) requires more explanation on our web page.
1192    
1193    ----------------------------------------------------------------------
1194    Name: Matthias Blume
1195    Date: 2001/07/19 15:00:00 EDT
1196    Tag: blume-20010719-libreorg
1197    Description:
1198    
1199    This update puts together a fairly extensive but straightforward change
1200    to the way the libraries that implement the interactive system are
1201    organized:
1202    
1203       The biggest change is the elimination of structure Compiler.  As a
1204       replacement for this structure, there is now a CM library
1205       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1206       that exports all the substructures of the original structure Compiler
1207       directly.  So instead of saying Compiler.Foo.bar one now simply
1208       says Foo.bar.  (The CM libraries actually export a collection of
1209       structures that is richer than the collection of substructures of
1210       structure Compiler.)
1211    
1212       To make the transition smooth, there is a separate library called
1213       $smlnj/compiler/compiler.cm which puts together and exports the
1214       original structure Compiler (or at least something very close to it).
1215    
1216       There are five members of the original structure Compiler
1217       that are not exported directly but which instead became members
1218       of a new structure Backend (described by signature BACKEND).  These are:
1219       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1220       Interact (: INTERACT), structure Machine (: MACHINE), and val
1221       architecture (: string).
1222    
1223       Structure Compiler.Version has become structure CompilerVersion.
1224    
1225       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1226       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1227       or x86, respectively.
1228       Each of these exports the same frontend structures that
1229       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1230       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1231       Hppa, PPC, Sparc, or X86, respectively.
1232    
1233       Library $smlnj/compiler/all.cm exports the union of the exports of
1234       $smlnj/compiler/<arch>.cm
1235    
1236       There are no structures <Arch>Compiler anymore, use
1237       $smlnj/compiler/<arch>.cm instead.
1238    
1239       Library host-compiler-0.cm is gone.  Instead, the internal library
1240       that instantiates CM is now called cm0.cm.  Selection of the host
1241       compiler (backend) is no longer done here but. (Responsibility for it
1242       now lies with $smlnj/compiler/current.cm.  This seems to be more
1243       logical.)
1244    
1245       Many individual files have been moved or renamed.  Some files have
1246       been split into multiple files, and some "dead" files have been deleted.
1247    
1248    Aside from these changes to library organization, there are also changes
1249    to the way the code itself is organized:
1250    
1251       Structure Binfile has been re-implemented in such a way that it no
1252       longer needs any knowledge of the compiler.  It exclusively deals
1253       with the details of binfile layout.  It no longer invokes the
1254       compiler (for the purpose of creating new prospective binfile
1255       content), and it no longer has any knowledge of how to interpret
1256       pickles.
1257    
1258       Structure Compile (: COMPILE) has been stripped down to the bare
1259       essentials of compilation.  It no longer deals with linking/execution.
1260       The interface has been cleaned up considerably.
1261    
1262       Utility routines for dealing with linking and execution have been
1263       moved into their own substructures.
1264    
1265       (The ultimate goal of these changes is to provide a light-weight
1266       binfile loader/linker (at least for, e.g., stable libraries) that
1267       does not require CM or the compiler to be present.)
1268    
1269    CM documentation has been updated to reflect the changes to library
1270    organization.
1271    
1272    ----------------------------------------------------------------------
1273    Name: Matthias Blume
1274    Date: 2001/07/10 17:30:00 EDT
1275    Tag: Release_110_34
1276    Description:
1277    
1278    Minor tweak to 110.34 (re-tagged):
1279    
1280      - README.html file added to CVS repository
1281      - runtime compiles properly under FreeBSD 3.X and 4.X
1282    
1283    ----------------------------------------------------------------------
1284    Name: Matthias Blume
1285    Date: 2001/07/10 17:30:00 EDT
1286    Tag: Release_110_34
1287    Description:
1288    
1289    New version number (110.34). New bootfiles.
1290    
1291    ----------------------------------------------------------------------
1292    Name: Matthias Blume
1293    Date: 2001/07/09 16:00:00 EDT
1294    Tag: blume-20010709-more-varargs
1295    Description:
1296    
1297    I changed the handling of varargs in ml-nlffigen again:
1298    The ellipsis ... will now simply be ignored (with an accompanying warning).
1299    
1300    The immediate effect is that you can actually call a varargs function
1301    from ML -- but you can't actually supply any arguments beyond the ones
1302    specified explicitly.  (For example, you can call printf with its format
1303    string, but you cannot pass additional arguments.)
1304    
1305    This behavior is only marginally more useful than the one before, but
1306    it has the advantage that a function or, more importantly, a function
1307    type never gets dropped on the floor, thus avoiding follow-up problems with
1308    other types that refer to the offending one.
1309    
1310    ----------------------------------------------------------------------
1311    Name: Matthias Blume
1312    Date: 2001/07/09 11:25:00 EDT
1313    Tag: blume-20010709-varargs
1314    Description:
1315    
1316    1. ckit-lib.cm now exports structure Error
1317    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1318       with a warning accompanied by a source location.  Moreover, it
1319       merely skips the offending function or type and proceeds with the
1320       rest of its work.u  As a result, one can safely feed C code containing
1321       "..." to ml-nlffigen.
1322    3. There are some internal improvements to CM, providing slightly
1323       more general string substitutions in the tools subsystem.
1324    
1325    ----------------------------------------------------------------------
1326    Name: Matthias Blume
1327    Date: 2001/06/27 15:10:00 EDT
1328    Tag: blume-20010627-concur
1329    Description:
1330    
1331    Fixed a small bug in CM's handling of parallel compilation.
1332    (You could observe the bug by Control-C-interrupting an ordinary
1333    CMB.make or CM.stabilize and then attaching some compile servers.
1334    The result was that all of a sudden the previously interrupted
1335    compilation would continue on its own.  This was because of
1336    an over-optimization: CM did not bother to clean out certain queues
1337    when no servers were attached "anyway", resulting in the contents
1338    of these queues to grab control when new servers did get attached.)
1339    
1340    There is also another minor update to the CM manual.
1341    
1342    ----------------------------------------------------------------------
1343    Name: Matthias Blume
1344    Date: 2001/06/26 16:15:00 EDT
1345    Tag: blume-20010626-cmdoc
1346    Description:
1347    
1348    Minor typo fixed in CM manual (syntax diagram for libraries).
1349    
1350    ----------------------------------------------------------------------
1351    Name: Matthias Blume
1352    Date: 2001/06/25 22:55:00 EDT
1353    Tag: blume-20010625-x86pc
1354    Description:
1355    
1356    Fixed a nasty bug in the X86 assembly code that caused signal
1357    handlers to fail (crash) randomly.
1358    
1359    ----------------------------------------------------------------------
1360    Name: Matthias Blume
1361    Date: 2001/06/25 12:05:00 EDT
1362    Tag: blume-20010625-nlffigen
1363    Description:
1364    
1365    This update fixes a number of minor bugs in ml-nlffigen as reported by
1366    Nick Carter <nbc@andrew.cmu.edu>.
1367    
1368      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1369      2. Default names for generated files are now derived from the name of
1370         the C file *without its directory*.  In particular, this causes generated
1371         files to be placed locally even if the C file is in some system directory.
1372      3. Default names for generated signatures and structures are also derived
1373         from the C file name without its directory.  This avoids silly things
1374         like "structure GL/GL".
1375         (Other silly names are still possible because ml-nlffigen does not do
1376          a thorough check of whether generated names are legal ML identifiers.
1377          When in doubt, use command line arguments to force particular names.)
1378    
1379    ----------------------------------------------------------------------
1380    Name: Matthias Blume
1381    Date: 2001/06/21 12:25:00 EDT
1382    Tag: blume-20010621-eXene
1383    Description:
1384    
1385    eXene now compiles and (sort of) works again.
1386    
1387    The library name (for version > 110.33) is $/eXene.cm.
1388    
1389    I also added an new example in src/eXene/examples/nbody.  See the
1390    README file there for details.
1391    
1392    ----------------------------------------------------------------------
1393    Name: Matthias Blume
1394    Date: 2001/06/20 16:40:00 EDT
1395    Tag: blume-20010620-cml
1396    Description:
1397    
1398    CML now compiles and works again.
1399    
1400    Libraries (for version > 110.33):
1401    
1402      $cml/cml.cm            Main CML library.
1403      $cml/basis.cm          CML's version of $/basis.cm.
1404      $cml/cml-internal.cm   Internal helper library.
1405      $cml/core-cml.cm       Internal helper library.
1406      $cml-lib/trace-cml.cm  Tracing facility.
1407      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1408    
1409    The installer (config/install.sh) has been taught how to properly
1410    install this stuff.
1411    
1412    ----------------------------------------------------------------------
1413    Name: Matthias Blume
1414    Date: 2001/06/19 17:55:00 EDT
1415    Tag: blume-20010619-instantiate
1416    Description:
1417    
1418    This un-breaks the fix for bug 1432.
1419    (The bug was originally fixed in 110.9 but I broke it again some
1420    time after that.)
1421    
1422    ----------------------------------------------------------------------
1423    Name: Matthias Blume
1424    Date: 2001/06/19 17:25:00 EDT
1425    Tag: blume-20010619-signals
1426    Description:
1427    
1428    This should (hopefully) fix the long-standing signal handling bug.
1429    (The runtime system was constructing a continuation record with an
1430    incorrect descriptor which would cause the GC to drop data on the floor...)
1431    
1432    ----------------------------------------------------------------------
1433    Name: Matthias Blume
1434    Date: 2001/06/15 15:05:00 EDT
1435    Tag: blume-20010615-moresparc
1436    Description:
1437    
1438    Here is a short late-hour update related to Sparc c-calls:
1439    
1440     -- made handling of double-word arguments a bit smarter
1441    
1442     -- instruction selection phase tries to collapse certain clumsily
1443        constructed ML-Trees; typical example:
1444    
1445            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1446    
1447        This currently has no further impact on SML/NJ since mlriscGen does
1448        not seem to generate such patterns in the first place, and c-calls
1449        (which did generate them in the beginning) has meanwhile been fixed
1450        so as to avoid them as well.
1451    
1452    ----------------------------------------------------------------------
1453    Name: Matthias Blume
1454    Date: 2001/06/15 15:05:00 EDT
1455    Tag: blume-20010615-sparc
1456    Description:
1457    
1458    The purpose of this update is to provide an implementation of NLFFI
1459    on Sparc machines.
1460    
1461    Here are the changes in detail:
1462    
1463       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1464       the Sparc implementation of the c-calls API.
1465       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1466       for accessing the ML frame.  Thus, we have a real frame pointer and
1467       can freely modify %sp without need for an omit-frame-ptr phase.
1468       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1469       this case.
1470       * ml-nlffigen has been taught to produce code for different architectures
1471       and calling conventions.
1472       * In a way similar to what was done in the x86 case, the Sparc
1473       backend uses its own specific extension to mltree.  (For example,
1474       it needs to be able to generate UNIMP instructions which are part
1475       of the calling convention.)
1476       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1477       to make it easier to plug in new machine- and os-dependent parts).
1478    
1479    There are some other fairly unrelated bug fixes and cleanups as well:
1480    
1481       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1482       that they properly share their libraries with existing SML/NJ libraries.
1483       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1484       follow-up error messages.
1485       * Updates to CM/CMB documentation.
1486    
1487    TODO items:
1488    
1489       * MLRISC should use a different register as its asmTemp on the Sparc.
1490         (The current %o2 is a really bad choice because it is part of the
1491         calling conventions, so things might interfere in unexpected ways.)
1492    
1493    ----------------------------------------------------------------------
1494    Name: Matthias Blume
1495    Date: 2001/06/07
1496    Tag: blume-20010607-calls
1497    Description:
1498    
1499    A number of internal changes related to C calls and calling conventions:
1500    
1501    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1502       number of bytes popped implicitly (by the callee).  In most cases
1503       this field is 0 but on x86/win32 it is some non-zero value.  This
1504       is information provided for the benefit of the "omit-frameptr" pass.
1505    2. The CALL instruction on the x86 carries a similar "pops" field.
1506       The instruction selection phase copies its value from the ML-Tree
1507       CALL statement.
1508    3. On all other architectures, the instruction selection phase checks
1509       whether "pops=0" and complains if not.
1510    4. The c-calls implementation for x86 now accepts two calling conventions:
1511       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1512       up after the call and pops is set to 0.  For "stdcall", the caller
1513       does nothing, leaving the cleanup to the callee; pops is set to
1514       the number of bytes that were pushed onto the stack.
1515    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1516       distinguish between "ccall" and "stdcall".
1517    6. The UNIMP instruction has been added to the supported Sparc instruction
1518       set. (This is needed for implementing the official C calling convention
1519       on this architecture.)
1520    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1521       work with the latest CM.
1522    
1523    ----------------------------------------------------------------------
1524    Name: Matthias Blume
1525    Date: 2001/06/05 15:10:00 EDT
1526    Tag: blume-20010605-cm-index
1527    Description:
1528    
1529    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1530    
1531    1. CM can now generate "index files".  These are human-readable files
1532       that list on a per-.cm-file basis each toplevel symbol defined or
1533       imported.  The location of the index file for
1534       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1535       To enable index-file generation, set CM.Control.generate_index to true
1536       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1537    
1538       The CM manual has been updated accordingly.
1539    
1540    2. I made some slight modifications to the c-calls API in MLRISC.
1541    
1542         a) There is now a callback to support saving/restoring of
1543            dedicated but caller-save registers around the actual call
1544            instruction.
1545         b) One can optionally specify a comment-annotation for the
1546            call instruction.
1547    
1548    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1549       (For example, the comment annotation shows the C prototype of
1550        the function being called.)
1551    
1552    ----------------------------------------------------------------------
1553    Name: Matthias Blume
1554  Date: 2001/06/01 13:30:00 EDT  Date: 2001/06/01 13:30:00 EDT
1555  Tag: blume-20010601-nlffi-cleanup  Tag: blume-20010601-nlffi-cleanup
1556  Description:  Description:
# Line 2309  Line 3846 
3846       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3847       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3848       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
3849       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3850       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3851       Dave's hands. :)       Dave's hands. :)
3852    

Legend:
Removed from v.837  
changed lines
  Added in v.1073

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