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 800, Fri Mar 16 17:22:47 2001 UTC revision 1076, Tue Feb 19 15:47:18 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: 2002/02/19 10:48:00 EST
18    Tag: blume-20020219-pqfix
19    Description:
20    
21    Applied Chris Okasaki's bug fix for priority queues.
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume
25    Date: 2002/02/15 17:05:00
26    Tag: Release_110_39
27    Description:
28    
29    Last-minute retagging is becoming a tradition... :-(
30    
31    This is the working release 110.39.
32    
33    ----------------------------------------------------------------------
34    Name: Matthias Blume
35    Date: 2002/02/15 16:00:00 EST
36    Tag: Release_110_39-orig
37    Description:
38    
39    Working release 110.39.  New bootfiles.
40    
41    (Update: There was a small bug in the installer so it wouldn't work
42    with all shells.  So I retagged. -Matthias)
43    
44    ----------------------------------------------------------------------
45    Name: Matthias Blume
46    Date: 2002/02/15 14:17:00 EST
47    Tag: blume-20020215-showbindings
48    Description:
49    
50    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
51    the latter can be useful for exploring what bindings are available at
52    the interactive prompt.  (The first function returns only the list
53    of symbols that are really bound, the second prints those but also the
54    ones that CM's autoloading mechanism knows about.)
55    
56    ----------------------------------------------------------------------
57    Name: Matthias Blume
58    Date: 2002/02/15 12:08:00 EST
59    Tag: blume-20020215-iptrs
60    Description:
61    
62    Two improvements to ml-nlffigen:
63    
64      1. Write files only if they do not exist or if their current contents
65         do not coincide with what's being written.  (That is, avoid messing
66         with the time stamps unless absolutely necessary.)
67    
68      2. Implement a "repository" mechanism for generated files related
69         to "incomplete pointer types".   See the README file for details.
70    
71    ----------------------------------------------------------------------
72    Name: Matthias Blume
73    Date: 2002/02/14 11:50:00 EST
74    Tag: blume-20020214-quote
75    Description:
76    
77    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
78    because of the new and improved tag generation scheme.  (Thanks to Allen
79    Leung for pointing it out.)
80    
81    ----------------------------------------------------------------------
82    Name: Lal George
83    Date: 2002/02/14 09:55:27 EST 2002
84    Tag: george-20020214-isabelle-bug
85    Description:
86    
87    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
88    of Isabelle on the x86.
89    
90    From Allen:
91    -----------
92     I've found the problem:
93    
94         in ra-core.sml, I use the counter "blocked" to keep track of the
95         true number of elements in the freeze queue.  When the counter goes
96         to zero, I skip examining the queue.  But I've messed up the
97         bookkeeping in combine():
98    
99             else ();
100             case !ucol of
101               PSEUDO => (if !cntv > 0 then
102                     (if !cntu > 0 then blocked := !blocked - 1 else ();
103                                        ^^^^^^^^^^^^^^^^^^^^^^^
104                      moveu := mergeMoveList(!movev, !moveu)
105                     )
106                  else ();
107    
108         combine() is called to coalesce two nodes u and v.
109         I think I was thinking that if the move counts of u and v are both
110         greater than zero then after they are coalesced then one node is
111         removed from the freeze queue.  Apparently I was thinking that
112         both u and v are of low degree, but that's clearly not necessarily true.
113    
114    
115    02/12/2002:
116        Here's the patch.  HOL now compiles.
117    
118        I don't know how this impact on performance (compile
119        time or runtime).  This bug caused the RA (especially on the x86)
120        to go thru the potential spill phase when there are still nodes on the
121        freeze queue.
122    
123    
124    
125    
126    ----------------------------------------------------------------------
127    Name: Matthias Blume
128    Date: 2002/02/13 22:40:00 EST
129    Tag: blume-20020213-fptr-rtti
130    Description:
131    
132    Fixed a bug in ml-nlffigen that was introduced with one of the previous
133    updates.
134    
135    ----------------------------------------------------------------------
136    Name: Matthias Blume
137    Date: 2002/02/13 16:41:00 EST
138    Tag: blume-20020213-cmlpq
139    Description:
140    
141    Added new priority queue export symbols (which have just been added to
142    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
143    not compile and the installer would choke.)
144    
145    ----------------------------------------------------------------------
146    Name: Matthias Blume
147    Date: 2002/02/13 16:15:00 EST
148    Tag: blume-20020213-various
149    Description:
150    
151    1. More tweaks to ml-nlffigen:
152    
153       - better internal datastructures (resulting in slight speedup)
154       - "-match" option requires exact match
155       - "localized" gensym counters (untagged structs/unions nested within
156         other structs/unions or within typedefs get a fresh counter; their
157         tag will be prefixed by a concatenation of their parents' tags)
158       - bug fixes (related to calculation of transitive closure of types
159         to be included in the output)
160    
161    2. Minor Basis updates:
162    
163       - added implementations for List.collate and Option.app
164    
165    ----------------------------------------------------------------------
166    Name: Matthias Blume
167    Date: 2002/02/11 15:55:00 EST
168    Tag: blume-20020211-gensym
169    Description:
170    
171    Added a "-gensym" option to command line of ml-nlffigen.  This can be
172    used to specify a "stem" -- a string that is inserted in all "gensym'd"
173    names (ML structure names that correspond to unnamed C structs, unions,
174    and enums), so that separate runs of ml-nlffigen do not clash.
175    
176    ----------------------------------------------------------------------
177    Name: Matthias Blume
178    Date: 2002/02/11 12:05:00 EST
179    Tag: blume-20020211-gensml
180    Description:
181    
182    A quick fix for a problem with GenSML (in the pgraph-util library):
183    Make generation of toplevel "local" optional.  (Strictly speaking,
184    signature definitions within "local" are not legal SML.)
185    
186    Other than that: updates to INSTALL and cm/TODO.
187    
188    ----------------------------------------------------------------------
189    Name: Matthias Blume
190    Date: 2002/02/08 15:00:00 EST
191    Tag: blume-20020208-uniquepid
192    Description:
193    
194    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
195    
196    1. The installer (config/install.sh) has gotten smarter:
197    
198         - Configuration options are a bit easier to specify now
199           (in config/targets).
200         - Bug in recognizing .tar.bz2 files fixed.
201         - Installer automatically resolves dependencies between
202           configuration options (e.g., if you ask for eXene, you will
203           also get cml -- regardless whether you asked for it or not).
204         - Installer can run in "quieter mode" by setting the environment
205           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
206           "completely silent", though.
207         - Build HashCons library as part of smlnj-lib.
208    
209    2. A new scheme for assigning persistent identifiers to compilation
210       units (and, by extension, to types etc.) has been put into place.
211       This fixes a long-standing bug where types and even dynamic values
212       can get internally confused, thereby compromising type safety
213       (abstraction) and dynamic correctness.  See
214    
215         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
216    
217       for an example of how things could go wrong until now.
218    
219       The downside of the new scheme is that pids are not quite as
220       persistent as they used to be: CM will generate a fresh pid
221       for every compilation unit that it thinks it sees for the first
222       time.  That means that if you compile starting from a clean, fresh
223       source tree at two different times, you end up with different
224       binaries.
225    
226       Cutoff recompilation, however, has not been compromised because
227       CM keeps pid information in special caches between runs.
228    
229    ----------------------------------------------------------------------
230    Name: Lal George
231    Date: 2002/02/07 15:34:13 EST 2002
232    Tag: <none>
233    Description:
234    
235    Compilers that generate assembly code may produce  global labels
236    whose value is resolved at link time. The various peephole optimization
237    modules did not take this in account.
238    
239    TODO. The Labels.addrOf function should really return an option
240    type so that clients are forced to deal with this issue, rather
241    than an exception being raised.
242    
243    ----------------------------------------------------------------------
244    Name: Lal George
245    Date: 2002/02/06 13:55:02 EST
246    Tag: george-20020206-ra-breakup
247    Description:
248    
249    1. A bug fix from Allen.
250    
251        A typo causes extra fstp %st(0)'s to be generated at compensation
252        edges, which might cause stack underflow traps at runtime.  This
253        occurs in fft where there are extraneous fstps right before the 'into'
254        trap instruction (in this case they are harmless since none of the
255        integers overflow.)
256    
257    2. Pulled out various utility modules that were embedded in the modules
258       of the register allocator. I need these modules for other purposes, but
259       they are not complete enough to put into a library (just yet).
260    ----------------------------------------------------------------------
261    Name: Matthias Blume
262    Date: 2002/01/31 16:05:00 EST
263    Tag: blume-20020131-sparc-ccalls
264    Description:
265    
266    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
267       of extra stack space by mistake.  Fixed.
268    
269    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
270    
271    ----------------------------------------------------------------------
272    Name: Allen Leung
273    Date: 2002/01/30
274    Tag: leunga-20020130-nowhere-bug-fix
275    Description:
276    
277       MLRISC bug fixes:
278       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
279       2. MachineInt.fromString was negating its value.
280    
281    ----------------------------------------------------------------------
282    Name: Matthias Blume
283    Date: 2002/01/29
284    Tag: blume-20020129-INSTALL
285    Description:
286    
287    - Added somewhat detailed installation instructions (file INSTALL).
288    - Fixed curl-detection bug in config/install.sh.
289    - It is now possible to select the URL getter using the URLGETTER
290      environment variable:
291    
292          not set / "unknown"      --> automatic detection (script tries wget,
293                                       curl, and lynx)
294          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
295                                       how to properly invoke them)
296          other                    --> use $URLGETTER directly, it must take
297                                       precisely two command-line arguments
298                                       (source URL and destination file name)
299    
300    ----------------------------------------------------------------------
301    Name: Matthias Blume
302    Date: 2002/01/28
303    Tag: blume-20020128-sparc-ccalls
304    Description:
305    
306    - Fixed problem with calculation of "used" registers in sparc-c-calls.
307    - Make use of the allocParam argument in sparc-c-calls.
308    
309    ----------------------------------------------------------------------
310    Name: Matthias Blume
311    Date: 2002/01/28
312    Tag: blume-20020128-allocParam
313    Description:
314    
315    John Reppy:  Changes c-calls API to accept client-callback for
316    allocating extra stack space.
317    me: Corresponding changes to mlriscGen (using a dummy argument that
318        does not change the current behavior).
319    
320    ----------------------------------------------------------------------
321    Name: Matthias Blume
322    Date: 2002/01/28 12:00:00
323    Tag: Release_110_38
324    Description:
325    
326    This time for real!!!
327    
328    ----------------------------------------------------------------------
329    Name: Matthias Blume
330    Date: 2002/01/28 10:56:00 EST
331    Tag: blume-20020128-retraction
332    Description:
333    
334    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
335       with blume-Release_110_38-retracted.)
336    
337    1. Fixed a problem with incorrect rounding modes in real64.sml.
338       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
339    
340    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
341       and enums fixed.  The naming of corresponding ML identifiers should
342       now be consistent again.
343    
344    ----------------------------------------------------------------------
345    Name: Allen Leung
346    Date: 2002/01/27
347    Tag: leunga-20020127-nowhere
348    Description:
349    
350       Added a target called nowhere in the configuration scripts.
351       Enabling this will build the MLRISC 'nowhere' tool (for translating
352       programs with where-clauses into legal SML code) during installation.
353    
354    ----------------------------------------------------------------------
355    Name: Matthias Blume
356    Date: 2002/01/25 21:27:00 EST
357    Tag: blume-Release_110_38-retracted
358    Description:
359    
360    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
361    
362    README will be added later.
363    
364    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
365    (Release_110_38) removed.  Reason: Last-minute bug fixes.
366    
367    ----------------------------------------------------------------------
368    Name: Matthias Blume
369    Date: 2002/01/25
370    Tag: blume-20020125-ffi
371    Description:
372    
373    A large number of tweaks and improvements to ml-nlffi-lib and
374    ml-nlffigen:
375    
376       - ML represenation types have been streamlined
377       - getter and setter functions work with concrete values, not abstract
378         ones where possible
379       - ml-nlffigen command line more flexible (see README file there)
380       - some bugs have been fixed (hopefully)
381    
382    ----------------------------------------------------------------------
383    Name: Lal George
384    Date: 2002/01/24
385    Tag: george-20020124-risc-ra-interface
386    Description:
387    
388       There is a dramatic simplification in the interface to the
389       register allocator for RISC architectures as a result of making
390       parallel copy instructions explicit.
391    
392    ----------------------------------------------------------------------
393    Name: Matthias Blume
394    Date: 2002/01/22
395    Tag: blume-20020122-x86-ccalls
396    Description:
397    
398    Bug fix for c-calls on x86 (having to do with how char- and
399    short-arguments are being handled).
400    
401    ----------------------------------------------------------------------
402    Name: Matthias Blume
403    Date: 2002/01/21
404    Tag: blume-20020121-ff
405    Description:
406    
407    Another day of fiddling with the FFI...
408    
409    1. Bug fix/workaround:  CKIT does not complain about negative array
410       dimensions, so ml-nlffigen has to guard itself against this possibility.
411       (Otherwise a negative dimension would send it into an infinite loop.)
412    
413    2. Some of the abstract types (light objects, light pointers, most "base"
414       types) in structure C are now eqtypes.
415    
416    3. Added constructors and test functions for NULL function pointers.
417    
418    ----------------------------------------------------------------------
419    Name: Matthias Blume
420    Date: 2002/01/18
421    Tag: blume-20020118-ready-for-new-release
422    Description:
423    
424    Made config/srcarchiveurl point to a new place.  (Will provide boot
425    files shortly.)
426    
427    Maybe we christen this to be 110.38?
428    
429    ----------------------------------------------------------------------
430    Name: Matthias Blume
431    Date: 2002/01/18
432    Tag: blume-20020118-more-ffifiddle
433    Description:
434    
435    Today's FFI fiddling:
436    
437      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
438        C.Set.* that use concrete (MLRep.*) arguments and results instead
439        of abstract ones.
440    
441      - Provided word-style bit operations etc. for "int" representation
442        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
443        and Long.
444    
445    ----------------------------------------------------------------------
446    Name: Matthias Blume
447    Date: 2002/01/18
448    Tag: blume-20020118-use-x86-fp
449    Description:
450    
451    Now that x86-fast-fp seems to be working, I turned it back on again
452    by default.  (Seems to work fine now, even with the FFI.)
453    
454    Other than that, I added some documentation about the FFI to
455    src/ml-nlffigen/README and updated the FFI test examples in
456    src/ml-nlffi-lib/Tests/*.
457    
458    ----------------------------------------------------------------------
459    Name: Allen Leung
460    Date: 2002/01/17
461    Tag: leunga-20020117-x86-fast-fp-call
462    Description:
463    
464       1. Fixed a problem with handling return fp values when x86's fast fp
465          mode is turned on.
466    
467       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
468          of %f32.
469    
470       3. Added a constructor INT32lit to the ast of MLRISC tools.
471    
472    ----------------------------------------------------------------------
473    Name: Matthias Blume
474    Date: 2002/01/16
475    Tag: blume-20020116-ffifiddle
476    Description:
477    
478    More fiddling with the FFI interface:
479    
480     - Make constness 'c instead of rw wherever possible.  This eliminates
481       the need for certain explicit coercions.  (However, due to ML's
482       value polymorphism, there will still be many cases where explicit
483       coercions are necessary.  Phantom types are not the whole answer
484       to modeling a subtyping relationship in ML.)
485    
486     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
487    
488     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
489       clumsy workaround for varargs problem.)
490    
491    ----------------------------------------------------------------------
492    Name: Lal George
493    Date: 2002/01/15
494    Tag: <none>
495    Description:
496    
497    1. Since COPY instructions are no longer native to the architecture,
498       a generic functor can be used to implement the expandCopies function.
499    
500    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
501       TEXT segment.
502    
503    ----------------------------------------------------------------------
504    Name: Matthias Blume
505    Date: 2002/01/15
506    Tag: blume-20020115-ffiupdates
507    Description:
508    
509    1. Fix for bug resulting in single-precision float values being returned
510       incorrectly from FFI calls.
511    
512    2. Small modifications to C FFI API:
513    
514        - memory-allocation routines return straight objects (no options)
515          and raise an exception in out-of-memory situations
516        - unsafe extensions to cast between function pointers and pointers
517          from/to ints
518        - added structure C_Debug as an alternative to structure C where
519          pointer-dereferencing (|*| and |*!) always check for null-pointers
520        - added open_lib' to DynLinkage;  open_lib' works like open_lib
521          but also takes a (possibly empty) list of existing library handles
522          that the current library depends on
523    
524    ----------------------------------------------------------------------
525    Name: Matthias Blume
526    Date: 2002/01/10
527    Tag: blume-20020110-newffigen
528    Description:
529    
530    1. Updates to portable graph code.
531    
532    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
533       more scalable now so that even huge interfaces such as the one
534       for GTK compile in finite time and space. :-)
535       See src/ml-nlffigen/README for details on what's new.
536    
537    ----------------------------------------------------------------------
538    Name: Lal George
539    Date: 2001/01/09 14:31:35 EST 2002
540    Tag: george-20011206-rm-native-copy
541    Description:
542    
543            Removed the native COPY and FCOPY instructions
544            from all the architectures and replaced it with the
545            explicit COPY instruction from the previous commit.
546    
547            It is now possible to simplify many of the optimizations
548            modules that manipulate copies. This has not been
549            done in this change.
550    
551    ----------------------------------------------------------------------
552    Name: Lal George
553    Date: 2001/12/06 16:50:13 EST 2001
554    Tag: george-20011206-mlrisc-instruction
555    Description:
556    
557    Changed the representation of instructions from being fully abstract
558    to being partially concrete. That is to say:
559    
560      from
561            type instruction
562    
563      to
564            type instr                              (* machine instruction *)
565    
566            datatype instruction =
567                LIVE of {regs: C.cellset, spilled: C.cellset}
568              | KILL of {regs: C.cellset, spilled: C.cellset}
569              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
570              | ANNOTATION of {i: instruction, a: Annotations.annotation}
571              | INSTR of instr
572    
573    This makes the handling of certain special instructions that appear on
574    all architectures easier and uniform.
575    
576    LIVE and KILL say that a list of registers are live or killed at the
577    program point where they appear. No spill code is generated when an
578    element of the 'regs' field is spilled, but the register is moved to
579    the 'spilled' (which is present, more for debugging than anything else).
580    
581    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
582    We used to generate:
583    
584            DEFFREG f1
585            f1 := f2 + f3
586            trapb
587    
588    but now generate:
589    
590            f1 := f2 + f3
591            trapb
592            LIVE {regs=[f1,f2,f3], spilled=[]}
593    
594    Furthermore, the DEFFREG (hack) required that all floating point instruction
595    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
596    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
597    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
598    this architecture).
599    
600    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
601    all the architectures. This will result in further simplification of the
602    register allocator that must be aware of them for coalescing purposes, and
603    will also simplify certain aspects of the machine description that provides
604    callbacks related to parallel copies.
605    
606    ANNOTATION should be obvious, and now INSTR represents the honest to God
607    machine instruction set!
608    
609    The <arch>/instructions/<arch>Instr.sml files define certain utility
610    functions for making porting easier -- essentially converting upper case
611    to lower case. All machine instructions (of type instr) are in upper case,
612    and the lower case form generates an MLRISC instruction. For example on
613    the alpha we have:
614    
615      datatype instr =
616         LDA of {r:cell, b:cell, d:operand}
617       | ...
618    
619      val lda : {r:cell, b:cell, d:operand} -> instruction
620        ...
621    
622    where lda is just (INSTR o LDA), etc.
623    
624    ----------------------------------------------------------------------
625    Name: Matthias Blume
626    Date: 2001/11/22 21:40:00 EST
627    Tag: Release_110_37
628    Description:
629    
630    Release 110.37.  This time for real.
631    
632    ----------------------------------------------------------------------
633    Name: Matthias Blume
634    Date: 2001/11/21 16:35:00 EST
635    Tag: blume-20011121-foot-in-mouth
636    Description:
637    
638    Removed the "Release_110_37" tag because of a serious bug.
639    This will be re-tagged once the bug is fixed.
640    
641    ----------------------------------------------------------------------
642    Name: Matthias Blume
643    Date: 2001/11/21 16:14:00 EST
644    Tag: blume-20011121-forgottenfile
645    Description:
646    
647    Forgot to add a file.  (Just a .tex-file -- part of
648    the CM manual source.)
649    
650    ----------------------------------------------------------------------
651    Name: Matthias Blume
652    Date: 2001/11/21 16:10:00 EST
653    Tag: blume-20011121-invalid_110_37
654    Description:
655    
656    Note: I removed the original tag "Release_110_37" from this commit
657          because we found a serious bug in all non-x86 backends.
658          - Matthias
659    
660    1. Modifications to the SML/NJ code generator and to the runtime system
661       so that code object name strings are directly inserted into code
662       objects at code generation time.  The only business the runtime system
663       has with this is now to read the name strings on occasions.
664       (The encoding of the name string has also changed somewhat.)
665    
666    2. CM now implements a simple "set calculus" for specifying export lists.
667       In particular, it is now possible to refer to the export lists of
668       other libraries/groups/sources and form unions as well as differences.
669       See the latest CM manual for details.
670    
671    3. An separate notion of "proxy" libraries has again be eliminated from
672       CM's model.  (Proxy libraries are now simply a special case of using
673       the export list calculus.)
674    
675    4. Some of the existing libraries now take advantage of the new set
676       calculus.
677       (Notice that not all libraries have been converted because some
678       of the existing .cm-files are supposed to be backward compatible
679       with 110.0.x.)
680    
681    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
682       "General.exnMessage"!  The former relies on a certain hook to be
683       initialized, and that often does not happen in the stand-alone case.)
684    
685    ----------------------------------------------------------------------
686    Name: Lal George
687    Date: 2001/11/21  13:56:18 EST
688    Tag: george-2001121-pseudo-ops
689    Description:
690    
691      Implemented a complete redesign of MLRISC pseudo-ops. Now there
692      ought to never be any question of incompatabilities with
693      pseudo-op syntax expected by host assemblers.
694    
695      For now, only modules supporting GAS syntax are implemented
696      but more should follow, such as MASM, and vendor assembler
697      syntax, e.g. IBM as, Sun as, etc.
698    
699    ----------------------------------------------------------------------
700    Name: Matthias Blume
701    Date: 2001/11/14 11:52:00 EST
702    Tag: blume-20011114-srcname
703    Description:
704    
705    1. Routed the name of the current source file to mlriscgen where it
706       should be directly emitted into the code object.  (This last part
707       is yet to be done.)
708    
709    2. Some cleanup of the pgraph code to make it match the proposal that
710       I put out the other day.  (The proposal notwithstanding, things are
711       still in flux here.)
712    
713    ----------------------------------------------------------------------
714    Name: Lal George
715    Date: 2001/11/14 09:44:04 EST
716    Tag:
717    Description:
718    
719      Fix for a backpatching bug reported by Allen.
720    
721      Because the boundary between short and long span-dependent
722      instructions is +/- 128, there are an astounding number of
723      span-dependent instructions whose size is over estimated.
724    
725      Allen came up with the idea of letting the size of span
726      dependent instructions be non-monotonic, for a maxIter
727      number of times, after which the size must be monotonically
728      increasing.
729    
730      This table shows the number of span-dependent instructions
731      whose size was over-estimated as a function of maxIter, for the
732      file Parse/parse/ml.grm.sml:
733    
734         maxIter            # of instructions:
735            10                      687
736            20                      438
737            30                      198
738            40                        0
739    
740      In compiling the compiler, there is no significant difference in
741      compilation speed between maxIter=10 and maxIter=40. Actually,
742      my measurements showed that maxIter=40 was a tad faster than
743      maxIter=10! Also 96% of the  files in the compiler reach a fix
744      point within 13 iterations, so fixing maxIter at 40, while high,
745      is okay.
746    
747    ----------------------------------------------------------------------
748    Name: Matthias Blume
749    Date: 2001/10/31 15:25:00 EST
750    Tag: blume-20011031-pgraph
751    Description:
752    
753    CKIT:
754    * Changed the "Function" constructor of type Ast.ctype to carry optional
755      argument identifiers.
756    * Changed the return type of TypeUtil.getFunction accordingly.
757    * Type equality ignores the argument names.
758    * TypeUtil.composite tries to preserve argument names but gives up quickly
759      if there is a mismatch.
760    
761    installation script:
762    * attempts to use "curl" if available (unless "wget" is available as well)
763    
764    CM:
765    * has an experimental implementation of "portable graphs" which I will
766      soon propose as an implementation-independent library format
767    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
768    
769    NLFFI-LIB:
770    * some cleanup (all cosmetic)
771    
772    NLFFIGEN:
773    * temporarily disabled the mechanism that suppresses ML output for
774      C definitions whose identifiers start with an underscore character
775    * generate val bindings for enum constants
776    * user can request that only one style (light or heavy) is being used;
777      default is to use both (command-line arguments: -heavy and -light)
778    * fixed bug in handling of function types involving incomplete pointers
779    * generate ML entry points that take record arguments (i.e., using
780      named arguments) for C functions that have a prototype with named
781      arguments
782      (see changes to CKIT)
783    
784    ----------------------------------------------------------------------
785    Name: Allen Leung
786    Date: 2001/10/27 20:34:00 EDT
787    Tag: leunga-20011027-x86-fast-fp-call
788    Description:
789    
790       Fixed the bug described in blume-20010920-slowfp.
791    
792       The fix involves
793          1. generating FCOPYs in FSTP in ia32-svid
794          2. marking a CALL with the appropriate annotation
795    
796    ----------------------------------------------------------------------
797    Name: Matthias Blume
798    Date: 2001/10/16 11:32:00 EDT
799    Tag: blume-20011016-netbsd
800    Description:
801    
802    Underscore patch from Chris Richards (fixing problem with compiling
803    runtime system under recent NetBSD).
804    
805    ----------------------------------------------------------------------
806    Name: Allen Leung
807    Date: 2001/10/12 17:18:32 EDT 2001
808    Tag: leung-20011012-x86-printflowgraph
809    Description:
810    
811    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
812    
813    ----------------------------------------------------------------------
814    Name: Lal George
815    Date: 2001/10/11 23:51:34 EDT
816    Tag: george-20011011-too-many-instrs
817    Description:
818    
819    The representation of a program point never expected to see more
820    than 65536 instructions in a basic block!
821    
822    ----------------------------------------------------------------------
823    Name: Lal George
824    Date: 2001/10/09 09:41:37 EDT
825    Tag: george-20011008-mlrisc-labels
826    Description:
827    
828    Changed the machine description files to support printing of
829    local and global labels in assembly code, based on host assembler
830    conventions.
831    
832    ----------------------------------------------------------------------
833    Name: Matthias Blume
834    Date: 2001/09/25 15:25:00 EDT
835    Tag: blume-20010925-exninfo
836    Description:
837    
838    I provided a non-hook implementation of exnName (at the toplevel) and
839    made the "dummy" implementation of exnMessage (at the toplevel) more
840    useful: if nothing gets "hooked in", then at least you are going to
841    see the exception name and a message indicating why you don't see more.
842    
843    [For the time being, programs that need exnMessage and want to use
844    ml-build should either use General.exnMessage (strongly recommended) or
845    refer to structure General at some other point so that CM sees a
846    static dependency.]
847    
848    [Similar remarks go for "print" and "use":  If you want to use their
849    functionality in stand-alone programs generated by ml-build, then use
850    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
851    
852    ----------------------------------------------------------------------
853    Name: Matthias Blume
854    Date: 2001/09/20 17:28:00 EDT
855    Tag: blume-20010920-slowfp
856    Description:
857    
858    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
859    
860    ----------------------------------------------------------------------
861    Name: Matthias Blume
862    Date: 2001/09/20 17:20:00 EDT
863    Tag: blume-20010920-canonicalpaths
864    Description:
865    
866    0. Updated the BOOT file (something that I forgot to do earlier).
867    
868    1. Small internal change to CM so that it avoids "/../" in filenames
869       as much as possible (but only where it is safe).
870    
871    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
872       that arguments that contain delimiters are passed through correctly.
873       This change also means that all "special" arguments of the form
874       @SMLxxx... must come first.
875    
876    3. Changed install script to put relative anchor names for tool commands
877       into pathconfig.
878    
879    ----------------------------------------------------------------------
880    Name: Matthias Blume
881    Date: 2001/09/18 15:35:00 EDT
882    Tag: blume-20010918-readme11036
883    Description:
884    
885    Added README files.
886    
887    ----------------------------------------------------------------------
888    Name: Matthias Blume
889    Date: 2001/09/18 11:45:00 EDT
890    Tag: Release_110_36 (retag)
891    Description:
892    
893    Fixed mistake in config/preloads. Retagged as 110.36.
894    
895    ----------------------------------------------------------------------
896    Name: Matthias Blume
897    Date: 2001/09/18 09:40:00 EDT
898    Tag: Release_110_36_orig (tag changed)
899    Description:
900    
901    New version (110.36).  New bootfiles.
902    
903    ----------------------------------------------------------------------
904    Name: Matthias Blume
905    Date: 2001/09/14 16:15:00 EDT
906    Tag: blume-20010914-x86fastfp
907    Description:
908    
909    John committed some changes that Allen made, in particular a (hopefully)
910    correctly working version of the x86-fp module.
911    
912    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
913    flag to "true".  Everything seems to compile to a fixpoint ok, and
914    "mandelbrot" speeds up by about 15%.
915    
916    ----------------------------------------------------------------------
917    Name: Matthias Blume
918    Date: 2001/09/13 11:20:00 EDT
919    Tag: blume-20010913-minimal
920    Description:
921    
922    1. Stefan Monnier's patch to fix a miscompilation problem that
923       was brought to light by John Reppy's work on Moby.
924    
925    2. Implemented a minimal "structure Compiler" that contains just
926       "version" and "architecture".  The minimal version will be
927       available when the full version is not.  This is for backward-
928       compatibility with code that wants to test Compiler.version.
929    
930    ----------------------------------------------------------------------
931    Name: Matthias Blume
932    Date: 2001/08/28 14:03:00 EDT
933    Tag: blume-20010828-ml-lex
934    Description:
935    
936    Fix for bug 1581, received from Neophytos Michael.
937    
938    ----------------------------------------------------------------------
939    Name: Matthias Blume
940    Date: 2001/08/27 11:20:00 EDT
941    Tag: blume-20010827-readme11035
942    Description:
943    
944    Fleshed out the README file for 110.35.
945    
946    ----------------------------------------------------------------------
947    Name: Matthias Blume
948    Date: 2001/08/24 17:10:00 EDT
949    Tag: Release_110_35
950    Description:
951    
952    New version number (110.35).  New bootfiles.
953    
954    ----------------------------------------------------------------------
955    Name: Lal George
956    Date: 2001/08/24 13:47:18 EDT 2001
957    Tag: george-20010824-MLRISC-graphs
958    Description:
959    
960     removed clusters from MLRISC completely and replaced with graphs.
961    
962    ----------------------------------------------------------------------
963    Name: Matthias Blume
964    Date: 2001/08/23 17:50:00 EDT
965    Tag: blume-20010823-toplevel
966    Description:
967    
968    - some reorganization of the code that implements various kinds of
969      environments in the compiler (static, dynamic, symbolic, combined)
970    - re-implemented the EnvRef module so that evalStream works properly
971      (if the stream contains references to "use", "CM.make", etc.)
972    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
973    
974    ----------------------------------------------------------------------
975    Name: Matthias Blume
976    Date: 2001/08/20 15:50 EDT
977    Tag: blume20010820-slipup
978    Description:
979    
980    I forgot to commit a few files.  Here they are...
981    
982    ----------------------------------------------------------------------
983    Name: Matthias Blume
984    Date: 2001/08/20 15:35:00 EDT
985    Tag: blume-20010820-debugprof
986    Description:
987    
988    !!!! NEW BOOTFILES !!!!
989    
990    This is another round of reorganizing the compiler sources.  This
991    time the main goal was to factor out all the "instrumentation"
992    passes (for profiling and backtracing) into their own library.
993    The difficulty was to do it in such a way that it does not depend
994    on elaborate.cm but only on elabdata.cm.
995    
996    Therefore there have been further changes to both elaborate.cm and
997    elabdata.cm -- more "generic" things have been moved from the former
998    to the latter.  As a result, I was forced to split the assignment
999    of numbers indicating "primtyc"s into two portions: SML-generic and
1000    SML/NJ-specific.  Since it would have been awkward to maintain,
1001    I bit the bullet and actually _changed_ the mapping between these
1002    numbers and primtycs.  The bottom line of this is that you need
1003    a new set of bin- and bootfiles.
1004    
1005    I have built new bootfiles for all architectures, so doing a fresh
1006    checkout and config/install.sh should be all you need.
1007    
1008    The newly created library's name is
1009    
1010        $smlnj/viscomp/debugprof.cm
1011    
1012    and its sources live under
1013    
1014        src/compiler/DebugProf
1015    
1016    ----------------------------------------------------------------------
1017    Name: Matthias Blume
1018    Date: 2001/08/15 17:15:00 EDT
1019    Tag: blume-20010815-compreorg
1020    Description:
1021    
1022    This is a first cut at reorganizing the CM libraries that make up the
1023    core of the compiler.  The idea is to separate out pieces that could
1024    be used independently by tools, e.g., the parser, the typechecker, etc.
1025    
1026    The current status is a step in this direction, but it is not quite
1027    satisfactory yet.  Expect more changes in the future.
1028    
1029    Here is the current (new) organization...
1030    
1031        What used to be $smlnj/viscomp/core.cm is now divided into
1032        six CM libraries:
1033    
1034             $smlnj/viscomp/basics.cm
1035                           /parser.cm
1036                           /elabdata.cm
1037                           /elaborate.cm
1038                           /execute.cm
1039                           /core.cm
1040    
1041        The CM files for these libraries live under src/system/smlnj/viscomp.
1042        All these libraries are proxy libraries that contain precisely
1043        one CM library component.  Here are the locations of the components
1044        (all within the src/compiler tree):
1045    
1046             Basics/basics.cm
1047             Parse/parser.cm
1048             ElabData/elabdata.cm
1049             Elaborator/elaborate.cm
1050             Execution/execute.cm
1051             core.cm
1052    
1053         [This organization is the same that has been used already
1054         for a while for the architecture-specific parts of the visible
1055         compiler and for the old version of core.cm.]
1056    
1057         As you will notice, many source files have been moved from their
1058         respective original locations to a new home in one of the above
1059         subtrees.
1060    
1061         The division of labor between the new libraries is the following:
1062    
1063             basics.cm:
1064                - Simple, basic definitions that pertain to many (or all) of
1065                  the other libraries.
1066             parser.cm:
1067                - The SML parser, producing output of type Ast.dec.
1068                - The type family for Ast is also defined and exported here.
1069             elabdata.cm:
1070                - The datatypes that describe input and output of the elaborator.
1071                  This includes types, absyn, and static environments.
1072             elaborator.cm:
1073                - The SML/NJ type checker and elaborator.
1074                  This maps an Ast.dec (with a given static environment) to
1075                  an Absyn.dec (with a new static environment).
1076                - This libraries implements certain modules that used to be
1077                  structures as functors (to remove dependencies on FLINT).
1078             execute.cm:
1079                - Everything having to do with executing binary code objects.
1080                - Dynamic environments.
1081             core.cm:
1082                - SML/NJ-specific instantiations of the elaborator and MLRISC.
1083                - Top-level modules.
1084                - FLINT (this should eventually become its own library)
1085    
1086    Notes:
1087    
1088    I am not 100% happy with the way I separated the elaborator (and its
1089    data structures) from FLINT.  Two instances of the same problem:
1090    
1091        1. Data structures contain certain fields that carry FLINT-specific
1092           information.  I hacked around this using exn and the property list
1093           module from smlnj-lib.  But the fact that there are middle-end
1094           specific fields around at all is a bit annoying.
1095    
1096        2. The elaborator calculates certain FLINT-related information.  I tried
1097           to make this as abstract as I could using functorization, but, again,
1098           the fact that the elaborator has to perform calculations on behalf
1099           of the middle-end at all is not nice.
1100    
1101        3. Having to used exn and property lists is unfortunate because it
1102           weakens type checking.  The other alternative (parameterizing
1103           nearly *everything*) is not appealing, though.
1104    
1105    I removed the "rebinding =" warning hack because due to the new organization
1106    it was awkward to maintain it.  As a result, the compiler now issues some of
1107    these warnings when compiling init.cmi during bootstrap compilation. On
1108    the plus side, you also get a warning when you do, for example:
1109       val op = = Int32.+
1110    which was not the case up to now.
1111    
1112    I placed "assign" and "deref" into the _Core structure so that the
1113    code that deals with the "lazy" keyword can find them there.  This
1114    removes the need for having access to the primitive environment
1115    during elaboration.
1116    
1117    ----------------------------------------------------------------------
1118    Name: Matthias Blume
1119    Date: 2001/08/13
1120    Tag: blume-20010813-closures
1121    Description:
1122    
1123    This fix was sent to us by Zhong Shao.  It is supposed to improve the
1124    performance of certain loops by avoiding needless closure allocation.
1125    
1126    ----------------------------------------------------------------------
1127    Name: Lal George
1128    Date: 2001/07/31 10:03:23 EDT 2001
1129    Tag: george-20010731-x86-fmalloc
1130    Description: Fixed bug in x86 calls
1131    
1132        There was a bug where call instructions would mysteriously
1133        vanish. The call instruction had to be one that returned
1134        a floating point value.
1135    
1136    ----------------------------------------------------------------------
1137    Name: Lal George
1138    Date: 2001/07/19 16:36:29 EDT 2001
1139    Tag: george-20010719-simple-cells
1140    Description:
1141    
1142    I have dramatically simplified the interface for CELLS in MLRISC.
1143    
1144    In summary, the cells interface is broken up into three parts:
1145    
1146      1. CellsBasis : CELLS_BASIS
1147    
1148            CellsBasis is a top level structure and common for all
1149            architectures.  it contains the definitions of basic datatypes
1150            and utility  functions over these types.
1151    
1152      2. functor Cells() : CELLS
1153    
1154            Cells generates an interface for CELLS that incorporates the
1155            specific resources on the target architecture, such as the
1156            presence of special register classes, their number and size,
1157            and various useful substructures.
1158    
1159      3. <ARCH>CELLS
1160    
1161            e.g. SparcCells: SPARCCELLS
1162    
1163            <ARCH>CELLS usually contains additional bindings for special
1164            registers  on the architecture, such as:
1165    
1166                    val r0 : cell           (* register zero *)
1167                    val y : cell            (* Y register *)
1168                    val psr : cell          (* processor status register *)
1169                    ...
1170    
1171            The structure returned by applying the Cells functor is opened
1172            in this interface.
1173    
1174    The main implication of all this is that the datatypes for cells is
1175    split between CellsBasis and CELLS -- a fairly simple change for user
1176    code.
1177    
1178    In the old scheme the CELLS interface had a definitional binding of
1179    the form:
1180    
1181            signature CELLS = sig
1182    
1183               structure CellsBasis = CellsBasis
1184    
1185               ...
1186    
1187            end
1188    
1189    With all the sharing constraints that goes on in MLRISC, this old
1190    design  quickly leads to errors such as:
1191    
1192            "structure definition spec inside of sharing ... "
1193    
1194    
1195    and appears to require an unacceptable amount of sharing and where
1196    constraint hackery.
1197    
1198    I think this error message (the interaction of definitional specs and
1199    sharing) requires more explanation on our web page.
1200    
1201    ----------------------------------------------------------------------
1202    Name: Matthias Blume
1203    Date: 2001/07/19 15:00:00 EDT
1204    Tag: blume-20010719-libreorg
1205    Description:
1206    
1207    This update puts together a fairly extensive but straightforward change
1208    to the way the libraries that implement the interactive system are
1209    organized:
1210    
1211       The biggest change is the elimination of structure Compiler.  As a
1212       replacement for this structure, there is now a CM library
1213       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1214       that exports all the substructures of the original structure Compiler
1215       directly.  So instead of saying Compiler.Foo.bar one now simply
1216       says Foo.bar.  (The CM libraries actually export a collection of
1217       structures that is richer than the collection of substructures of
1218       structure Compiler.)
1219    
1220       To make the transition smooth, there is a separate library called
1221       $smlnj/compiler/compiler.cm which puts together and exports the
1222       original structure Compiler (or at least something very close to it).
1223    
1224       There are five members of the original structure Compiler
1225       that are not exported directly but which instead became members
1226       of a new structure Backend (described by signature BACKEND).  These are:
1227       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1228       Interact (: INTERACT), structure Machine (: MACHINE), and val
1229       architecture (: string).
1230    
1231       Structure Compiler.Version has become structure CompilerVersion.
1232    
1233       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1234       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1235       or x86, respectively.
1236       Each of these exports the same frontend structures that
1237       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1238       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1239       Hppa, PPC, Sparc, or X86, respectively.
1240    
1241       Library $smlnj/compiler/all.cm exports the union of the exports of
1242       $smlnj/compiler/<arch>.cm
1243    
1244       There are no structures <Arch>Compiler anymore, use
1245       $smlnj/compiler/<arch>.cm instead.
1246    
1247       Library host-compiler-0.cm is gone.  Instead, the internal library
1248       that instantiates CM is now called cm0.cm.  Selection of the host
1249       compiler (backend) is no longer done here but. (Responsibility for it
1250       now lies with $smlnj/compiler/current.cm.  This seems to be more
1251       logical.)
1252    
1253       Many individual files have been moved or renamed.  Some files have
1254       been split into multiple files, and some "dead" files have been deleted.
1255    
1256    Aside from these changes to library organization, there are also changes
1257    to the way the code itself is organized:
1258    
1259       Structure Binfile has been re-implemented in such a way that it no
1260       longer needs any knowledge of the compiler.  It exclusively deals
1261       with the details of binfile layout.  It no longer invokes the
1262       compiler (for the purpose of creating new prospective binfile
1263       content), and it no longer has any knowledge of how to interpret
1264       pickles.
1265    
1266       Structure Compile (: COMPILE) has been stripped down to the bare
1267       essentials of compilation.  It no longer deals with linking/execution.
1268       The interface has been cleaned up considerably.
1269    
1270       Utility routines for dealing with linking and execution have been
1271       moved into their own substructures.
1272    
1273       (The ultimate goal of these changes is to provide a light-weight
1274       binfile loader/linker (at least for, e.g., stable libraries) that
1275       does not require CM or the compiler to be present.)
1276    
1277    CM documentation has been updated to reflect the changes to library
1278    organization.
1279    
1280    ----------------------------------------------------------------------
1281    Name: Matthias Blume
1282    Date: 2001/07/10 17:30:00 EDT
1283    Tag: Release_110_34
1284    Description:
1285    
1286    Minor tweak to 110.34 (re-tagged):
1287    
1288      - README.html file added to CVS repository
1289      - runtime compiles properly under FreeBSD 3.X and 4.X
1290    
1291    ----------------------------------------------------------------------
1292    Name: Matthias Blume
1293    Date: 2001/07/10 17:30:00 EDT
1294    Tag: Release_110_34
1295    Description:
1296    
1297    New version number (110.34). New bootfiles.
1298    
1299    ----------------------------------------------------------------------
1300    Name: Matthias Blume
1301    Date: 2001/07/09 16:00:00 EDT
1302    Tag: blume-20010709-more-varargs
1303    Description:
1304    
1305    I changed the handling of varargs in ml-nlffigen again:
1306    The ellipsis ... will now simply be ignored (with an accompanying warning).
1307    
1308    The immediate effect is that you can actually call a varargs function
1309    from ML -- but you can't actually supply any arguments beyond the ones
1310    specified explicitly.  (For example, you can call printf with its format
1311    string, but you cannot pass additional arguments.)
1312    
1313    This behavior is only marginally more useful than the one before, but
1314    it has the advantage that a function or, more importantly, a function
1315    type never gets dropped on the floor, thus avoiding follow-up problems with
1316    other types that refer to the offending one.
1317    
1318    ----------------------------------------------------------------------
1319    Name: Matthias Blume
1320    Date: 2001/07/09 11:25:00 EDT
1321    Tag: blume-20010709-varargs
1322    Description:
1323    
1324    1. ckit-lib.cm now exports structure Error
1325    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1326       with a warning accompanied by a source location.  Moreover, it
1327       merely skips the offending function or type and proceeds with the
1328       rest of its work.u  As a result, one can safely feed C code containing
1329       "..." to ml-nlffigen.
1330    3. There are some internal improvements to CM, providing slightly
1331       more general string substitutions in the tools subsystem.
1332    
1333    ----------------------------------------------------------------------
1334    Name: Matthias Blume
1335    Date: 2001/06/27 15:10:00 EDT
1336    Tag: blume-20010627-concur
1337    Description:
1338    
1339    Fixed a small bug in CM's handling of parallel compilation.
1340    (You could observe the bug by Control-C-interrupting an ordinary
1341    CMB.make or CM.stabilize and then attaching some compile servers.
1342    The result was that all of a sudden the previously interrupted
1343    compilation would continue on its own.  This was because of
1344    an over-optimization: CM did not bother to clean out certain queues
1345    when no servers were attached "anyway", resulting in the contents
1346    of these queues to grab control when new servers did get attached.)
1347    
1348    There is also another minor update to the CM manual.
1349    
1350    ----------------------------------------------------------------------
1351    Name: Matthias Blume
1352    Date: 2001/06/26 16:15:00 EDT
1353    Tag: blume-20010626-cmdoc
1354    Description:
1355    
1356    Minor typo fixed in CM manual (syntax diagram for libraries).
1357    
1358    ----------------------------------------------------------------------
1359    Name: Matthias Blume
1360    Date: 2001/06/25 22:55:00 EDT
1361    Tag: blume-20010625-x86pc
1362    Description:
1363    
1364    Fixed a nasty bug in the X86 assembly code that caused signal
1365    handlers to fail (crash) randomly.
1366    
1367    ----------------------------------------------------------------------
1368    Name: Matthias Blume
1369    Date: 2001/06/25 12:05:00 EDT
1370    Tag: blume-20010625-nlffigen
1371    Description:
1372    
1373    This update fixes a number of minor bugs in ml-nlffigen as reported by
1374    Nick Carter <nbc@andrew.cmu.edu>.
1375    
1376      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1377      2. Default names for generated files are now derived from the name of
1378         the C file *without its directory*.  In particular, this causes generated
1379         files to be placed locally even if the C file is in some system directory.
1380      3. Default names for generated signatures and structures are also derived
1381         from the C file name without its directory.  This avoids silly things
1382         like "structure GL/GL".
1383         (Other silly names are still possible because ml-nlffigen does not do
1384          a thorough check of whether generated names are legal ML identifiers.
1385          When in doubt, use command line arguments to force particular names.)
1386    
1387    ----------------------------------------------------------------------
1388    Name: Matthias Blume
1389    Date: 2001/06/21 12:25:00 EDT
1390    Tag: blume-20010621-eXene
1391    Description:
1392    
1393    eXene now compiles and (sort of) works again.
1394    
1395    The library name (for version > 110.33) is $/eXene.cm.
1396    
1397    I also added an new example in src/eXene/examples/nbody.  See the
1398    README file there for details.
1399    
1400    ----------------------------------------------------------------------
1401    Name: Matthias Blume
1402    Date: 2001/06/20 16:40:00 EDT
1403    Tag: blume-20010620-cml
1404    Description:
1405    
1406    CML now compiles and works again.
1407    
1408    Libraries (for version > 110.33):
1409    
1410      $cml/cml.cm            Main CML library.
1411      $cml/basis.cm          CML's version of $/basis.cm.
1412      $cml/cml-internal.cm   Internal helper library.
1413      $cml/core-cml.cm       Internal helper library.
1414      $cml-lib/trace-cml.cm  Tracing facility.
1415      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1416    
1417    The installer (config/install.sh) has been taught how to properly
1418    install this stuff.
1419    
1420    ----------------------------------------------------------------------
1421    Name: Matthias Blume
1422    Date: 2001/06/19 17:55:00 EDT
1423    Tag: blume-20010619-instantiate
1424    Description:
1425    
1426    This un-breaks the fix for bug 1432.
1427    (The bug was originally fixed in 110.9 but I broke it again some
1428    time after that.)
1429    
1430    ----------------------------------------------------------------------
1431    Name: Matthias Blume
1432    Date: 2001/06/19 17:25:00 EDT
1433    Tag: blume-20010619-signals
1434    Description:
1435    
1436    This should (hopefully) fix the long-standing signal handling bug.
1437    (The runtime system was constructing a continuation record with an
1438    incorrect descriptor which would cause the GC to drop data on the floor...)
1439    
1440    ----------------------------------------------------------------------
1441    Name: Matthias Blume
1442    Date: 2001/06/15 15:05:00 EDT
1443    Tag: blume-20010615-moresparc
1444    Description:
1445    
1446    Here is a short late-hour update related to Sparc c-calls:
1447    
1448     -- made handling of double-word arguments a bit smarter
1449    
1450     -- instruction selection phase tries to collapse certain clumsily
1451        constructed ML-Trees; typical example:
1452    
1453            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1454    
1455        This currently has no further impact on SML/NJ since mlriscGen does
1456        not seem to generate such patterns in the first place, and c-calls
1457        (which did generate them in the beginning) has meanwhile been fixed
1458        so as to avoid them as well.
1459    
1460    ----------------------------------------------------------------------
1461    Name: Matthias Blume
1462    Date: 2001/06/15 15:05:00 EDT
1463    Tag: blume-20010615-sparc
1464    Description:
1465    
1466    The purpose of this update is to provide an implementation of NLFFI
1467    on Sparc machines.
1468    
1469    Here are the changes in detail:
1470    
1471       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1472       the Sparc implementation of the c-calls API.
1473       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1474       for accessing the ML frame.  Thus, we have a real frame pointer and
1475       can freely modify %sp without need for an omit-frame-ptr phase.
1476       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1477       this case.
1478       * ml-nlffigen has been taught to produce code for different architectures
1479       and calling conventions.
1480       * In a way similar to what was done in the x86 case, the Sparc
1481       backend uses its own specific extension to mltree.  (For example,
1482       it needs to be able to generate UNIMP instructions which are part
1483       of the calling convention.)
1484       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1485       to make it easier to plug in new machine- and os-dependent parts).
1486    
1487    There are some other fairly unrelated bug fixes and cleanups as well:
1488    
1489       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1490       that they properly share their libraries with existing SML/NJ libraries.
1491       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1492       follow-up error messages.
1493       * Updates to CM/CMB documentation.
1494    
1495    TODO items:
1496    
1497       * MLRISC should use a different register as its asmTemp on the Sparc.
1498         (The current %o2 is a really bad choice because it is part of the
1499         calling conventions, so things might interfere in unexpected ways.)
1500    
1501    ----------------------------------------------------------------------
1502    Name: Matthias Blume
1503    Date: 2001/06/07
1504    Tag: blume-20010607-calls
1505    Description:
1506    
1507    A number of internal changes related to C calls and calling conventions:
1508    
1509    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1510       number of bytes popped implicitly (by the callee).  In most cases
1511       this field is 0 but on x86/win32 it is some non-zero value.  This
1512       is information provided for the benefit of the "omit-frameptr" pass.
1513    2. The CALL instruction on the x86 carries a similar "pops" field.
1514       The instruction selection phase copies its value from the ML-Tree
1515       CALL statement.
1516    3. On all other architectures, the instruction selection phase checks
1517       whether "pops=0" and complains if not.
1518    4. The c-calls implementation for x86 now accepts two calling conventions:
1519       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1520       up after the call and pops is set to 0.  For "stdcall", the caller
1521       does nothing, leaving the cleanup to the callee; pops is set to
1522       the number of bytes that were pushed onto the stack.
1523    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1524       distinguish between "ccall" and "stdcall".
1525    6. The UNIMP instruction has been added to the supported Sparc instruction
1526       set. (This is needed for implementing the official C calling convention
1527       on this architecture.)
1528    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1529       work with the latest CM.
1530    
1531    ----------------------------------------------------------------------
1532    Name: Matthias Blume
1533    Date: 2001/06/05 15:10:00 EDT
1534    Tag: blume-20010605-cm-index
1535    Description:
1536    
1537    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1538    
1539    1. CM can now generate "index files".  These are human-readable files
1540       that list on a per-.cm-file basis each toplevel symbol defined or
1541       imported.  The location of the index file for
1542       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1543       To enable index-file generation, set CM.Control.generate_index to true
1544       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1545    
1546       The CM manual has been updated accordingly.
1547    
1548    2. I made some slight modifications to the c-calls API in MLRISC.
1549    
1550         a) There is now a callback to support saving/restoring of
1551            dedicated but caller-save registers around the actual call
1552            instruction.
1553         b) One can optionally specify a comment-annotation for the
1554            call instruction.
1555    
1556    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1557       (For example, the comment annotation shows the C prototype of
1558        the function being called.)
1559    
1560    ----------------------------------------------------------------------
1561    Name: Matthias Blume
1562    Date: 2001/06/01 13:30:00 EDT
1563    Tag: blume-20010601-nlffi-cleanup
1564    Description:
1565    
1566    This is mostly a cleanup of MLFFI stuff:
1567    
1568       - some signature files have been put into a more exposed place
1569       - the ugly 'f type parameter is gone (simplifies types tremendously!)
1570       - ml-nlffigen changed accordingly
1571       - tutorial updated
1572    
1573    Other changes:
1574    
1575       - author's affiliation in CM manual(s) updated
1576       - some more recognized keywords added to Allen's sml.sty
1577    
1578    ----------------------------------------------------------------------
1579    Name: Matthias Blume
1580    Date: 2001/05/25 15:30:00 EDT
1581    Tag: blume-20010525-iptr
1582    Description:
1583    
1584      - put the official 110.33-README (as it appears on the ftp server) under
1585        CVS
1586      - fixed a small bug related to incomplete pointer types in
1587        ml-nlffigen
1588      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1589        (it does not need the 'f type parameter)
1590    
1591    ----------------------------------------------------------------------
1592    Name: Matthias Blume
1593    Date: 2001/05/23 14:30:00 EDT
1594    Tag: Release_110_33
1595    Description:
1596    
1597    New version number (110.33).  New bootfiles.
1598    
1599    ----------------------------------------------------------------------
1600    Name: Matthias Blume
1601    Date: 2001/05/22 18:06:00 EDT
1602    Tag: blume-20010522-targets
1603    Description:
1604    
1605    Made install.sh use file config/targets.customized if it exists, falling
1606    back to config/targets if it doesn't.  This way one can have a customized
1607    version of the targets file without touching the "real thing", thus
1608    eliminating the constant fear of accidentally checking something bogus
1609    back into the CVS repository...  (File config/targets.customized must
1610    not be added to the repository!)
1611    
1612    ----------------------------------------------------------------------
1613    Name: Matthias Blume
1614    Date: 2001/05/22 16:30:00 EDT
1615    Tag: blume-20010522-minitut
1616    Description:
1617    
1618    1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
1619       struct returns.
1620    2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
1621       incomplete, preliminary documentation for NLFFI.
1622    
1623    ----------------------------------------------------------------------
1624    Name: Matthias Blume
1625    Date: 2001/05/14 11:30:00 EDT
1626    Tag: blume-20010514-script
1627    Description:
1628    
1629    Some bugs in install script fixed.
1630    
1631    In addition to that I also made a slight change to the NLFFI API:
1632    Functors generated by ml-nlffigen now take the dynamic library as a
1633    straight functor argument, not as a suspended one.  (The original
1634    functor code used to force the suspension right away anyway, so there
1635    was nothing gained by this complication of the interface.)
1636    
1637    ----------------------------------------------------------------------
1638    Name: Matthias Blume
1639    Date: 2001/05/11 14:35:00 EDT
1640    Tag: blume-20010511-ml-nlffi
1641    Description:
1642    
1643    I finally took the plunge and added my new FFI code to the main
1644    repository.  For x86-linux it is now ready for prime-time.
1645    
1646    There are two new subdirectories of "src":
1647    
1648      - ml-nlffi-lib:
1649           The utility library for programs using the FFI interface.
1650           Here is the implementation of $/c.cm and its associated low-level
1651           partners $/c-int.cm and $/memory.cm.
1652      - ml-nlffigen:
1653           A stand-alone program for generating ML glue code from C source
1654           code.
1655    
1656    Building ml-nlffigen requires $/ckit-lib.cm.
1657    
1658    The config/install.sh script has been updates to do the Right Thing
1659    (hopefully).
1660    
1661    Notice that the source tree for the C-Kit will not be put under "src"
1662    but directly under the installation root directory.  (This is the
1663    structure that currently exists on the CVS server when you check out
1664    module "sml".)  Fortunately, config/install.sh knows about this oddity.
1665    
1666    Bugs: No documentation yet.
1667    
1668    ----------------------------------------------------------------------
1669    Name: Matthias Blume
1670    Date: 2001/05/09 16:35:00 EDT
1671    Tag: blume-20010509-cpscontract
1672    Description:
1673    
1674    Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
1675    wrapper/unwrapper elimination did not decrement usage counts and some
1676    dead variables got overlooked by the dead-up logic.)
1677    
1678    ----------------------------------------------------------------------
1679    Name: Lal George
1680    Date: 2001/05/08  17:26:09 EDT
1681    Tag: george-20010508-omit-frameptr
1682    Description:
1683    
1684    Changes to implement the omit-frame-pointer optimization to support
1685    raw C calls. For now, there is only support on the Intel x86, but
1686    other architectures will follow as more experience is gained with this.
1687    
1688    
1689    ----------------------------------------------------------------------
1690    Name: Matthias Blume
1691    Date: 2001/05/07 14:40:00 EDT
1692    Tag: blume-20010507-proxies
1693    Description:
1694    
1695    I made into "proxy libraries" all libraries that qualify for such a
1696    change.  (A qualifying library is a library that has another library or
1697    groups as its sole member and repeats that member's export list
1698    verbatim.  A proxy library avoids this repetition by omitting its export
1699    list, effectively inheriting the list that its (only) member exports.
1700    See the CM manual for more explanation.)
1701    The main effect is that explicit export lists for these libraries
1702    do not have to be kepts in sync, making maintenance a bit easier.
1703    
1704    I also added copyright notices to many .cm-files.
1705    
1706    Last but not least, I made a new set of bootfiles.
1707    
1708    ----------------------------------------------------------------------
1709    Name: Matthias Blume
1710    Date: 2001/05/04 17:00:00 EDT
1711    Tag: blume-20010504-cm-lsplit
1712    Description:
1713    
1714    0. John merged pending changes to $/smlnj-lib.cm
1715    
1716    1. Allen's previous change accidentally backed out of one of Lal's
1717       earlier changes.  I undid this mistake (re-introducing Lal's change).
1718    
1719    2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
1720       within the compiler where applicable.  There is some code simplification
1721       because of that.
1722    
1723    3. The "split" phase (in FLINT) is now part of the default list of phases.
1724       Compiler.Control.LambdaSplitting.* can be used to globally control the
1725       lambda-splitting (cross-module-inlining) engine.  In addition to that,
1726       it can now also be controlled on a per-source basis: CM has been taught
1727       a new tool parameter applicable to ML source files.
1728    
1729       - To turn lambda-splitting off completely:
1730            local open Compiler.Control.LambdaSplitting in
1731                val _ = set Off
1732            end
1733       - To make "no lambda-splitting" the global default (but allow per-source
1734         overriding); this is the initial setting:
1735            local open Compiler.Control.LambdaSplitting in
1736                val _ = set (Default NONE)
1737            end
1738       - To make "lambda-splitting with aggressiveness a" the global default
1739         (and allow per-source overriding):
1740            local open Compiler.Control.LambdaSplitting in
1741                val _ = set (Default (SOME a))
1742            end
1743    
1744       - To turn lambda-splitting off for a given ML souce file (say: a.sml)
1745         write (in the respective .cm-file):
1746            a.sml (lambdasplitting:off)
1747       - To turn lambda-splitting for a.sml on with minimal aggressiveness:
1748            a.sml (lambdasplitting:on)
1749       - To turn lambda-splitting for a.sml on with aggressiveness <a> (where
1750         <a> is a decimal non-negative integer):
1751            a.sml (lambdasplitting:<a>)
1752       - To turn lambda-splitting for a.sml on with maximal aggressiveness:
1753            a.sml (lambdasplitting:infinity)
1754       - To use the global default for a.sml:
1755            a.sml (lambdasplitting:default)
1756         or simply
1757            a.sml
1758    
1759    ----------------------------------------------------------------------
1760    Name: Allen Leung
1761    Date: 2001/05/04 01:57:00 EDT
1762    Tag: leunga-20010504-sync
1763    Description:
1764    
1765      MLRISC features.
1766    
1767      1. Fix to CMPXCHG instructions.
1768      2. Changed RA interface to allow annotations in callbacks.
1769      3. Added a new method to the stream interface to allow annotations updates.
1770    
1771    ----------------------------------------------------------------------
1772    Name: Matthias Blume
1773    Date: 2001/05/01 11:45:00 EDT
1774    Tag: blume-20010501-pcedittmp
1775    Description:
1776    
1777    Changed install.sh to use the current working directory instead of
1778    /usr/tmp for a temporary file (pcedittmp).  The previous choice
1779    of /usr/tmp caused trouble with MacOS X because of file premission
1780    problems.
1781    
1782    ----------------------------------------------------------------------
1783    Name: Matthias Blume
1784    Date: 2001/04/20 11:10:00 EDT
1785    Tag: blume-20010420-inMLflag
1786    Description:
1787    
1788     - added vp_limitPtrMask to vproc-state.h
1789       (for use by the raw-C-calls mechanism to implement proper interrupt
1790        handling)
1791     - made the ML compiler aware of various data-structure offsets so it
1792       can generate code for accessing the vp_inML flag and vp_limitPtrMask
1793     - tweaked mlriscGen.sml to have it emit interrupt-handling code for
1794       raw C-calls
1795    
1796    ----------------------------------------------------------------------
1797    Name: Lal George
1798    Date: 2001/04/20 09:15:28 EDT
1799    Tag: george-20010420-macosX
1800    Description:
1801    
1802     - Changes to port to Mac OS X; Darwin.
1803    
1804     - In the process I found that sqrt was broken on the PPC, because the
1805       fsqrt instruction is not implemented.
1806    
1807    ----------------------------------------------------------------------
1808    Name: Matthias Blume
1809    Date: 2001/04/18 12:45:00 EDT
1810    Tag: blume-20010418-ccalls
1811    Description:
1812    
1813     - fixed two off-by-4 errors in the x86-specific c-calls implementation
1814       (this bug prevented structure arguments containing pointers from being
1815        passed correctly)
1816     - changed the raw-C-call code in mlriscGen.sml in such a way that
1817       structure arguments are represented as a pointer to the beginning
1818       of the structure (instead of having a series of synthesized arguments,
1819       one for each structure member)
1820    
1821     - made makeml script's verbosity level configurable via environment
1822       variable (MAKEML_VERBOSITY)
1823    
1824     - eliminated placeholder implementations for f32l, w16s, i16s, and f32s
1825       in rawmem-x86.sml; we are now using the real thing
1826    
1827    ----------------------------------------------------------------------
1828    Name: Matthias Blume
1829    Date: 2001/03/22 16:25:00 EST
1830    Tag: blume-20010322-bootfiles
1831    Description:
1832    
1833    Created a new set of bootfiles (for your automatic installation convenience).
1834    
1835    ----------------------------------------------------------------------
1836    Name: Matthias Blume
1837    Date: 2001/03/22 15:10:00 EST
1838    Tag: blume-20010322-rawmem-parcm
1839    Description:
1840    
1841    1. All "raw memory access" primitives for the new FFI are implemented now
1842       (at least on the x86).
1843    2. Some further cleanup of CM's parallel make mechanism.
1844    
1845    ----------------------------------------------------------------------
1846    Name: Matthias Blume
1847    Date: 2001/03/19 17:53:00 EST
1848    Tag: blume-20010319-parallel
1849    Description:
1850    
1851    Parallel make (using compile servers) now works again.
1852    
1853    To this end, CM.stabilize and CMB.make have been modified to work in
1854    two passes when compile servers are attached:
1855       1. Compile everything, do not perform stabilization; this pass
1856          uses compile servers
1857       2. Stabilize everything; this pass does not use compile servers
1858    If there are no compile servers, the two passes are combined into one
1859    (as before).  Splitting the passes increases the inherent parallelism
1860    in the dependency graph because the entire graph including all
1861    libraries is available at the same time.  This, in turn, improves
1862    server utilization.  The downside is that the master process will
1863    have to do some extra work after compilation is done (because for
1864    technical reasons it must re-read all the binfiles during stabilization).
1865    
1866    ----------------------------------------------------------------------
1867    Name: Matthias Blume
1868  Date: 2001/03/16 12:22:00 EST  Date: 2001/03/16 12:22:00 EST
1869  Tag: blume-20010316-bootfiles  Tag: blume-20010316-bootfiles
1870  Description:  Description:
# Line 2003  Line 3854 
3854       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3855       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3856       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
3857       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3858       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3859       Dave's hands. :)       Dave's hands. :)
3860    

Legend:
Removed from v.800  
changed lines
  Added in v.1076

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