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 833, Tue May 22 22:06:12 2001 UTC revision 1063, Wed Feb 13 21:41:53 2002 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16  Name: Matthias Blume  Name: Matthias Blume
17    Date: 2002/02/13 16:41:00 EST
18    Tag: blume-20020213-cmlpq
19    Description:
20    
21    Added new priority queue export symbols (which have just been added to
22    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
23    not compile and the installer would choke.)
24    
25    ----------------------------------------------------------------------
26    Name: Matthias Blume
27    Date: 2002/02/13 16:15:00 EST
28    Tag: blume-20020213-various
29    Description:
30    
31    1. More tweaks to ml-nlffigen:
32    
33       - better internal datastructures (resulting in slight speedup)
34       - "-match" option requires exact match
35       - "localized" gensym counters (untagged structs/unions nested within
36         other structs/unions or within typedefs get a fresh counter; their
37         tag will be prefixed by a concatenation of their parents' tags)
38       - bug fixes (related to calculation of transitive closure of types
39         to be included in the output)
40    
41    2. Minor Basis updates:
42    
43       - added implementations for List.collate and Option.app
44    
45    ----------------------------------------------------------------------
46    Name: Matthias Blume
47    Date: 2002/02/11 15:55:00 EST
48    Tag: blume-20020211-gensym
49    Description:
50    
51    Added a "-gensym" option to command line of ml-nlffigen.  This can be
52    used to specify a "stem" -- a string that is inserted in all "gensym'd"
53    names (ML structure names that correspond to unnamed C structs, unions,
54    and enums), so that separate runs of ml-nlffigen do not clash.
55    
56    ----------------------------------------------------------------------
57    Name: Matthias Blume
58    Date: 2002/02/11 12:05:00 EST
59    Tag: blume-20020211-gensml
60    Description:
61    
62    A quick fix for a problem with GenSML (in the pgraph-util library):
63    Make generation of toplevel "local" optional.  (Strictly speaking,
64    signature definitions within "local" are not legal SML.)
65    
66    Other than that: updates to INSTALL and cm/TODO.
67    
68    ----------------------------------------------------------------------
69    Name: Matthias Blume
70    Date: 2002/02/08 15:00:00 EST
71    Tag: blume-20020208-uniquepid
72    Description:
73    
74    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
75    
76    1. The installer (config/install.sh) has gotten smarter:
77    
78         - Configuration options are a bit easier to specify now
79           (in config/targets).
80         - Bug in recognizing .tar.bz2 files fixed.
81         - Installer automatically resolves dependencies between
82           configuration options (e.g., if you ask for eXene, you will
83           also get cml -- regardless whether you asked for it or not).
84         - Installer can run in "quieter mode" by setting the environment
85           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
86           "completely silent", though.
87         - Build HashCons library as part of smlnj-lib.
88    
89    2. A new scheme for assigning persistent identifiers to compilation
90       units (and, by extension, to types etc.) has been put into place.
91       This fixes a long-standing bug where types and even dynamic values
92       can get internally confused, thereby compromising type safety
93       (abstraction) and dynamic correctness.  See
94    
95         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
96    
97       for an example of how things could go wrong until now.
98    
99       The downside of the new scheme is that pids are not quite as
100       persistent as they used to be: CM will generate a fresh pid
101       for every compilation unit that it thinks it sees for the first
102       time.  That means that if you compile starting from a clean, fresh
103       source tree at two different times, you end up with different
104       binaries.
105    
106       Cutoff recompilation, however, has not been compromised because
107       CM keeps pid information in special caches between runs.
108    
109    ----------------------------------------------------------------------
110    Name: Lal George
111    Date: 2002/02/07 15:34:13 EST 2002
112    Tag: <none>
113    Description:
114    
115    Compilers that generate assembly code may produce  global labels
116    whose value is resolved at link time. The various peephole optimization
117    modules did not take this in account.
118    
119    TODO. The Labels.addrOf function should really return an option
120    type so that clients are forced to deal with this issue, rather
121    than an exception being raised.
122    
123    ----------------------------------------------------------------------
124    Name: Lal George
125    Date: 2002/02/06 13:55:02 EST
126    Tag: george-20020206-ra-breakup
127    Description:
128    
129    1. A bug fix from Allen.
130    
131        A typo causes extra fstp %st(0)'s to be generated at compensation
132        edges, which might cause stack underflow traps at runtime.  This
133        occurs in fft where there are extraneous fstps right before the 'into'
134        trap instruction (in this case they are harmless since none of the
135        integers overflow.)
136    
137    2. Pulled out various utility modules that were embedded in the modules
138       of the register allocator. I need these modules for other purposes, but
139       they are not complete enough to put into a library (just yet).
140    ----------------------------------------------------------------------
141    Name: Matthias Blume
142    Date: 2002/01/31 16:05:00 EST
143    Tag: blume-20020131-sparc-ccalls
144    Description:
145    
146    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
147       of extra stack space by mistake.  Fixed.
148    
149    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
150    
151    ----------------------------------------------------------------------
152    Name: Allen Leung
153    Date: 2002/01/30
154    Tag: leunga-20020130-nowhere-bug-fix
155    Description:
156    
157       MLRISC bug fixes:
158       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
159       2. MachineInt.fromString was negating its value.
160    
161    ----------------------------------------------------------------------
162    Name: Matthias Blume
163    Date: 2002/01/29
164    Tag: blume-20020129-INSTALL
165    Description:
166    
167    - Added somewhat detailed installation instructions (file INSTALL).
168    - Fixed curl-detection bug in config/install.sh.
169    - It is now possible to select the URL getter using the URLGETTER
170      environment variable:
171    
172          not set / "unknown"      --> automatic detection (script tries wget,
173                                       curl, and lynx)
174          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
175                                       how to properly invoke them)
176          other                    --> use $URLGETTER directly, it must take
177                                       precisely two command-line arguments
178                                       (source URL and destination file name)
179    
180    ----------------------------------------------------------------------
181    Name: Matthias Blume
182    Date: 2002/01/28
183    Tag: blume-20020128-sparc-ccalls
184    Description:
185    
186    - Fixed problem with calculation of "used" registers in sparc-c-calls.
187    - Make use of the allocParam argument in sparc-c-calls.
188    
189    ----------------------------------------------------------------------
190    Name: Matthias Blume
191    Date: 2002/01/28
192    Tag: blume-20020128-allocParam
193    Description:
194    
195    John Reppy:  Changes c-calls API to accept client-callback for
196    allocating extra stack space.
197    me: Corresponding changes to mlriscGen (using a dummy argument that
198        does not change the current behavior).
199    
200    ----------------------------------------------------------------------
201    Name: Matthias Blume
202    Date: 2002/01/28 12:00:00
203    Tag: Release_110_38
204    Description:
205    
206    This time for real!!!
207    
208    ----------------------------------------------------------------------
209    Name: Matthias Blume
210    Date: 2002/01/28 10:56:00 EST
211    Tag: blume-20020128-retraction
212    Description:
213    
214    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
215       with blume-Release_110_38-retracted.)
216    
217    1. Fixed a problem with incorrect rounding modes in real64.sml.
218       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
219    
220    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
221       and enums fixed.  The naming of corresponding ML identifiers should
222       now be consistent again.
223    
224    ----------------------------------------------------------------------
225    Name: Allen Leung
226    Date: 2002/01/27
227    Tag: leunga-20020127-nowhere
228    Description:
229    
230       Added a target called nowhere in the configuration scripts.
231       Enabling this will build the MLRISC 'nowhere' tool (for translating
232       programs with where-clauses into legal SML code) during installation.
233    
234    ----------------------------------------------------------------------
235    Name: Matthias Blume
236    Date: 2002/01/25 21:27:00 EST
237    Tag: blume-Release_110_38-retracted
238    Description:
239    
240    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
241    
242    README will be added later.
243    
244    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
245    (Release_110_38) removed.  Reason: Last-minute bug fixes.
246    
247    ----------------------------------------------------------------------
248    Name: Matthias Blume
249    Date: 2002/01/25
250    Tag: blume-20020125-ffi
251    Description:
252    
253    A large number of tweaks and improvements to ml-nlffi-lib and
254    ml-nlffigen:
255    
256       - ML represenation types have been streamlined
257       - getter and setter functions work with concrete values, not abstract
258         ones where possible
259       - ml-nlffigen command line more flexible (see README file there)
260       - some bugs have been fixed (hopefully)
261    
262    ----------------------------------------------------------------------
263    Name: Lal George
264    Date: 2002/01/24
265    Tag: george-20020124-risc-ra-interface
266    Description:
267    
268       There is a dramatic simplification in the interface to the
269       register allocator for RISC architectures as a result of making
270       parallel copy instructions explicit.
271    
272    ----------------------------------------------------------------------
273    Name: Matthias Blume
274    Date: 2002/01/22
275    Tag: blume-20020122-x86-ccalls
276    Description:
277    
278    Bug fix for c-calls on x86 (having to do with how char- and
279    short-arguments are being handled).
280    
281    ----------------------------------------------------------------------
282    Name: Matthias Blume
283    Date: 2002/01/21
284    Tag: blume-20020121-ff
285    Description:
286    
287    Another day of fiddling with the FFI...
288    
289    1. Bug fix/workaround:  CKIT does not complain about negative array
290       dimensions, so ml-nlffigen has to guard itself against this possibility.
291       (Otherwise a negative dimension would send it into an infinite loop.)
292    
293    2. Some of the abstract types (light objects, light pointers, most "base"
294       types) in structure C are now eqtypes.
295    
296    3. Added constructors and test functions for NULL function pointers.
297    
298    ----------------------------------------------------------------------
299    Name: Matthias Blume
300    Date: 2002/01/18
301    Tag: blume-20020118-ready-for-new-release
302    Description:
303    
304    Made config/srcarchiveurl point to a new place.  (Will provide boot
305    files shortly.)
306    
307    Maybe we christen this to be 110.38?
308    
309    ----------------------------------------------------------------------
310    Name: Matthias Blume
311    Date: 2002/01/18
312    Tag: blume-20020118-more-ffifiddle
313    Description:
314    
315    Today's FFI fiddling:
316    
317      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
318        C.Set.* that use concrete (MLRep.*) arguments and results instead
319        of abstract ones.
320    
321      - Provided word-style bit operations etc. for "int" representation
322        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
323        and Long.
324    
325    ----------------------------------------------------------------------
326    Name: Matthias Blume
327    Date: 2002/01/18
328    Tag: blume-20020118-use-x86-fp
329    Description:
330    
331    Now that x86-fast-fp seems to be working, I turned it back on again
332    by default.  (Seems to work fine now, even with the FFI.)
333    
334    Other than that, I added some documentation about the FFI to
335    src/ml-nlffigen/README and updated the FFI test examples in
336    src/ml-nlffi-lib/Tests/*.
337    
338    ----------------------------------------------------------------------
339    Name: Allen Leung
340    Date: 2002/01/17
341    Tag: leunga-20020117-x86-fast-fp-call
342    Description:
343    
344       1. Fixed a problem with handling return fp values when x86's fast fp
345          mode is turned on.
346    
347       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
348          of %f32.
349    
350       3. Added a constructor INT32lit to the ast of MLRISC tools.
351    
352    ----------------------------------------------------------------------
353    Name: Matthias Blume
354    Date: 2002/01/16
355    Tag: blume-20020116-ffifiddle
356    Description:
357    
358    More fiddling with the FFI interface:
359    
360     - Make constness 'c instead of rw wherever possible.  This eliminates
361       the need for certain explicit coercions.  (However, due to ML's
362       value polymorphism, there will still be many cases where explicit
363       coercions are necessary.  Phantom types are not the whole answer
364       to modeling a subtyping relationship in ML.)
365    
366     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
367    
368     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
369       clumsy workaround for varargs problem.)
370    
371    ----------------------------------------------------------------------
372    Name: Lal George
373    Date: 2002/01/15
374    Tag: <none>
375    Description:
376    
377    1. Since COPY instructions are no longer native to the architecture,
378       a generic functor can be used to implement the expandCopies function.
379    
380    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
381       TEXT segment.
382    
383    ----------------------------------------------------------------------
384    Name: Matthias Blume
385    Date: 2002/01/15
386    Tag: blume-20020115-ffiupdates
387    Description:
388    
389    1. Fix for bug resulting in single-precision float values being returned
390       incorrectly from FFI calls.
391    
392    2. Small modifications to C FFI API:
393    
394        - memory-allocation routines return straight objects (no options)
395          and raise an exception in out-of-memory situations
396        - unsafe extensions to cast between function pointers and pointers
397          from/to ints
398        - added structure C_Debug as an alternative to structure C where
399          pointer-dereferencing (|*| and |*!) always check for null-pointers
400        - added open_lib' to DynLinkage;  open_lib' works like open_lib
401          but also takes a (possibly empty) list of existing library handles
402          that the current library depends on
403    
404    ----------------------------------------------------------------------
405    Name: Matthias Blume
406    Date: 2002/01/10
407    Tag: blume-20020110-newffigen
408    Description:
409    
410    1. Updates to portable graph code.
411    
412    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
413       more scalable now so that even huge interfaces such as the one
414       for GTK compile in finite time and space. :-)
415       See src/ml-nlffigen/README for details on what's new.
416    
417    ----------------------------------------------------------------------
418    Name: Lal George
419    Date: 2001/01/09 14:31:35 EST 2002
420    Tag: george-20011206-rm-native-copy
421    Description:
422    
423            Removed the native COPY and FCOPY instructions
424            from all the architectures and replaced it with the
425            explicit COPY instruction from the previous commit.
426    
427            It is now possible to simplify many of the optimizations
428            modules that manipulate copies. This has not been
429            done in this change.
430    
431    ----------------------------------------------------------------------
432    Name: Lal George
433    Date: 2001/12/06 16:50:13 EST 2001
434    Tag: george-20011206-mlrisc-instruction
435    Description:
436    
437    Changed the representation of instructions from being fully abstract
438    to being partially concrete. That is to say:
439    
440      from
441            type instruction
442    
443      to
444            type instr                              (* machine instruction *)
445    
446            datatype instruction =
447                LIVE of {regs: C.cellset, spilled: C.cellset}
448              | KILL of {regs: C.cellset, spilled: C.cellset}
449              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
450              | ANNOTATION of {i: instruction, a: Annotations.annotation}
451              | INSTR of instr
452    
453    This makes the handling of certain special instructions that appear on
454    all architectures easier and uniform.
455    
456    LIVE and KILL say that a list of registers are live or killed at the
457    program point where they appear. No spill code is generated when an
458    element of the 'regs' field is spilled, but the register is moved to
459    the 'spilled' (which is present, more for debugging than anything else).
460    
461    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
462    We used to generate:
463    
464            DEFFREG f1
465            f1 := f2 + f3
466            trapb
467    
468    but now generate:
469    
470            f1 := f2 + f3
471            trapb
472            LIVE {regs=[f1,f2,f3], spilled=[]}
473    
474    Furthermore, the DEFFREG (hack) required that all floating point instruction
475    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
476    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
477    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
478    this architecture).
479    
480    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
481    all the architectures. This will result in further simplification of the
482    register allocator that must be aware of them for coalescing purposes, and
483    will also simplify certain aspects of the machine description that provides
484    callbacks related to parallel copies.
485    
486    ANNOTATION should be obvious, and now INSTR represents the honest to God
487    machine instruction set!
488    
489    The <arch>/instructions/<arch>Instr.sml files define certain utility
490    functions for making porting easier -- essentially converting upper case
491    to lower case. All machine instructions (of type instr) are in upper case,
492    and the lower case form generates an MLRISC instruction. For example on
493    the alpha we have:
494    
495      datatype instr =
496         LDA of {r:cell, b:cell, d:operand}
497       | ...
498    
499      val lda : {r:cell, b:cell, d:operand} -> instruction
500        ...
501    
502    where lda is just (INSTR o LDA), etc.
503    
504    ----------------------------------------------------------------------
505    Name: Matthias Blume
506    Date: 2001/11/22 21:40:00 EST
507    Tag: Release_110_37
508    Description:
509    
510    Release 110.37.  This time for real.
511    
512    ----------------------------------------------------------------------
513    Name: Matthias Blume
514    Date: 2001/11/21 16:35:00 EST
515    Tag: blume-20011121-foot-in-mouth
516    Description:
517    
518    Removed the "Release_110_37" tag because of a serious bug.
519    This will be re-tagged once the bug is fixed.
520    
521    ----------------------------------------------------------------------
522    Name: Matthias Blume
523    Date: 2001/11/21 16:14:00 EST
524    Tag: blume-20011121-forgottenfile
525    Description:
526    
527    Forgot to add a file.  (Just a .tex-file -- part of
528    the CM manual source.)
529    
530    ----------------------------------------------------------------------
531    Name: Matthias Blume
532    Date: 2001/11/21 16:10:00 EST
533    Tag: blume-20011121-invalid_110_37
534    Description:
535    
536    Note: I removed the original tag "Release_110_37" from this commit
537          because we found a serious bug in all non-x86 backends.
538          - Matthias
539    
540    1. Modifications to the SML/NJ code generator and to the runtime system
541       so that code object name strings are directly inserted into code
542       objects at code generation time.  The only business the runtime system
543       has with this is now to read the name strings on occasions.
544       (The encoding of the name string has also changed somewhat.)
545    
546    2. CM now implements a simple "set calculus" for specifying export lists.
547       In particular, it is now possible to refer to the export lists of
548       other libraries/groups/sources and form unions as well as differences.
549       See the latest CM manual for details.
550    
551    3. An separate notion of "proxy" libraries has again be eliminated from
552       CM's model.  (Proxy libraries are now simply a special case of using
553       the export list calculus.)
554    
555    4. Some of the existing libraries now take advantage of the new set
556       calculus.
557       (Notice that not all libraries have been converted because some
558       of the existing .cm-files are supposed to be backward compatible
559       with 110.0.x.)
560    
561    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
562       "General.exnMessage"!  The former relies on a certain hook to be
563       initialized, and that often does not happen in the stand-alone case.)
564    
565    ----------------------------------------------------------------------
566    Name: Lal George
567    Date: 2001/11/21  13:56:18 EST
568    Tag: george-2001121-pseudo-ops
569    Description:
570    
571      Implemented a complete redesign of MLRISC pseudo-ops. Now there
572      ought to never be any question of incompatabilities with
573      pseudo-op syntax expected by host assemblers.
574    
575      For now, only modules supporting GAS syntax are implemented
576      but more should follow, such as MASM, and vendor assembler
577      syntax, e.g. IBM as, Sun as, etc.
578    
579    ----------------------------------------------------------------------
580    Name: Matthias Blume
581    Date: 2001/11/14 11:52:00 EST
582    Tag: blume-20011114-srcname
583    Description:
584    
585    1. Routed the name of the current source file to mlriscgen where it
586       should be directly emitted into the code object.  (This last part
587       is yet to be done.)
588    
589    2. Some cleanup of the pgraph code to make it match the proposal that
590       I put out the other day.  (The proposal notwithstanding, things are
591       still in flux here.)
592    
593    ----------------------------------------------------------------------
594    Name: Lal George
595    Date: 2001/11/14 09:44:04 EST
596    Tag:
597    Description:
598    
599      Fix for a backpatching bug reported by Allen.
600    
601      Because the boundary between short and long span-dependent
602      instructions is +/- 128, there are an astounding number of
603      span-dependent instructions whose size is over estimated.
604    
605      Allen came up with the idea of letting the size of span
606      dependent instructions be non-monotonic, for a maxIter
607      number of times, after which the size must be monotonically
608      increasing.
609    
610      This table shows the number of span-dependent instructions
611      whose size was over-estimated as a function of maxIter, for the
612      file Parse/parse/ml.grm.sml:
613    
614         maxIter            # of instructions:
615            10                      687
616            20                      438
617            30                      198
618            40                        0
619    
620      In compiling the compiler, there is no significant difference in
621      compilation speed between maxIter=10 and maxIter=40. Actually,
622      my measurements showed that maxIter=40 was a tad faster than
623      maxIter=10! Also 96% of the  files in the compiler reach a fix
624      point within 13 iterations, so fixing maxIter at 40, while high,
625      is okay.
626    
627    ----------------------------------------------------------------------
628    Name: Matthias Blume
629    Date: 2001/10/31 15:25:00 EST
630    Tag: blume-20011031-pgraph
631    Description:
632    
633    CKIT:
634    * Changed the "Function" constructor of type Ast.ctype to carry optional
635      argument identifiers.
636    * Changed the return type of TypeUtil.getFunction accordingly.
637    * Type equality ignores the argument names.
638    * TypeUtil.composite tries to preserve argument names but gives up quickly
639      if there is a mismatch.
640    
641    installation script:
642    * attempts to use "curl" if available (unless "wget" is available as well)
643    
644    CM:
645    * has an experimental implementation of "portable graphs" which I will
646      soon propose as an implementation-independent library format
647    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
648    
649    NLFFI-LIB:
650    * some cleanup (all cosmetic)
651    
652    NLFFIGEN:
653    * temporarily disabled the mechanism that suppresses ML output for
654      C definitions whose identifiers start with an underscore character
655    * generate val bindings for enum constants
656    * user can request that only one style (light or heavy) is being used;
657      default is to use both (command-line arguments: -heavy and -light)
658    * fixed bug in handling of function types involving incomplete pointers
659    * generate ML entry points that take record arguments (i.e., using
660      named arguments) for C functions that have a prototype with named
661      arguments
662      (see changes to CKIT)
663    
664    ----------------------------------------------------------------------
665    Name: Allen Leung
666    Date: 2001/10/27 20:34:00 EDT
667    Tag: leunga-20011027-x86-fast-fp-call
668    Description:
669    
670       Fixed the bug described in blume-20010920-slowfp.
671    
672       The fix involves
673          1. generating FCOPYs in FSTP in ia32-svid
674          2. marking a CALL with the appropriate annotation
675    
676    ----------------------------------------------------------------------
677    Name: Matthias Blume
678    Date: 2001/10/16 11:32:00 EDT
679    Tag: blume-20011016-netbsd
680    Description:
681    
682    Underscore patch from Chris Richards (fixing problem with compiling
683    runtime system under recent NetBSD).
684    
685    ----------------------------------------------------------------------
686    Name: Allen Leung
687    Date: 2001/10/12 17:18:32 EDT 2001
688    Tag: leung-20011012-x86-printflowgraph
689    Description:
690    
691    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
692    
693    ----------------------------------------------------------------------
694    Name: Lal George
695    Date: 2001/10/11 23:51:34 EDT
696    Tag: george-20011011-too-many-instrs
697    Description:
698    
699    The representation of a program point never expected to see more
700    than 65536 instructions in a basic block!
701    
702    ----------------------------------------------------------------------
703    Name: Lal George
704    Date: 2001/10/09 09:41:37 EDT
705    Tag: george-20011008-mlrisc-labels
706    Description:
707    
708    Changed the machine description files to support printing of
709    local and global labels in assembly code, based on host assembler
710    conventions.
711    
712    ----------------------------------------------------------------------
713    Name: Matthias Blume
714    Date: 2001/09/25 15:25:00 EDT
715    Tag: blume-20010925-exninfo
716    Description:
717    
718    I provided a non-hook implementation of exnName (at the toplevel) and
719    made the "dummy" implementation of exnMessage (at the toplevel) more
720    useful: if nothing gets "hooked in", then at least you are going to
721    see the exception name and a message indicating why you don't see more.
722    
723    [For the time being, programs that need exnMessage and want to use
724    ml-build should either use General.exnMessage (strongly recommended) or
725    refer to structure General at some other point so that CM sees a
726    static dependency.]
727    
728    [Similar remarks go for "print" and "use":  If you want to use their
729    functionality in stand-alone programs generated by ml-build, then use
730    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
731    
732    ----------------------------------------------------------------------
733    Name: Matthias Blume
734    Date: 2001/09/20 17:28:00 EDT
735    Tag: blume-20010920-slowfp
736    Description:
737    
738    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
739    
740    ----------------------------------------------------------------------
741    Name: Matthias Blume
742    Date: 2001/09/20 17:20:00 EDT
743    Tag: blume-20010920-canonicalpaths
744    Description:
745    
746    0. Updated the BOOT file (something that I forgot to do earlier).
747    
748    1. Small internal change to CM so that it avoids "/../" in filenames
749       as much as possible (but only where it is safe).
750    
751    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
752       that arguments that contain delimiters are passed through correctly.
753       This change also means that all "special" arguments of the form
754       @SMLxxx... must come first.
755    
756    3. Changed install script to put relative anchor names for tool commands
757       into pathconfig.
758    
759    ----------------------------------------------------------------------
760    Name: Matthias Blume
761    Date: 2001/09/18 15:35:00 EDT
762    Tag: blume-20010918-readme11036
763    Description:
764    
765    Added README files.
766    
767    ----------------------------------------------------------------------
768    Name: Matthias Blume
769    Date: 2001/09/18 11:45:00 EDT
770    Tag: Release_110_36 (retag)
771    Description:
772    
773    Fixed mistake in config/preloads. Retagged as 110.36.
774    
775    ----------------------------------------------------------------------
776    Name: Matthias Blume
777    Date: 2001/09/18 09:40:00 EDT
778    Tag: Release_110_36_orig (tag changed)
779    Description:
780    
781    New version (110.36).  New bootfiles.
782    
783    ----------------------------------------------------------------------
784    Name: Matthias Blume
785    Date: 2001/09/14 16:15:00 EDT
786    Tag: blume-20010914-x86fastfp
787    Description:
788    
789    John committed some changes that Allen made, in particular a (hopefully)
790    correctly working version of the x86-fp module.
791    
792    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
793    flag to "true".  Everything seems to compile to a fixpoint ok, and
794    "mandelbrot" speeds up by about 15%.
795    
796    ----------------------------------------------------------------------
797    Name: Matthias Blume
798    Date: 2001/09/13 11:20:00 EDT
799    Tag: blume-20010913-minimal
800    Description:
801    
802    1. Stefan Monnier's patch to fix a miscompilation problem that
803       was brought to light by John Reppy's work on Moby.
804    
805    2. Implemented a minimal "structure Compiler" that contains just
806       "version" and "architecture".  The minimal version will be
807       available when the full version is not.  This is for backward-
808       compatibility with code that wants to test Compiler.version.
809    
810    ----------------------------------------------------------------------
811    Name: Matthias Blume
812    Date: 2001/08/28 14:03:00 EDT
813    Tag: blume-20010828-ml-lex
814    Description:
815    
816    Fix for bug 1581, received from Neophytos Michael.
817    
818    ----------------------------------------------------------------------
819    Name: Matthias Blume
820    Date: 2001/08/27 11:20:00 EDT
821    Tag: blume-20010827-readme11035
822    Description:
823    
824    Fleshed out the README file for 110.35.
825    
826    ----------------------------------------------------------------------
827    Name: Matthias Blume
828    Date: 2001/08/24 17:10:00 EDT
829    Tag: Release_110_35
830    Description:
831    
832    New version number (110.35).  New bootfiles.
833    
834    ----------------------------------------------------------------------
835    Name: Lal George
836    Date: 2001/08/24 13:47:18 EDT 2001
837    Tag: george-20010824-MLRISC-graphs
838    Description:
839    
840     removed clusters from MLRISC completely and replaced with graphs.
841    
842    ----------------------------------------------------------------------
843    Name: Matthias Blume
844    Date: 2001/08/23 17:50:00 EDT
845    Tag: blume-20010823-toplevel
846    Description:
847    
848    - some reorganization of the code that implements various kinds of
849      environments in the compiler (static, dynamic, symbolic, combined)
850    - re-implemented the EnvRef module so that evalStream works properly
851      (if the stream contains references to "use", "CM.make", etc.)
852    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
853    
854    ----------------------------------------------------------------------
855    Name: Matthias Blume
856    Date: 2001/08/20 15:50 EDT
857    Tag: blume20010820-slipup
858    Description:
859    
860    I forgot to commit a few files.  Here they are...
861    
862    ----------------------------------------------------------------------
863    Name: Matthias Blume
864    Date: 2001/08/20 15:35:00 EDT
865    Tag: blume-20010820-debugprof
866    Description:
867    
868    !!!! NEW BOOTFILES !!!!
869    
870    This is another round of reorganizing the compiler sources.  This
871    time the main goal was to factor out all the "instrumentation"
872    passes (for profiling and backtracing) into their own library.
873    The difficulty was to do it in such a way that it does not depend
874    on elaborate.cm but only on elabdata.cm.
875    
876    Therefore there have been further changes to both elaborate.cm and
877    elabdata.cm -- more "generic" things have been moved from the former
878    to the latter.  As a result, I was forced to split the assignment
879    of numbers indicating "primtyc"s into two portions: SML-generic and
880    SML/NJ-specific.  Since it would have been awkward to maintain,
881    I bit the bullet and actually _changed_ the mapping between these
882    numbers and primtycs.  The bottom line of this is that you need
883    a new set of bin- and bootfiles.
884    
885    I have built new bootfiles for all architectures, so doing a fresh
886    checkout and config/install.sh should be all you need.
887    
888    The newly created library's name is
889    
890        $smlnj/viscomp/debugprof.cm
891    
892    and its sources live under
893    
894        src/compiler/DebugProf
895    
896    ----------------------------------------------------------------------
897    Name: Matthias Blume
898    Date: 2001/08/15 17:15:00 EDT
899    Tag: blume-20010815-compreorg
900    Description:
901    
902    This is a first cut at reorganizing the CM libraries that make up the
903    core of the compiler.  The idea is to separate out pieces that could
904    be used independently by tools, e.g., the parser, the typechecker, etc.
905    
906    The current status is a step in this direction, but it is not quite
907    satisfactory yet.  Expect more changes in the future.
908    
909    Here is the current (new) organization...
910    
911        What used to be $smlnj/viscomp/core.cm is now divided into
912        six CM libraries:
913    
914             $smlnj/viscomp/basics.cm
915                           /parser.cm
916                           /elabdata.cm
917                           /elaborate.cm
918                           /execute.cm
919                           /core.cm
920    
921        The CM files for these libraries live under src/system/smlnj/viscomp.
922        All these libraries are proxy libraries that contain precisely
923        one CM library component.  Here are the locations of the components
924        (all within the src/compiler tree):
925    
926             Basics/basics.cm
927             Parse/parser.cm
928             ElabData/elabdata.cm
929             Elaborator/elaborate.cm
930             Execution/execute.cm
931             core.cm
932    
933         [This organization is the same that has been used already
934         for a while for the architecture-specific parts of the visible
935         compiler and for the old version of core.cm.]
936    
937         As you will notice, many source files have been moved from their
938         respective original locations to a new home in one of the above
939         subtrees.
940    
941         The division of labor between the new libraries is the following:
942    
943             basics.cm:
944                - Simple, basic definitions that pertain to many (or all) of
945                  the other libraries.
946             parser.cm:
947                - The SML parser, producing output of type Ast.dec.
948                - The type family for Ast is also defined and exported here.
949             elabdata.cm:
950                - The datatypes that describe input and output of the elaborator.
951                  This includes types, absyn, and static environments.
952             elaborator.cm:
953                - The SML/NJ type checker and elaborator.
954                  This maps an Ast.dec (with a given static environment) to
955                  an Absyn.dec (with a new static environment).
956                - This libraries implements certain modules that used to be
957                  structures as functors (to remove dependencies on FLINT).
958             execute.cm:
959                - Everything having to do with executing binary code objects.
960                - Dynamic environments.
961             core.cm:
962                - SML/NJ-specific instantiations of the elaborator and MLRISC.
963                - Top-level modules.
964                - FLINT (this should eventually become its own library)
965    
966    Notes:
967    
968    I am not 100% happy with the way I separated the elaborator (and its
969    data structures) from FLINT.  Two instances of the same problem:
970    
971        1. Data structures contain certain fields that carry FLINT-specific
972           information.  I hacked around this using exn and the property list
973           module from smlnj-lib.  But the fact that there are middle-end
974           specific fields around at all is a bit annoying.
975    
976        2. The elaborator calculates certain FLINT-related information.  I tried
977           to make this as abstract as I could using functorization, but, again,
978           the fact that the elaborator has to perform calculations on behalf
979           of the middle-end at all is not nice.
980    
981        3. Having to used exn and property lists is unfortunate because it
982           weakens type checking.  The other alternative (parameterizing
983           nearly *everything*) is not appealing, though.
984    
985    I removed the "rebinding =" warning hack because due to the new organization
986    it was awkward to maintain it.  As a result, the compiler now issues some of
987    these warnings when compiling init.cmi during bootstrap compilation. On
988    the plus side, you also get a warning when you do, for example:
989       val op = = Int32.+
990    which was not the case up to now.
991    
992    I placed "assign" and "deref" into the _Core structure so that the
993    code that deals with the "lazy" keyword can find them there.  This
994    removes the need for having access to the primitive environment
995    during elaboration.
996    
997    ----------------------------------------------------------------------
998    Name: Matthias Blume
999    Date: 2001/08/13
1000    Tag: blume-20010813-closures
1001    Description:
1002    
1003    This fix was sent to us by Zhong Shao.  It is supposed to improve the
1004    performance of certain loops by avoiding needless closure allocation.
1005    
1006    ----------------------------------------------------------------------
1007    Name: Lal George
1008    Date: 2001/07/31 10:03:23 EDT 2001
1009    Tag: george-20010731-x86-fmalloc
1010    Description: Fixed bug in x86 calls
1011    
1012        There was a bug where call instructions would mysteriously
1013        vanish. The call instruction had to be one that returned
1014        a floating point value.
1015    
1016    ----------------------------------------------------------------------
1017    Name: Lal George
1018    Date: 2001/07/19 16:36:29 EDT 2001
1019    Tag: george-20010719-simple-cells
1020    Description:
1021    
1022    I have dramatically simplified the interface for CELLS in MLRISC.
1023    
1024    In summary, the cells interface is broken up into three parts:
1025    
1026      1. CellsBasis : CELLS_BASIS
1027    
1028            CellsBasis is a top level structure and common for all
1029            architectures.  it contains the definitions of basic datatypes
1030            and utility  functions over these types.
1031    
1032      2. functor Cells() : CELLS
1033    
1034            Cells generates an interface for CELLS that incorporates the
1035            specific resources on the target architecture, such as the
1036            presence of special register classes, their number and size,
1037            and various useful substructures.
1038    
1039      3. <ARCH>CELLS
1040    
1041            e.g. SparcCells: SPARCCELLS
1042    
1043            <ARCH>CELLS usually contains additional bindings for special
1044            registers  on the architecture, such as:
1045    
1046                    val r0 : cell           (* register zero *)
1047                    val y : cell            (* Y register *)
1048                    val psr : cell          (* processor status register *)
1049                    ...
1050    
1051            The structure returned by applying the Cells functor is opened
1052            in this interface.
1053    
1054    The main implication of all this is that the datatypes for cells is
1055    split between CellsBasis and CELLS -- a fairly simple change for user
1056    code.
1057    
1058    In the old scheme the CELLS interface had a definitional binding of
1059    the form:
1060    
1061            signature CELLS = sig
1062    
1063               structure CellsBasis = CellsBasis
1064    
1065               ...
1066    
1067            end
1068    
1069    With all the sharing constraints that goes on in MLRISC, this old
1070    design  quickly leads to errors such as:
1071    
1072            "structure definition spec inside of sharing ... "
1073    
1074    
1075    and appears to require an unacceptable amount of sharing and where
1076    constraint hackery.
1077    
1078    I think this error message (the interaction of definitional specs and
1079    sharing) requires more explanation on our web page.
1080    
1081    ----------------------------------------------------------------------
1082    Name: Matthias Blume
1083    Date: 2001/07/19 15:00:00 EDT
1084    Tag: blume-20010719-libreorg
1085    Description:
1086    
1087    This update puts together a fairly extensive but straightforward change
1088    to the way the libraries that implement the interactive system are
1089    organized:
1090    
1091       The biggest change is the elimination of structure Compiler.  As a
1092       replacement for this structure, there is now a CM library
1093       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1094       that exports all the substructures of the original structure Compiler
1095       directly.  So instead of saying Compiler.Foo.bar one now simply
1096       says Foo.bar.  (The CM libraries actually export a collection of
1097       structures that is richer than the collection of substructures of
1098       structure Compiler.)
1099    
1100       To make the transition smooth, there is a separate library called
1101       $smlnj/compiler/compiler.cm which puts together and exports the
1102       original structure Compiler (or at least something very close to it).
1103    
1104       There are five members of the original structure Compiler
1105       that are not exported directly but which instead became members
1106       of a new structure Backend (described by signature BACKEND).  These are:
1107       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1108       Interact (: INTERACT), structure Machine (: MACHINE), and val
1109       architecture (: string).
1110    
1111       Structure Compiler.Version has become structure CompilerVersion.
1112    
1113       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1114       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1115       or x86, respectively.
1116       Each of these exports the same frontend structures that
1117       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1118       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1119       Hppa, PPC, Sparc, or X86, respectively.
1120    
1121       Library $smlnj/compiler/all.cm exports the union of the exports of
1122       $smlnj/compiler/<arch>.cm
1123    
1124       There are no structures <Arch>Compiler anymore, use
1125       $smlnj/compiler/<arch>.cm instead.
1126    
1127       Library host-compiler-0.cm is gone.  Instead, the internal library
1128       that instantiates CM is now called cm0.cm.  Selection of the host
1129       compiler (backend) is no longer done here but. (Responsibility for it
1130       now lies with $smlnj/compiler/current.cm.  This seems to be more
1131       logical.)
1132    
1133       Many individual files have been moved or renamed.  Some files have
1134       been split into multiple files, and some "dead" files have been deleted.
1135    
1136    Aside from these changes to library organization, there are also changes
1137    to the way the code itself is organized:
1138    
1139       Structure Binfile has been re-implemented in such a way that it no
1140       longer needs any knowledge of the compiler.  It exclusively deals
1141       with the details of binfile layout.  It no longer invokes the
1142       compiler (for the purpose of creating new prospective binfile
1143       content), and it no longer has any knowledge of how to interpret
1144       pickles.
1145    
1146       Structure Compile (: COMPILE) has been stripped down to the bare
1147       essentials of compilation.  It no longer deals with linking/execution.
1148       The interface has been cleaned up considerably.
1149    
1150       Utility routines for dealing with linking and execution have been
1151       moved into their own substructures.
1152    
1153       (The ultimate goal of these changes is to provide a light-weight
1154       binfile loader/linker (at least for, e.g., stable libraries) that
1155       does not require CM or the compiler to be present.)
1156    
1157    CM documentation has been updated to reflect the changes to library
1158    organization.
1159    
1160    ----------------------------------------------------------------------
1161    Name: Matthias Blume
1162    Date: 2001/07/10 17:30:00 EDT
1163    Tag: Release_110_34
1164    Description:
1165    
1166    Minor tweak to 110.34 (re-tagged):
1167    
1168      - README.html file added to CVS repository
1169      - runtime compiles properly under FreeBSD 3.X and 4.X
1170    
1171    ----------------------------------------------------------------------
1172    Name: Matthias Blume
1173    Date: 2001/07/10 17:30:00 EDT
1174    Tag: Release_110_34
1175    Description:
1176    
1177    New version number (110.34). New bootfiles.
1178    
1179    ----------------------------------------------------------------------
1180    Name: Matthias Blume
1181    Date: 2001/07/09 16:00:00 EDT
1182    Tag: blume-20010709-more-varargs
1183    Description:
1184    
1185    I changed the handling of varargs in ml-nlffigen again:
1186    The ellipsis ... will now simply be ignored (with an accompanying warning).
1187    
1188    The immediate effect is that you can actually call a varargs function
1189    from ML -- but you can't actually supply any arguments beyond the ones
1190    specified explicitly.  (For example, you can call printf with its format
1191    string, but you cannot pass additional arguments.)
1192    
1193    This behavior is only marginally more useful than the one before, but
1194    it has the advantage that a function or, more importantly, a function
1195    type never gets dropped on the floor, thus avoiding follow-up problems with
1196    other types that refer to the offending one.
1197    
1198    ----------------------------------------------------------------------
1199    Name: Matthias Blume
1200    Date: 2001/07/09 11:25:00 EDT
1201    Tag: blume-20010709-varargs
1202    Description:
1203    
1204    1. ckit-lib.cm now exports structure Error
1205    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1206       with a warning accompanied by a source location.  Moreover, it
1207       merely skips the offending function or type and proceeds with the
1208       rest of its work.u  As a result, one can safely feed C code containing
1209       "..." to ml-nlffigen.
1210    3. There are some internal improvements to CM, providing slightly
1211       more general string substitutions in the tools subsystem.
1212    
1213    ----------------------------------------------------------------------
1214    Name: Matthias Blume
1215    Date: 2001/06/27 15:10:00 EDT
1216    Tag: blume-20010627-concur
1217    Description:
1218    
1219    Fixed a small bug in CM's handling of parallel compilation.
1220    (You could observe the bug by Control-C-interrupting an ordinary
1221    CMB.make or CM.stabilize and then attaching some compile servers.
1222    The result was that all of a sudden the previously interrupted
1223    compilation would continue on its own.  This was because of
1224    an over-optimization: CM did not bother to clean out certain queues
1225    when no servers were attached "anyway", resulting in the contents
1226    of these queues to grab control when new servers did get attached.)
1227    
1228    There is also another minor update to the CM manual.
1229    
1230    ----------------------------------------------------------------------
1231    Name: Matthias Blume
1232    Date: 2001/06/26 16:15:00 EDT
1233    Tag: blume-20010626-cmdoc
1234    Description:
1235    
1236    Minor typo fixed in CM manual (syntax diagram for libraries).
1237    
1238    ----------------------------------------------------------------------
1239    Name: Matthias Blume
1240    Date: 2001/06/25 22:55:00 EDT
1241    Tag: blume-20010625-x86pc
1242    Description:
1243    
1244    Fixed a nasty bug in the X86 assembly code that caused signal
1245    handlers to fail (crash) randomly.
1246    
1247    ----------------------------------------------------------------------
1248    Name: Matthias Blume
1249    Date: 2001/06/25 12:05:00 EDT
1250    Tag: blume-20010625-nlffigen
1251    Description:
1252    
1253    This update fixes a number of minor bugs in ml-nlffigen as reported by
1254    Nick Carter <nbc@andrew.cmu.edu>.
1255    
1256      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1257      2. Default names for generated files are now derived from the name of
1258         the C file *without its directory*.  In particular, this causes generated
1259         files to be placed locally even if the C file is in some system directory.
1260      3. Default names for generated signatures and structures are also derived
1261         from the C file name without its directory.  This avoids silly things
1262         like "structure GL/GL".
1263         (Other silly names are still possible because ml-nlffigen does not do
1264          a thorough check of whether generated names are legal ML identifiers.
1265          When in doubt, use command line arguments to force particular names.)
1266    
1267    ----------------------------------------------------------------------
1268    Name: Matthias Blume
1269    Date: 2001/06/21 12:25:00 EDT
1270    Tag: blume-20010621-eXene
1271    Description:
1272    
1273    eXene now compiles and (sort of) works again.
1274    
1275    The library name (for version > 110.33) is $/eXene.cm.
1276    
1277    I also added an new example in src/eXene/examples/nbody.  See the
1278    README file there for details.
1279    
1280    ----------------------------------------------------------------------
1281    Name: Matthias Blume
1282    Date: 2001/06/20 16:40:00 EDT
1283    Tag: blume-20010620-cml
1284    Description:
1285    
1286    CML now compiles and works again.
1287    
1288    Libraries (for version > 110.33):
1289    
1290      $cml/cml.cm            Main CML library.
1291      $cml/basis.cm          CML's version of $/basis.cm.
1292      $cml/cml-internal.cm   Internal helper library.
1293      $cml/core-cml.cm       Internal helper library.
1294      $cml-lib/trace-cml.cm  Tracing facility.
1295      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1296    
1297    The installer (config/install.sh) has been taught how to properly
1298    install this stuff.
1299    
1300    ----------------------------------------------------------------------
1301    Name: Matthias Blume
1302    Date: 2001/06/19 17:55:00 EDT
1303    Tag: blume-20010619-instantiate
1304    Description:
1305    
1306    This un-breaks the fix for bug 1432.
1307    (The bug was originally fixed in 110.9 but I broke it again some
1308    time after that.)
1309    
1310    ----------------------------------------------------------------------
1311    Name: Matthias Blume
1312    Date: 2001/06/19 17:25:00 EDT
1313    Tag: blume-20010619-signals
1314    Description:
1315    
1316    This should (hopefully) fix the long-standing signal handling bug.
1317    (The runtime system was constructing a continuation record with an
1318    incorrect descriptor which would cause the GC to drop data on the floor...)
1319    
1320    ----------------------------------------------------------------------
1321    Name: Matthias Blume
1322    Date: 2001/06/15 15:05:00 EDT
1323    Tag: blume-20010615-moresparc
1324    Description:
1325    
1326    Here is a short late-hour update related to Sparc c-calls:
1327    
1328     -- made handling of double-word arguments a bit smarter
1329    
1330     -- instruction selection phase tries to collapse certain clumsily
1331        constructed ML-Trees; typical example:
1332    
1333            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1334    
1335        This currently has no further impact on SML/NJ since mlriscGen does
1336        not seem to generate such patterns in the first place, and c-calls
1337        (which did generate them in the beginning) has meanwhile been fixed
1338        so as to avoid them as well.
1339    
1340    ----------------------------------------------------------------------
1341    Name: Matthias Blume
1342    Date: 2001/06/15 15:05:00 EDT
1343    Tag: blume-20010615-sparc
1344    Description:
1345    
1346    The purpose of this update is to provide an implementation of NLFFI
1347    on Sparc machines.
1348    
1349    Here are the changes in detail:
1350    
1351       * src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
1352       the Sparc implementation of the c-calls API.
1353       * The Sparc backend of SML/NJ has been modified to uniformely use %fp
1354       for accessing the ML frame.  Thus, we have a real frame pointer and
1355       can freely modify %sp without need for an omit-frame-ptr phase.
1356       The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
1357       this case.
1358       * ml-nlffigen has been taught to produce code for different architectures
1359       and calling conventions.
1360       * In a way similar to what was done in the x86 case, the Sparc
1361       backend uses its own specific extension to mltree.  (For example,
1362       it needs to be able to generate UNIMP instructions which are part
1363       of the calling convention.)
1364       * ml-nlffi-lib was reorganized to make it more modular (in particular,
1365       to make it easier to plug in new machine- and os-dependent parts).
1366    
1367    There are some other fairly unrelated bug fixes and cleanups as well:
1368    
1369       * I further hacked the .cm files for MLRISC tools (like MDLGen) so
1370       that they properly share their libraries with existing SML/NJ libraries.
1371       * I fixed a minor cosmetic bug in CM, supressing certain spurious
1372       follow-up error messages.
1373       * Updates to CM/CMB documentation.
1374    
1375    TODO items:
1376    
1377       * MLRISC should use a different register as its asmTemp on the Sparc.
1378         (The current %o2 is a really bad choice because it is part of the
1379         calling conventions, so things might interfere in unexpected ways.)
1380    
1381    ----------------------------------------------------------------------
1382    Name: Matthias Blume
1383    Date: 2001/06/07
1384    Tag: blume-20010607-calls
1385    Description:
1386    
1387    A number of internal changes related to C calls and calling conventions:
1388    
1389    1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
1390       number of bytes popped implicitly (by the callee).  In most cases
1391       this field is 0 but on x86/win32 it is some non-zero value.  This
1392       is information provided for the benefit of the "omit-frameptr" pass.
1393    2. The CALL instruction on the x86 carries a similar "pops" field.
1394       The instruction selection phase copies its value from the ML-Tree
1395       CALL statement.
1396    3. On all other architectures, the instruction selection phase checks
1397       whether "pops=0" and complains if not.
1398    4. The c-calls implementation for x86 now accepts two calling conventions:
1399       "ccall" and "stdcall".  When "ccall" is selected, the caller cleans
1400       up after the call and pops is set to 0.  For "stdcall", the caller
1401       does nothing, leaving the cleanup to the callee; pops is set to
1402       the number of bytes that were pushed onto the stack.
1403    5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
1404       distinguish between "ccall" and "stdcall".
1405    6. The UNIMP instruction has been added to the supported Sparc instruction
1406       set. (This is needed for implementing the official C calling convention
1407       on this architecture.)
1408    7. I fixed some of the .cm files under src/MLRISC/Tools to make them
1409       work with the latest CM.
1410    
1411    ----------------------------------------------------------------------
1412    Name: Matthias Blume
1413    Date: 2001/06/05 15:10:00 EDT
1414    Tag: blume-20010605-cm-index
1415    Description:
1416    
1417    0. The "lambdasplit" parameter for class "sml" in CM has been documented.
1418    
1419    1. CM can now generate "index files".  These are human-readable files
1420       that list on a per-.cm-file basis each toplevel symbol defined or
1421       imported.  The location of the index file for
1422       <p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
1423       To enable index-file generation, set CM.Control.generate_index to true
1424       or export an environment-symbol: export CM_GENERATE_INDEX=true.
1425    
1426       The CM manual has been updated accordingly.
1427    
1428    2. I made some slight modifications to the c-calls API in MLRISC.
1429    
1430         a) There is now a callback to support saving/restoring of
1431            dedicated but caller-save registers around the actual call
1432            instruction.
1433         b) One can optionally specify a comment-annotation for the
1434            call instruction.
1435    
1436    3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
1437       (For example, the comment annotation shows the C prototype of
1438        the function being called.)
1439    
1440    ----------------------------------------------------------------------
1441    Name: Matthias Blume
1442    Date: 2001/06/01 13:30:00 EDT
1443    Tag: blume-20010601-nlffi-cleanup
1444    Description:
1445    
1446    This is mostly a cleanup of MLFFI stuff:
1447    
1448       - some signature files have been put into a more exposed place
1449       - the ugly 'f type parameter is gone (simplifies types tremendously!)
1450       - ml-nlffigen changed accordingly
1451       - tutorial updated
1452    
1453    Other changes:
1454    
1455       - author's affiliation in CM manual(s) updated
1456       - some more recognized keywords added to Allen's sml.sty
1457    
1458    ----------------------------------------------------------------------
1459    Name: Matthias Blume
1460    Date: 2001/05/25 15:30:00 EDT
1461    Tag: blume-20010525-iptr
1462    Description:
1463    
1464      - put the official 110.33-README (as it appears on the ftp server) under
1465        CVS
1466      - fixed a small bug related to incomplete pointer types in
1467        ml-nlffigen
1468      - small cosmetic change to the ml-nlffi-lib's "arr" type constructor
1469        (it does not need the 'f type parameter)
1470    
1471    ----------------------------------------------------------------------
1472    Name: Matthias Blume
1473    Date: 2001/05/23 14:30:00 EDT
1474    Tag: Release_110_33
1475    Description:
1476    
1477    New version number (110.33).  New bootfiles.
1478    
1479    ----------------------------------------------------------------------
1480    Name: Matthias Blume
1481  Date: 2001/05/22 18:06:00 EDT  Date: 2001/05/22 18:06:00 EDT
1482  Tag: blume-20010522-targets  Tag: blume-20010522-targets
1483  Description:  Description:
# Line 2270  Line 3734 
3734       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3735       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3736       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
3737       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3738       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3739       Dave's hands. :)       Dave's hands. :)
3740    

Legend:
Removed from v.833  
changed lines
  Added in v.1063

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