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 840, Fri Jun 15 19:05:19 2001 UTC revision 1060, Mon Feb 11 20:55:09 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: 2002/02/11 15:55:00 EST
18    Tag: blume-20020211-gensym
19    Description:
20    
21    Added a "-gensym" option to command line of ml-nlffigen.  This can be
22    used to specify a "stem" -- a string that is inserted in all "gensym'd"
23    names (ML structure names that correspond to unnamed C structs, unions,
24    and enums), so that separate runs of ml-nlffigen do not clash.
25    
26    ----------------------------------------------------------------------
27    Name: Matthias Blume
28    Date: 2002/02/11 12:05:00 EST
29    Tag: blume-20020211-gensml
30    Description:
31    
32    A quick fix for a problem with GenSML (in the pgraph-util library):
33    Make generation of toplevel "local" optional.  (Strictly speaking,
34    signature definitions within "local" are not legal SML.)
35    
36    Other than that: updates to INSTALL and cm/TODO.
37    
38    ----------------------------------------------------------------------
39    Name: Matthias Blume
40    Date: 2002/02/08 15:00:00 EST
41    Tag: blume-20020208-uniquepid
42    Description:
43    
44    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
45    
46    1. The installer (config/install.sh) has gotten smarter:
47    
48         - Configuration options are a bit easier to specify now
49           (in config/targets).
50         - Bug in recognizing .tar.bz2 files fixed.
51         - Installer automatically resolves dependencies between
52           configuration options (e.g., if you ask for eXene, you will
53           also get cml -- regardless whether you asked for it or not).
54         - Installer can run in "quieter mode" by setting the environment
55           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
56           "completely silent", though.
57         - Build HashCons library as part of smlnj-lib.
58    
59    2. A new scheme for assigning persistent identifiers to compilation
60       units (and, by extension, to types etc.) has been put into place.
61       This fixes a long-standing bug where types and even dynamic values
62       can get internally confused, thereby compromising type safety
63       (abstraction) and dynamic correctness.  See
64    
65         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
66    
67       for an example of how things could go wrong until now.
68    
69       The downside of the new scheme is that pids are not quite as
70       persistent as they used to be: CM will generate a fresh pid
71       for every compilation unit that it thinks it sees for the first
72       time.  That means that if you compile starting from a clean, fresh
73       source tree at two different times, you end up with different
74       binaries.
75    
76       Cutoff recompilation, however, has not been compromised because
77       CM keeps pid information in special caches between runs.
78    
79    ----------------------------------------------------------------------
80    Name: Lal George
81    Date: 2002/02/07 15:34:13 EST 2002
82    Tag: <none>
83    Description:
84    
85    Compilers that generate assembly code may produce  global labels
86    whose value is resolved at link time. The various peephole optimization
87    modules did not take this in account.
88    
89    TODO. The Labels.addrOf function should really return an option
90    type so that clients are forced to deal with this issue, rather
91    than an exception being raised.
92    
93    ----------------------------------------------------------------------
94    Name: Lal George
95    Date: 2002/02/06 13:55:02 EST
96    Tag: george-20020206-ra-breakup
97    Description:
98    
99    1. A bug fix from Allen.
100    
101        A typo causes extra fstp %st(0)'s to be generated at compensation
102        edges, which might cause stack underflow traps at runtime.  This
103        occurs in fft where there are extraneous fstps right before the 'into'
104        trap instruction (in this case they are harmless since none of the
105        integers overflow.)
106    
107    2. Pulled out various utility modules that were embedded in the modules
108       of the register allocator. I need these modules for other purposes, but
109       they are not complete enough to put into a library (just yet).
110    ----------------------------------------------------------------------
111    Name: Matthias Blume
112    Date: 2002/01/31 16:05:00 EST
113    Tag: blume-20020131-sparc-ccalls
114    Description:
115    
116    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
117       of extra stack space by mistake.  Fixed.
118    
119    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
120    
121    ----------------------------------------------------------------------
122    Name: Allen Leung
123    Date: 2002/01/30
124    Tag: leunga-20020130-nowhere-bug-fix
125    Description:
126    
127       MLRISC bug fixes:
128       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
129       2. MachineInt.fromString was negating its value.
130    
131    ----------------------------------------------------------------------
132    Name: Matthias Blume
133    Date: 2002/01/29
134    Tag: blume-20020129-INSTALL
135    Description:
136    
137    - Added somewhat detailed installation instructions (file INSTALL).
138    - Fixed curl-detection bug in config/install.sh.
139    - It is now possible to select the URL getter using the URLGETTER
140      environment variable:
141    
142          not set / "unknown"      --> automatic detection (script tries wget,
143                                       curl, and lynx)
144          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
145                                       how to properly invoke them)
146          other                    --> use $URLGETTER directly, it must take
147                                       precisely two command-line arguments
148                                       (source URL and destination file name)
149    
150    ----------------------------------------------------------------------
151    Name: Matthias Blume
152    Date: 2002/01/28
153    Tag: blume-20020128-sparc-ccalls
154    Description:
155    
156    - Fixed problem with calculation of "used" registers in sparc-c-calls.
157    - Make use of the allocParam argument in sparc-c-calls.
158    
159    ----------------------------------------------------------------------
160    Name: Matthias Blume
161    Date: 2002/01/28
162    Tag: blume-20020128-allocParam
163    Description:
164    
165    John Reppy:  Changes c-calls API to accept client-callback for
166    allocating extra stack space.
167    me: Corresponding changes to mlriscGen (using a dummy argument that
168        does not change the current behavior).
169    
170    ----------------------------------------------------------------------
171    Name: Matthias Blume
172    Date: 2002/01/28 12:00:00
173    Tag: Release_110_38
174    Description:
175    
176    This time for real!!!
177    
178    ----------------------------------------------------------------------
179    Name: Matthias Blume
180    Date: 2002/01/28 10:56:00 EST
181    Tag: blume-20020128-retraction
182    Description:
183    
184    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
185       with blume-Release_110_38-retracted.)
186    
187    1. Fixed a problem with incorrect rounding modes in real64.sml.
188       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
189    
190    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
191       and enums fixed.  The naming of corresponding ML identifiers should
192       now be consistent again.
193    
194    ----------------------------------------------------------------------
195    Name: Allen Leung
196    Date: 2002/01/27
197    Tag: leunga-20020127-nowhere
198    Description:
199    
200       Added a target called nowhere in the configuration scripts.
201       Enabling this will build the MLRISC 'nowhere' tool (for translating
202       programs with where-clauses into legal SML code) during installation.
203    
204    ----------------------------------------------------------------------
205    Name: Matthias Blume
206    Date: 2002/01/25 21:27:00 EST
207    Tag: blume-Release_110_38-retracted
208    Description:
209    
210    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
211    
212    README will be added later.
213    
214    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
215    (Release_110_38) removed.  Reason: Last-minute bug fixes.
216    
217    ----------------------------------------------------------------------
218    Name: Matthias Blume
219    Date: 2002/01/25
220    Tag: blume-20020125-ffi
221    Description:
222    
223    A large number of tweaks and improvements to ml-nlffi-lib and
224    ml-nlffigen:
225    
226       - ML represenation types have been streamlined
227       - getter and setter functions work with concrete values, not abstract
228         ones where possible
229       - ml-nlffigen command line more flexible (see README file there)
230       - some bugs have been fixed (hopefully)
231    
232    ----------------------------------------------------------------------
233    Name: Lal George
234    Date: 2002/01/24
235    Tag: george-20020124-risc-ra-interface
236    Description:
237    
238       There is a dramatic simplification in the interface to the
239       register allocator for RISC architectures as a result of making
240       parallel copy instructions explicit.
241    
242    ----------------------------------------------------------------------
243    Name: Matthias Blume
244    Date: 2002/01/22
245    Tag: blume-20020122-x86-ccalls
246    Description:
247    
248    Bug fix for c-calls on x86 (having to do with how char- and
249    short-arguments are being handled).
250    
251    ----------------------------------------------------------------------
252    Name: Matthias Blume
253    Date: 2002/01/21
254    Tag: blume-20020121-ff
255    Description:
256    
257    Another day of fiddling with the FFI...
258    
259    1. Bug fix/workaround:  CKIT does not complain about negative array
260       dimensions, so ml-nlffigen has to guard itself against this possibility.
261       (Otherwise a negative dimension would send it into an infinite loop.)
262    
263    2. Some of the abstract types (light objects, light pointers, most "base"
264       types) in structure C are now eqtypes.
265    
266    3. Added constructors and test functions for NULL function pointers.
267    
268    ----------------------------------------------------------------------
269    Name: Matthias Blume
270    Date: 2002/01/18
271    Tag: blume-20020118-ready-for-new-release
272    Description:
273    
274    Made config/srcarchiveurl point to a new place.  (Will provide boot
275    files shortly.)
276    
277    Maybe we christen this to be 110.38?
278    
279    ----------------------------------------------------------------------
280    Name: Matthias Blume
281    Date: 2002/01/18
282    Tag: blume-20020118-more-ffifiddle
283    Description:
284    
285    Today's FFI fiddling:
286    
287      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
288        C.Set.* that use concrete (MLRep.*) arguments and results instead
289        of abstract ones.
290    
291      - Provided word-style bit operations etc. for "int" representation
292        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
293        and Long.
294    
295    ----------------------------------------------------------------------
296    Name: Matthias Blume
297    Date: 2002/01/18
298    Tag: blume-20020118-use-x86-fp
299    Description:
300    
301    Now that x86-fast-fp seems to be working, I turned it back on again
302    by default.  (Seems to work fine now, even with the FFI.)
303    
304    Other than that, I added some documentation about the FFI to
305    src/ml-nlffigen/README and updated the FFI test examples in
306    src/ml-nlffi-lib/Tests/*.
307    
308    ----------------------------------------------------------------------
309    Name: Allen Leung
310    Date: 2002/01/17
311    Tag: leunga-20020117-x86-fast-fp-call
312    Description:
313    
314       1. Fixed a problem with handling return fp values when x86's fast fp
315          mode is turned on.
316    
317       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
318          of %f32.
319    
320       3. Added a constructor INT32lit to the ast of MLRISC tools.
321    
322    ----------------------------------------------------------------------
323    Name: Matthias Blume
324    Date: 2002/01/16
325    Tag: blume-20020116-ffifiddle
326    Description:
327    
328    More fiddling with the FFI interface:
329    
330     - Make constness 'c instead of rw wherever possible.  This eliminates
331       the need for certain explicit coercions.  (However, due to ML's
332       value polymorphism, there will still be many cases where explicit
333       coercions are necessary.  Phantom types are not the whole answer
334       to modeling a subtyping relationship in ML.)
335    
336     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
337    
338     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
339       clumsy workaround for varargs problem.)
340    
341    ----------------------------------------------------------------------
342    Name: Lal George
343    Date: 2002/01/15
344    Tag: <none>
345    Description:
346    
347    1. Since COPY instructions are no longer native to the architecture,
348       a generic functor can be used to implement the expandCopies function.
349    
350    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
351       TEXT segment.
352    
353    ----------------------------------------------------------------------
354    Name: Matthias Blume
355    Date: 2002/01/15
356    Tag: blume-20020115-ffiupdates
357    Description:
358    
359    1. Fix for bug resulting in single-precision float values being returned
360       incorrectly from FFI calls.
361    
362    2. Small modifications to C FFI API:
363    
364        - memory-allocation routines return straight objects (no options)
365          and raise an exception in out-of-memory situations
366        - unsafe extensions to cast between function pointers and pointers
367          from/to ints
368        - added structure C_Debug as an alternative to structure C where
369          pointer-dereferencing (|*| and |*!) always check for null-pointers
370        - added open_lib' to DynLinkage;  open_lib' works like open_lib
371          but also takes a (possibly empty) list of existing library handles
372          that the current library depends on
373    
374    ----------------------------------------------------------------------
375    Name: Matthias Blume
376    Date: 2002/01/10
377    Tag: blume-20020110-newffigen
378    Description:
379    
380    1. Updates to portable graph code.
381    
382    2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
383       more scalable now so that even huge interfaces such as the one
384       for GTK compile in finite time and space. :-)
385       See src/ml-nlffigen/README for details on what's new.
386    
387    ----------------------------------------------------------------------
388    Name: Lal George
389    Date: 2001/01/09 14:31:35 EST 2002
390    Tag: george-20011206-rm-native-copy
391    Description:
392    
393            Removed the native COPY and FCOPY instructions
394            from all the architectures and replaced it with the
395            explicit COPY instruction from the previous commit.
396    
397            It is now possible to simplify many of the optimizations
398            modules that manipulate copies. This has not been
399            done in this change.
400    
401    ----------------------------------------------------------------------
402    Name: Lal George
403    Date: 2001/12/06 16:50:13 EST 2001
404    Tag: george-20011206-mlrisc-instruction
405    Description:
406    
407    Changed the representation of instructions from being fully abstract
408    to being partially concrete. That is to say:
409    
410      from
411            type instruction
412    
413      to
414            type instr                              (* machine instruction *)
415    
416            datatype instruction =
417                LIVE of {regs: C.cellset, spilled: C.cellset}
418              | KILL of {regs: C.cellset, spilled: C.cellset}
419              | COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
420              | ANNOTATION of {i: instruction, a: Annotations.annotation}
421              | INSTR of instr
422    
423    This makes the handling of certain special instructions that appear on
424    all architectures easier and uniform.
425    
426    LIVE and KILL say that a list of registers are live or killed at the
427    program point where they appear. No spill code is generated when an
428    element of the 'regs' field is spilled, but the register is moved to
429    the 'spilled' (which is present, more for debugging than anything else).
430    
431    LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
432    We used to generate:
433    
434            DEFFREG f1
435            f1 := f2 + f3
436            trapb
437    
438    but now generate:
439    
440            f1 := f2 + f3
441            trapb
442            LIVE {regs=[f1,f2,f3], spilled=[]}
443    
444    Furthermore, the DEFFREG (hack) required that all floating point instruction
445    use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
446    defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
447    in a cleaner alpha implementation. (Hopefully, intel will not get rid of
448    this architecture).
449    
450    COPYXXX is intended to replace the parallel COPY and FCOPY  available on
451    all the architectures. This will result in further simplification of the
452    register allocator that must be aware of them for coalescing purposes, and
453    will also simplify certain aspects of the machine description that provides
454    callbacks related to parallel copies.
455    
456    ANNOTATION should be obvious, and now INSTR represents the honest to God
457    machine instruction set!
458    
459    The <arch>/instructions/<arch>Instr.sml files define certain utility
460    functions for making porting easier -- essentially converting upper case
461    to lower case. All machine instructions (of type instr) are in upper case,
462    and the lower case form generates an MLRISC instruction. For example on
463    the alpha we have:
464    
465      datatype instr =
466         LDA of {r:cell, b:cell, d:operand}
467       | ...
468    
469      val lda : {r:cell, b:cell, d:operand} -> instruction
470        ...
471    
472    where lda is just (INSTR o LDA), etc.
473    
474    ----------------------------------------------------------------------
475    Name: Matthias Blume
476    Date: 2001/11/22 21:40:00 EST
477    Tag: Release_110_37
478    Description:
479    
480    Release 110.37.  This time for real.
481    
482    ----------------------------------------------------------------------
483    Name: Matthias Blume
484    Date: 2001/11/21 16:35:00 EST
485    Tag: blume-20011121-foot-in-mouth
486    Description:
487    
488    Removed the "Release_110_37" tag because of a serious bug.
489    This will be re-tagged once the bug is fixed.
490    
491    ----------------------------------------------------------------------
492    Name: Matthias Blume
493    Date: 2001/11/21 16:14:00 EST
494    Tag: blume-20011121-forgottenfile
495    Description:
496    
497    Forgot to add a file.  (Just a .tex-file -- part of
498    the CM manual source.)
499    
500    ----------------------------------------------------------------------
501    Name: Matthias Blume
502    Date: 2001/11/21 16:10:00 EST
503    Tag: blume-20011121-invalid_110_37
504    Description:
505    
506    Note: I removed the original tag "Release_110_37" from this commit
507          because we found a serious bug in all non-x86 backends.
508          - Matthias
509    
510    1. Modifications to the SML/NJ code generator and to the runtime system
511       so that code object name strings are directly inserted into code
512       objects at code generation time.  The only business the runtime system
513       has with this is now to read the name strings on occasions.
514       (The encoding of the name string has also changed somewhat.)
515    
516    2. CM now implements a simple "set calculus" for specifying export lists.
517       In particular, it is now possible to refer to the export lists of
518       other libraries/groups/sources and form unions as well as differences.
519       See the latest CM manual for details.
520    
521    3. An separate notion of "proxy" libraries has again be eliminated from
522       CM's model.  (Proxy libraries are now simply a special case of using
523       the export list calculus.)
524    
525    4. Some of the existing libraries now take advantage of the new set
526       calculus.
527       (Notice that not all libraries have been converted because some
528       of the existing .cm-files are supposed to be backward compatible
529       with 110.0.x.)
530    
531    5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
532       "General.exnMessage"!  The former relies on a certain hook to be
533       initialized, and that often does not happen in the stand-alone case.)
534    
535    ----------------------------------------------------------------------
536    Name: Lal George
537    Date: 2001/11/21  13:56:18 EST
538    Tag: george-2001121-pseudo-ops
539    Description:
540    
541      Implemented a complete redesign of MLRISC pseudo-ops. Now there
542      ought to never be any question of incompatabilities with
543      pseudo-op syntax expected by host assemblers.
544    
545      For now, only modules supporting GAS syntax are implemented
546      but more should follow, such as MASM, and vendor assembler
547      syntax, e.g. IBM as, Sun as, etc.
548    
549    ----------------------------------------------------------------------
550    Name: Matthias Blume
551    Date: 2001/11/14 11:52:00 EST
552    Tag: blume-20011114-srcname
553    Description:
554    
555    1. Routed the name of the current source file to mlriscgen where it
556       should be directly emitted into the code object.  (This last part
557       is yet to be done.)
558    
559    2. Some cleanup of the pgraph code to make it match the proposal that
560       I put out the other day.  (The proposal notwithstanding, things are
561       still in flux here.)
562    
563    ----------------------------------------------------------------------
564    Name: Lal George
565    Date: 2001/11/14 09:44:04 EST
566    Tag:
567    Description:
568    
569      Fix for a backpatching bug reported by Allen.
570    
571      Because the boundary between short and long span-dependent
572      instructions is +/- 128, there are an astounding number of
573      span-dependent instructions whose size is over estimated.
574    
575      Allen came up with the idea of letting the size of span
576      dependent instructions be non-monotonic, for a maxIter
577      number of times, after which the size must be monotonically
578      increasing.
579    
580      This table shows the number of span-dependent instructions
581      whose size was over-estimated as a function of maxIter, for the
582      file Parse/parse/ml.grm.sml:
583    
584         maxIter            # of instructions:
585            10                      687
586            20                      438
587            30                      198
588            40                        0
589    
590      In compiling the compiler, there is no significant difference in
591      compilation speed between maxIter=10 and maxIter=40. Actually,
592      my measurements showed that maxIter=40 was a tad faster than
593      maxIter=10! Also 96% of the  files in the compiler reach a fix
594      point within 13 iterations, so fixing maxIter at 40, while high,
595      is okay.
596    
597    ----------------------------------------------------------------------
598    Name: Matthias Blume
599    Date: 2001/10/31 15:25:00 EST
600    Tag: blume-20011031-pgraph
601    Description:
602    
603    CKIT:
604    * Changed the "Function" constructor of type Ast.ctype to carry optional
605      argument identifiers.
606    * Changed the return type of TypeUtil.getFunction accordingly.
607    * Type equality ignores the argument names.
608    * TypeUtil.composite tries to preserve argument names but gives up quickly
609      if there is a mismatch.
610    
611    installation script:
612    * attempts to use "curl" if available (unless "wget" is available as well)
613    
614    CM:
615    * has an experimental implementation of "portable graphs" which I will
616      soon propose as an implementation-independent library format
617    * there are also new libraries $/pgraph.cm and $/pgraph-util.cm
618    
619    NLFFI-LIB:
620    * some cleanup (all cosmetic)
621    
622    NLFFIGEN:
623    * temporarily disabled the mechanism that suppresses ML output for
624      C definitions whose identifiers start with an underscore character
625    * generate val bindings for enum constants
626    * user can request that only one style (light or heavy) is being used;
627      default is to use both (command-line arguments: -heavy and -light)
628    * fixed bug in handling of function types involving incomplete pointers
629    * generate ML entry points that take record arguments (i.e., using
630      named arguments) for C functions that have a prototype with named
631      arguments
632      (see changes to CKIT)
633    
634    ----------------------------------------------------------------------
635    Name: Allen Leung
636    Date: 2001/10/27 20:34:00 EDT
637    Tag: leunga-20011027-x86-fast-fp-call
638    Description:
639    
640       Fixed the bug described in blume-20010920-slowfp.
641    
642       The fix involves
643          1. generating FCOPYs in FSTP in ia32-svid
644          2. marking a CALL with the appropriate annotation
645    
646    ----------------------------------------------------------------------
647    Name: Matthias Blume
648    Date: 2001/10/16 11:32:00 EDT
649    Tag: blume-20011016-netbsd
650    Description:
651    
652    Underscore patch from Chris Richards (fixing problem with compiling
653    runtime system under recent NetBSD).
654    
655    ----------------------------------------------------------------------
656    Name: Allen Leung
657    Date: 2001/10/12 17:18:32 EDT 2001
658    Tag: leung-20011012-x86-printflowgraph
659    Description:
660    
661    X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
662    
663    ----------------------------------------------------------------------
664    Name: Lal George
665    Date: 2001/10/11 23:51:34 EDT
666    Tag: george-20011011-too-many-instrs
667    Description:
668    
669    The representation of a program point never expected to see more
670    than 65536 instructions in a basic block!
671    
672    ----------------------------------------------------------------------
673    Name: Lal George
674    Date: 2001/10/09 09:41:37 EDT
675    Tag: george-20011008-mlrisc-labels
676    Description:
677    
678    Changed the machine description files to support printing of
679    local and global labels in assembly code, based on host assembler
680    conventions.
681    
682    ----------------------------------------------------------------------
683    Name: Matthias Blume
684    Date: 2001/09/25 15:25:00 EDT
685    Tag: blume-20010925-exninfo
686    Description:
687    
688    I provided a non-hook implementation of exnName (at the toplevel) and
689    made the "dummy" implementation of exnMessage (at the toplevel) more
690    useful: if nothing gets "hooked in", then at least you are going to
691    see the exception name and a message indicating why you don't see more.
692    
693    [For the time being, programs that need exnMessage and want to use
694    ml-build should either use General.exnMessage (strongly recommended) or
695    refer to structure General at some other point so that CM sees a
696    static dependency.]
697    
698    [Similar remarks go for "print" and "use":  If you want to use their
699    functionality in stand-alone programs generated by ml-build, then use
700    TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
701    
702    ----------------------------------------------------------------------
703    Name: Matthias Blume
704    Date: 2001/09/20 17:28:00 EDT
705    Tag: blume-20010920-slowfp
706    Description:
707    
708    Allen says that x86-fast-fp is not safe yet, so I turned it off again...
709    
710    ----------------------------------------------------------------------
711    Name: Matthias Blume
712    Date: 2001/09/20 17:20:00 EDT
713    Tag: blume-20010920-canonicalpaths
714    Description:
715    
716    0. Updated the BOOT file (something that I forgot to do earlier).
717    
718    1. Small internal change to CM so that it avoids "/../" in filenames
719       as much as possible (but only where it is safe).
720    
721    2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
722       that arguments that contain delimiters are passed through correctly.
723       This change also means that all "special" arguments of the form
724       @SMLxxx... must come first.
725    
726    3. Changed install script to put relative anchor names for tool commands
727       into pathconfig.
728    
729    ----------------------------------------------------------------------
730    Name: Matthias Blume
731    Date: 2001/09/18 15:35:00 EDT
732    Tag: blume-20010918-readme11036
733    Description:
734    
735    Added README files.
736    
737    ----------------------------------------------------------------------
738    Name: Matthias Blume
739    Date: 2001/09/18 11:45:00 EDT
740    Tag: Release_110_36 (retag)
741    Description:
742    
743    Fixed mistake in config/preloads. Retagged as 110.36.
744    
745    ----------------------------------------------------------------------
746    Name: Matthias Blume
747    Date: 2001/09/18 09:40:00 EDT
748    Tag: Release_110_36_orig (tag changed)
749    Description:
750    
751    New version (110.36).  New bootfiles.
752    
753    ----------------------------------------------------------------------
754    Name: Matthias Blume
755    Date: 2001/09/14 16:15:00 EDT
756    Tag: blume-20010914-x86fastfp
757    Description:
758    
759    John committed some changes that Allen made, in particular a (hopefully)
760    correctly working version of the x86-fp module.
761    
762    I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
763    flag to "true".  Everything seems to compile to a fixpoint ok, and
764    "mandelbrot" speeds up by about 15%.
765    
766    ----------------------------------------------------------------------
767    Name: Matthias Blume
768    Date: 2001/09/13 11:20:00 EDT
769    Tag: blume-20010913-minimal
770    Description:
771    
772    1. Stefan Monnier's patch to fix a miscompilation problem that
773       was brought to light by John Reppy's work on Moby.
774    
775    2. Implemented a minimal "structure Compiler" that contains just
776       "version" and "architecture".  The minimal version will be
777       available when the full version is not.  This is for backward-
778       compatibility with code that wants to test Compiler.version.
779    
780    ----------------------------------------------------------------------
781    Name: Matthias Blume
782    Date: 2001/08/28 14:03:00 EDT
783    Tag: blume-20010828-ml-lex
784    Description:
785    
786    Fix for bug 1581, received from Neophytos Michael.
787    
788    ----------------------------------------------------------------------
789    Name: Matthias Blume
790    Date: 2001/08/27 11:20:00 EDT
791    Tag: blume-20010827-readme11035
792    Description:
793    
794    Fleshed out the README file for 110.35.
795    
796    ----------------------------------------------------------------------
797    Name: Matthias Blume
798    Date: 2001/08/24 17:10:00 EDT
799    Tag: Release_110_35
800    Description:
801    
802    New version number (110.35).  New bootfiles.
803    
804    ----------------------------------------------------------------------
805    Name: Lal George
806    Date: 2001/08/24 13:47:18 EDT 2001
807    Tag: george-20010824-MLRISC-graphs
808    Description:
809    
810     removed clusters from MLRISC completely and replaced with graphs.
811    
812    ----------------------------------------------------------------------
813    Name: Matthias Blume
814    Date: 2001/08/23 17:50:00 EDT
815    Tag: blume-20010823-toplevel
816    Description:
817    
818    - some reorganization of the code that implements various kinds of
819      environments in the compiler (static, dynamic, symbolic, combined)
820    - re-implemented the EnvRef module so that evalStream works properly
821      (if the stream contains references to "use", "CM.make", etc.)
822    - cleaned up evalloop.sml and interact.sml (but they need more cleaning)
823    
824    ----------------------------------------------------------------------
825    Name: Matthias Blume
826    Date: 2001/08/20 15:50 EDT
827    Tag: blume20010820-slipup
828    Description:
829    
830    I forgot to commit a few files.  Here they are...
831    
832    ----------------------------------------------------------------------
833    Name: Matthias Blume
834    Date: 2001/08/20 15:35:00 EDT
835    Tag: blume-20010820-debugprof
836    Description:
837    
838    !!!! NEW BOOTFILES !!!!
839    
840    This is another round of reorganizing the compiler sources.  This
841    time the main goal was to factor out all the "instrumentation"
842    passes (for profiling and backtracing) into their own library.
843    The difficulty was to do it in such a way that it does not depend
844    on elaborate.cm but only on elabdata.cm.
845    
846    Therefore there have been further changes to both elaborate.cm and
847    elabdata.cm -- more "generic" things have been moved from the former
848    to the latter.  As a result, I was forced to split the assignment
849    of numbers indicating "primtyc"s into two portions: SML-generic and
850    SML/NJ-specific.  Since it would have been awkward to maintain,
851    I bit the bullet and actually _changed_ the mapping between these
852    numbers and primtycs.  The bottom line of this is that you need
853    a new set of bin- and bootfiles.
854    
855    I have built new bootfiles for all architectures, so doing a fresh
856    checkout and config/install.sh should be all you need.
857    
858    The newly created library's name is
859    
860        $smlnj/viscomp/debugprof.cm
861    
862    and its sources live under
863    
864        src/compiler/DebugProf
865    
866    ----------------------------------------------------------------------
867    Name: Matthias Blume
868    Date: 2001/08/15 17:15:00 EDT
869    Tag: blume-20010815-compreorg
870    Description:
871    
872    This is a first cut at reorganizing the CM libraries that make up the
873    core of the compiler.  The idea is to separate out pieces that could
874    be used independently by tools, e.g., the parser, the typechecker, etc.
875    
876    The current status is a step in this direction, but it is not quite
877    satisfactory yet.  Expect more changes in the future.
878    
879    Here is the current (new) organization...
880    
881        What used to be $smlnj/viscomp/core.cm is now divided into
882        six CM libraries:
883    
884             $smlnj/viscomp/basics.cm
885                           /parser.cm
886                           /elabdata.cm
887                           /elaborate.cm
888                           /execute.cm
889                           /core.cm
890    
891        The CM files for these libraries live under src/system/smlnj/viscomp.
892        All these libraries are proxy libraries that contain precisely
893        one CM library component.  Here are the locations of the components
894        (all within the src/compiler tree):
895    
896             Basics/basics.cm
897             Parse/parser.cm
898             ElabData/elabdata.cm
899             Elaborator/elaborate.cm
900             Execution/execute.cm
901             core.cm
902    
903         [This organization is the same that has been used already
904         for a while for the architecture-specific parts of the visible
905         compiler and for the old version of core.cm.]
906    
907         As you will notice, many source files have been moved from their
908         respective original locations to a new home in one of the above
909         subtrees.
910    
911         The division of labor between the new libraries is the following:
912    
913             basics.cm:
914                - Simple, basic definitions that pertain to many (or all) of
915                  the other libraries.
916             parser.cm:
917                - The SML parser, producing output of type Ast.dec.
918                - The type family for Ast is also defined and exported here.
919             elabdata.cm:
920                - The datatypes that describe input and output of the elaborator.
921                  This includes types, absyn, and static environments.
922             elaborator.cm:
923                - The SML/NJ type checker and elaborator.
924                  This maps an Ast.dec (with a given static environment) to
925                  an Absyn.dec (with a new static environment).
926                - This libraries implements certain modules that used to be
927                  structures as functors (to remove dependencies on FLINT).
928             execute.cm:
929                - Everything having to do with executing binary code objects.
930                - Dynamic environments.
931             core.cm:
932                - SML/NJ-specific instantiations of the elaborator and MLRISC.
933                - Top-level modules.
934                - FLINT (this should eventually become its own library)
935    
936    Notes:
937    
938    I am not 100% happy with the way I separated the elaborator (and its
939    data structures) from FLINT.  Two instances of the same problem:
940    
941        1. Data structures contain certain fields that carry FLINT-specific
942           information.  I hacked around this using exn and the property list
943           module from smlnj-lib.  But the fact that there are middle-end
944           specific fields around at all is a bit annoying.
945    
946        2. The elaborator calculates certain FLINT-related information.  I tried
947           to make this as abstract as I could using functorization, but, again,
948           the fact that the elaborator has to perform calculations on behalf
949           of the middle-end at all is not nice.
950    
951        3. Having to used exn and property lists is unfortunate because it
952           weakens type checking.  The other alternative (parameterizing
953           nearly *everything*) is not appealing, though.
954    
955    I removed the "rebinding =" warning hack because due to the new organization
956    it was awkward to maintain it.  As a result, the compiler now issues some of
957    these warnings when compiling init.cmi during bootstrap compilation. On
958    the plus side, you also get a warning when you do, for example:
959       val op = = Int32.+
960    which was not the case up to now.
961    
962    I placed "assign" and "deref" into the _Core structure so that the
963    code that deals with the "lazy" keyword can find them there.  This
964    removes the need for having access to the primitive environment
965    during elaboration.
966    
967    ----------------------------------------------------------------------
968    Name: Matthias Blume
969    Date: 2001/08/13
970    Tag: blume-20010813-closures
971    Description:
972    
973    This fix was sent to us by Zhong Shao.  It is supposed to improve the
974    performance of certain loops by avoiding needless closure allocation.
975    
976    ----------------------------------------------------------------------
977    Name: Lal George
978    Date: 2001/07/31 10:03:23 EDT 2001
979    Tag: george-20010731-x86-fmalloc
980    Description: Fixed bug in x86 calls
981    
982        There was a bug where call instructions would mysteriously
983        vanish. The call instruction had to be one that returned
984        a floating point value.
985    
986    ----------------------------------------------------------------------
987    Name: Lal George
988    Date: 2001/07/19 16:36:29 EDT 2001
989    Tag: george-20010719-simple-cells
990    Description:
991    
992    I have dramatically simplified the interface for CELLS in MLRISC.
993    
994    In summary, the cells interface is broken up into three parts:
995    
996      1. CellsBasis : CELLS_BASIS
997    
998            CellsBasis is a top level structure and common for all
999            architectures.  it contains the definitions of basic datatypes
1000            and utility  functions over these types.
1001    
1002      2. functor Cells() : CELLS
1003    
1004            Cells generates an interface for CELLS that incorporates the
1005            specific resources on the target architecture, such as the
1006            presence of special register classes, their number and size,
1007            and various useful substructures.
1008    
1009      3. <ARCH>CELLS
1010    
1011            e.g. SparcCells: SPARCCELLS
1012    
1013            <ARCH>CELLS usually contains additional bindings for special
1014            registers  on the architecture, such as:
1015    
1016                    val r0 : cell           (* register zero *)
1017                    val y : cell            (* Y register *)
1018                    val psr : cell          (* processor status register *)
1019                    ...
1020    
1021            The structure returned by applying the Cells functor is opened
1022            in this interface.
1023    
1024    The main implication of all this is that the datatypes for cells is
1025    split between CellsBasis and CELLS -- a fairly simple change for user
1026    code.
1027    
1028    In the old scheme the CELLS interface had a definitional binding of
1029    the form:
1030    
1031            signature CELLS = sig
1032    
1033               structure CellsBasis = CellsBasis
1034    
1035               ...
1036    
1037            end
1038    
1039    With all the sharing constraints that goes on in MLRISC, this old
1040    design  quickly leads to errors such as:
1041    
1042            "structure definition spec inside of sharing ... "
1043    
1044    
1045    and appears to require an unacceptable amount of sharing and where
1046    constraint hackery.
1047    
1048    I think this error message (the interaction of definitional specs and
1049    sharing) requires more explanation on our web page.
1050    
1051    ----------------------------------------------------------------------
1052    Name: Matthias Blume
1053    Date: 2001/07/19 15:00:00 EDT
1054    Tag: blume-20010719-libreorg
1055    Description:
1056    
1057    This update puts together a fairly extensive but straightforward change
1058    to the way the libraries that implement the interactive system are
1059    organized:
1060    
1061       The biggest change is the elimination of structure Compiler.  As a
1062       replacement for this structure, there is now a CM library
1063       (known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
1064       that exports all the substructures of the original structure Compiler
1065       directly.  So instead of saying Compiler.Foo.bar one now simply
1066       says Foo.bar.  (The CM libraries actually export a collection of
1067       structures that is richer than the collection of substructures of
1068       structure Compiler.)
1069    
1070       To make the transition smooth, there is a separate library called
1071       $smlnj/compiler/compiler.cm which puts together and exports the
1072       original structure Compiler (or at least something very close to it).
1073    
1074       There are five members of the original structure Compiler
1075       that are not exported directly but which instead became members
1076       of a new structure Backend (described by signature BACKEND).  These are:
1077       structure Profile (: PROFILE), structure Compile (: COMPILE), structure
1078       Interact (: INTERACT), structure Machine (: MACHINE), and val
1079       architecture (: string).
1080    
1081       Structure Compiler.Version has become structure CompilerVersion.
1082    
1083       Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
1084       by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
1085       or x86, respectively.
1086       Each of these exports the same frontend structures that
1087       $smlnj/compiler.cm exports.  But they do not have a structure Backend
1088       and instead export some structure <Arch>Backend where <Arch> is Alpha32,
1089       Hppa, PPC, Sparc, or X86, respectively.
1090    
1091       Library $smlnj/compiler/all.cm exports the union of the exports of
1092       $smlnj/compiler/<arch>.cm
1093    
1094       There are no structures <Arch>Compiler anymore, use
1095       $smlnj/compiler/<arch>.cm instead.
1096    
1097       Library host-compiler-0.cm is gone.  Instead, the internal library
1098       that instantiates CM is now called cm0.cm.  Selection of the host
1099       compiler (backend) is no longer done here but. (Responsibility for it
1100       now lies with $smlnj/compiler/current.cm.  This seems to be more
1101       logical.)
1102    
1103       Many individual files have been moved or renamed.  Some files have
1104       been split into multiple files, and some "dead" files have been deleted.
1105    
1106    Aside from these changes to library organization, there are also changes
1107    to the way the code itself is organized:
1108    
1109       Structure Binfile has been re-implemented in such a way that it no
1110       longer needs any knowledge of the compiler.  It exclusively deals
1111       with the details of binfile layout.  It no longer invokes the
1112       compiler (for the purpose of creating new prospective binfile
1113       content), and it no longer has any knowledge of how to interpret
1114       pickles.
1115    
1116       Structure Compile (: COMPILE) has been stripped down to the bare
1117       essentials of compilation.  It no longer deals with linking/execution.
1118       The interface has been cleaned up considerably.
1119    
1120       Utility routines for dealing with linking and execution have been
1121       moved into their own substructures.
1122    
1123       (The ultimate goal of these changes is to provide a light-weight
1124       binfile loader/linker (at least for, e.g., stable libraries) that
1125       does not require CM or the compiler to be present.)
1126    
1127    CM documentation has been updated to reflect the changes to library
1128    organization.
1129    
1130    ----------------------------------------------------------------------
1131    Name: Matthias Blume
1132    Date: 2001/07/10 17:30:00 EDT
1133    Tag: Release_110_34
1134    Description:
1135    
1136    Minor tweak to 110.34 (re-tagged):
1137    
1138      - README.html file added to CVS repository
1139      - runtime compiles properly under FreeBSD 3.X and 4.X
1140    
1141    ----------------------------------------------------------------------
1142    Name: Matthias Blume
1143    Date: 2001/07/10 17:30:00 EDT
1144    Tag: Release_110_34
1145    Description:
1146    
1147    New version number (110.34). New bootfiles.
1148    
1149    ----------------------------------------------------------------------
1150    Name: Matthias Blume
1151    Date: 2001/07/09 16:00:00 EDT
1152    Tag: blume-20010709-more-varargs
1153    Description:
1154    
1155    I changed the handling of varargs in ml-nlffigen again:
1156    The ellipsis ... will now simply be ignored (with an accompanying warning).
1157    
1158    The immediate effect is that you can actually call a varargs function
1159    from ML -- but you can't actually supply any arguments beyond the ones
1160    specified explicitly.  (For example, you can call printf with its format
1161    string, but you cannot pass additional arguments.)
1162    
1163    This behavior is only marginally more useful than the one before, but
1164    it has the advantage that a function or, more importantly, a function
1165    type never gets dropped on the floor, thus avoiding follow-up problems with
1166    other types that refer to the offending one.
1167    
1168    ----------------------------------------------------------------------
1169    Name: Matthias Blume
1170    Date: 2001/07/09 11:25:00 EDT
1171    Tag: blume-20010709-varargs
1172    Description:
1173    
1174    1. ckit-lib.cm now exports structure Error
1175    2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
1176       with a warning accompanied by a source location.  Moreover, it
1177       merely skips the offending function or type and proceeds with the
1178       rest of its work.u  As a result, one can safely feed C code containing
1179       "..." to ml-nlffigen.
1180    3. There are some internal improvements to CM, providing slightly
1181       more general string substitutions in the tools subsystem.
1182    
1183    ----------------------------------------------------------------------
1184    Name: Matthias Blume
1185    Date: 2001/06/27 15:10:00 EDT
1186    Tag: blume-20010627-concur
1187    Description:
1188    
1189    Fixed a small bug in CM's handling of parallel compilation.
1190    (You could observe the bug by Control-C-interrupting an ordinary
1191    CMB.make or CM.stabilize and then attaching some compile servers.
1192    The result was that all of a sudden the previously interrupted
1193    compilation would continue on its own.  This was because of
1194    an over-optimization: CM did not bother to clean out certain queues
1195    when no servers were attached "anyway", resulting in the contents
1196    of these queues to grab control when new servers did get attached.)
1197    
1198    There is also another minor update to the CM manual.
1199    
1200    ----------------------------------------------------------------------
1201    Name: Matthias Blume
1202    Date: 2001/06/26 16:15:00 EDT
1203    Tag: blume-20010626-cmdoc
1204    Description:
1205    
1206    Minor typo fixed in CM manual (syntax diagram for libraries).
1207    
1208    ----------------------------------------------------------------------
1209    Name: Matthias Blume
1210    Date: 2001/06/25 22:55:00 EDT
1211    Tag: blume-20010625-x86pc
1212    Description:
1213    
1214    Fixed a nasty bug in the X86 assembly code that caused signal
1215    handlers to fail (crash) randomly.
1216    
1217    ----------------------------------------------------------------------
1218    Name: Matthias Blume
1219    Date: 2001/06/25 12:05:00 EDT
1220    Tag: blume-20010625-nlffigen
1221    Description:
1222    
1223    This update fixes a number of minor bugs in ml-nlffigen as reported by
1224    Nick Carter <nbc@andrew.cmu.edu>.
1225    
1226      1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
1227      2. Default names for generated files are now derived from the name of
1228         the C file *without its directory*.  In particular, this causes generated
1229         files to be placed locally even if the C file is in some system directory.
1230      3. Default names for generated signatures and structures are also derived
1231         from the C file name without its directory.  This avoids silly things
1232         like "structure GL/GL".
1233         (Other silly names are still possible because ml-nlffigen does not do
1234          a thorough check of whether generated names are legal ML identifiers.
1235          When in doubt, use command line arguments to force particular names.)
1236    
1237    ----------------------------------------------------------------------
1238    Name: Matthias Blume
1239    Date: 2001/06/21 12:25:00 EDT
1240    Tag: blume-20010621-eXene
1241    Description:
1242    
1243    eXene now compiles and (sort of) works again.
1244    
1245    The library name (for version > 110.33) is $/eXene.cm.
1246    
1247    I also added an new example in src/eXene/examples/nbody.  See the
1248    README file there for details.
1249    
1250    ----------------------------------------------------------------------
1251    Name: Matthias Blume
1252    Date: 2001/06/20 16:40:00 EDT
1253    Tag: blume-20010620-cml
1254    Description:
1255    
1256    CML now compiles and works again.
1257    
1258    Libraries (for version > 110.33):
1259    
1260      $cml/cml.cm            Main CML library.
1261      $cml/basis.cm          CML's version of $/basis.cm.
1262      $cml/cml-internal.cm   Internal helper library.
1263      $cml/core-cml.cm       Internal helper library.
1264      $cml-lib/trace-cml.cm  Tracing facility.
1265      $cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
1266    
1267    The installer (config/install.sh) has been taught how to properly
1268    install this stuff.
1269    
1270    ----------------------------------------------------------------------
1271    Name: Matthias Blume
1272    Date: 2001/06/19 17:55:00 EDT
1273    Tag: blume-20010619-instantiate
1274    Description:
1275    
1276    This un-breaks the fix for bug 1432.
1277    (The bug was originally fixed in 110.9 but I broke it again some
1278    time after that.)
1279    
1280    ----------------------------------------------------------------------
1281    Name: Matthias Blume
1282    Date: 2001/06/19 17:25:00 EDT
1283    Tag: blume-20010619-signals
1284    Description:
1285    
1286    This should (hopefully) fix the long-standing signal handling bug.
1287    (The runtime system was constructing a continuation record with an
1288    incorrect descriptor which would cause the GC to drop data on the floor...)
1289    
1290    ----------------------------------------------------------------------
1291    Name: Matthias Blume
1292    Date: 2001/06/15 15:05:00 EDT
1293    Tag: blume-20010615-moresparc
1294    Description:
1295    
1296    Here is a short late-hour update related to Sparc c-calls:
1297    
1298     -- made handling of double-word arguments a bit smarter
1299    
1300     -- instruction selection phase tries to collapse certain clumsily
1301        constructed ML-Trees; typical example:
1302    
1303            ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
1304    
1305        This currently has no further impact on SML/NJ since mlriscGen does
1306        not seem to generate such patterns in the first place, and c-calls
1307        (which did generate them in the beginning) has meanwhile been fixed
1308        so as to avoid them as well.
1309    
1310    ----------------------------------------------------------------------
1311    Name: Matthias Blume
1312  Date: 2001/06/15 15:05:00 EDT  Date: 2001/06/15 15:05:00 EDT
1313  Tag: blume-20010615-sparc  Tag: blume-20010615-sparc
1314  Description:  Description:
# Line 2409  Line 3704 
3704       elaborator). There were a lot of changes during my "linkpath" trials       elaborator). There were a lot of changes during my "linkpath" trials
3705       that could have been reverted to their original state but weren't.       that could have been reverted to their original state but weren't.
3706       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
3707       than what was strictly necessary...  (I _did_ resist the tempation       than what was strictly necessary...  (I _did_ resist the temptation
3708       of doing any "global reformatting" to avoid an untimely death at       of doing any "global reformatting" to avoid an untimely death at
3709       Dave's hands. :)       Dave's hands. :)
3710    

Legend:
Removed from v.840  
changed lines
  Added in v.1060

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